Graham Markall
Graham Markall
Seems like the most recent fail was just a CI hiccup, so I've restarted that one.
The other failure is just a lack of changelog entry for the release - see https://numba.readthedocs.io/en/stable/developer/contributing.html#release-notes
Because of the old / new split and re-join, it was hard to merge `main`, so out of interest I rebased it in https://github.com/numba/numba/pull/10125 @ZzEeKkAa Feel free to use /...
It would be good to get this in one day - commenting to make it un-stale.
It's still not possible, and it would be a large effort to implement - for example, I note that the LLVM `LLParser.cpp` file is approximately 10,000 lines of handcrafted parser.
It's not clear that this issue is related to Numba. If you think it is related to Numba, could you please post a small sample of code that can be...
Thanks for the report and nice reproducer. I can reproduce the issue locally. @DrTodd13 Do you have any thoughts on this one?
Many thanks for the report / reproducer. I can also reproduce the issue with Numba `main` and Python 3.11.9.
Minimal reproducer: ```python import numba as nb import numpy as np @nb.njit(fastmath=True) def f(n, v): try: for i in range(n): return 1.0 / v except: return 1.0 / v f(1,...
Actually we don't need the for loop, fastmath, a division by zero, or any arguments - we can make this much simpler: ```python import numba as nb @nb.njit def f():...