pcbjam/wasm/bindings/pcbjam_libs_reload.h

177 lines
7.2 KiB
C
Raw Normal View History

/*
* JS editor library reload (r2-idb-sync realtime): drop a PCBJAM lib's
* in-plugin cache and re-sync the open editor's library tree after a remote
* (peer) edit landed in the page's SyncStack/IDB.
*
* The lib plugins (SCH_IO_PCBJAM_LIB / PCB_IO_PCBJAM_FP) cache parsed items for
* the plugin instance's lifetime and pin GetModifyHash()/GetLibraryTimestamp()
* to a constant, so nothing upstream ever re-reads the provider. The one full
* invalidation is LIBRARY_MANAGER::ReloadLibraryEntry it erases the LIB_DATA
* entry INCLUDING the plugin instance (and with it the fat-load cache), the
* exact pattern the remote-symbol import flows already use
* (eeschema/widgets/panel_remote_symbol.cpp). LoadLibraryEntry then refetches
* synchronously (the fat "bodies" crossing served from the now-fresh IDB, no
* network), and MAIL_RELOAD_LIB makes an open editor frame rebuild its tree.
*
* Header-only (the collab_common.h pattern); common-code includes only, so the
* merged kicad_editor TU (deliberately eeschema/pcbnew-header-free) can use it.
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
* Runs on the apply coroutine: LoadLibraryEntry suspends in the JS bridge, so
* the reload must serialize with the collab applies and local edits.
*/
#pragma once
#ifdef __EMSCRIPTEN__
#include <string>
#include <wx/app.h>
#include <wx/string.h>
#include <kiway.h>
#include <kiway_player.h>
#include <libraries/library_manager.h>
#include <libraries/library_table.h>
#include <mail_type.h>
#include <pgm_base.h>
#include "collab_common.h"
namespace pcbjam_libs {
/**
* Reload one library from its provider and refresh any open editor tree.
* @param aKind "symbol" | "footprint" which lib table the nickname lives in.
* @param aNickname the lib-table row name (LibInfo.name on the JS side).
*/
inline void reloadLibrary( std::string aKind, std::string aNickname )
{
KIWAY_PLAYER* top =
wxTheApp ? dynamic_cast<KIWAY_PLAYER*>( wxTheApp->GetTopWindow() ) : nullptr;
if( !top )
return;
const bool fp = aKind == "footprint";
const wxString nick = wxString::FromUTF8( aNickname.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( top, [top, fp, nick]()
{
LIBRARY_MANAGER& mgr = Pgm().GetLibraryManager();
const LIBRARY_TABLE_TYPE type =
fp ? LIBRARY_TABLE_TYPE::FOOTPRINT : LIBRARY_TABLE_TYPE::SYMBOL;
// Drop the LIB_DATA entry (plugin instance + its fat-load cache)…
mgr.ReloadLibraryEntry( type, nick );
// …and bring the lib back to LOADED synchronously, so the tree sync
// below reads fresh items instead of re-listing a LOADING stub.
mgr.LoadLibraryEntry( type, nick );
// An open editor frame re-syncs its tree from the (now fresh) adapter.
// The payload names the lib to FORCE-refresh: the symbol tree gates
// re-enumeration on a modify hash that is a pinned constant for PCBJAM
// plugins (and its mtime component never moves for a remote edit), so a
// plain sync would skip the lib; the named node is rebuilt instead.
// ExpressMail only delivers to frames that exist — no frame is created.
std::string payload( nick.utf8_str() );
top->Kiway().ExpressMail( fp ? FRAME_FOOTPRINT_EDITOR : FRAME_SCH_SYMBOL_EDITOR,
MAIL_RELOAD_LIB, payload );
} );
}
/**
* Cheap counterpart of reloadLibrary (libs 0019): drop the lib's LIB_DATA
* entry (plugin instance + parsed cache) so the NEXT access re-reads the
* provider no eager LoadLibraryEntry fat-load, no tree mail. A remote edit
* calls this; the fat-load happens lazily when the user looks at the lib,
* or explicitly through reloadLibrary from "Update from library".
* (pcbnew's FOOTPRINT_LIBRARY_ADAPTER::PreloadedFootprints is invalidated by
* the pcbnew-side caller this header stays common-code only.)
*/
inline void invalidateLibrary( std::string aKind, std::string aNickname )
{
KIWAY_PLAYER* top =
wxTheApp ? dynamic_cast<KIWAY_PLAYER*>( wxTheApp->GetTopWindow() ) : nullptr;
if( !top )
return;
const bool fp = aKind == "footprint";
const wxString nick = wxString::FromUTF8( aNickname.c_str() );
pcbjam_collab::runOnCoroutine( top, [fp, nick]()
{
Pgm().GetLibraryManager().ReloadLibraryEntry(
fp ? LIBRARY_TABLE_TYPE::FOOTPRINT : LIBRARY_TABLE_TYPE::SYMBOL, nick );
} );
}
/**
* Add one PCBJAM lib-table row at RUNTIME and load it the lib SET is
* otherwise frozen at boot (sym/fp-lib-table are written once in preRun).
* Used by the "a new team library appeared" flow: the JS side has already
* created the `/mnt/pcbjam/<id>` MEMFS placeholder; this inserts the row into
* the LIVE in-memory global table (the table FILE is only re-read by
* LoadGlobalTables, which would drop every adapter cache deliberately not
* used), loads the entry through the provider bridge, and re-syncs any open
* editor tree. Idempotent: an existing nickname is a no-op.
*
* @param aKind "symbol" | "footprint" which lib table gets the row.
* @param aNickname the lib-table row name (LibInfo.name on the JS side).
* @param aUri the virtual mount uri ("/mnt/pcbjam/<libId>").
*/
inline void addLibraryEntry( std::string aKind, std::string aNickname, std::string aUri )
{
KIWAY_PLAYER* top =
wxTheApp ? dynamic_cast<KIWAY_PLAYER*>( wxTheApp->GetTopWindow() ) : nullptr;
if( !top )
return;
const bool fp = aKind == "footprint";
const wxString nick = wxString::FromUTF8( aNickname.c_str() );
const wxString uri = wxString::FromUTF8( aUri.c_str() );
pcbjam_collab::runOnCoroutine( top, [top, fp, nick, uri]()
{
LIBRARY_MANAGER& mgr = Pgm().GetLibraryManager();
const LIBRARY_TABLE_TYPE type =
fp ? LIBRARY_TABLE_TYPE::FOOTPRINT : LIBRARY_TABLE_TYPE::SYMBOL;
if( std::optional<LIBRARY_MANAGER_ADAPTER*> adapter = mgr.Adapter( type ); adapter )
{
LIBRARY_TABLE* table = ( *adapter )->GlobalTable();
if( !table )
return;
// Idempotent-cheap: a nickname that already has a row is done —
// content refresh is kicadLibsReload's job, and doing it here
// would let a re-announce trigger a pointless re-fat-load.
if( table->HasRow( nick ) )
return;
LIBRARY_TABLE_ROW& row = table->InsertRow();
row.SetNickname( nick );
row.SetURI( uri );
row.SetType( fp ? wxS( "PCBJAM_FP" ) : wxS( "PCBJAM" ) );
}
else
{
return;
}
// Bring the new row to LOADED through the JS provider bridge (GetRow
// walks the table on a cache miss, so the inserted row is found).
mgr.LoadLibraryEntry( type, nick );
// Open editor frames re-sync their tree; the nickname payload forces
// the named node's rebuild past the pinned-modify-hash gate (same
// reasoning as reloadLibrary above).
std::string payload( nick.utf8_str() );
top->Kiway().ExpressMail( fp ? FRAME_FOOTPRINT_EDITOR : FRAME_SCH_SYMBOL_EDITOR,
MAIL_RELOAD_LIB, payload );
} );
}
} // namespace pcbjam_libs
#endif // __EMSCRIPTEN__