lukewagner opened issue #11582:
The following two component-model async tests fail on the current (Sept 28)
devrelease:The problem seems to be that the call to
waitable-set.waiton line 51 which is defined withoutcancellableset (on line 64) returns theTASK_CANCELLEDcode (= 6) whensubtask.cancelis called on line 144. Instead, I believe that, since the subtask isn't cancellable,subtask.cancelshould immediately return theBLOCKEDcode (= -1).The problem seems to be that
$consume_cb(defined on line 147) is called with$event_code=STREAM_WRITE(= 3) instead of the expectedSTREAM_READ(= 2) for thestream.readissued on line 139. Additionally, the$indexisn't$insr(= 2), which was passed tostream.read; it's3. Maybe there's a mixup between the readable and writable ends here?
lukewagner added the bug label to Issue #11582.
alexcrichton added the wasm-proposal:component-model-async label to Issue #11582.
alexcrichton assigned dicej to issue #11582.
dicej closed issue #11582:
The following two component-model async tests fail on the current (Sept 28)
devrelease:The problem seems to be that the call to
waitable-set.waiton line 51 which is defined withoutcancellableset (on line 64) returns theTASK_CANCELLEDcode (= 6) whensubtask.cancelis called on line 144. Instead, I believe that, since the subtask isn't cancellable,subtask.cancelshould immediately return theBLOCKEDcode (= -1).The problem seems to be that
$consume_cb(defined on line 147) is called with$event_code=STREAM_WRITE(= 3) instead of the expectedSTREAM_READ(= 2) for thestream.readissued on line 139. Additionally, the$indexisn't$insr(= 2), which was passed tostream.read; it's3. Maybe there's a mixup between the readable and writable ends here?
dicej commented on issue #11582:
The first failure is addressed by https://github.com/bytecodealliance/wasmtime/pull/11671 and the second one has apparently been fixed by some other change since this issue was opened :tada:
Last updated: Dec 06 2025 at 06:05 UTC