Stream: git-wasmtime

Topic: wasmtime / PR #13996 [stack-switching] Tests

Wasmtime GitHub notifications bot (Jul 27 2026 at 11:28):

dhil opened PR #13996 from dhil:continuation-tests to bytecodealliance:main:

This patch imports the testsuite from wasmfx/wasmfxtime for stack switching.

Depends on #11717.

Wasmtime GitHub notifications bot (Jul 27 2026 at 11:28):

dhil requested cfallin for a review on PR #13996.

Wasmtime GitHub notifications bot (Jul 27 2026 at 11:28):

dhil requested wasmtime-compiler-reviewers for a review on PR #13996.

Wasmtime GitHub notifications bot (Jul 27 2026 at 11:28):

dhil requested wasmtime-core-reviewers for a review on PR #13996.

Wasmtime GitHub notifications bot (Jul 27 2026 at 11:29):

dhil converted PR #13996 [stack-switching] Tests to a draft.

Wasmtime GitHub notifications bot (Jul 27 2026 at 15:35):

github-actions[bot] added the label wasmtime:api on PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 12:28):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:11):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:12):

dhil has marked PR #13996 as ready for review.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:13):

:memo: dhil submitted PR review.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:13):

:speech_balloon: dhil created PR review comment:

This is the interesting change introduced by this patch. It implements the WAST directive assert_suspension. I am not sure this is the best way to implement it.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:19):

:memo: cfallin submitted PR review.

Wasmtime GitHub notifications bot (Jul 28 2026 at 15:19):

:speech_balloon: cfallin created PR review comment:

Rather than match on an error message, could we try to downcast the trap (see e.g. the assert_exception case above)?

I see at least UnhandledTag; a quick skim didn't find the suspend-outside-cont but I guess you'd know what kind of trap that raises (and we can add a variant if it is not distinguished currently).

Wasmtime GitHub notifications bot (Jul 28 2026 at 16:06):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 16:38):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 16:40):

:memo: dhil submitted PR review.

Wasmtime GitHub notifications bot (Jul 28 2026 at 16:40):

:speech_balloon: dhil created PR review comment:

Fixed in https://github.com/bytecodealliance/wasmtime/pull/13996/commits/73b99fcffa49462d16724e2fdd98b8c4a049ba92. Thanks!

I realised that we map unhandled suspension and suspension outside a continuation to the same trap UnhandledTag. I think the difference only manifests in the spec interpreter's testsuite, meaning I think we can freely ignore the distinction here.

Wasmtime GitHub notifications bot (Jul 28 2026 at 16:40):

:speech_balloon: dhil edited PR review comment.

Wasmtime GitHub notifications bot (Jul 28 2026 at 17:04):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 17:34):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 17:46):

dhil updated PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 17:46):

dhil edited PR #13996:

This patch imports the testsuite from wasmfx/wasmfxtime for stack switching.

~Depends on #11717~.

Wasmtime GitHub notifications bot (Jul 28 2026 at 20:49):

:thumbs_up: cfallin submitted PR review:

Looks good with the more precise matching -- thanks!

Wasmtime GitHub notifications bot (Jul 28 2026 at 20:49):

cfallin added PR #13996 [stack-switching] Tests to the merge queue.

Wasmtime GitHub notifications bot (Jul 28 2026 at 21:14):

:check: cfallin merged PR #13996.

Wasmtime GitHub notifications bot (Jul 28 2026 at 21:14):

cfallin removed PR #13996 [stack-switching] Tests from the merge queue.

Last updated: Jul 29 2026 at 05:03 UTC