Repository navigation

[BUG]: Errors occurring in `Buffer.close` are not raised #1118

Closed

Milestone

cuda.core backlog

opened

Is this a duplicate?

I confirmed there appear to be no duplicate issues for this bug and that I agree to the Code of Conduct

Type of Bug

Silent Failure

Component

cuda.core

Describe the bug

The description says it all. Errors appear to be suppressed by Cython "noexcept" decorations.

How to Reproduce

See #1117 and tests under memory_ipc/test_errors.py

Expected behavior

Errors in Buffer.close should be raised. Suppressions should occur during implicit calls from, e.g., __del__ or __dealloc__.

Operating System

No response

nvidia-smi output

No response

Andy-Jost
added
bugSomething isn't working
on Oct 9, 2025
Andy-Jost
changed the title ~~[-][BUG]: Error occurring in `Buffer.close` are not raised[/-]~~ [+][BUG]: Errors occurring in `Buffer.close` are not raised[/+] on Oct 9, 2025
rparolin
added the
Bug
issue type on Oct 13, 2025
rparolin
added this to the cuda-python 13-next, 12-next milestone on Oct 13, 2025
leofang
removed the
Bug
issue type on Oct 16, 2025
leofang
added
triageNeeds the team's attention
on Oct 16, 2025
leofang
modified the milestones: cuda-python 13.0.3 & 12.9.4, cuda.core parking lot on Oct 16, 2025
leofang
removed
bugSomething isn't working
on Oct 23, 2025
rparolin
removed
triageNeeds the team's attention
on Jan 19, 2026
Andy-Jost commented on Feb 23, 2026
Andy-Jost
on Feb 23, 2026
ContributorAuthor
Fixed with switch to RAII resource handles in cuda.core.
Andy-Jost
closed this as completed on Feb 23, 2026

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

Metadata

Assignees

No one assigned

Labels

No labels

No labels

Type

No type

Projects

No projects

Milestone

cuda.core backlog

Relationships

None yet

Development

No branches or pull requests