numba · Issues· 1791 open
Open on GitHubLocally synced open issues (discussions stay on GitHub)
- #10564
`TimingListener` may not be thread-safe
threadingbug - incorrect behaviorfree-threadingUpdated Sep 17, 2026 - #10368
isinstance in jit code yields an unsafe cast warning
bugeasyUpdated Sep 16, 2026 - #10297
Dependency Dashboard
TaskUpdated Sep 16, 2026 - #6559
RISC-V Support
feature_requestllvmISA: RISC-VUpdated Sep 16, 2026 - #6264
Caching: functions captured in closure
feature_requestcachingUpdated Sep 15, 2026 - #10836
Add `kind='heapsort'` to `ndarray.sort` and `argsort`
feature_requestperformanceperformance - run timeUpdated Sep 15, 2026 - #10522
Function overloads of some NumPy functions with scalar input are not exhaustive
buggood first issueUpdated Sep 14, 2026 - #10485
Disk cache crashes for exec()'d functions with `__module__=None`
good first issuebug - incorrect behaviorUpdated Sep 14, 2026 - #10581
Python 3.15
feature_requestPython 3.15Updated Sep 10, 2026 - #7179
Support jit decorator flags in cc.export (e.g. fastmath)
feature_requestgood first issueAOTUpdated Sep 10, 2026 - #10821
cache=True reuses artifacts across different NUMBA_SLP_VECTORIZE / NUMBA_OPT settings
cachingbug - incorrect behaviorUpdated Sep 9, 2026 - #10015
AssertionError: Unreachable condition reached (op code RERAISE executed) with batch_size > 1
bug - failure to compileUpdated Sep 8, 2026 - #10563
Test failure: `test_random_concurrent_mix_use_tbb`
bug - segfaultUpdated Sep 8, 2026 - #10753
Regression (0.66.0): two-argument max/min fail on CUDA after #10543 refactored them into *args overloads
bugCUDAUpdated Sep 8, 2026 - #6549
CUDA: Allow dynamic allocation of local arrays
feature_requestCUDAgood second issueUpdated Sep 6, 2026