AI_Assistant/pcbjam
Watch
1
0
Fork
You've already forked pcbjam
 0
Code Issues Pull requests Projects Releases 7 Packages Wiki Activity Actions

docs 21+22: name the tool-body park site that blocks Phase D

Browse source 
Traced, not inferred: TOOL_MANAGER::RunSynchronousAction spins
`while(synchronousControl == STS_RUNNING) { wxYield(); wxMilliSleep(1); }`
(tool_manager.cpp:370-371), and on wasm wxMilliSleep -> nanosleep ->
__wasm_main_thread_yield_ms is an ASYNCIFY PARK OF THE STACK IT STANDS ON —
in a loop, inside a tool body, with a nested wxYield() dispatch running on
top of the parked stack.

The case closes on its callers: they are exactly the tools whose specs died
at D-on — edit_tool_move_fct (move-with-m, presence-locks move),
sch_drawing_tools (draw wires), the drawing/edit tools behind draw-lines.
The pcbnew spec comments already described the symptom from outside ("the
asyncified pointer-move handler") without naming the park; this is it.

So doc 21's K7 "anywhere" class has ONE caller that blocks Phase D, and it
moves first: the wait must yield the owning context instead of sleeping in
place, with the atomic's transition marking it ready. Recorded with the fix
shape, the upstream-divergence question it raises, and the gate that matters
(the KiCad suite's four canvas-tool specs — the harness has no
RunSynchronousAction and stays green either way).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LBjomQfKyRa3jBdeAKpmTw
This commit is contained in:
Gergő Törcsvári 2026-08-07 19:28:39 +02:00
commit f6171a7d34
No known key found for this signature in database
GPG key ID: 8E75F2CDE64E5322
2 changed files with 51 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

2
docs/features/async/21-park-site-audit.md
View file

@ -40,7 +40,7 @@ the ngspice pair as 2); after D-1 no site has a legacy twin — one code path pe
| K4 | `js_occExportRequest` — STEP export | `wasm/stubs/exporter_step_stub.cpp:59` | **entry** (export dialog flow) | occ-service worker promise | **D4** |
| K5 | `js_occLoadModelRequest` — OCE model load | `wasm/stubs/oce_plugin_stub.cpp:64` | **entry or fiber** (3D cache load path) | occ-service worker promise | **D4** |
| K6 | `js_ngspice_request` / `js_ngspice_get_vec` | `wasm/stubs/sharedspice_client.cpp:60/82` | **entry** (sim frame actions; `get_vec` has the known pre-existing asyncify crash) | ngspice-service worker promise | **D4** |
| K7 | `__wasm_main_thread_yield_ms` — the nanosleep shadow | `wasm/shims/nanosleep_yield.c:32` | **whatever main-thread stack calls `sleep_for`** — the "anywhere" class (raytracer join loops today) | setTimeout | **D4**, plus a caller sweep: each `sleep_for` reached on the main thread is a park site of its own |
| K7 | `__wasm_main_thread_yield_ms` — the nanosleep shadow | `wasm/shims/nanosleep_yield.c:32` | **whatever main-thread stack calls `sleep_for`** — the "anywhere" class (raytracer join loops today), and **`TOOL_MANAGER::RunSynchronousAction`'s spin loop, `tool_manager.cpp:370-371` — an in-place park INSIDE a tool body, in a loop, with a nested `wxYield()` dispatch on top of it** (identified 2026-08-07; doc 22 §10 "The tool-body park site, NAMED") | setTimeout | **D4**, plus a caller sweep: each `sleep_for` reached on the main thread is a park site of its own. The `RunSynchronousAction` caller is the one that BLOCKS Phase D — it kills four canvas-tool specs at D-on with the blue screen — and it moves first |
| T1 | open-gate test park (`testParkMs` sleeps) | `wasm/bindings/{pcbnew,eeschema,pl_editor,kicad_editor}_embind.cpp` ×8 | **entry** (deliberate) | timeout | keep — stages the open-window collisions |
| T2 | timer park lever | `wasm/bindings/timer_park.h:73` | mailbox-delivered timer entry (deliberate) | timeout | keep |
| T3 | fiber park lever | `wasm/bindings/fiber_park.h:78` | **fiber mid-body (deliberate — stages exactly the doc-19 strand)** | timeout | keep; its refusal pin flips red→green at D6 |