test(allocator): skip failing test on 32-bit platforms#21291
Conversation
How to use the Graphite Merge QueueAdd either label to this PR to merge it via the merge queue:
You must have a Graphite account in order to use the merge queue. Sign up using this link. An organization admin has enabled the Graphite Merge Queue in this repository. Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue. This stack of pull requests is managed by Graphite. Learn more about stacking. |
github-actions
Bot
added
A-allocator
Merge activity
|
There was a problem hiding this comment.
Pull request overview
This PR makes 32-bit CI green by disabling a bump-allocator OOM/pointer-overflow panic test that is consistently failing on 32-bit platforms.
Changes:
- Gate
oom_instead_of_bump_pointer_overflowto only run on 64-bit targets (and not under Miri).
This test is consistently failing on 32-bit CI. e.g. https://github.com/oxc-project/oxc/actions/runs/24210394836/job/70677275438 I don't understand the logic of this test. At first glance it looks like the test is flawed, and it's only passing on 64-bit for incidental reasons - due to the memory regions that allocators tend to use. I'll investigate further, but in meantime just skip the test on 32-bit to make CI green again.
5a7d409 to
bdac624
Compare
Merging this PR will not alter performance
Comparing 
Footnotes
|
Fix a test for `Bump` which was failing on 32-bit in CI. #21291 temporarily disabled the test. Turn out, there's no easy way to make this test work on 32-bit platforms, and #21510 added ample unit tests which effectively test the same thing. So keep it disabled on 32-bit, but fix the test to make it work properly on 64-bit. Rationale for the fix is explained in fitzgen/bumpalo#320.
This test is consistently failing on 32-bit CI.
e.g. https://github.com/oxc-project/oxc/actions/runs/24210394836/job/70677275438
I don't understand the logic of this test. At first glance it looks like the test is flawed, and it's only passing on 64-bit for incidental reasons - due to the memory regions that allocators tend to use.
I'll investigate further, but in meantime just skip the test on 32-bit to make CI green again.