# Numba Meeting: 2024-12-03

Attendees:
FPOC (last week): Kaustubh
FPOC (incoming): Guilherme

NOTE: All communication is subject to the [Numba Code of Conduct](https://github.com/numba/numba-governance/blob/accepted/code-of-conduct.md).

Please refer to [this calendar](https://calendar.google.com/calendar/embed?src=5rqnddm4gjrdfjm31fsv182epk%40group.calendar.google.com) for the next meeting date.

## Discussion

- Click the spanner at the bottom of HackMD to turn off snow
- https://github.com/numba/llvmlite/pull/1097
- joint review with Siu & Stu
- azure->GHA blocked due to using third party action- `maxim-lobanov/setup-xcode@v1`
TODO:
- consult best practice from GH or Anaconda
- potentially put specific action into the allowed-list.
- Moving Numba to new pass manager: https://github.com/numba/numba/pull/9676
- Can proceed to review
- release status update
- RC1 artifacts uploaded
- conda-forge build pending review on llvmlite recipe before numba recipe can be reviewed
- https://github.com/conda-forge/llvmlite-feedstock/pull/95
- RC2 likely needed for harderror issues
- harderror issues (problems caused by switch to new_style error capturing)
- 0.60.0's using new_style as default is proof that the problem is not too problematic
- Should we add some "helper" function things to llvmlite to help with opaque pointer migration? Example: I've found a ptr_of_type(x) function which sorts out opaque vs typed ptr useful locally.
- Yes, following the 0.44 release
- Could add another llvmlite release prior to next Numba release to make this available if it's helpful.
- llvmlite 0.44.0 patches
- https://github.com/numba/llvmlite/pull/1098
- ~~thinking ahead to 0.62. (CUDA, CI changes)~~
- https://github.com/numba/llvmlite/pull/1103
- peer2peer comm required
- https://github.com/Olympus-HPC/proteus
- Stu to plan a talk from the proteus devs
- https://github.com/numba/numba/pull/9807 -- TypeError: 'mappingproxy' object does not support item assignment -- what to do about this find, could it lead to subtle bugs?
- suggestion

      def func(locals=MappingProxyType({})):
          locals = dict(locals)

## New "Ready for Review" PRs

## 1. New Issues
- [numba#9812](https://github.com/numba/numba/issues/9812) - numba 0.61.0rc1 fails on macOS 12
- [numba#9813](https://github.com/numba/numba/issues/9813) - numba v0.61.0rc1 installation fails because missing MS Visual C++
- [numba#9814](https://github.com/numba/numba/issues/9814) - Possible race condition in or around `snt_check()`
- [numba#9817](https://github.com/numba/numba/issues/9817) - debuginfo output doesn't seem to work with gdb v15
- [numba#9819](https://github.com/numba/numba/issues/9819) - Sum is one order of magnitude slower than numpy for non c-contiguous arrays
- [numba#9820](https://github.com/numba/numba/issues/9820) - Slightly ease level-one type checking rule
- [numba#9822](https://github.com/numba/numba/issues/9822) - `sys.tracebacklimit = 0` causes error for simple numba script
- [llvmlite#1105](https://github.com/numba/llvmlite/issues/1105) - Tight pins in conda package for 0.44.0rc1

### Closed Issues
- [numba#9815](https://github.com/numba/numba/issues/9815) - CUDA Simulator not supporting cuda.default_stream()

## 2. New PRs
- [numba#9816](https://github.com/numba/numba/pull/9816) - Use NumbaTypeError instead of TypeError in reshape, transpose, astype
- [numba#9818](https://github.com/numba/numba/pull/9818) - [DO NOT MERGE] Trigger for pycc tests
- [numba#9821](https://github.com/numba/numba/pull/9821) - Add debug print to debug test_monitoring_multiple_threads failing on …
- [llvmlite#1104](https://github.com/numba/llvmlite/pull/1104) - Target 0.44.0: Fix osx wheel build

### Closed PRs

(last numba: 9822; llvmlite 1105)

## Short-term Roadmap

2024-gantt: https://github.com/numba/numba/issues/9540