fix: drop SCOPED_TRACE that crashes the pixi Windows job - #33
dv-picknik wants to merge 3 commits into
Conversation
The test crashes with SEH 0xc0000005 in the pixi_conda_build (windows-latest) job on every PR since #30, while the cmake Windows job, which builds gtest from source through conan, passes. It is the only test in the suite that uses SCOPED_TRACE. Stream the child tag into the assertion instead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The push triggers name master, which here mirrors upstream, while PRs merge to main, the default branch. No workflow runs on merge to main, so the pixi Windows crash introduced by #30 never showed there. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Team Run ID: 📒 Files selected for processing (7)
Included review availability: This review used your included allowance. 7 included reviews remain after this review. Your included PR review attempts over the past 7 days set your current allowance at 8 reviews per hour. 📝 SummarySummary by CodeRabbit
WalkthroughSix GitHub Actions workflows now run on pushes to ChangesWorkflow Push Triggers
Reactive Test Failure Output
Priority: ⬇️ Low Merge Risk: ⚪ Minimal · up to The workflow and test changes match their stated intent; no merge-blocking behavior change is evident. Caution Pre-merge checks failedPlease resolve all errors before merging. Addressing warnings is optional.
❌ Failed checks (1 error)
✅ Passed checks (3 passed)
Full details: Human Review CheckExplanation The PR changes six GitHub Actions workflows from push triggers on
Comment |
Under pixi's Windows job, the two tests that use SCOPED_TRACE crash with 0xc0000005: this one and Reactive.MissingOrEmptyGenericChildIdIsRejected, which #33 changes the same way. Stream the case name into each assertion instead, and wrap the build in ASSERT_NO_THROW so an unexpected throw still names its case. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Workflow changes need a human reviewer to override CodeRabbit's pre-merge check, which would hold up the crash fix. The trigger change moves to its own PR. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
[written by AI]
pixi_conda_build (windows-latest)has failed on every PR since #30.Reactive.MissingOrEmptyGenericChildIdIsRejectedcrashes with SEH0xc0000005. It's the only test in the suite that usesSCOPED_TRACE, and the newSCOPED_TRACEtest in #32 crashed the same way. #32 drops it too. Thecmake Windowsjob passes the same tests. It builds gtest from source through conan, while pixi links conda-forge's prebuilt gtest DLL, so I think the prebuilt DLL is the trigger. I have no Windows machine to confirm the mechanism. This PR's own pixi Windows run passes, the first green one since #30.This PR streams the child tag into
EXPECT_THROWinstead of usingSCOPED_TRACE, so a failure still names the case. The push-trigger fix that was here moved to #34, because workflow changes trip CodeRabbit's Human Review Check.On Linux all 216 tests pass. PickNikRobotics/moveit_pro#21028 and #28 fixed an earlier Windows failure, a toolset mismatch that broke the build before any test ran.
🤖 Generated with Claude Code