Graham Markall

Now that [numba/llvmlite#1009](https://github.com/numba/llvmlite/pull/1009) (which essentially implements the same changes made here, just inside llvmlite) is merged and an RC has been produced, I've received various reports that this fixes issues...

Just to add, we've been using this implementation in Numba / llvmlite for a few months now: - It appears to solve the ARM relocation overflow error - We haven't...

I'd add that we've continued to use this with no issues reported in Numba / llvmlite.

If the flow of the program involves compiling more and more functions as time goes by, it will eventually run out of memory because jitted functions can never be fully...

Similar discussions: https://github.com/numba/numba/issues/9672 https://github.com/numba/numba/issues/570 https://github.com/numba/numba/issues/2290

Thanks for the feature request! This seems like a good feature to be able to use. We would like to implement it, but probably won't have the time to do...

Thanks for the report and analysis! Are you able to provide a PR with the changes that worked for you, or a diff of the changes? We don't have a...

In RAPIDS we have https://github.com/rapidsai/gha-tools/blob/main/tools/rapids-conda-retry to make network failures much less likely to turn into a fatal error - perhaps this is worth incorporating into Numba / llvmlite scripts?

@a-gn Thanks for sharing this. I'll reopen the issue. Do you have any thoughts / ideas about what the correct fix is?

Changing labels to keep this open. To provide appropriate expectations, I think it's going to be hard for a fix for this to come to the top of the priorities...