Skip to content

Quarantine TypeIntrinsics on WASM ReadyToRun - #135967

Closed
agocke with Copilot wants to merge 3 commits into
mainfrom
copilot/wasmr2r-fix-stale-dynamic-helper-frame
Closed

agocke with Copilot wants to merge 3 commits into
mainfrom
copilot/wasmr2r-fix-stale-dynamic-helper-frame

Conversation

Copilot AI commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor
  • Failure: On WASM ReadyToRun, the TypeIntrinsics test can hit a runtime assertion after catching a NullReferenceException.
  • Change: Conditionally skip TestEntryPoint in both _r and _ro via the shared ActiveIssue attribute. Non-R2R coverage remains enabled; the runtime defect is unchanged.

@azure-pipelines

Copy link
Copy Markdown
Azure Pipelines:
16 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

@dotnet-policy-service

Copy link
Copy Markdown
Contributor

Tagging subscribers to this area: @dotnet/runtime-infrastructure
See info in area-owners.md if you want to be subscribed.

Co-authored-by: agocke <515774+agocke@users.noreply.github.com>
@agocke agocke changed the title [WIP] Fix stale DynamicHelperFrame after caught NullReferenceException [WIP] Disable failing intrinsics test for WASM R2R Oct 16, 2026
Co-authored-by: agocke <515774+agocke@users.noreply.github.com>
Copilot AI changed the title [WIP] Disable failing intrinsics test for WASM R2R Quarantine TypeIntrinsics on WASM ReadyToRun Oct 16, 2026
Copilot AI requested a review from agocke October 16, 2026 19:03
@lewing

lewing commented Oct 16, 2026

Copy link
Copy Markdown
Member

Superseded by #135970, which fixes the root cause. On wasm R2R the JIT skipped prolog zeroing for GC locals stored after a throwing null check (e.g. Object.GetType), so a GC reported stale stack contents. If a quarantine is still needed before #135970 merges, this should reference #135966 instead of "Fixes", so the issue stays open.

Note

This comment was drafted with GitHub Copilot.

@agocke

agocke commented Oct 16, 2026

Copy link
Copy Markdown
Member

@lewing I'm fine with taking that PR right now, but in general we should avoid enabling tests only to mark them Known Build Error. If tests don't pass on the given platform yet they shouldn't be enabled and we shouldn't wait support for the test is implemented.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

Status: No status
Status: Done

Development

Successfully merging this pull request may close these issues.

[wasm/R2R] TypeIntrinsics_r: GC reports uninitialized frame slots during throw-helper calls (pMT->Validate assert)

3 participants