stuartarchibald

> Yes, building from master. > > Just found #502 and am currently working around with CXXFLAGS. Great, if that doesn't work please shout. Also, wheels are up for 0.30.0.

Thanks for the report. I think this will require making `compiler-rt` available to `llvmlite`, which also involves building it as part of the `llvmdev` package. IIRC this is on the...

Note: Issue title updated to reflect required changes/help with searching/reference.

Thanks for the idea/patches. This is something that's been a struggle for e.g. the AMD GPU target compilation chain before, AFAIAA there is no programmatic API for `lld` and this...

@certik thanks for the responses. 1. Replying to the points > * No runtime dependency on LLD (on my machine the LLD package is 16MB), but if it is binded...

@certik no problem. I think that if you would find such a tool useful and would use it then it is worth doing. I would anticipate that other users would...

This this happens: ``` (qiskit0.30) [root@rhel84 ~]# numba -s Segmentation fault (core dumped) ``` suggests that the DSOs are probably segfaulting on `dlopen`. IIRC the `numba -s` command doesn't attempt...

xref: https://github.com/conda-forge/llvmlite-feedstock/pull/62#issuecomment-1210408912 suggest discussing a point release to coincide with Numba 0.56.1 to patch this test.

Thanks for the patch @kc611, I've provided some initial feedback inline.

Thanks for the demonstration @luk-f-a. Current thoughts/issues. 1. If a function is implemented outside of the call/overload system, for example `len()` it will not appear in `calltypes` as it's directly...