pcbjam/wasm/bindings/eeschema_embind.cpp

2538 lines
93 KiB
C++
Raw Normal View History

/*
* Embind bindings for KiCad eeschema WASM.
*
* Picked up automatically by scripts/kicad/build-kicad-target.sh when building
* the eeschema app (it compiles wasm/bindings/<app>_embind.cpp if present).
*/
#ifdef __EMSCRIPTEN__
2026-06-03 17:49:11 +02:00
#include <emscripten.h>
#include <emscripten/bind.h>
#include <kiway_player.h>
#include <kiway.h>
#include <map>
2026-06-03 17:49:11 +02:00
#include <memory>
#include <set>
#include <string>
#include <vector>
#include <wx/app.h>
#include <wx/filename.h>
#include <wx/string.h>
#include <wx/window.h>
2026-06-03 17:49:11 +02:00
#include <nlohmann/json.hpp>
#include <kiid.h>
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
#include <layer_ids.h>
#include <pcbjam_read_only.h>
#include <project.h>
2026-06-03 17:49:11 +02:00
#include <schematic.h>
#include <sch_edit_frame.h>
#include <symbol_edit_frame.h>
#include <sch_io/kicad_sexpr/sch_io_kicad_sexpr.h>
#include <sch_sheet.h>
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
#include <richio.h>
#include <lib_symbol.h>
#include <tools/sch_selection.h>
#include <tools/sch_selection_tool.h>
2026-06-03 17:49:11 +02:00
#include <sch_commit.h>
#include <sch_draw_panel.h>
#include <geometry/eda_angle.h>
#include <math/util.h>
#include <tool/tool_manager.h>
#include <view/view.h>
#include <view/view_overlay.h>
#include <chrono>
#include <wx/event.h>
2026-06-03 17:49:11 +02:00
#include <sch_item.h>
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
#include <sch_line.h>
#include <sch_junction.h>
#include <sch_no_connect.h>
#include <sch_text.h>
#include <sch_label.h>
#include <sch_symbol.h>
#include <sch_field.h>
#include <sch_shape.h>
#include <eda_shape.h>
#include <stroke_params.h>
2026-06-03 17:49:11 +02:00
#include <sch_screen.h>
#include <sch_sheet_path.h>
#include <schematic_settings.h>
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
#include <tool/actions.h>
#include <tools/sch_actions.h>
#include <tool/coroutine.h>
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
#include <pcbjam_remote_lock.h>
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
#include "collab_common.h"
#include "open_gate.h"
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
#include "collab_presence_core.h"
#include "collab_presence_style.h"
#include "pcbjam_theme.h"
#include "pcbjam_libs_reload.h"
jspi: migration phases 0-7 — build knob, scheduler shim, test successor suite Toolchain: emsdk 6.0.6 (versions.sh; cache-hash keys on it). Build knob PCBJAM_ASYNC_BACKEND=jspi|asyncify: build-kicad-target.sh links editors with -sJSPI + -sJSPI_EXPORTS=@scripts/common/jspi-exports.txt + --pre-js jspi-scheduler.js (no DYNCALLS, no post-link asyncify pipeline); wx build stamps the backend and forces clean on flip or unknown provenance; docker/build.sh passes the knob, seeds the emscripten ports cache from the volume every launch, jspi postprocess = patch-env-shim only. scripts/common/shims/jspi-scheduler.js: the JSPI successor scheduler — token-wait registry, resume turnstile (one armed resume between engine re-entries, SP swaps only at microtask boundaries), green-region spill stacks (16-aligned tops), S1 embind mutator FIFO lane + parker wraps, S6 shutdown, libctx integration hooks (suspend/end/quarantine + g_current arm/clear), SuspendError attributor, lost-wake + stuck-window watchdogs, __wxWaitDump observability. Embind: PARKER registrations get emscripten::async() under PCBJAM_JSPI (wasm/bindings/pcbjam_async_policy.h). nanosleep yields route via the shim. Tests: tests/asyncify -> tests/jspi successor suite (jspi-stack red/green shadow-stack battery, jspi-coroutine MiniCoro harness, suspend-races semantic scenarios + __wxWaitDump books coherence); projects jspi-firefox/ jspi-chrome (asyncify-webkit retired — no JSPI in WebKit); unconditional Firefox JSPI pref; guard-beacons -> wait-beacons (+wxScheduler/libctxJspi families); Makefile.wasm links test apps against JSPI with the shim as a tracked link prerequisite. Web: WasmTool setRo await + __wxWaitDump forensics, open-flow contained promise, scheduler-shim.test.ts retargeted (8 green). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 07:06:24 +02:00
#include "pcbjam_async_policy.h"
#include <algorithm>
using namespace emscripten;
2026-06-03 17:49:11 +02:00
using json = nlohmann::json;
// Programmatically open a project file (schematic) in the running editor frame,
// without UI automation. Mirrors single_top.cpp's MacOpenFile path: the editor
// frame is the app's top window and is a KIWAY_PLAYER. Returns the result of
// OpenProjectFiles, or false if no frame is available — letting the JS caller
// fall back to driving File→Open.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
//
// KICAD_MERGED_EMBIND (kicad_editor, editor-unification Part 2): pcbnew_embind.cpp
// defines the identical function and registers the same JS names — in the merged image
// the frame-agnostic duplicates (this + kicadCollabOnSave) and the shared-name
// registrations live once in kicad_editor_embind.cpp, which dispatches the per-editor
// entries (renamed schCollab*/pcbCollab* below; JS-facing names are unchanged).
#ifndef KICAD_MERGED_EMBIND
bool kicadOpenFile( std::string path )
{
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// Held across every suspension of the load; see open_gate.h.
pcbjam_open::BusyGuard busy;
if( pcbjam_open::testParkMs() > 0 )
emscripten_sleep( pcbjam_open::testParkMs() );
KIWAY_PLAYER* frame =
wxTheApp ? static_cast<KIWAY_PLAYER*>( wxTheApp->GetTopWindow() ) : nullptr;
if( !frame )
return false;
if( wxWindow* blocking = frame->Kiway().GetBlockingDialog() )
blocking->Close( true );
bool ok = frame->OpenProjectFiles(
std::vector<wxString>( 1, wxString::FromUTF8( path.c_str() ) ) );
// Test-only post-load park (open_gate.h): model fully loaded, gate still
// closed — the deterministic window the collab-load-fuzz spec hammers.
if( pcbjam_open::testParkMs() > 0 )
emscripten_sleep( pcbjam_open::testParkMs() );
return ok;
}
// JS-pollable open-in-flight probe (open_gate.h): the web shell defers the
// collab/presence attach until the open chain has truly completed.
bool kicadOpenFileBusy()
{
return pcbjam_open::busy();
}
// Test-only (collab-load-fuzz): arm the deterministic open parks.
void kicadTestSetOpenPark( int aMs )
{
pcbjam_open::testParkMs() = aMs;
}
// Read-only viewer lock (read-only-viewer): flips the process-global
// PCBJAM_READ_ONLY flag consumed by TOOL_MANAGER (view-only action allowlist)
// and the selection tools (nothing selectable), and mirrors it onto the
// project so the setup dialogs grey out. Returns false until the editor frame
// exists so JS polls; the shell fails CLOSED if it never applies.
bool kicadSetReadOnly( bool aReadOnly )
{
KIWAY_PLAYER* frame =
wxTheApp ? dynamic_cast<KIWAY_PLAYER*>( wxTheApp->GetTopWindow() ) : nullptr;
if( !frame )
return false;
PCBJAM_READ_ONLY::Set( aReadOnly );
frame->Prj().SetReadOnly( aReadOnly );
return true;
}
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
#endif // !KICAD_MERGED_EMBIND
2026-06-03 17:49:11 +02:00
// ───────────────────────────── Yjs collaborative bridge ─────────────────────────────
//
// eeschema's half of the unified bridge contract (features/yjs-bridge/0001, 0003).
// Unlike pl_editor it needs NO kicad-fork change: SCH_ITEM already carries a stable
// KIID, and eeschema has native change machinery, so the adapter is a thin re-use:
// ChangeSource (emit) = a SCHEMATIC_LISTENER subclass (SCH_COMMIT::Push fires it)
// apply = SCH_COMMIT Modify/Remove + Push (drives connectivity recompute)
// The generic JS reconciler / transport / WasmTool wiring are reused unchanged.
//
// Scope of this first commit (0003 §"first PoC"): position-level sync of existing
// items — changed (move/edit) and removed. Decomposed field coverage via reflection,
// constructing arbitrary new item types on `added`, and symbol-instance / multi-sheet
// scoping are deferred (see TODOs). Items are resolved by globally-unique uuid, so
// changed/removed already work across the whole hierarchy without sheet scoping.
namespace {
// Guard so SCH_COMMIT::Push's listener callbacks during apply() aren't re-emitted.
bool s_applyingRemote = false;
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
using pcbjam_collab::toUtf8;
2026-06-03 17:49:11 +02:00
SCH_EDIT_FRAME* schFrame()
{
return wxTheApp ? dynamic_cast<SCH_EDIT_FRAME*>( wxTheApp->GetTopWindow() ) : nullptr;
}
// The screen of the sheet the editor is currently showing. Per-sheet collab keys a room
// to each .kicad_sch, so the snapshot/diff that feeds a room must cover ONLY this screen,
// never the whole Hierarchy(). GetCurrentSheet().LastScreen() is the active sheet's screen
// (GetScreen() tracks the same screen and is the fallback before a sheet path exists).
SCH_SCREEN* currentScreen( SCH_EDIT_FRAME* aFrame )
{
if( !aFrame )
return nullptr;
if( SCH_SCREEN* screen = aFrame->GetCurrentSheet().LastScreen() )
return screen;
return aFrame->GetScreen();
}
2026-06-03 17:49:11 +02:00
json itemToJson( SCH_ITEM* aItem )
{
VECTOR2I p = aItem->GetPosition();
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
json j = {
2026-06-03 17:49:11 +02:00
{ "id", toUtf8( aItem->m_Uuid.AsString() ) },
{ "type", toUtf8( aItem->GetClass() ) },
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
{ "x", p.x }, // internal units; integral, so no quantization needed
2026-06-03 17:49:11 +02:00
{ "y", p.y },
};
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
// Per-type fields needed to reconstruct the item on `added` (0003 converters).
// Hand-mapped for the wire-editing types; other types sync position-only (move) and
// are skipped on `added` (logged) until their converters are added.
if( aItem->Type() == SCH_LINE_T )
{
auto* line = static_cast<SCH_LINE*>( aItem );
j["sx"] = line->GetStartPoint().x;
j["sy"] = line->GetStartPoint().y;
j["ex"] = line->GetEndPoint().x;
j["ey"] = line->GetEndPoint().y;
j["layer"] = (int) line->GetLayer();
}
if( EDA_TEXT* txt = dynamic_cast<EDA_TEXT*>( aItem ) )
j["text"] = toUtf8( txt->GetText() );
if( SCH_LABEL_BASE* lbl = dynamic_cast<SCH_LABEL_BASE*>( aItem ) )
j["shape"] = (int) lbl->GetShape();
// Graphic shapes (circle / rectangle / arc / line / bezier): geometry is start/end plus,
// for an arc, the center, and for a bezier the two control points. Stroke + fill complete it.
if( aItem->Type() == SCH_SHAPE_T )
{
auto* shp = static_cast<SCH_SHAPE*>( aItem );
j["stype"] = (int) shp->GetShape(); // SHAPE_T
j["sx"] = shp->GetStart().x;
j["sy"] = shp->GetStart().y;
j["ex"] = shp->GetEnd().x;
j["ey"] = shp->GetEnd().y;
j["layer"] = (int) shp->GetLayer();
j["width"] = shp->GetStroke().GetWidth();
j["fill"] = (int) shp->GetFillMode();
if( shp->GetShape() == SHAPE_T::ARC )
{
VECTOR2I c = shp->GetCenter();
j["cx"] = c.x;
j["cy"] = c.y;
}
else if( shp->GetShape() == SHAPE_T::BEZIER )
{
j["c1x"] = shp->GetBezierC1().x;
j["c1y"] = shp->GetBezierC1().y;
j["c2x"] = shp->GetBezierC2().x;
j["c2y"] = shp->GetBezierC2().y;
}
}
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
return j;
}
// Construct a new SCH_ITEM from a delta item (for `added`), with the delta's uuid
// (m_Uuid is const → const_cast, exactly as the s-expr parser does). Returns nullptr for
// types without a converter yet.
SCH_ITEM* makeItem( const json& j )
{
std::string type = j.value( "type", "" );
SCH_ITEM* item = nullptr;
if( type == "SCH_LINE" )
{
int layer = j.value( "layer", (int) LAYER_NOTES );
auto* line = new SCH_LINE( VECTOR2I( j.value( "sx", 0 ), j.value( "sy", 0 ) ), layer );
line->SetEndPoint( VECTOR2I( j.value( "ex", 0 ), j.value( "ey", 0 ) ) );
item = line;
}
else if( type == "SCH_JUNCTION" )
{
item = new SCH_JUNCTION( VECTOR2I( j.value( "x", 0 ), j.value( "y", 0 ) ) );
}
else if( type == "SCH_NO_CONNECT" )
{
item = new SCH_NO_CONNECT( VECTOR2I( j.value( "x", 0 ), j.value( "y", 0 ) ) );
}
else if( type == "SCH_TEXT" )
{
auto* txt = new SCH_TEXT( VECTOR2I( j.value( "x", 0 ), j.value( "y", 0 ) ),
wxString::FromUTF8( j.value( "text", "" ).c_str() ) );
// Mirror the interactive text tool (sch_drawing_tools.cpp createNewText): parent the
// item to the schematic and apply the project's default text size, so a remotely-added
// text resolves variables / renders identically to a locally-placed one.
if( SCH_EDIT_FRAME* fr = schFrame() )
{
txt->SetParent( &fr->Schematic() );
int sz = fr->Schematic().Settings().m_DefaultTextSize;
txt->SetTextSize( VECTOR2I( sz, sz ) );
}
item = txt;
}
else if( type == "SCH_LABEL" || type == "SCH_GLOBALLABEL" || type == "SCH_HIERLABEL" )
{
VECTOR2I pos( j.value( "x", 0 ), j.value( "y", 0 ) );
wxString text = wxString::FromUTF8( j.value( "text", "" ).c_str() );
SCH_LABEL_BASE* lbl = ( type == "SCH_LABEL" ) ? (SCH_LABEL_BASE*) new SCH_LABEL( pos, text )
: ( type == "SCH_GLOBALLABEL" ) ? (SCH_LABEL_BASE*) new SCH_GLOBALLABEL( pos, text )
: (SCH_LABEL_BASE*) new SCH_HIERLABEL( pos, text );
if( j.contains( "shape" ) )
lbl->SetShape( (LABEL_FLAG_SHAPE) j["shape"].get<int>() );
item = lbl;
}
else if( type == "SCH_SHAPE" )
{
// Reconstruct from the geometry itemToJson emits. Committing a *new* SCH_SHAPE used to
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// trap in SCH_COMMIT::Push's CHT_ADD path under the retired asyncify runtime because
// doApply ran off the coroutine context; doApply runs inside a COROUTINE
// (kicadCollabApply), serialized with local edits, exactly as a native draw does.
// (0006/0007.) NB FILL_T::NO_FILL == 1, not 0.
SHAPE_T st = (SHAPE_T) j.value( "stype", (int) SHAPE_T::RECTANGLE );
int layer = j.value( "layer", (int) LAYER_NOTES );
int width = j.value( "width", 0 );
FILL_T fill = (FILL_T) j.value( "fill", (int) FILL_T::NO_FILL );
auto* shp = new SCH_SHAPE( st, (SCH_LAYER_ID) layer, width, fill );
// Rectangle: two corners. Circle: start = center, end = a point on the radius. Both are
// fully defined by start+end (what itemToJson emits via GetStart()/GetEnd()).
shp->SetStart( VECTOR2I( j.value( "sx", 0 ), j.value( "sy", 0 ) ) );
shp->SetEnd( VECTOR2I( j.value( "ex", 0 ), j.value( "ey", 0 ) ) );
if( st == SHAPE_T::ARC && j.contains( "cx" ) )
{
shp->SetCenterX( j["cx"].get<int>() );
shp->SetCenterY( j["cy"].get<int>() );
}
else if( st == SHAPE_T::BEZIER )
{
if( j.contains( "c1x" ) )
shp->SetBezierC1( VECTOR2I( j["c1x"].get<int>(), j["c1y"].get<int>() ) );
if( j.contains( "c2x" ) )
shp->SetBezierC2( VECTOR2I( j["c2x"].get<int>(), j["c2y"].get<int>() ) );
}
if( SCH_EDIT_FRAME* fr = schFrame() )
shp->SetParent( &fr->Schematic() );
item = shp;
}
// SCH_SYMBOL `added` is still deferred (needs the s-expr clipboard-blob emit + LoadContent
// reconstruction; symbol PLACEMENT is also natively blocked until symbol libraries are
// bundled — see features/yjs-bridge tasks). Symbol move/position already syncs via `changed`.
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
if( item )
const_cast<KIID&>( item->m_Uuid ) = KIID( wxString::FromUTF8( j.value( "id", "" ).c_str() ) );
return item;
2026-06-03 17:49:11 +02:00
}
// Current model of the ACTIVE screen as an array of item json. One collab room == one
// .kicad_sch screen, so we never fold in the rest of the hierarchy (uuids are unique
// within a screen, so no cross-sheet dedup is needed).
json snapshotItems( SCH_EDIT_FRAME* aFrame )
2026-06-03 17:49:11 +02:00
{
json arr = json::array();
2026-06-03 17:49:11 +02:00
if( SCH_SCREEN* screen = currentScreen( aFrame ) )
2026-06-03 17:49:11 +02:00
{
for( SCH_ITEM* item : screen->Items() )
arr.push_back( itemToJson( item ) );
2026-06-03 17:49:11 +02:00
}
return arr;
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// Wire emitters (legacy scalar delta + v2 items): shared, collab_common.h.
using pcbjam_collab::emitDelta;
using pcbjam_collab::emitItemsWire;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// Notify the standalone that the editor switched to a different sheet — a different
// .kicad_sch == a different collab room (ysync subschemas). The path is the active
// screen's load path: the same absolute MEMFS form kicadCollabOnSave emits, so the JS
// side strips the project prefix the same way (relativeProjectPath). No-op without a
// JS listener.
void emitSheetChanged()
{
SCH_SCREEN* screen = currentScreen( schFrame() );
if( !screen )
return;
std::string s = toUtf8( screen->GetFileName() );
EM_ASM( {
if( window.kicadCollab && window.kicadCollab.onSheetChanged )
staging CI green-up: P-1 apply-slot wedge under JSPI, popup P-4 regression, ysync double-seed header, spec fixes - P-1 (findings-p.spec P-1c red on CI + locally): a JS exception thrown from a window.kicadCollab wire callback does NOT reach the C++ catch(...) — under JSPI it REJECTS the promising coroutine entry, slot.done never flips and the apply slot is busy for the page's life. Three layers: libcontext flags the rejected record dead (kicad → d0afbc4bd3), drainApplies() reaps a !done && !CanResume() body, and every EM_ASM window.kicadCollab.onX(...) call is try/catch-guarded (collab_common.h, kicad_editor/eeschema/pl_editor embind). - popup.spec 'dismisses on outside click': wx P-4 SetFocus delegation excluded for wxPopupWindow (wxwidgets → a7436d1052). - ysync-two-tab concurrent seed: repairLayoutY + seedNonce retraction deleted both headers (pcbjam-shared → 8dabf2e). - dark-mode.spec: stale light-mode reference (pre wx fa46f1822c DOM re-projection) replaced with the CI chromium render. - items-wire-batch-loss.spec: precondition obsolete after P-5 — poisoned entry is now synthetic, plus a P-5 contract assertion (field blob == ''). - presence-ghost-peer.spec: skip early when no /parties/ socket (CI web e2e runs the BroadcastChannel provider); findings-p1.spec: polled click-select precondition. Local: kicad-chromium findings/presence/ysync/collab 59/59, popup 7/7, dark-mode 1/1, pcbjam-shared vitest 249/249. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LvvARNMxQrq17vogn3REMU
2026-08-28 22:25:21 +02:00
{
// A throwing listener must never unwind the wasm frame that called it: under
// JSPI that rejects the running coroutine's entry (findings P-1).
try { window.kicadCollab.onSheetChanged( UTF8ToString( $0 ) ); }
catch( e ) { console.error( '[pcbjam collab] onSheetChanged listener threw', e ); }
}
}, s.c_str() );
}
// ───────────────────────── sheet navigator bridge (sheet-panel) ─────────────────────────
//
// The React stand-in for the docked wx HIERARCHY_PANE that kicadSetChrome(false)
// hides: kicadSheetsGetTree() returns the loaded hierarchy (every SCH_SHEET_PATH
// instance — a sheet file placed twice is two rows, like the wx pane) plus the
// current path; kicadSheetsEnter(path) navigates. Same shape as pcbnew's layer
// bridge: validate synchronously, apply on the frame's coroutine, then push the
// fresh state to window.kicadCollab.onSheetsState so the panel is event-driven.
// Rows are page-number ordered (the wx pane's order); `parent` is the KIID path
// of the enclosing instance ("" for the root) so JS can rebuild the tree.
json sheetsStateJson( SCH_EDIT_FRAME* aFrame )
{
SCH_SHEET_LIST list = aFrame->Schematic().Hierarchy();
list.SortByPageNumbers( /* aUpdateVirtualPageNums */ false );
json sheets = json::array();
for( const SCH_SHEET_PATH& path : list )
{
SCH_SHEET_PATH parent = path;
if( parent.size() > 1 )
parent.pop_back();
SCH_SCREEN* screen = path.LastScreen();
SCH_SHEET* sheet = path.Last();
wxString name = sheet ? sheet->GetName() : wxString();
// The root sheet carries no Sheetname — show its file like the wx pane does.
if( name.IsEmpty() && screen )
name = wxFileName( screen->GetFileName() ).GetName();
sheets.push_back( {
{ "path", toUtf8( path.PathAsString() ) },
{ "parent", path.size() > 1 ? toUtf8( parent.PathAsString() ) : std::string() },
{ "name", toUtf8( name ) },
{ "file", screen ? toUtf8( screen->GetFileName() ) : std::string() },
{ "page", toUtf8( path.GetPageNumber() ) },
{ "depth", static_cast<int>( path.size() ) - 1 },
} );
}
return { { "current", toUtf8( aFrame->GetCurrentSheet().PathAsString() ) },
{ "sheets", sheets } };
}
void emitSheetsState( SCH_EDIT_FRAME* aFrame )
{
if( !aFrame )
return;
std::string s = sheetsStateJson( aFrame ).dump();
EM_ASM( {
if( window.kicadCollab && window.kicadCollab.onSheetsState )
{
// Never let a throwing listener unwind the wasm frame (findings P-1).
try { window.kicadCollab.onSheetsState( UTF8ToString( $0 ) ); }
catch( e ) { console.error( '[pcbjam sheets] onSheetsState listener threw', e ); }
}
}, s.c_str() );
}
fix(ysync): wire dialect == file dialect, uuid churn, drift noise The Y.Doc is the source of truth for the FILE, but both live wires wrote KiCad's CLIPBOARD dialect — a lossy, paste-oriented format. Every difference was permanent, unfixable drift. - pcbnew: serialize footprint blobs with CTL_FOR_BOARD, not CLIPBOARD_IO's CTL_FOR_CLIPBOARD, which emitted (version)(generator)(generator_version) inside every (footprint …). Keep (locked yes). - eeschema: aForClipboard=false — clipboard mode collapsed every symbol's (instances … (path "/sheet")) to (path ""). - Re-supply (version) at PARSE time only (withFootprintVersion): the token is invalid file content but load-bearing on decode — without it the parser starts at m_requiredVersion=0 and stamps (hide yes) on every mandatory field. - FOOTPRINT copy ctor: restore mandatory-field uuids (EDA_ITEM::operator= keeps the target's const m_Uuid, so Clone() rerolled all four). - drift: classify order-only diffs as `reordered` — y-sexpr v2 reorders legitimately; excluded from counts, report-worthiness and dedupe hashes. Migration 0017. - fpedit from eeschema: AsyncLoad()+BlockUntilLoaded() in initLibraryTree — FACE_PCB starts lazily there and never preloaded its libraries. Guards: wire-vs-file round-trip tests (pcbnew + eeschema), fpedit-from-eeschema (verified red without the fix), symedit-from-eeschema. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016p9kjdGBdcpwUSjJ3q5xg2
2026-07-20 18:29:52 +02:00
// Serialize one live schematic item to its native s-expr via a one-item
// SCH_SELECTION through SCH_IO_KICAD_SEXPR::Format. For a symbol the output also
// carries its (lib_symbols …) definition (that prelude is emitted for any symbol
// in the selection, independent of aForClipboard).
//
// aForClipboard MUST be false. Clipboard mode is a LOSSY, paste-oriented dialect:
// it rewrites `(instances (project … (path …)))` relative to aRelativePath — so a
// symbol on the current sheet collapses to `(path "")` — takes the REFERENCE field
// from the per-sheet instance instead of the ordinal one, and keeps orphaned
// instance data (sch_io_kicad_sexpr.cpp saveSymbol: ~758-766, ~791-806, ~903).
// The Y.Doc is the source of truth for the FILE, so a wire blob must be byte-equal
// to that item's subtree in a full file save; clipboard form would silently strip
// every symbol's sheet path and unit/reference on materialize.
//
// aRelativePath is still required (Format wxCHECKs it non-null) but is unread on
// the aForClipboard=false path.
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
std::string itemBlob( SCH_EDIT_FRAME* aFrame, SCH_ITEM* aItem )
{
SCH_SELECTION sel;
sel.SetScreen( aFrame->GetScreen() );
sel.Add( aItem );
STRING_FORMATTER fmt;
SCH_IO_KICAD_SEXPR plugin;
plugin.Format( &sel, &aFrame->GetCurrentSheet(), aFrame->Schematic(), &fmt,
fix(ysync): wire dialect == file dialect, uuid churn, drift noise The Y.Doc is the source of truth for the FILE, but both live wires wrote KiCad's CLIPBOARD dialect — a lossy, paste-oriented format. Every difference was permanent, unfixable drift. - pcbnew: serialize footprint blobs with CTL_FOR_BOARD, not CLIPBOARD_IO's CTL_FOR_CLIPBOARD, which emitted (version)(generator)(generator_version) inside every (footprint …). Keep (locked yes). - eeschema: aForClipboard=false — clipboard mode collapsed every symbol's (instances … (path "/sheet")) to (path ""). - Re-supply (version) at PARSE time only (withFootprintVersion): the token is invalid file content but load-bearing on decode — without it the parser starts at m_requiredVersion=0 and stamps (hide yes) on every mandatory field. - FOOTPRINT copy ctor: restore mandatory-field uuids (EDA_ITEM::operator= keeps the target's const m_Uuid, so Clone() rerolled all four). - drift: classify order-only diffs as `reordered` — y-sexpr v2 reorders legitimately; excluded from counts, report-worthiness and dedupe hashes. Migration 0017. - fpedit from eeschema: AsyncLoad()+BlockUntilLoaded() in initLibraryTree — FACE_PCB starts lazily there and never preloaded its libraries. Guards: wire-vs-file round-trip tests (pcbnew + eeschema), fpedit-from-eeschema (verified red without the fix), symedit-from-eeschema. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016p9kjdGBdcpwUSjJ3q5xg2
2026-07-20 18:29:52 +02:00
/*aForClipboard*/ false );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
return fmt.GetString();
}
// ── Emit via post-settle snapshot diff ───────────────────────────────────────────────────
//
// A local edit is a single SCH_COMMIT::Push that fires OnItemsAdded/Removed/Changed
// synchronously and THEN runs RecalculateConnections (sch_commit.cpp ~402-430). So the native
// listener only ever sees the *pre-cleanup* (raw) geometry, while the connectivity cleanup
// that follows — merging collinear wires, dropping redundant junctions, splitting at new
// crossings — is never reported. Broadcasting those raw per-category lists made the peer
// reconstruct the edit from the raw state and run ITS OWN cleanup, over a different "dirty"
// scope, so on a big connected drag the two peers cleaned up differently and the peer lost
// segments/junctions.
//
// Instead, treat the listener purely as a "something changed" trigger and broadcast a DIFF of
// the full model taken AFTER the edit settles — a CallAfter, which runs once Push (cleanup
// included) has fully returned. That captures tab A's FINAL, already-clean geometry; the peer
// applies it and re-cleaning already-clean geometry is idempotent, so the two converge. (This
// mirrors pl_editor's snapshot-differ.) g_baseline is the last-broadcast state.
// Diff baseline of the ACTIVE screen only (per-sheet collab room scope), keyed by uuid.
std::map<std::string, json> snapshotByUuid( SCH_EDIT_FRAME* aFrame )
2026-06-03 17:49:11 +02:00
{
std::map<std::string, json> m;
2026-06-03 17:49:11 +02:00
if( SCH_SCREEN* screen = currentScreen( aFrame ) )
2026-06-03 17:49:11 +02:00
{
for( SCH_ITEM* item : screen->Items() )
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
{
// ERC markers are screen items but not document content: the file
// writer never saves them and the selection writer cannot format
// one (empty blob → the JS side skips the entry with a warning).
// Keep them out of the collab model entirely.
if( item->Type() == SCH_MARKER_T )
continue;
m[toUtf8( item->m_Uuid.AsString() )] = itemToJson( item );
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
}
2026-06-03 17:49:11 +02:00
}
return m;
}
2026-06-03 17:49:11 +02:00
std::map<std::string, json> g_baseline;
bool g_flushScheduled = false;
// Roots the listener saw change since the last flush (uuids, captured at callback
// time — removed items may be freed before the flush runs). The scalar snapshot
// diff is a LOSSY projection: rotations/mirrors (anchor unchanged), field text
// edits (the commit stages the SCH_FIELD, which is not a screen item), stroke
// properties etc. never move it (bug 04). Dirty roots emit their v2 blob
// unconditionally; the apply is an idempotent upsert and the TS layer drops
// no-op bodies, so a false positive costs one local serialization.
std::set<std::string> g_dirty;
// Lift a commit-staged item to the SCREEN item the differ tracks (a field/pin/cell
// lifts to its symbol/sheet/label/table — same promotion sch_commit's undo uses).
void noteDirty( SCH_ITEM* aItem )
{
if( !aItem )
return;
while( EDA_ITEM* p = aItem->GetParent() )
{
if( !p->IsType( { SCH_SYMBOL_T, SCH_TABLE_T, SCH_SHEET_T, SCH_LABEL_LOCATE_ANY_T } ) )
break;
aItem = static_cast<SCH_ITEM*>( p );
}
g_dirty.insert( toUtf8( aItem->m_Uuid.AsString() ) );
}
// Re-seed the diff baseline to the current model — after handing out a seed snapshot, or after
// applying a remote delta (so those items aren't re-broadcast as a spurious local diff/echo).
// Declares "current model == broadcast state", so pending dirty marks are stale too — on a
// sheet switch they'd otherwise emit the OLD sheet's items into the new sheet's room.
void rebaseline()
{
if( SCH_EDIT_FRAME* fr = schFrame() )
g_baseline = snapshotByUuid( fr );
g_dirty.clear();
}
// TARGETED rebaseline (bug 05): refresh baseline entries ONLY for the uuids a remote
// apply touched. A global rebaseline() would fold a concurrently-committed local edit
// (its flush is queued BEHIND the apply on the same pending-event list) into the
// baseline and silently swallow it; targeted, the edit still diffs and emits. The
// connectivity cleanup the apply's Push produced likewise stays diffable — the
// post-apply flush broadcasts it (idempotent on the original sender).
void rebaselineTouched( SCH_EDIT_FRAME* aFrame, const std::vector<std::string>& aIds )
{
for( const std::string& id : aIds )
{
g_baseline.erase( id );
KIID kid( wxString::FromUTF8( id.c_str() ) );
if( SCH_ITEM* live = aFrame->Schematic().ResolveItem( kid, nullptr, /*allowNull*/ true ) )
g_baseline[id] = itemToJson( live );
}
}
// Diff the current (settled, post-cleanup) model against the baseline and broadcast the change.
void flushDiff()
{
g_flushScheduled = false;
SCH_EDIT_FRAME* fr = schFrame();
2026-06-03 17:49:11 +02:00
if( !fr )
return;
std::map<std::string, json> cur = snapshotByUuid( fr );
json added = json::array(), changed = json::array(), removed = json::array();
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// v2 items wire (per-item s-expr blobs), built from the same diff. Screen items
// are already root-level (fields live inside their symbols), so no lifting.
json wAdded = json::array(), wChanged = json::array();
std::set<std::string> wDone;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
auto blobFor = [&]( const std::string& id, json& aArr )
{
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
KIID kid( wxString::FromUTF8( id.c_str() ) );
SCH_ITEM* item = fr->Schematic().ResolveItem( kid, nullptr, /*allowNull*/ true );
if( !item )
return;
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
// A child that reached the dirty set unlifted (field/pin/sheet-pin —
// ResolveItem walks direct children too) serializes to an EMPTY blob:
// the selection writer has no case for it, so the entry used to reach
// JS as envelope furniture and be skipped, silently losing the edit.
// Lift to the screen item the differ tracks (same promotion as
// noteDirty) so the parent's re-blob carries the child's content.
while( EDA_ITEM* p = item->GetParent() )
{
if( !p->IsType( { SCH_SYMBOL_T, SCH_TABLE_T, SCH_SHEET_T, SCH_LABEL_LOCATE_ANY_T } ) )
break;
item = static_cast<SCH_ITEM*>( p );
}
std::string rootId = toUtf8( item->m_Uuid.AsString() );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
if( !wDone.insert( rootId ).second )
return;
std::string sexpr = itemBlob( fr, item );
// Still empty after lifting (a marker that slipped in, an unformattable
// type): never put a hollow envelope on the wire — mirror pcbnew's P-5
// guard and leave a breadcrumb instead.
if( sexpr.empty() )
{
std::string type = toUtf8( item->GetClass() );
EM_ASM( {
console.warn( '[pcbjam collab] eeschema: skipped un-serializable dirty root',
{ uuid: UTF8ToString( $0 ), type: UTF8ToString( $1 ) } );
}, rootId.c_str(), type.c_str() );
return;
}
aArr.push_back( json{ { "sexpr", sexpr }, { "parent", nullptr } } );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
};
for( const auto& [id, j] : cur )
{
auto it = g_baseline.find( id );
if( it == g_baseline.end() )
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
{
added.push_back( j );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
blobFor( id, wAdded );
}
else if( it->second != j )
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
{
changed.push_back( j );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
blobFor( id, wChanged );
}
}
for( const auto& [id, j] : g_baseline )
{
if( !cur.count( id ) )
removed.push_back( id );
}
// Dirty roots (bug 04): whatever the listener saw commit emits its blob on the
// v2 wire even when the scalar projection didn't move (rotation, field text,
// stroke edits). wDone dedups against the scalar-diff emits; deleted ids
// resolve null inside blobFor and skip (the removal loop covered them).
for( const std::string& id : g_dirty )
blobFor( id, wChanged );
g_dirty.clear();
g_baseline = std::move( cur );
2026-06-03 17:49:11 +02:00
if( !added.empty() || !changed.empty() || !removed.empty() )
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
emitDelta( json{ { "added", added }, { "changed", changed }, { "removed", removed } } );
if( !wAdded.empty() || !wChanged.empty() || !removed.empty() )
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
emitItemsWire( json{ { "added", wAdded }, { "changed", wChanged }, { "removed", removed } } );
}
// Coalesce all the listener callbacks of one commit (and any other edits in the same loop
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// turn) into a single post-settle diff. flushDiff runs inside a COROUTINE via the apply
// queue: its v2 items emit serializes items via SCH_IO_KICAD_SEXPR::Format (itemBlob), and
// queueing it with the applies keeps emits from interleaving with a suspended apply body
// (under the retired asyncify runtime the bare CallAfter stack additionally trapped here —
// same lesson as doApply, 0007).
void scheduleFlush()
{
if( g_flushScheduled )
return;
g_flushScheduled = true;
if( SCH_EDIT_FRAME* fr = schFrame() )
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, []() { flushDiff(); } );
else
flushDiff();
}
// A hierarchical sheet was just created locally ("Add Sheet"): write its new child screen
// to the child .kicad_sch file and tell the standalone (window.kicadCollab.onSheetCreated),
// so the child is persisted/registered the moment it's created — without waiting for the
// user to enter it or save the project. Otherwise the parent's `(sheet … child)` reference
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// dangles for peers / on reload. Deferred onto the apply coroutine (CallAfter + COROUTINE),
// same as flushDiff/doApply, so the save serializes with any in-flight apply.
// The sheet is re-resolved by uuid in the deferred body so a
// since-deleted sheet (e.g. an immediate undo) is a no-op rather than a dangling pointer.
void scheduleSheetSave( SCH_SHEET* aSheet )
{
SCH_EDIT_FRAME* fr = schFrame();
SCH_SCREEN* parent = currentScreen( fr );
if( !fr || !parent || !aSheet->GetScreen() )
return;
wxFileName childFn( aSheet->GetFileName() ); // relative "Sheetfile"
childFn.MakeAbsolute( wxFileName( parent->GetFileName() ).GetPath() );
std::string childAbs = toUtf8( childFn.GetFullPath() );
std::string uuid = toUtf8( aSheet->m_Uuid.AsString() );
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, childAbs, uuid]() {
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
KIID kid( wxString::FromUTF8( uuid.c_str() ) );
SCH_ITEM* item = fr->Schematic().ResolveItem( kid, nullptr, /*allowNull*/ true );
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
if( !item || item->Type() != SCH_SHEET_T )
return;
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
try
{
SCH_IO_KICAD_SEXPR io;
io.SaveSchematicFile( wxString::FromUTF8( childAbs.c_str() ),
static_cast<SCH_SHEET*>( item ), &fr->Schematic() );
}
catch( ... )
{
return; // a write failure must not abort the runtime
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
EM_ASM( {
if( window.kicadCollab && window.kicadCollab.onSheetCreated )
staging CI green-up: P-1 apply-slot wedge under JSPI, popup P-4 regression, ysync double-seed header, spec fixes - P-1 (findings-p.spec P-1c red on CI + locally): a JS exception thrown from a window.kicadCollab wire callback does NOT reach the C++ catch(...) — under JSPI it REJECTS the promising coroutine entry, slot.done never flips and the apply slot is busy for the page's life. Three layers: libcontext flags the rejected record dead (kicad → d0afbc4bd3), drainApplies() reaps a !done && !CanResume() body, and every EM_ASM window.kicadCollab.onX(...) call is try/catch-guarded (collab_common.h, kicad_editor/eeschema/pl_editor embind). - popup.spec 'dismisses on outside click': wx P-4 SetFocus delegation excluded for wxPopupWindow (wxwidgets → a7436d1052). - ysync-two-tab concurrent seed: repairLayoutY + seedNonce retraction deleted both headers (pcbjam-shared → 8dabf2e). - dark-mode.spec: stale light-mode reference (pre wx fa46f1822c DOM re-projection) replaced with the CI chromium render. - items-wire-batch-loss.spec: precondition obsolete after P-5 — poisoned entry is now synthetic, plus a P-5 contract assertion (field blob == ''). - presence-ghost-peer.spec: skip early when no /parties/ socket (CI web e2e runs the BroadcastChannel provider); findings-p1.spec: polled click-select precondition. Local: kicad-chromium findings/presence/ysync/collab 59/59, popup 7/7, dark-mode 1/1, pcbjam-shared vitest 249/249. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LvvARNMxQrq17vogn3REMU
2026-08-28 22:25:21 +02:00
{
// A throwing listener must never unwind the wasm frame that called it: under
// JSPI that rejects the running coroutine's entry (findings P-1).
try { window.kicadCollab.onSheetCreated( UTF8ToString( $0 ) ); }
catch( e ) { console.error( '[pcbjam collab] onSheetCreated listener threw', e ); }
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
}, childAbs.c_str() );
} );
}
// ChangeSource: the native SCHEMATIC_LISTENER is just a trigger — the actual change set comes
// from the post-settle snapshot diff above. Skipped while applying a remote delta (no echo);
// doApply rebaselines instead.
//
// Presence (collab-presence 0003): schematic changes often change the selection
// too (delete, paste) with no closing canvas event — the trigger below also
// piggybacks a selection re-check. Defined in the presence section further down.
void schedulePresenceSelCheck();
// Findings Y-1/Y-3: ANY document change (local commit or remote apply) repaints
// the peers' overlay from the live document and re-checks the local selection,
// post-settle on the apply coroutine. Runs BEFORE the applying-remote early
// return below — the remote apply is exactly the case that left peers' boxes on
// deleted items.
void schedulePresenceDocChanged();
class COLLAB_LISTENER : public SCHEMATIC_LISTENER
{
public:
void OnSchItemsAdded( SCHEMATIC&, std::vector<SCH_ITEM*>& aItems ) override
{
// A newly-added hierarchical sheet → persist + register its child file (above).
if( !s_applyingRemote )
{
for( SCH_ITEM* item : aItems )
if( item->Type() == SCH_SHEET_T )
scheduleSheetSave( static_cast<SCH_SHEET*>( item ) );
}
trigger( aItems );
}
void OnSchItemsChanged( SCHEMATIC&, std::vector<SCH_ITEM*>& v ) override { trigger( v ); }
void OnSchItemsRemoved( SCHEMATIC&, std::vector<SCH_ITEM*>& v ) override { trigger( v ); }
// The editor switched to a different sheet (a different .kicad_sch == a different
// collab room). Re-baseline so the first edit on the new sheet diffs against ITS
// screen, not the previous one, then tell the standalone to rebind its room to the
// now-active sheet file. Fires from SCH_EDIT_FRAME::DisplayCurrentSheet, by which
// point GetCurrentSheet()/GetScreen() already point at the new sheet.
void OnSchSheetChanged( SCHEMATIC& ) override
{
rebaseline();
emitSheetChanged();
// The floating sheet panel follows wx-driven navigation too (double-
// clicking a sheet symbol, Alt+Backspace, the toolbar arrows).
emitSheetsState( schFrame() );
}
private:
// Capture the touched roots at callback time (fields/pins lift to their
// symbol — noteDirty), then coalesce into one post-settle flush.
void trigger( const std::vector<SCH_ITEM*>& aItems )
{
schedulePresenceDocChanged();
if( s_applyingRemote )
return;
for( SCH_ITEM* item : aItems )
noteDirty( item );
scheduleFlush();
schedulePresenceSelCheck();
}
2026-06-03 17:49:11 +02:00
};
COLLAB_LISTENER* g_listener = nullptr;
// Get the live SCHEMATIC and ensure our listener is registered on it (idempotent).
SCHEMATIC* ensureBridge()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return nullptr;
SCHEMATIC& sch = fr->Schematic();
if( !g_listener )
{
g_listener = new COLLAB_LISTENER();
sch.AddListener( g_listener );
}
return &sch;
}
// ───────────────────────── collab presence (collab-presence 0003) ─────────────────────────
//
// eeschema port of pcbnew's presence layer (0002 — see pcbnew_embind.cpp for the full
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// design rationale). The state + event/scheduling machinery live in the shared
// pcbjam_presence::CORE (collab_presence_core.h); this TU supplies only the
// eeschema-specific hooks: frame/tool lookup, KIID resolution via
// SCHEMATIC::ResolveItem, the bare-uuid selection payload, and the per-peer draw
// (cross-app ghosts gated to the CURRENT sheet). Zero kicad-fork changes: wx-layer
// Bind() triggers on the GAL canvas + the COLLAB_LISTENER piggyback, selection read
// post-settle from SCH_SELECTION_TOOL. Rooms are per-sheet (warm pool), so peers
// publishing cursor/selection in the bound room are BY CONSTRUCTION on this same
// sheet file — no sheet filtering is needed for same-room selections; the JS side
// rebinds the whole presence layer on sheet navigation (onSheetChanged) and clears
// the overlay in between.
// Cross-app selection (0006): the schematic items a peer's xsel (symbol uuids
// from a pcbnew peer) resolves to on the CURRENT sheet. Unlike same-room
// selections (per-sheet rooms scope those by construction) xsel arrives
// project-wide, so only items that resolve onto the sheet THIS canvas is
// showing count — a bbox from another sheet's screen would land at meaningless
// coordinates. ONE resolver shared by the ghost render and the test probe
// (kicadCollabTestGetCrossMapped), so the assertion can't drift from the pixels.
std::vector<SCH_ITEM*> resolveXsel( SCH_EDIT_FRAME* aFrame, const pcbjam_presence::PEER& aPeer )
{
std::vector<SCH_ITEM*> items;
for( const KIID& id : aPeer.xsel )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
SCH_SHEET_PATH path;
SCH_ITEM* item = aFrame->Schematic().ResolveItem( id, &path, /*allowNull*/ true );
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
if( item && path.LastScreen() == aFrame->GetScreen() )
items.push_back( item );
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return items;
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
pcbjam_presence::CORE& presenceCore()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
static pcbjam_presence::CORE core = []()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
pcbjam_presence::CORE c;
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// eeschema ships its own defaults (hairline outline, subtler fill/cursor).
c.style = pcbjam_presence::eeschemaDefaultStyle();
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
c.frame = []() -> EDA_DRAW_FRAME* { return schFrame(); };
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
c.selectionTool = []( EDA_DRAW_FRAME* fr ) -> SELECTION_TOOL*
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return fr->GetToolManager()->GetTool<SCH_SELECTION_TOOL>();
};
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// Selection emit = the bare uuid array (0006: eeschema uuids ARE the
// symbol uuids; pcbnew's counterpart adds fpPaths).
c.selectionEmitPayload = []( EDA_DRAW_FRAME* fr ) -> json
{
return presenceCore().selectionUuids( fr );
};
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
c.resolveItem = []( EDA_DRAW_FRAME* fr, const KIID& id ) -> EDA_ITEM*
{
return static_cast<SCH_EDIT_FRAME*>( fr )->Schematic()
.ResolveItem( id, nullptr, /*allowNull*/ true );
};
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
c.drawPeerShapes = []( pcbjam_presence::CORE& aCore, EDA_DRAW_FRAME* aFrame,
const pcbjam_presence::PEER& peer, const KIGFX::COLOR4D& color,
double px )
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
SCH_EDIT_FRAME* fr = static_cast<SCH_EDIT_FRAME*>( aFrame );
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
for( const KIID& id : peer.selection )
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
{
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( id, &path, /*allowNull*/ true );
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
if( !item )
continue; // not in this schematic (yet) — skip silently
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
feat(collab): follow-user (collab-presence 0008) + chip depth-layer fix Follow-user: click a peer's roster avatar to mirror their viewport until local input breaks it. - collab_presence_core.h: CORE::fitViewport(cx, cy, halfW, halfH) — fit the leader's world rect with CONTAIN semantics (zoom derived from the follower's own canvas via the ToScreen ratio; GetScale is the zoom, not px/IU). Exported as kicadCollabFitViewport from both editor TUs + the merged dispatcher. - presence-kicad.ts: publish the visible world rect (viewportRect) into awareness, 100 ms trailing throttle; guarded for pre-0008 handles. - follow-user.ts: createFollow — follows an awareness CLIENT (a tab, not a user); applies leader rect changes via FitViewport, dedupes unchanged republishes; break-on-interact compares local onViewport echoes against the last applied rect (2% rel tolerance, echo-grace before the first fit lands); unfollows on leader-left; pauses on eeschema sheet mismatch. - PresenceRoster: avatars are follow toggles (ring on the followed peer); WasmTool renders the "Following <name> — move to stop" banner. - tests: 7 controller units (85/85 collab), fitViewport round-trip e2e in both kicad presence specs (20/20), two-tab tests/web/follow.spec.ts (converge → track → wheel-zoom breaks → subsequent moves ignored). Chip depth-layer fix (user-reported): name chips washed out inside low-alpha selection fills — chip rects shared the shapes overlay's single depth, and same-depth fragments drawn LATER lose the depth test, so an earlier-painted fill rejected the chip's pixels. Now three layers via the fork's VIEW_OVERLAY::SetDepthOffset: text (0) < chips + pin dots (1) < selection shapes (2). drawLabel/drawCursor/drawSelectionBox take the chip overlay explicitly; comment-pin dots move to the chip layer too (the 0005 "drawn last so pins sit above" comment had the rule backwards). Verified with a chip-inside-30%-fill pixel repro + the full presence suite. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-09 09:30:40 +02:00
pcbjam_presence::drawSelectionBox( aCore.overlay.get(), aCore.chipOverlay.get(),
aCore.textOverlay.get(),
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
item->ViewBBox(), peer.name, color, px,
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
aCore.style );
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// Cross-app ghosts (0006) — see resolveXsel for the sheet gating.
if( !peer.xsel.empty() )
{
pcbjam_presence::STYLE ghost = pcbjam_presence::ghostStyle( aCore.style );
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
for( SCH_ITEM* item : resolveXsel( fr, peer ) )
{
feat(collab): follow-user (collab-presence 0008) + chip depth-layer fix Follow-user: click a peer's roster avatar to mirror their viewport until local input breaks it. - collab_presence_core.h: CORE::fitViewport(cx, cy, halfW, halfH) — fit the leader's world rect with CONTAIN semantics (zoom derived from the follower's own canvas via the ToScreen ratio; GetScale is the zoom, not px/IU). Exported as kicadCollabFitViewport from both editor TUs + the merged dispatcher. - presence-kicad.ts: publish the visible world rect (viewportRect) into awareness, 100 ms trailing throttle; guarded for pre-0008 handles. - follow-user.ts: createFollow — follows an awareness CLIENT (a tab, not a user); applies leader rect changes via FitViewport, dedupes unchanged republishes; break-on-interact compares local onViewport echoes against the last applied rect (2% rel tolerance, echo-grace before the first fit lands); unfollows on leader-left; pauses on eeschema sheet mismatch. - PresenceRoster: avatars are follow toggles (ring on the followed peer); WasmTool renders the "Following <name> — move to stop" banner. - tests: 7 controller units (85/85 collab), fitViewport round-trip e2e in both kicad presence specs (20/20), two-tab tests/web/follow.spec.ts (converge → track → wheel-zoom breaks → subsequent moves ignored). Chip depth-layer fix (user-reported): name chips washed out inside low-alpha selection fills — chip rects shared the shapes overlay's single depth, and same-depth fragments drawn LATER lose the depth test, so an earlier-painted fill rejected the chip's pixels. Now three layers via the fork's VIEW_OVERLAY::SetDepthOffset: text (0) < chips + pin dots (1) < selection shapes (2). drawLabel/drawCursor/drawSelectionBox take the chip overlay explicitly; comment-pin dots move to the chip layer too (the 0005 "drawn last so pins sit above" comment had the rule backwards). Verified with a chip-inside-30%-fill pixel repro + the full presence suite. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-09 09:30:40 +02:00
pcbjam_presence::drawSelectionBox( aCore.overlay.get(), aCore.chipOverlay.get(),
aCore.textOverlay.get(),
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
item->ViewBBox(), peer.name, color, px,
ghost );
}
}
};
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return c;
}();
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return core;
}
void schedulePresenceSelCheck()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().scheduleSelCheck();
}
void schedulePresenceDocChanged()
{
presenceCore().onDocChanged();
}
2026-06-03 17:49:11 +02:00
} // namespace
namespace {
// SCH_SYMBOL::Move() / SCH_LABEL_BASE::Move() move their child fields (reference, value, …) via
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// an inner `field.Move()`. Under the retired asyncify runtime that inner call silently no-oped
// in the apply context — the field text was left behind at its old position while the body
// moved — so the fields are re-moved here with a devirtualized call. Kept as-is through the
// JSPI migration; the collab move/drift suites pin the peer's field positions.
void moveFields( std::vector<SCH_FIELD>& aFields, const VECTOR2I& aDelta )
{
for( SCH_FIELD& field : aFields )
field.SCH_FIELD::Move( aDelta );
}
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// Move an item to an absolute position for the `changed` path. Under the retired asyncify
// runtime the virtual SCH_ITEM::Move(), dispatched from the apply/CallAfter context, silently
// NO-OPED (symbols/junctions/labels never moved on the peer — only SCH_LINE worked, via its
// direct SetStart/EndPoint path; see 0003 / eeschema_collab_asyncify_apply). Move() is
// therefore devirtualized with explicit class-qualified calls — statically bound plain wasm
// calls, correct on any runtime, so they stay. Composite items additionally need their child
// fields moved (see moveFields).
void moveItemTo( SCH_ITEM* aItem, const VECTOR2I& aNewPos )
{
VECTOR2I delta = aNewPos - aItem->GetPosition();
if( delta == VECTOR2I( 0, 0 ) )
return;
switch( aItem->Type() )
{
case SCH_SYMBOL_T:
{
auto* sym = static_cast<SCH_SYMBOL*>( aItem );
sym->SCH_SYMBOL::Move( delta );
moveFields( sym->GetFields(), delta );
break;
}
case SCH_JUNCTION_T: static_cast<SCH_JUNCTION*>( aItem )->SCH_JUNCTION::Move( delta ); break;
case SCH_NO_CONNECT_T: static_cast<SCH_NO_CONNECT*>( aItem )->SCH_NO_CONNECT::Move( delta ); break;
case SCH_TEXT_T: static_cast<SCH_TEXT*>( aItem )->SCH_TEXT::Move( delta ); break;
case SCH_LABEL_T:
case SCH_GLOBAL_LABEL_T:
case SCH_HIER_LABEL_T:
{
auto* lbl = static_cast<SCH_LABEL_BASE*>( aItem );
lbl->SCH_LABEL_BASE::Move( delta );
moveFields( lbl->GetFields(), delta );
break;
}
default:
aItem->Move( delta ); // virtual fallback (may no-op in the apply context)
break;
}
}
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
// The actual model mutation, via SCH_COMMIT so connectivity/ERC recompute as for a UI
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// edit. (0003 decoded an asyncify-era dynCall trap on this path; the shim that healed
// it retired with that runtime.)
2026-06-03 17:49:11 +02:00
void doApply( SCH_EDIT_FRAME* aFrame, const json& aDelta )
{
SCHEMATIC& sch = aFrame->Schematic();
s_applyingRemote = true;
SCH_COMMIT commit( aFrame );
bool staged = false;
// With SKIP_UNDO no undo picker takes ownership of removed items; the commit
// detaches them from the screen and we free them after Push. Fields are hidden
// by CHT_REMOVE, not detached (still owned by their parent), so never freed.
std::vector<SCH_ITEM*> removedItems;
2026-06-03 17:49:11 +02:00
for( const json& rid : aDelta.value( "removed", json::array() ) )
{
SCH_SHEET_PATH path;
KIID id( wxString::FromUTF8( rid.get<std::string>().c_str() ) );
if( SCH_ITEM* item = sch.ResolveItem( id, &path, /*allowNull*/ true ) )
{
commit.Remove( item, path.LastScreen() );
if( item->Type() != SCH_FIELD_T )
removedItems.push_back( item );
2026-06-03 17:49:11 +02:00
staged = true;
}
}
for( const json& j : aDelta.value( "changed", json::array() ) )
{
SCH_SHEET_PATH path;
KIID id( wxString::FromUTF8( j.value( "id", "" ).c_str() ) );
if( SCH_ITEM* item = sch.ResolveItem( id, &path, /*allowNull*/ true ) )
{
commit.Modify( item, path.LastScreen() );
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
if( item->Type() == SCH_LINE_T && j.contains( "sx" ) )
{
// Wires reshape (endpoints move independently), so set both points.
auto* line = static_cast<SCH_LINE*>( item );
line->SetStartPoint( VECTOR2I( j["sx"].get<int>(), j["sy"].get<int>() ) );
line->SetEndPoint( VECTOR2I( j["ex"].get<int>(), j["ey"].get<int>() ) );
}
else if( j.contains( "x" ) && j.contains( "y" ) )
2026-06-03 17:49:11 +02:00
{
moveItemTo( item, VECTOR2I( j["x"].get<int>(), j["y"].get<int>() ) );
2026-06-03 17:49:11 +02:00
}
staged = true;
}
}
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
for( const json& j : aDelta.value( "added", json::array() ) )
{
KIID id( wxString::FromUTF8( j.value( "id", "" ).c_str() ) );
if( sch.ResolveItem( id, nullptr, /*allowNull*/ true ) )
continue; // already present (our own echo)
if( SCH_ITEM* item = makeItem( j ) )
{
commit.Add( item, aFrame->GetScreen() );
staged = true;
}
else
{
EM_ASM( { console.log( "[collab] eeschema apply: no converter for added type " + UTF8ToString( $0 ) ); },
j.value( "type", "?" ).c_str() );
}
}
2026-06-03 17:49:11 +02:00
// SKIP_UNDO: a peer's edit must never land on this editor's undo stack — Ctrl+Z
// would revert (and re-broadcast) the peer's work. Undo is local-ops-only; stale
// local undo entries are dropped/re-resolved by UUID at undo time (miss 09).
2026-06-03 17:49:11 +02:00
if( staged )
commit.Push( wxT( "Collaborative edit" ), SKIP_UNDO );
for( SCH_ITEM* item : removedItems )
delete item;
2026-06-03 17:49:11 +02:00
// The applied remote changes (and any connectivity cleanup they triggered) are now the
// shared state — fold them into the baseline so the post-apply listener flush doesn't
// re-broadcast them as a local diff (echo).
rebaseline();
2026-06-03 17:49:11 +02:00
s_applyingRemote = false;
}
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// v2 items apply: removed by uuid; added/changed are an idempotent per-item upsert.
// Each blob is parsed through the clipboard-paste path — LoadContent into a throwaway
// sheet (sch_editor_control.cpp Paste pattern) — then the loaded items are detached,
// matched by uuid against the live model (replace), lib-relinked for symbols, and
// committed. Runs inside the apply COROUTINE (see kicadCollabApplyItems).
// ysync bug 07 (UP side): an apply is queued for the room of ONE sheet but runs
// deferred on whatever screen is active when it runs. The JS binding stamps the
// envelope with its sheet's project-relative path; if that is not the screen
// shown now, adding the items here would put another sheet's items on this
// screen (8/28: the root's `(sheet …)` entries landed in a subsheet, which then
// referenced itself and failed to load). Drop it — the binding for the shown
// sheet reconciles on its own bind.
static bool applyTargetsShownSheet( SCH_EDIT_FRAME* aFrame, const json& aWire )
{
auto it = aWire.find( "sheet" );
if( it == aWire.end() || !it->is_string() )
return true; // untagged (single-file tools, older clients): legacy behaviour
const std::string want = it->get<std::string>();
SCH_SCREEN* cur = currentScreen( aFrame );
const std::string have = cur ? toUtf8( cur->GetFileName() ) : std::string();
if( have == want )
return true;
// `have` is absolute (MEMFS project dir + relative path); `want` is relative.
if( have.size() > want.size() && have[have.size() - want.size() - 1] == '/'
&& have.compare( have.size() - want.size(), want.size(), want ) == 0 )
return true;
EM_ASM( { console.warn( "[collab] applyItems dropped: envelope sheet " + UTF8ToString( $0 )
+ " != shown sheet " + UTF8ToString( $1 ) ); },
want.c_str(), have.c_str() );
return false;
}
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
void doApplyItems( SCH_EDIT_FRAME* aFrame, const json& aWire )
{
if( !applyTargetsShownSheet( aFrame, aWire ) )
return;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
SCHEMATIC& sch = aFrame->Schematic();
s_applyingRemote = true;
SCH_COMMIT commit( aFrame );
bool staged = false;
std::vector<std::string> touched; // uuids this apply acts on (targeted rebaseline)
// Owned by nobody once the SKIP_UNDO commit detaches them — freed after Push
// (fields are hidden, not detached, so excluded). See doApply.
std::vector<SCH_ITEM*> removedItems;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
for( const json& rid : aWire.value( "removed", json::array() ) )
{
SCH_SHEET_PATH path;
KIID id( wxString::FromUTF8( rid.get<std::string>().c_str() ) );
touched.push_back( rid.get<std::string>() );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
if( SCH_ITEM* item = sch.ResolveItem( id, &path, /*allowNull*/ true ) )
{
commit.Remove( item, path.LastScreen() );
if( item->Type() != SCH_FIELD_T )
removedItems.push_back( item );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
staged = true;
}
}
auto upsert = [&]( const json& w )
{
std::string sexpr = w.value( "sexpr", "" );
if( sexpr.find_first_not_of( " \t\r\n" ) == std::string::npos )
return;
// Parse into a throwaway sheet exactly like clipboard paste does. The screen
// is heap-allocated and owned by the sheet (freed with it).
SCH_SHEET tempSheet;
SCH_SCREEN* tempScreen = new SCH_SCREEN( &sch );
tempSheet.SetScreen( tempScreen );
STRING_LINE_READER reader( sexpr, wxT( "collab-items" ) );
SCH_IO_KICAD_SEXPR plugin;
try
{
plugin.LoadContent( reader, &tempSheet );
}
catch( ... )
{
EM_ASM( { console.log( "[collab] eeschema applyItems: blob parse failed" ); } );
return;
}
// Resolve a symbol's LIB_SYMBOL: prefer the blob's own (lib_symbols …) cache
// (a clipboard-style blob carries it), else the live screen's — peers share
// the same document so the definition is normally already present.
auto findLib = [&]( SCH_SYMBOL* aSym ) -> LIB_SYMBOL*
{
wxString lookup = aSym->GetLibId().Format().wx_str();
if( !aSym->UseLibIdLookup() )
lookup = aSym->GetSchSymbolLibraryName();
auto& tlibs = tempScreen->GetLibSymbols();
auto ti = tlibs.find( lookup );
if( ti != tlibs.end() )
return new LIB_SYMBOL( *ti->second );
auto& libs = aFrame->GetScreen()->GetLibSymbols();
auto li = libs.find( lookup );
if( li != libs.end() )
return new LIB_SYMBOL( *li->second );
return nullptr;
};
std::vector<SCH_ITEM*> loaded;
for( SCH_ITEM* item : tempScreen->Items() )
loaded.push_back( item );
for( SCH_ITEM* item : loaded )
{
tempScreen->Remove( item ); // detach: tempSheet's dtor must not free it
SCH_SHEET_PATH path;
if( SCH_ITEM* existing = sch.ResolveItem( item->m_Uuid, &path, /*allowNull*/ true ) )
{
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
commit.Remove( existing, path.LastScreen() );
if( existing->Type() != SCH_FIELD_T )
removedItems.push_back( existing );
}
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
if( item->Type() == SCH_SYMBOL_T )
{
auto* sym = static_cast<SCH_SYMBOL*>( item );
if( LIB_SYMBOL* lib = findLib( sym ) )
sym->SetLibSymbol( lib );
}
item->SetParent( &sch );
touched.push_back( toUtf8( item->m_Uuid.AsString() ) );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
commit.Add( item, aFrame->GetScreen() );
staged = true;
}
};
for( const json& w : aWire.value( "added", json::array() ) )
upsert( w );
for( const json& w : aWire.value( "changed", json::array() ) )
upsert( w );
// SKIP_UNDO: remote applies never land on the local undo stack (see doApply).
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
if( staged )
commit.Push( wxT( "Collaborative edit (items)" ), SKIP_UNDO );
for( SCH_ITEM* item : removedItems )
delete item;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// Fold ONLY the applied uuids into the baseline (echo suppression), then flush:
// anything else that now differs — a concurrent local edit, the connectivity
// cleanup this apply's Push produced — broadcasts as a normal local diff
// instead of being swallowed (bug 05).
rebaselineTouched( aFrame, touched );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
s_applyingRemote = false;
scheduleFlush();
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
}
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
// Test/PoC move (the SCH_COMMIT body for kicadCollabTestMoveFirst, deferred via CallAfter).
void collabTestMove( SCH_EDIT_FRAME* aFrame, SCH_ITEM* aItem, SCH_SCREEN* aScreen, int aDx,
int aDy )
{
SCH_COMMIT commit( aFrame );
commit.Modify( aItem, aScreen );
aItem->Move( VECTOR2I( aDx, aDy ) );
commit.Push( wxT( "Collab test move" ) );
}
2026-06-03 17:49:11 +02:00
} // namespace
// JS → C++. Apply a remote per-item delta by uuid, through SCH_COMMIT so connectivity/
// ERC/hierarchy recompute the same way a UI edit would (0003 §apply).
//
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// SCH_COMMIT must run on the editor's main loop, not on this embind ccall or an
// emscripten_async_call/setTimeout callback: wxEvtHandler::CallAfter queues onto the
// app's pending-event list, which the wasm main loop drains every frame via
// ProcessPendingEvents() (src/wasm/evtloop.cpp) — i.e. the exact context real UI edits
// run in. So defer the whole mutation there.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
void schCollabApply( std::string aJson )
2026-06-03 17:49:11 +02:00
{
// Open-in-flight guard (open_gate.h): never touch the model while a
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// kicadOpenFile chain is suspended mid-load — commits/virtuals would walk
// a half-built schematic mid-mutation.
// Callers gate on kicadOpenFileBusy; fuzzed by tests/kicad/collab-load-fuzz.spec.ts.
if( pcbjam_open::busy() )
return;
2026-06-03 17:49:11 +02:00
json delta = json::parse( aJson, nullptr, /*allow_exceptions*/ false );
if( delta.is_discarded() )
return;
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return;
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// Defer to the editor's main loop + apply coroutine (runOnCoroutine) so SCH_COMMIT runs
// like a normal edit: a commit body that suspends (connectivity/GAL work) returns early
// from COROUTINE::Call, so applies must serialize with each other and with local edits
// or they interleave on shared commit/listener state. (0007, drift-trio #10.)
pcbjam_collab::runOnCoroutine( fr, [fr, delta]() { doApply( fr, delta ); } );
2026-06-03 17:49:11 +02:00
}
// JS pull of the full current model as an all-"added" delta (seed/baseline). Also
// registers the change listener on first call.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
std::string schCollabSnapshot()
2026-06-03 17:49:11 +02:00
{
if( pcbjam_open::busy() ) // open in flight (open_gate.h) — see schCollabApply
return json{ { "added", json::array() }, { "changed", json::array() },
{ "removed", json::array() } }.dump();
ensureBridge();
json added = snapshotItems( schFrame() );
2026-06-03 17:49:11 +02:00
// Seed the diff baseline to exactly the model we're handing out, so the first local edit
// diffs against this snapshot (and we don't re-broadcast the whole model).
rebaseline();
2026-06-03 17:49:11 +02:00
return json{ { "added", added }, { "changed", json::array() },
{ "removed", json::array() } }.dump();
}
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// JS → C++, v2 items wire. Same CallAfter + COROUTINE context as kicadCollabApply
// (LoadContent + SCH_COMMIT must run where native edits run).
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
void schCollabApplyItems( std::string aJson )
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
{
if( pcbjam_open::busy() ) // open in flight (open_gate.h) — see schCollabApply
return;
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
json wire = json::parse( aJson, nullptr, /*allow_exceptions*/ false );
if( wire.is_discarded() )
return;
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return;
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, wire]() { doApplyItems( fr, wire ); } );
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
}
// JS pull of the ACTIVE screen's model as an all-"added" v2 items wire: one clipboard-
// style blob per item on the current sheet (one collab room == one .kicad_sch screen).
// Registers the listener + rebaselines exactly like kicadCollabSnapshot.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
std::string schCollabSnapshotItems()
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
{
if( pcbjam_open::busy() ) // open in flight (open_gate.h) — see schCollabApply
return json{ { "added", json::array() }, { "changed", json::array() },
{ "removed", json::array() } }.dump();
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
SCH_EDIT_FRAME* fr = schFrame();
json added = json::array();
if( fr )
{
ensureBridge();
if( SCH_SCREEN* screen = currentScreen( fr ) )
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
{
for( SCH_ITEM* item : screen->Items() )
collab: paste-collision sync-delete fix — lenient sibling restage, forced removals for deleted dirty roots, eeschema child-blob lifting Field bug (2026-08-31): copy-pasting a symbol whose pins kept their source uuids corrupted the sheet's ydoc on delete (pcbjam-shared: cross-parent re-key fix, bumped here); the pcbnew tab's sibling mirror then silently froze on the dangling refs, so "Update PCB from Schematic" with delete enabled never removed the deleted symbols' footprints — until any later edit resynced the sheet. - standalone sibling-restage: render tolerantly past dangling item refs (docToFile onMissingItem) and console.warn on drops/failures instead of only the gated debug log — a frozen mirror is no longer silent. - pcbnew_embind flushDiff: dirty roots that are already STRUCT_DELETED emit forced removals on both wires instead of being P-5-skipped (observed: 8 footprints deleted on the board stayed in the board room forever); they are erased from the next baseline so a redo re-adds them. - eeschema_embind blobFor: lift an unlifted child (field/pin/sheet-pin) to its screen root before serializing — the selection writer emits nothing for such a child standalone, so the entry used to reach JS as an empty envelope and be skipped, silently dropping the edit; residual empty blobs now warn instead of shipping hollow envelopes. SCH_MARKER_T (ERC artifacts, never file content) stays out of the snapshot and seed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MLPKYptXFrxHj5Gu7rhToz
2026-08-31 12:50:12 +02:00
{
// Same exclusions as snapshotByUuid/blobFor: markers are not
// document content, and an empty blob (unformattable type)
// must never seed a hollow envelope.
if( item->Type() == SCH_MARKER_T )
continue;
std::string sexpr = itemBlob( fr, item );
if( sexpr.empty() )
continue;
added.push_back( json{ { "sexpr", sexpr }, { "parent", nullptr } } );
}
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
}
rebaseline();
}
return json{ { "added", added }, { "changed", json::array() },
{ "removed", json::array() } }.dump();
}
2026-06-03 17:49:11 +02:00
// Test/PoC helper: move the first schematic item by (dx,dy) IU via a real SCH_COMMIT,
// firing the listener — a deterministic local edit for the two-tab demo / e2e.
// Returns the moved item's uuid.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
std::string schCollabTestMoveFirst( int aDx, int aDy )
2026-06-03 17:49:11 +02:00
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
SCHEMATIC& sch = fr->Schematic();
for( const SCH_SHEET_PATH& path : sch.Hierarchy() )
{
SCH_SCREEN* screen = const_cast<SCH_SHEET_PATH&>( path ).LastScreen();
if( !screen )
continue;
for( SCH_ITEM* item : screen->Items() )
{
fix(wasm): dynCall signature-mismatch fallback + eeschema collab wire converters Two things, both verified in the real web app (two-tab eeschema collab). 1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl. Programmatic editor edits trapped with 'indirect call signature mismatch': the asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a stale type for some table indices (post-asyncify+O2) even though the table entry is valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS where the trampoline fails. Fix: the shim now catches the 'signature mismatch' RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move). 2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp). doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are skipped (next batch). eeschema re-enabled in the web app collab gate. Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 19:37:15 +02:00
fr->CallAfter( [fr, item, screen, aDx, aDy]() { collabTestMove( fr, item, screen, aDx, aDy ); } );
2026-06-03 17:49:11 +02:00
return toUtf8( item->m_Uuid.AsString() );
}
}
return "";
}
// Outcome of a deferred update-from-library run → window event
// `pcbjam:lib-update-done` {ok, updated, missing[]} (libs 0017 §2c).
static void emitLibUpdateDone( int aUpdated, const std::vector<std::string>& aMissing )
{
nlohmann::json j;
j["ok"] = true;
j["updated"] = aUpdated;
j["missing"] = aMissing;
std::string s = j.dump();
EM_ASM( {
if( typeof window !== 'undefined' )
window.dispatchEvent( new CustomEvent( 'pcbjam:lib-update-done',
{ detail: JSON.parse( UTF8ToString( $0 ) ) } ) );
}, s.c_str() );
}
// How many placed instances of a library symbol the open schematic holds —
// the JS lib-sync bridge asks after a remote lib update so the editor chrome
// can warn "a symbol you are using changed" (placed SCH_SYMBOLs keep their
// embedded copy across a lib reload, so the user must update explicitly).
// Counts across all unique screens of the hierarchy; 0 without a schematic
// frame (symbol editor / viewer sessions).
// Re-read the named library symbols into every placed instance — the headless
// core of Tools ▸ Update Symbols from Library… (DIALOG_CHANGE_SYMBOLS::
// processSymbols with the dialog's defaults: keep field text/positions, take
// the new body/pins/attributes), scoped to `aNamesJson` of `aLibNickname`.
// Runs as a normal SCH_COMMIT on the frame's coroutine so peers receive it as
// an ordinary edit. Returns {ok, updated, missing[]}. (libs 0017 §2c)
std::string schUpdateFromLibrary( std::string aLibNickname, std::string aNamesJson )
{
nlohmann::json out;
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
{
out["ok"] = false;
out["error"] = "no schematic frame";
return out.dump();
}
std::set<wxString> names;
try
{
for( const auto& n : nlohmann::json::parse( aNamesJson ) )
names.insert( wxString::FromUTF8( n.get<std::string>().c_str() ) );
}
catch( ... )
{
out["ok"] = false;
out["error"] = "bad names";
return out.dump();
}
const wxString lib = wxString::FromUTF8( aLibNickname.c_str() );
pcbjam_collab::runOnCoroutine( fr, [fr, lib, names]()
{
int updated = 0;
std::vector<std::string> missing;
SCH_COMMIT commit( fr );
SCH_SCREENS screens( fr->Schematic().Root() );
for( SCH_SCREEN* screen = screens.GetFirst(); screen; screen = screens.GetNext() )
{
std::vector<SCH_SYMBOL*> targets;
for( SCH_ITEM* item : screen->Items().OfType( SCH_SYMBOL_T ) )
{
SCH_SYMBOL* sym = static_cast<SCH_SYMBOL*>( item );
const LIB_ID& id = sym->GetLibId();
if( wxString( id.GetLibNickname() ) == lib && names.count( wxString( id.GetLibItemName() ) ) )
targets.push_back( sym );
}
for( SCH_SYMBOL* sym : targets )
{
LIB_SYMBOL* libSymbol = fr->GetLibSymbol( sym->GetLibId() );
if( !libSymbol )
{
missing.push_back( std::string( sym->GetLibId().Format().c_str() ) );
continue;
}
std::unique_ptr<LIB_SYMBOL> flattened = libSymbol->Flatten();
if( flattened->GetUnitCount() < sym->GetUnit() )
{
missing.push_back( std::string( sym->GetLibId().Format().c_str() ) );
continue;
}
// Same order as the dialog: remove, record, swap the lib symbol,
// re-append (the screen's RTree + connectivity re-index).
screen->Remove( sym );
commit.Modified( sym, static_cast<SCH_SYMBOL*>( sym->Clone() ), screen );
sym->SetLibSymbol( flattened.release() );
sym->SetExcludedFromSim( sym->GetLibSymbolRef()->GetExcludedFromSim() );
sym->SetExcludedFromBOM( sym->GetLibSymbolRef()->GetExcludedFromBOM() );
sym->SetExcludedFromBoard( sym->GetLibSymbolRef()->GetExcludedFromBoard() );
sym->SetShowPinNames( sym->GetLibSymbolRef()->GetShowPinNames() );
sym->SetShowPinNumbers( sym->GetLibSymbolRef()->GetShowPinNumbers() );
sym->SetSchSymbolLibraryName( wxEmptyString );
screen->Append( sym );
updated++;
}
}
commit.Push( wxT( "Update symbols from library" ) );
fr->GetCanvas()->Refresh();
// The Symbol Editor's own copy (libs 0019 F3): re-open it from the
// fresh library when unmodified; report and leave it with local edits.
if( SYMBOL_EDIT_FRAME* se = dynamic_cast<SYMBOL_EDIT_FRAME*>(
fr->Kiway().Player( FRAME_SCH_SYMBOL_EDITOR, false ) ) )
{
if( LIB_SYMBOL* cur = se->GetCurSymbol() )
{
const wxString curName = cur->GetName();
if( se->GetCurLib() == lib && names.count( curName ) )
{
const LIB_ID id( lib, curName );
if( se->IsContentModified() )
missing.push_back( std::string( id.Format().c_str() )
+ " (open in the Symbol Editor with unsaved edits — save or revert it first)" );
else
se->LoadSymbol( id, 1, 1 );
}
}
}
emitLibUpdateDone( updated, missing );
} );
// The body runs deferred on the frame's coroutine (runOnCoroutine =
// CallAfter): the caller awaits the `pcbjam:lib-update-done` window event
// for the outcome.
out["ok"] = true;
out["queued"] = true;
return out.dump();
}
// Standalone-eeschema shape of kicadUpdateFromLibrary(kind, lib, namesJson).
static std::string schUpdateFromLibraryShim( std::string aKind, std::string aLib, std::string aNames )
{
if( aKind != "symbol" )
return "{\"ok\":false,\"error\":\"kind not handled by this editor\"}";
return schUpdateFromLibrary( aLib, aNames );
}
int schLibsSymbolUsage( std::string aLibNickname, std::string aSymbolName )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return 0;
const LIB_ID target( wxString::FromUTF8( aLibNickname.c_str() ),
wxString::FromUTF8( aSymbolName.c_str() ) );
int count = 0;
SCH_SCREENS screens( fr->Schematic().Root() );
for( SCH_SCREEN* screen = screens.GetFirst(); screen; screen = screens.GetNext() )
{
for( SCH_ITEM* item : screen->Items().OfType( SCH_SYMBOL_T ) )
{
if( static_cast<SCH_SYMBOL*>( item )->GetLibId() == target )
count++;
}
}
// The Symbol Editor's open copy counts as "used" too (libs 0019 F3).
if( SYMBOL_EDIT_FRAME* se = dynamic_cast<SYMBOL_EDIT_FRAME*>(
fr->Kiway().Player( FRAME_SCH_SYMBOL_EDITOR, false ) ) )
{
if( LIB_SYMBOL* cur = se->GetCurSymbol() )
{
if( se->GetCurLib() == wxString( target.GetLibNickname() )
&& cur->GetName() == wxString( target.GetLibItemName() ) )
count++;
}
}
return count;
}
2026-06-03 17:49:11 +02:00
// Test helper: read an item's position by uuid as "x,y" (internal units).
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
std::string schCollabGetPos( std::string aId )
2026-06-03 17:49:11 +02:00
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
KIID id( wxString::FromUTF8( aId.c_str() ) );
if( SCH_ITEM* item = fr->Schematic().ResolveItem( id, nullptr, /*allowNull*/ true ) )
{
VECTOR2I p = item->GetPosition();
return std::to_string( p.x ) + "," + std::to_string( p.y );
}
return "";
}
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
// ── ysync-review repro hooks ─────────────────────────────────────────────────
// Local-edit test hooks for the ysync-review repro e2e (docs/features/
// ysync-review on the ysync-review branch): each drives a REAL SCH_COMMIT via
// CallAfter + COROUTINE (the doApply wrapping), so the SCHEMATIC_LISTENER →
// flushDiff emit path runs exactly as for a UI edit. Each returns false when
// the uuid doesn't resolve, letting the spec distinguish "hook missed the
// item" from "differ missed the edit" (bug 04).
// Delete an item by uuid via a real SCH_COMMIT.
bool schCollabTestRemoveItem( std::string aId )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !item )
return false;
SCH_SCREEN* screen = path.LastScreen();
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, item, screen]() {
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
SCH_COMMIT commit( fr );
commit.Remove( item, screen );
commit.Push( wxT( "Collab test remove" ) );
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
} );
return true;
}
// Rotate an item in place (aDeg snapped to 90° CCW steps) — bug 04: a symbol's
// GetPosition() is unchanged by an in-place rotation and its json carries no
// orientation, so the rotation is invisible to the scalar differ.
bool schCollabTestRotateItem( std::string aId, double aDeg )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !item )
return false;
SCH_SCREEN* screen = path.LastScreen();
int steps = ( (int) ( aDeg / 90.0 + ( aDeg >= 0 ? 0.5 : -0.5 ) ) % 4 + 4 ) % 4;
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, item, screen, steps]() {
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
SCH_COMMIT commit( fr );
commit.Modify( item, screen );
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
for( int i = 0; i < steps; ++i )
item->Rotate( item->GetPosition(), /*aRotateCCW*/ true );
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
commit.Push( wxT( "Collab test rotate" ) );
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
} );
return true;
}
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// Run Edit>Undo / read the undo depth — miss 09; frame-generic, collab_common.h.
bool schCollabTestUndo()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return pcbjam_collab::testUndo( schFrame() );
}
int schCollabTestUndoDepth()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return pcbjam_collab::testUndoDepth( schFrame() );
}
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
// Set a symbol's Value field text — bug 04: fields live inside the symbol (not
// in screen->Items()) and the symbol json carries no field text, so the most
// common schematic edit after moving things never syncs.
bool schCollabTestSetFieldText( std::string aId, std::string aText )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !item || item->Type() != SCH_SYMBOL_T )
return false;
SCH_SYMBOL* sym = static_cast<SCH_SYMBOL*>( item );
SCH_SCREEN* screen = path.LastScreen();
wxString text = wxString::FromUTF8( aText.c_str() );
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, sym, screen, text]() {
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
SCH_COMMIT commit( fr );
commit.Modify( sym, screen );
sym->SetValueFieldText( text );
commit.Push( wxT( "Collab test field text" ) );
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
} );
return true;
}
// ── drift-trio phase B action hooks (standalone-hardening 0008 §5) ───────────
// Creation/mutation primitives for the trio harness's action catalog. Each
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
// drives a REAL SCH_COMMIT on the apply coroutine, so the SCHEMATIC_LISTENER →
// flushDiff emit path runs exactly as for a UI edit. Names are tool-unique
// (registered outside the KICAD_MERGED_EMBIND guard — same convention as
// kicadCollabTestSetFieldText), so the merged image needs no dispatcher.
// Commit a freshly-built item onto the current screen; returns its uuid.
static std::string schCollabTestCommitAdd( SCH_ITEM* aItem, const wxChar* aMsg )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
{
delete aItem;
return "";
}
SCH_SCREEN* screen = fr->GetScreen();
// SCH_COMMIT::Push finds the SCHEMATIC via the ITEM's parent chain and
// silently skips every listener notification (OnItemsAdded → our emit)
// when it comes up null — a bare `new SCH_ITEM` has no parent yet, so
// parent it like the drawing tools do before staging.
aItem->SetParent( screen );
std::string id = toUtf8( aItem->m_Uuid.AsString() );
wxString msg( aMsg );
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, aItem, screen, msg]() {
SCH_COMMIT commit( fr );
commit.Add( aItem, screen );
commit.Push( msg );
} );
return id;
}
std::string schCollabTestAddWire( int aX1, int aY1, int aX2, int aY2 )
{
SCH_LINE* line = new SCH_LINE( VECTOR2I( aX1, aY1 ), LAYER_WIRE );
line->SetEndPoint( VECTOR2I( aX2, aY2 ) );
return schCollabTestCommitAdd( line, wxT( "Collab test add wire" ) );
}
std::string schCollabTestAddJunction( int aX, int aY )
{
return schCollabTestCommitAdd( new SCH_JUNCTION( VECTOR2I( aX, aY ) ),
wxT( "Collab test add junction" ) );
}
std::string schCollabTestAddNoConnect( int aX, int aY )
{
return schCollabTestCommitAdd( new SCH_NO_CONNECT( VECTOR2I( aX, aY ) ),
wxT( "Collab test add no-connect" ) );
}
// aKind: "label" | "global" | "hier".
std::string schCollabTestAddLabel( std::string aKind, std::string aText, int aX, int aY )
{
wxString txt = wxString::FromUTF8( aText.c_str() );
VECTOR2I pos( aX, aY );
SCH_LABEL_BASE* label;
if( aKind == "global" )
label = new SCH_GLOBALLABEL( pos, txt );
else if( aKind == "hier" )
label = new SCH_HIERLABEL( pos, txt );
else
label = new SCH_LABEL( pos, txt );
return schCollabTestCommitAdd( label, wxT( "Collab test add label" ) );
}
// Place another instance of a symbol ALREADY in the schematic's lib_symbols
// cache (the harness fixtures embed their defs — no external lib needed).
std::string schCollabTestAddSymbol( std::string aLibId, int aX, int aY, std::string aRef )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
SCH_SCREEN* screen = fr->GetScreen();
wxString libIdStr = wxString::FromUTF8( aLibId.c_str() );
LIB_SYMBOL* libSym = nullptr;
for( const auto& [name, sym] : screen->GetLibSymbols() )
{
if( name == libIdStr )
{
libSym = sym;
break;
}
}
if( !libSym )
return "";
LIB_ID libId;
if( libId.Parse( libIdStr ) >= 0 )
return "";
SCH_SHEET_PATH& sheet = fr->GetCurrentSheet();
SCH_SYMBOL* sym = new SCH_SYMBOL( *libSym, libId, &sheet, 1, 1, VECTOR2I( aX, aY ) );
sym->SetParent( screen ); // Push's SCHEMATIC lookup — see schCollabTestCommitAdd
sym->SetRef( &sheet, wxString::FromUTF8( aRef.c_str() ) );
std::string id = toUtf8( sym->m_Uuid.AsString() );
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, sym, screen]() {
SCH_COMMIT commit( fr );
commit.Add( sym, screen );
commit.Push( wxT( "Collab test add symbol" ) );
} );
return id;
}
// By-uuid variant of MoveFirst (same CallAfter + devirtualized move path).
bool schCollabTestMoveSchItem( std::string aId, int aDx, int aDy )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SHEET_PATH probe;
if( !fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ), &probe,
/*allowNull*/ true ) )
return false;
// Re-resolve when the deferred body runs: a remote remove can apply in
// between and the captured pointer would be dangling — the commit would
// resurrect a deleted item (drift-trio S4 move-vs-delete). Vanished =>
// the move loses, silently.
fr->CallAfter( [fr, aId, aDx, aDy]() {
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( item )
collabTestMove( fr, item, path.LastScreen(), aDx, aDy );
} );
return true;
}
bool schCollabTestMirrorSchItem( std::string aId, bool aHorizontal )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !item )
return false;
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, aId, aHorizontal]() { // re-resolve on the coroutine (S4)
SCH_SHEET_PATH path;
SCH_ITEM* live = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !live )
return;
SCH_COMMIT commit( fr );
commit.Modify( live, path.LastScreen() );
if( aHorizontal )
live->MirrorHorizontally( live->GetPosition().x );
else
live->MirrorVertically( live->GetPosition().y );
commit.Push( wxT( "Collab test mirror" ) );
} );
return true;
}
// Duplicate (fresh uuids — exercises the Clone()/uuid-churn drift class).
std::string schCollabTestDuplicateSchItem( std::string aId, int aDx, int aDy )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
SCH_SHEET_PATH path;
SCH_ITEM* item = fr->Schematic().ResolveItem( KIID( wxString::FromUTF8( aId.c_str() ) ),
&path, /*allowNull*/ true );
if( !item )
return "";
SCH_SCREEN* screen = path.LastScreen();
SCH_ITEM* dup = item->Duplicate( /*addToParentGroup*/ false );
dup->Move( VECTOR2I( aDx, aDy ) );
std::string id = toUtf8( dup->m_Uuid.AsString() );
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
pcbjam_collab::runOnCoroutine( fr, [fr, dup, screen]() {
SCH_COMMIT commit( fr );
commit.Add( dup, screen );
commit.Push( wxT( "Collab test duplicate" ) );
} );
return id;
}
// Programmatically save the in-memory schematic to a .kicad_sch file, without
// driving the Save As dialog — eeschema's analogue of pl_editor's
// kicadSaveDrawingSheet. Serializes the root sheet via the same SCH_IO_KICAD_SEXPR
// writer eeschema uses, so a test can read the file back from MEMFS and assert the
// file ⇄ Y.Doc round trip (README §A; feature 0004). Single-sheet scope: the round-
// trip fixtures are flat schematics; saving the root sheet writes the whole model.
// C++ → JS save notification (standalone-hardening save routing). Called from the
// kicad fork's save chokepoint (SCH_EDIT_FRAME::saveSchematicFile) after a
// successful write to MEMFS, so the web app can route the saved bytes onward
// (API upload, local-disk write-back, download). No-op without a JS listener.
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
// KICAD_MERGED_EMBIND: identical definition in pcbnew_embind.cpp; the merged image
// gets the one in kicad_editor_embind.cpp (both fork save chokepoints call it).
#ifndef KICAD_MERGED_EMBIND
extern "C" void kicadCollabOnSave( const char* aPath )
{
EM_ASM( {
if( window.kicadCollab && window.kicadCollab.onSave )
staging CI green-up: P-1 apply-slot wedge under JSPI, popup P-4 regression, ysync double-seed header, spec fixes - P-1 (findings-p.spec P-1c red on CI + locally): a JS exception thrown from a window.kicadCollab wire callback does NOT reach the C++ catch(...) — under JSPI it REJECTS the promising coroutine entry, slot.done never flips and the apply slot is busy for the page's life. Three layers: libcontext flags the rejected record dead (kicad → d0afbc4bd3), drainApplies() reaps a !done && !CanResume() body, and every EM_ASM window.kicadCollab.onX(...) call is try/catch-guarded (collab_common.h, kicad_editor/eeschema/pl_editor embind). - popup.spec 'dismisses on outside click': wx P-4 SetFocus delegation excluded for wxPopupWindow (wxwidgets → a7436d1052). - ysync-two-tab concurrent seed: repairLayoutY + seedNonce retraction deleted both headers (pcbjam-shared → 8dabf2e). - dark-mode.spec: stale light-mode reference (pre wx fa46f1822c DOM re-projection) replaced with the CI chromium render. - items-wire-batch-loss.spec: precondition obsolete after P-5 — poisoned entry is now synthetic, plus a P-5 contract assertion (field blob == ''). - presence-ghost-peer.spec: skip early when no /parties/ socket (CI web e2e runs the BroadcastChannel provider); findings-p1.spec: polled click-select precondition. Local: kicad-chromium findings/presence/ysync/collab 59/59, popup 7/7, dark-mode 1/1, pcbjam-shared vitest 249/249. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LvvARNMxQrq17vogn3REMU
2026-08-28 22:25:21 +02:00
{
// A throwing listener must never unwind the wasm frame that called it: under
// JSPI that rejects the running coroutine's entry (findings P-1).
try { window.kicadCollab.onSave( UTF8ToString( $0 ) ); }
catch( e ) { console.error( '[pcbjam collab] onSave listener threw', e ); }
}
}, aPath );
}
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
#endif // !KICAD_MERGED_EMBIND
// ── presence entry points (collab-presence 0003 — eeschema port of the 0002 set) ────────────
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// Install the presence input hooks on the GAL canvas + the fork's soft-lock
// query (idempotent) — shared core. The canvas is the same window across
// sheet navigation, so one install serves the whole session.
void schCollabPresenceStart()
{
// Presence needs the document listener too (findings Y-1/Y-3: doc changes
// repaint peers' shapes + re-check the local selection) — collab normally
// registers it first via snapshot/apply, but presence must not depend on
// that ordering (test harnesses, presence-only sessions).
ensureBridge();
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().start();
}
// JS → C++: full remote-peers snapshot (same wire as pcbnew's kicadCollabSetRemote).
// Rooms are per-sheet, so the JS rebind pushes a fresh (or empty) snapshot on every
// sheet switch — peers here are always this sheet's.
void schCollabSetRemote( std::string aJson )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().setRemote( aJson );
}
// JS → C++: cursor-only update for the peers of the last snapshot (findings Y-4).
void schCollabSetRemoteCursors( std::string aJson )
{
presenceCore().setRemoteCursors( aJson );
}
// JS → C++ (collab-presence 0005): comment pin dots (same wire as pcbnew).
void schCollabSetPins( std::string aJson )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().setPins( aJson );
}
// JS → C++ (presence tuner): live-patch the overlay STYLE and repaint —
// see collab_presence_style.h + pcbnew's counterpart.
void schCollabSetStyle( std::string aJson )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().setStyle( aJson );
}
// JS → C++ (comments-ux 0002 F4): live color-theme switch (see pcbjam_theme.h).
void schSetColorTheme( std::string aTheme )
{
pcbjam_theme::setColorTheme( schFrame(), aTheme );
}
// Pre-main chrome appearance seed (called at onRuntimeInitialized).
void schSetDarkChrome( bool aDark )
{
pcbjam_theme::setDarkChromeFlag( aDark );
}
// Tuner helper: a VARIED demo-selection set for the current sheet — smallest +
// largest symbol and two bundles of wires (net-ish), mirroring pcbnew's
// pcbCollabTestDemoSet so the style preview shows the range of shapes.
std::string schCollabTestDemoSet()
{
SCH_EDIT_FRAME* fr = schFrame();
json groups = json::array();
if( fr )
{
if( SCH_SCREEN* screen = currentScreen( fr ) )
{
SCH_SYMBOL* smallest = nullptr;
SCH_SYMBOL* largest = nullptr;
double minA = 0, maxA = 0;
std::vector<std::string> wires;
for( SCH_ITEM* item : screen->Items() )
{
if( item->Type() == SCH_SYMBOL_T )
{
BOX2I bb = item->GetBoundingBox();
double a = (double) bb.GetWidth() * bb.GetHeight();
auto* sym = static_cast<SCH_SYMBOL*>( item );
if( !smallest || a < minA ) { smallest = sym; minA = a; }
if( !largest || a > maxA ) { largest = sym; maxA = a; }
}
else if( item->Type() == SCH_LINE_T && wires.size() < 8 )
{
wires.push_back( toUtf8( item->m_Uuid.AsString() ) );
}
}
if( smallest )
groups.push_back( { { "label", "symbol small" },
{ "ids", { toUtf8( smallest->m_Uuid.AsString() ) } } } );
if( largest && largest != smallest )
groups.push_back( { { "label", "symbol large" },
{ "ids", { toUtf8( largest->m_Uuid.AsString() ) } } } );
if( wires.size() >= 2 )
{
size_t half = wires.size() / 2;
groups.push_back( { { "label", "wires A" },
{ "ids", std::vector<std::string>( wires.begin(),
wires.begin() + half ) } } );
groups.push_back( { { "label", "wires B" },
{ "ids", std::vector<std::string>( wires.begin() + half,
wires.end() ) } } );
}
}
}
return json{ { "groups", groups } }.dump();
}
// Test/tuner helper: the first N item uuids of the CURRENT sheet — real,
// resolvable KIIDs for synthetic remote-selection previews.
std::string schCollabTestListItems( int aCount )
{
SCH_EDIT_FRAME* fr = schFrame();
json out = json::array();
if( fr )
{
if( SCH_SCREEN* screen = currentScreen( fr ) )
{
for( SCH_ITEM* item : screen->Items() )
{
if( (int) out.size() >= aCount )
break;
out.push_back( toUtf8( item->m_Uuid.AsString() ) );
}
}
}
return out.dump();
}
// JS → C++ (0005): pan the view to a world position (comment panel "jump to pin").
void schCollabSetViewport( double aCx, double aCy )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().panTo( aCx, aCy );
}
feat(collab): follow-user (collab-presence 0008) + chip depth-layer fix Follow-user: click a peer's roster avatar to mirror their viewport until local input breaks it. - collab_presence_core.h: CORE::fitViewport(cx, cy, halfW, halfH) — fit the leader's world rect with CONTAIN semantics (zoom derived from the follower's own canvas via the ToScreen ratio; GetScale is the zoom, not px/IU). Exported as kicadCollabFitViewport from both editor TUs + the merged dispatcher. - presence-kicad.ts: publish the visible world rect (viewportRect) into awareness, 100 ms trailing throttle; guarded for pre-0008 handles. - follow-user.ts: createFollow — follows an awareness CLIENT (a tab, not a user); applies leader rect changes via FitViewport, dedupes unchanged republishes; break-on-interact compares local onViewport echoes against the last applied rect (2% rel tolerance, echo-grace before the first fit lands); unfollows on leader-left; pauses on eeschema sheet mismatch. - PresenceRoster: avatars are follow toggles (ring on the followed peer); WasmTool renders the "Following <name> — move to stop" banner. - tests: 7 controller units (85/85 collab), fitViewport round-trip e2e in both kicad presence specs (20/20), two-tab tests/web/follow.spec.ts (converge → track → wheel-zoom breaks → subsequent moves ignored). Chip depth-layer fix (user-reported): name chips washed out inside low-alpha selection fills — chip rects shared the shapes overlay's single depth, and same-depth fragments drawn LATER lose the depth test, so an earlier-painted fill rejected the chip's pixels. Now three layers via the fork's VIEW_OVERLAY::SetDepthOffset: text (0) < chips + pin dots (1) < selection shapes (2). drawLabel/drawCursor/drawSelectionBox take the chip overlay explicitly; comment-pin dots move to the chip layer too (the 0005 "drawn last so pins sit above" comment had the rule backwards). Verified with a chip-inside-30%-fill pixel repro + the full presence suite. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-09 09:30:40 +02:00
// JS → C++ (0008 follow-user): fit a leader's world rect into this canvas.
void schCollabFitViewport( double aCx, double aCy, double aHalfW, double aHalfH )
{
presenceCore().fitViewport( aCx, aCy, aHalfW, aHalfH );
}
// JS pull of the current viewport transform: `{cx,cy,scale,w,h}` with scale = px per
// IU via the GAL matrix (GetScale() is the zoom, not px/IU — pcbnew 0002 lesson).
std::string schCollabGetViewport()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return presenceCore().viewportJson();
}
// JS pull of the CURRENT selection's uuids (presence seed + e2e no-leak probe).
std::string schCollabGetSelection()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "[]";
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return presenceCore().selectionUuids( fr ).dump();
}
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
// JS pull of the current selection in the 0006 payload shape. eeschema has no
// footprint paths — the uuids ARE the symbol uuids — but the export keeps the
// merged image's kicadCollabGetSelectionFull contract uniform across editors.
std::string schCollabGetSelectionFull()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "{\"uuids\":[],\"fpPaths\":[]}";
json payload;
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
payload["uuids"] = presenceCore().selectionUuids( fr );
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
payload["fpPaths"] = json::array();
return payload.dump();
}
// Test probe (0006): the schematic-item uuids the current peers' cross-app
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// selections resolve to ON THE CURRENT SHEET — same resolveXsel the render uses.
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
std::string schCollabTestGetCrossMapped()
{
json arr = json::array();
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return arr.dump();
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
for( const pcbjam_presence::PEER& peer : presenceCore().peers )
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
for( SCH_ITEM* item : resolveXsel( fr, peer ) )
arr.push_back( toUtf8( item->m_Uuid.AsString() ) );
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
}
return arr.dump();
}
// Test helper: REALLY select the current sheet's first item through the selection
// tool, then run the presence check (programmatic selects close no canvas event).
std::string schCollabTestSelectFirst()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
SCH_SCREEN* screen = currentScreen( fr );
if( !screen )
return "";
SCH_ITEM* target = nullptr;
for( SCH_ITEM* item : screen->Items() )
{
target = item;
break;
}
if( !target )
return "";
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().selectItem( target );
return toUtf8( target->m_Uuid.AsString() );
}
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
// Test helper (0006): REALLY select the first SYMBOL on the current sheet —
// the deterministic cross-app subject (TestSelectFirst may pick a wire, which
// legitimately maps to nothing in pcbnew). Returns the uuid, "" without one.
std::string schCollabTestSelectComponent()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
SCH_SCREEN* screen = currentScreen( fr );
if( !screen )
return "";
SCH_ITEM* target = nullptr;
for( SCH_ITEM* item : screen->Items().OfType( SCH_SYMBOL_T ) )
{
// Power symbols (PWR_FLAG, GND, …) legitimately have no footprint —
// they'd make the cross-app subject map to nothing by construction.
if( static_cast<SCH_SYMBOL*>( item )->IsPower() )
continue;
target = item;
break;
}
if( !target )
return "";
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().selectItem( target );
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
return toUtf8( target->m_Uuid.AsString() );
}
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
// Test helper (0007): select a SPECIFIC item by uuid — see pcbnew_embind.cpp
// for why the tiebreak specs must not rely on cross-tab iteration order.
bool schCollabTestSelectByUuid( std::string aUuid )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
SCH_SCREEN* screen = currentScreen( fr );
if( !screen )
return false;
const wxString want = wxString::FromUTF8( aUuid.c_str() );
for( SCH_ITEM* item : screen->Items() )
{
if( item->m_Uuid.AsString() == want )
{
presenceCore().selectItem( item );
return true;
}
}
return false;
}
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
// JS → C++ (0007): tiebreak release — see pcbnew_embind.cpp for the design
// (cancel-interactive-if-a-tool-holds-them → selective unselect → infobar →
// forced re-emit).
void schCollabReleaseSelection( std::string aUuidsJson, std::string aHolder )
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().releaseSelection( aUuidsJson, aHolder );
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
}
// Test probe (0007): the current remote soft-lock set as `[{uuid, name}]`.
std::string schCollabTestGetLocked()
{
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
return presenceCore().locksJson();
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
}
// Test helper: clear the selection through the tool + run the presence check.
bool schCollabTestClearSelection()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
fr->CallAfter( [fr]() {
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
// ClearSelection is not on the shared SELECTION_TOOL base — the one
// presence hook that stays editor-typed.
if( SCH_SELECTION_TOOL* st = fr->GetToolManager()->GetTool<SCH_SELECTION_TOOL>() )
{
st->ClearSelection();
refactor(collab): dedup embind collab/presence layer into shared headers (collab_common.h + collab_presence_core.h) The eeschema/pcbnew binding TUs had ~1000 lines of copy-pasted collab code. Factored into two header-only shared files (zero build-script changes — the collab_presence_style.h precedent): - collab_common.h (pcbjam_collab): toUtf8, runOnFiber (the CallAfter+COROUTINE fiber idiom — was ~25 inline copies), the window.kicadCollab wire emitters (onDelta/onItems/onCursor/onSelection/onViewport), frame-generic undo test hooks. - collab_presence_core.h (pcbjam_presence::CORE): PEER/PIN + all presence state and machinery (start/canvas binds/lock query, setRemote/setPins/ setStyle, selection check + dedupe, overlay redraw loop, viewport push/pull, releaseSelection, locks probe), written against the EDA_DRAW_FRAME + SELECTION_TOOL base classes. Per-editor hooks: frame, selectionTool, selectionEmitPayload, resolveItem, drawPeerShapes. One CORE instance per TU (anonymous-namespace presenceCore()) so the merged image keeps per-editor state separation. - NEW per-editor resolveXsel(frame, peer): ONE cross-app resolver shared by the ghost render AND kicadCollabTestGetCrossMapped — the mapping loop was duplicated within each TU, letting the test probe drift from the pixels. Deliberately NOT factored: the Yjs differ/apply halves (itemToJson/makeItem/ flushDiff/doApply*) — structurally parallel but the bodies encode per-editor sync semantics and editor-specific asyncify devirtualization workarounds that must stay visible. kicadOpenFile/kicadCollabOnSave keep the existing KICAD_MERGED_EMBIND mechanism. TestClearSelection stays editor-typed (ClearSelection is not on the SELECTION_TOOL base). eeschema_embind 2203→1721 lines, pcbnew_embind 2518→1993. JS-facing names, signatures and the kicad_editor_embind.cpp dispatcher are unchanged. Verified: kicad_editor image builds clean; tests/kicad presence+locks 18/18 (incl. ghost-render pixel compares), collab+ysync-repros 31 passed/2 skipped. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-08 15:18:58 +02:00
presenceCore().scheduleSelCheck();
}
} );
return true;
}
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
// Merged-image dispatch probe (kicad_editor_embind.cpp): is the active top window the
// schematic editor? Counterpart of pcbnew_embind.cpp's pcbEditorActive().
bool schEditorActive()
{
return schFrame() != nullptr;
}
void kicadSaveSchematic( std::string path )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return;
SCHEMATIC& sch = fr->Schematic();
// Save the CURRENT sheet, not Schematic().Root(): in the wasm open flow the
// opened document is displayed as the current sheet but can sit under an
// auto-created project root, so Root()'s own screen holds only a child-sheet
// symbol (not the loaded items). GetCurrentSheet() is the screen the editor is
// actually showing — the one whose items the snapshot/round-trip care about.
SCH_SHEET* sheet = fr->GetCurrentSheet().Last();
if( !sheet )
sheet = &sch.Root();
try
{
SCH_IO_KICAD_SEXPR io;
io.SaveSchematicFile( wxString::FromUTF8( path.c_str() ), sheet, &sch );
}
catch( ... )
{
// Don't abort the wasm runtime on a save failure; the JS caller detects it
// by the file being absent / empty.
}
}
// Sheet navigator bridge (sheet-panel) — see sheetsStateJson. ensureBridge()
// so the OnSchSheetChanged push is wired even for sessions that never bind
// presence (read-only viewers).
std::string schSheetsGetTree()
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return "";
ensureBridge();
return sheetsStateJson( fr ).dump();
}
// Navigate to the sheet instance whose KIID path is aPath (as reported by
// kicadSheetsGetTree). Validated synchronously against the live hierarchy; the
// switch itself runs on the coroutine through the navigate tool's changeSheet
// action — the same route the wx hierarchy pane takes, so back/forward history
// stays consistent. DisplayCurrentSheet fires OnSchSheetChanged → the collab
// sheet-manager rebinds its room and the panel gets its onSheetsState push.
// The action is on the read-only allowlist (view-only), so viewers can navigate.
bool schSheetsEnter( const std::string& aPath )
{
SCH_EDIT_FRAME* fr = schFrame();
if( !fr )
return false;
wxString wanted = wxString::FromUTF8( aPath.c_str() );
SCH_SHEET_LIST list = fr->Schematic().Hierarchy();
auto it = std::find_if( list.begin(), list.end(),
[&]( const SCH_SHEET_PATH& p )
{
return p.PathAsString() == wanted;
} );
if( it == list.end() )
return false;
if( *it == fr->GetCurrentSheet() )
{
emitSheetsState( fr );
return true;
}
SCH_SHEET_PATH target = *it;
pcbjam_collab::runOnCoroutine( fr, [fr, target]() {
SCH_SHEET_PATH path = target;
fr->GetToolManager()->RunAction<SCH_SHEET_PATH*>( SCH_ACTIONS::changeSheet, &path );
// changeSheet already pushed via OnSchSheetChanged; a no-op path
// (already there / rejected) still leaves the panel consistent.
emitSheetsState( fr );
} );
return true;
}
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
static bool kicadCollabBusyProbe()
{
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
return pcbjam_collab::applyBusy() || !pcbjam_collab::applyQueue().empty();
}
EMSCRIPTEN_BINDINGS(eeschema) {
// Programmatic save of the in-memory schematic (round-trip tests, README §A).
function("kicadSaveSchematic", &kicadSaveSchematic);
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
// eeschema-only ysync-review repro hook (name not shared with pcbnew).
function("kicadCollabTestSetFieldText", &schCollabTestSetFieldText);
// drift-trio phase B action hooks (tool-unique names, merged-image safe).
function("kicadCollabTestAddWire", &schCollabTestAddWire);
function("kicadCollabTestAddJunction", &schCollabTestAddJunction);
function("kicadCollabTestAddNoConnect", &schCollabTestAddNoConnect);
function("kicadCollabTestAddLabel", &schCollabTestAddLabel);
function("kicadCollabTestAddSymbol", &schCollabTestAddSymbol);
function("kicadCollabTestMoveSchItem", &schCollabTestMoveSchItem);
function("kicadCollabTestMirrorSchItem", &schCollabTestMirrorSchItem);
function("kicadCollabTestDuplicateSchItem", &schCollabTestDuplicateSchItem);
// Sheet navigator bridge (sheet-panel): hierarchy tree + navigate.
function("kicadSheetsGetTree", &schSheetsGetTree);
function("kicadSheetsEnter", &schSheetsEnter);
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
#ifndef KICAD_MERGED_EMBIND
// JS names ALSO registered by pcbnew_embind.cpp — in the merged image these are
// registered once by kicad_editor_embind.cpp, dispatching on the active frame.
// Programmatic file open (preferred over UI automation from the web app).
jspi: migration phases 0-7 — build knob, scheduler shim, test successor suite Toolchain: emsdk 6.0.6 (versions.sh; cache-hash keys on it). Build knob PCBJAM_ASYNC_BACKEND=jspi|asyncify: build-kicad-target.sh links editors with -sJSPI + -sJSPI_EXPORTS=@scripts/common/jspi-exports.txt + --pre-js jspi-scheduler.js (no DYNCALLS, no post-link asyncify pipeline); wx build stamps the backend and forces clean on flip or unknown provenance; docker/build.sh passes the knob, seeds the emscripten ports cache from the volume every launch, jspi postprocess = patch-env-shim only. scripts/common/shims/jspi-scheduler.js: the JSPI successor scheduler — token-wait registry, resume turnstile (one armed resume between engine re-entries, SP swaps only at microtask boundaries), green-region spill stacks (16-aligned tops), S1 embind mutator FIFO lane + parker wraps, S6 shutdown, libctx integration hooks (suspend/end/quarantine + g_current arm/clear), SuspendError attributor, lost-wake + stuck-window watchdogs, __wxWaitDump observability. Embind: PARKER registrations get emscripten::async() under PCBJAM_JSPI (wasm/bindings/pcbjam_async_policy.h). nanosleep yields route via the shim. Tests: tests/asyncify -> tests/jspi successor suite (jspi-stack red/green shadow-stack battery, jspi-coroutine MiniCoro harness, suspend-races semantic scenarios + __wxWaitDump books coherence); projects jspi-firefox/ jspi-chrome (asyncify-webkit retired — no JSPI in WebKit); unconditional Firefox JSPI pref; guard-beacons -> wait-beacons (+wxScheduler/libctxJspi families); Makefile.wasm links test apps against JSPI with the shim as a tracked link prerequisite. Web: WasmTool setRo await + __wxWaitDump forensics, open-flow contained promise, scheduler-shim.test.ts retargeted (8 green). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 07:06:24 +02:00
function("kicadOpenFile", &kicadOpenFile PCBJAM_PARKER_POLICY);
function("kicadOpenFileBusy", &kicadOpenFileBusy);
function("kicadTestSetOpenPark", &kicadTestSetOpenPark);
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose The runtime is JSPI-only; this removes everything that still pretended otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove the inventory; every deletion verified by grep closure + full gates. Broken-right-now fixes: - deploy-staging.yml passed the retired opt_level input — the workflow could not even start. Removed. - env.sh carried dead exports with a live -sASYNCIFY=1 inside (WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason. - docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone. Dead weight removed: - binaryen submodule (nothing builds or invokes it), wasm-opt-bench workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess scaffolding (existed to parallelize the deleted wasm-opt phase; the postprocess is a seconds-long node script and now runs inline), build-monitor's dead asyncify rows, sched-context orphan build output, dead .gitignore entries, the .jspi-assets spike dir (the two wf-result research JSONs moved to docs/features/async/migration-evidence/). - bindings: fiber_park.h + its 12 embind registrations (broken-if- called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route, main_stack_runner.h + 5 includes, the always-null context-sleep weak hook in nanosleep_yield.c. - shim: the backend field (installed-flag idempotency instead), noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the WasmTool fallback and string-dump normalize branch). - web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts (gerber-demo keeps it: it loads the deployed CDN release, which predates emscripten 6 — noted inline). Conditionals: all 'backend === jspi' checks reduced to scheduler- presence checks; races_quiescent re-keyed from Asyncify.state (vacuous) to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive, which is the probing activation's own window by definition). Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→ JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS, kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests), collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→ wasmTrapSignatures (lists byte-identical). Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused asserts re-keyed to live JSPI beacons; eeschema-load's failure message no longer sends the developer to a deleted script; wait-beacons' dead families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is unconstructible); the embind test.fail re-gated with the JSPI reason (plain embind invokers cannot suspend — verified still failing); lint-determinism now scans tests/jspi (166 files clean); eeschema-collab local-move gated to chromium (~50% flaky on FF even solo; pcbnew twin covers both engines). Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md describes the single-phase build; docs/features/async/README.md banner-marked historical and repointed at the NEW 23-jspi-runtime.md (current architecture: export census, turnstile, libcontext ownership + refusal contract, embind call shapes, the em-pthread service-wrapper trick, exception policy, known gaps). Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the quiescence-probe fix; the 3 other reds were verified contention flakes solo-green or the documented FF gate), web 76/0, jspi 18/18 both engines, vitest 295/295 + 17/17, all lints green, live-app census clean. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
function("kicadCollabBusy", &kicadCollabBusyProbe);
// Read-only viewer lock (read-only-viewer).
function("kicadSetReadOnly", &kicadSetReadOnly);
2026-06-03 17:49:11 +02:00
// Yjs collaborative bridge entry points (same contract as pl_editor).
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
function("kicadCollabApply", &schCollabApply);
function("kicadCollabSnapshot", &schCollabSnapshot);
feat: v2 per-item s-expr collab bridge in all three tools (ysync 0008 Stage C) Add kicadCollabSnapshotItems / kicadCollabApplyItems / window.kicadCollab.onItems to pl_editor, eeschema, pcbnew — per-item native-blob payloads ({added/changed:[{sexpr,parent}], removed:[uuid]}) alongside the untouched scalar wire (legacy collab specs stay green). - pl_editor: itemBlob per item; apply = SetPageLayout(append) + replace-by-uuid (pointer-snapshot safe); bare payloads get the kicad_wks envelope. Snapshot + apply + emit verified headless. - eeschema: clipboard Format per item (symbols carry their lib_symbols); apply mirrors the native paste — LoadContent into a throwaway sheet → detach → replace-by-uuid → symbol lib relink (blob's lib_symbols first, live screen's second) → SCH_COMMIT, in the CallAfter+COROUTINE context. Snapshot + apply verified headless (a "lost" lone junction turned out to be correct connection cleanup — test uses text). - pcbnew: blobForItem per ROOT item with child→footprint lifting in flushDiff; apply = makeFromBlob + commit replace-by-uuid on the fiber; bare non-footprint payloads get wrapInBoardEnvelope (live board layer table). Snapshot + footprint replace/add WITH children (the 0004 containment gap, closed) + removal verified headless. Track/via/zone/text blob-apply hits the documented asyncify-fragile envelope parse (reconfirmed empirically — a verbatim SaveSelection segment envelope dies silently in the commit) and stays on the legacy scalar apply; tracked in ysync 0008 status. - eeschema/pcbnew scheduleFlush now runs flushDiff inside a COROUTINE: the per-item Format in the v2 emit needs the fiber stack (0007 lesson). Their emit remains unverifiable headless (both legacy two-tab tests are test.skip: "open=false → SCH_COMMIT no-ops" / "harness can't PAINT") — verify in the real app at Stage D; pl_editor's emit IS verified. - tests/kicad/items-bridge.spec.ts: per-tool suite (snapshot uuids → local-edit emit (where drivable) → apply changed/added/removed via save-readback → no apply echo). 3/3 pass; roundtrip + collab suites unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-11 13:55:15 +02:00
// v2 items bridge: per-item s-expr payloads (ysync 0008).
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
function("kicadCollabApplyItems", &schCollabApplyItems);
function("kicadCollabSnapshotItems", &schCollabSnapshotItems);
function("kicadCollabTestMoveFirst", &schCollabTestMoveFirst);
function("kicadCollabGetPos", &schCollabGetPos);
test(ysync): repro tests for review bugs 01-07 + v2 items-wire e2e port (miss 11) The 2026-07-02 sync review (docs/features/ysync-review, ysync-review branch) found 7 bugs and that the two-tab e2e only exercised the DEAD legacy scalar wire. This lands plan doc 15 in full; results + empirical findings in doc 16. - tests/collab/browser-entry-v2.ts (+build.mjs): the PRODUCTION v2 stack bundled for e2e (connectKicadDoc + attachKicadCollab, kdoc_* keys), with in-page renderActiveDoc/singleSeedRender/driftReport helpers and yjs forced to ONE copy (the two web pnpm workspaces otherwise bundle two instanceof-incompatible instances). - tests/kicad/ysync-two-tab.spec.ts: pl_editor green baseline (A↔B edits, ITEM-level drift silence) + divergent-uuid adopt; bug-01 pcb/ee fresh-room repros (Chromium-only: two kicad_editor tabs exceed Firefox's per-process wasm budget); bug-06 concurrent-seed race; bug-03 Y-half. - tests/kicad/ysync-repros-{pcbnew,eeschema}.spec.ts: bugs 02/03/05 + the bug-04 matrix (anchor-centred fp rotation, pad resize, endpoint drag, symbol rotation, Value-field edit), each with green landed-preconditions; the "local move emits" controls double as headless-emit probes — GREEN on both tools, so every emit-dependent repro is a live test.fail. - wasm/bindings: 7 local-edit test hooks via real commits (CallAfter+COROUTINE) — TestRemoveItem/TestRotateItem (both tools, dispatched in the merged image), TestSetPadSize/TestMoveEndpoint (pcbnew), TestSetFieldText (eeschema). - web/standalone ysync-repros.test.ts: bug-01 units (C++-faithful fake gating emit on ensureBridge) + bug-07a/b (stale DOWN hook, real sheet-manager gap). - web/pcbjam-shared bump: bug-03/06 unit repros. Convention: every repro asserts the CORRECT behavior and is expected-fail (test.fail/it.fails) naming its bug doc; a fix flips it to "unexpected pass", forcing marker removal — the repro becomes the regression test. Every expected failure verified (JSON reporter) to fail at its documented assert. Suite state: 39 passed / 0 failed / 0 flaky / 5 skipped (2 firefox guards, 2 pre-existing legacy two-tab skips, 1 pre-existing roundtrip fixme). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DPfrVhfYgPPgtawjSssZfn
2026-07-03 12:43:35 +02:00
// ysync-review repro hooks shared with pcbnew (dispatched when merged).
function("kicadCollabTestRemoveItem", &schCollabTestRemoveItem);
function("kicadCollabTestRotateItem", &schCollabTestRotateItem);
function("kicadCollabTestUndo", &schCollabTestUndo);
function("kicadCollabTestUndoDepth", &schCollabTestUndoDepth);
// Presence (collab-presence 0003) — shared names with pcbnew's 0002 set.
function("kicadCollabPresenceStart", &schCollabPresenceStart);
function("kicadCollabSetRemote", &schCollabSetRemote);
function("kicadCollabSetRemoteCursors", &schCollabSetRemoteCursors);
function("kicadCollabSetPins", &schCollabSetPins);
function("kicadCollabSetViewport", &schCollabSetViewport);
feat(collab): follow-user (collab-presence 0008) + chip depth-layer fix Follow-user: click a peer's roster avatar to mirror their viewport until local input breaks it. - collab_presence_core.h: CORE::fitViewport(cx, cy, halfW, halfH) — fit the leader's world rect with CONTAIN semantics (zoom derived from the follower's own canvas via the ToScreen ratio; GetScale is the zoom, not px/IU). Exported as kicadCollabFitViewport from both editor TUs + the merged dispatcher. - presence-kicad.ts: publish the visible world rect (viewportRect) into awareness, 100 ms trailing throttle; guarded for pre-0008 handles. - follow-user.ts: createFollow — follows an awareness CLIENT (a tab, not a user); applies leader rect changes via FitViewport, dedupes unchanged republishes; break-on-interact compares local onViewport echoes against the last applied rect (2% rel tolerance, echo-grace before the first fit lands); unfollows on leader-left; pauses on eeschema sheet mismatch. - PresenceRoster: avatars are follow toggles (ring on the followed peer); WasmTool renders the "Following <name> — move to stop" banner. - tests: 7 controller units (85/85 collab), fitViewport round-trip e2e in both kicad presence specs (20/20), two-tab tests/web/follow.spec.ts (converge → track → wheel-zoom breaks → subsequent moves ignored). Chip depth-layer fix (user-reported): name chips washed out inside low-alpha selection fills — chip rects shared the shapes overlay's single depth, and same-depth fragments drawn LATER lose the depth test, so an earlier-painted fill rejected the chip's pixels. Now three layers via the fork's VIEW_OVERLAY::SetDepthOffset: text (0) < chips + pin dots (1) < selection shapes (2). drawLabel/drawCursor/drawSelectionBox take the chip overlay explicitly; comment-pin dots move to the chip layer too (the 0005 "drawn last so pins sit above" comment had the rule backwards). Verified with a chip-inside-30%-fill pixel repro + the full presence suite. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013u9h8fkQktH7KRECaHJmUG
2026-07-09 09:30:40 +02:00
// Follow-user (collab-presence 0008).
function("kicadCollabFitViewport", &schCollabFitViewport);
function("kicadCollabSetStyle", &schCollabSetStyle);
// Live color-theme switch (comments-ux 0002 F4).
function("kicadSetColorTheme", &schSetColorTheme);
function("kicadSetDarkChrome", &schSetDarkChrome);
function("kicadCollabTestListItems", &schCollabTestListItems);
function("kicadCollabTestDemoSet", &schCollabTestDemoSet);
function("kicadCollabGetViewport", &schCollabGetViewport);
function("kicadCollabGetSelection", &schCollabGetSelection);
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
// Cross-app selection (0006).
function("kicadCollabGetSelectionFull", &schCollabGetSelectionFull);
function("kicadCollabTestGetCrossMapped", &schCollabTestGetCrossMapped);
feat(collab): selection soft-locks — remote-selected items can't be dragged locally (collab-presence 0007) While a peer has an item selected, local users can still select it for inspection but move/drag/rotate/delete skip it with an infobar naming the holder (native locked-item UX; enforced via the fork's PCBJAM_REMOTE_LOCK query — kicad 81f9cd80fd, the epic's first fork-touching phase). Overlapping holds (both grabbed inside the awareness propagation window) tie-break deterministically: lowest (user.id, clientID) keeps the item, every losing client auto-releases it. - lock-tiebreak.ts: pure policy — beats(), remoteLocks() (union of ALL other clients' selections incl. own user's other tabs, minus own-held-and-winning uuids so the winner isn't blocked mid-release), contestedReleases() - presence.ts: clients() (per-client view, no user dedupe) + self(); FIX for a pre-existing flaky stack overflow — resolveCollision re-entered itself synchronously via its own patch's awareness 'change' and could ping-pong on stale same-user states (~1-in-3 unit runs); re-entrancy guard defers re-resolution to the next genuine delivery - presence-kicad.ts: locks ride the kicadCollabSetRemote snapshot (`locks:[{uuid,name}]`); losing overlaps call kicadCollabReleaseSelection - wasm bindings (both TUs + merged dispatch): g_locks map + fork query install; kicadCollabReleaseSelection (cancelInteractive only when a tool stack is live — bare ESC would clear the whole selection — then selective RemoveItemFromSel + infobar + forced re-emit); kicadCollabTestGetLocked - tests: lock-tiebreak unit suite; presence-locks e2e for both editors (real move veto — pcbnew click+M hotkey since its default left-drag is rubber-band select, eeschema real drag — each with an unlocked control); two-tab tests/web/locks.spec.ts (lock propagation + deterministic tiebreak release + unlock on clear, passing vs real partykit) Spec: docs/features/collab-presence/0007 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 20:35:26 +02:00
// Selection soft-locks (0007).
function("kicadCollabReleaseSelection", &schCollabReleaseSelection);
function("kicadCollabTestGetLocked", &schCollabTestGetLocked);
function("kicadCollabTestSelectFirst", &schCollabTestSelectFirst);
feat(collab): cross-app selection — eeschema symbol ⇄ pcbnew footprint ghost highlight (collab-presence 0006) Selecting a symbol in eeschema ghost-highlights the linked footprint(s) in every pcbnew tab of the project, and vice versa — across users AND one user's own two tabs. Native KIWAY cross-probe is inert in WASM (one frame per page); this rides the presence layer instead. - cross-app.ts: project-wide awareness-only room (presenceRoomId), publishes full PresenceState at selection rate (cursor always null); peers() = other- TOOL clients incl. own user's other tabs; window.__pcbjamCrossApp test handle - presence-kicad.ts: parseSelectionEmit (bare array | {uuids,fpPaths}), xselFromPeerState (pcbnew paths → symbol uuids; eeschema uuids verbatim), cross peers appended to the kicadCollabSetRemote snapshot as {id "<user>#x<client>", name "<user> · sch|pcb", xsel} - C++ (zero fork changes): pcbnew emits {uuids, fpPaths} (FOOTPRINT::GetPath) and ghost-renders xsel via path-tail suffix scan; eeschema resolves xsel via ResolveItem gated to the CURRENT sheet (xsel arrives project-wide, unlike room-scoped selections); ghostStyle = alphas × xselAlphaScale (0.55, tuner- patchable); new exports kicadCollabGetSelectionFull / TestGetCrossMapped / TestSelectComponent (skips power symbols — PWR_FLAG has no footprint) + merged-image dispatch - tests: presence suites extended (payload shape, ghost render pixel tests, 13/13) + new two-tab tests/web/cross-probe.spec.ts (passing vs real partykit); eeschema pixel compares now target the #glcanvas-* GAL panel (the whole-window #canvas compare flaked on the auto-dismissing version infobar — also fixes the long-known presence-eeschema restore flake); cross-app + presence-kicad vitest suites Spec: docs/features/collab-presence/0006 (closed repo). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Lg5jwWuhFH5dL8hEcBDuP2
2026-07-07 19:30:18 +02:00
function("kicadCollabTestSelectComponent", &schCollabTestSelectComponent);
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
function("kicadCollabTestSelectByUuid", &schCollabTestSelectByUuid);
function("kicadCollabTestClearSelection", &schCollabTestClearSelection);
// Library reload after a remote (synced) lib edit — r2-idb-sync realtime.
jspi: migration phases 0-7 — build knob, scheduler shim, test successor suite Toolchain: emsdk 6.0.6 (versions.sh; cache-hash keys on it). Build knob PCBJAM_ASYNC_BACKEND=jspi|asyncify: build-kicad-target.sh links editors with -sJSPI + -sJSPI_EXPORTS=@scripts/common/jspi-exports.txt + --pre-js jspi-scheduler.js (no DYNCALLS, no post-link asyncify pipeline); wx build stamps the backend and forces clean on flip or unknown provenance; docker/build.sh passes the knob, seeds the emscripten ports cache from the volume every launch, jspi postprocess = patch-env-shim only. scripts/common/shims/jspi-scheduler.js: the JSPI successor scheduler — token-wait registry, resume turnstile (one armed resume between engine re-entries, SP swaps only at microtask boundaries), green-region spill stacks (16-aligned tops), S1 embind mutator FIFO lane + parker wraps, S6 shutdown, libctx integration hooks (suspend/end/quarantine + g_current arm/clear), SuspendError attributor, lost-wake + stuck-window watchdogs, __wxWaitDump observability. Embind: PARKER registrations get emscripten::async() under PCBJAM_JSPI (wasm/bindings/pcbjam_async_policy.h). nanosleep yields route via the shim. Tests: tests/asyncify -> tests/jspi successor suite (jspi-stack red/green shadow-stack battery, jspi-coroutine MiniCoro harness, suspend-races semantic scenarios + __wxWaitDump books coherence); projects jspi-firefox/ jspi-chrome (asyncify-webkit retired — no JSPI in WebKit); unconditional Firefox JSPI pref; guard-beacons -> wait-beacons (+wxScheduler/libctxJspi families); Makefile.wasm links test apps against JSPI with the shim as a tracked link prerequisite. Web: WasmTool setRo await + __wxWaitDump forensics, open-flow contained promise, scheduler-shim.test.ts retargeted (8 green). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 07:06:24 +02:00
function("kicadLibsReload", &pcbjam_libs::reloadLibrary PCBJAM_PARKER_POLICY);
// Runtime lib-table row insert + load (a new team library appeared
// mid-session; the lib set is otherwise frozen at boot).
function("kicadLibsAddEntry", &pcbjam_libs::addLibraryEntry PCBJAM_PARKER_POLICY);
// Placed-instance count for a library symbol (drives the "a symbol you are
// using was updated" toast after a remote lib edit).
function("kicadLibsSymbolUsage", &schLibsSymbolUsage);
// Update placed symbols from the library (libs 0017 §2c).
function("kicadUpdateFromLibrary", &schUpdateFromLibraryShim);
feat(wasm): kicad_editor — merge the pcbnew+eeschema kifaces into ONE bundle (Part 2) All four editors (PCB / Footprint / Schematic / Symbol) are now runtime --frame choices of a single kicad_editor.wasm (178 MB at -O1 vs 147+82 separate; shared wx/common/boost linked once). One editor per page load, as before; frames pcb / fpedit / sch / symedit. - wasm/editor/: the merged executable target (single_top + both kiface library sets, whole-archive pcbcommon) + the safety-net focus-walk Kiface() dispatch TU. Gated by KICAD_WASM_MERGED_EDITOR (kicad submodule bump carries the fork side: per-engine Kiface/getter binding + ODR renames + dual-kiface launcher). - wasm/bindings/: per-editor collab entries renamed pcbCollab*/schCollab* (JS names unchanged); duplicate kicadOpenFile/kicadCollabOnSave + shared-name registrations guarded behind KICAD_MERGED_EMBIND; new kicad_editor_embind.cpp registers each shared JS name once, dispatching on the live frame. - Build: kicad_editor app (build wrapper, target case arms, 3-object embind compile with the ABI-critical flags, STUB_APP=pcbnew); docker/build.sh "all" = kicad_editor calculator pl_editor gerbview (pcbnew/eeschema stay as explicit debug apps); scripts/kicad/audit-merged-symbols.sh = repeatable ODR-collision audit (run on kicad bumps). - Frontend: Bundle type (bundle ≠ tool); TOOL_BUNDLE maps all four editors to kicad_editor; explicit --frame tokens for pcbnew (pcb) and eeschema (sch); publish list = the 4 real bundles. - Tests/CI: five harnesses load kicad_editor.js with explicit frame tokens; PCBNEW_FAMILY_SPECS renamed BIG_MODULE_SPECS + the 8 eeschema-family specs (they now boot the merged module — SpiderMonkey x86 CI OOM routing); frame-runtime spec covers all four frames from the one bundle. Validated so far: frame-runtime 4/4 (each frame boots with the right title, no aborts, no duplicate embind registration); 24-spec merged-module regression green; 3D raytracer renders. Known pre-existing failure: 3d-viewer title-bar drag deadlock, fixed on main by 7630c7e (2N+8 pthread pre-warm) — picked up by the follow-up rebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-02 14:48:11 +02:00
#endif // !KICAD_MERGED_EMBIND
}
#endif