Rohit Goswami

Could you open a different issue with a reproducer for the `--python=same` thing @lucascolley?

If the intent is to have `asv` not benchmark certain configurations you might consider the `SkipNotImplemented` exception as seen in this section on [Skipping Benchmarks](https://asv.readthedocs.io/en/stable/writing_benchmarks.html#skipping-benchmarks). Could you provide a minimum...

Following https://info.arxiv.org/help/submit_tex.html#minted-and-other-frozen-cache-packages correctly will work (make sure all the files under `_minted` are uploaded and not deleted).

Closing for now, it doesn't seem likely that such a function will make it into the NumPy API. A post to the mailing list would be a better starting point...

Also worth noting that as currently implemented, the version is sensitive to the formatting changes made to the suite, and can easily be replaced by an AST based hash mechanism...

Allow non-reproducible, `asv` oblivious way to track changes

FWIW this is likely the cause of: ```bash =================================== FAILURES =================================== _______________________________ test_run_publish _______________________________ capfd = basic_conf = ('/tmp/pytest-of-runner/pytest-0/test_run_publish0', '/home/runner/work/asv/asv/test', , '/tmp/pytest-of-runner/pytest-0/test_run_publish0/asv-machine.json') @pytest.mark.skipif( HAS_PYPY or (not HAS_VIRTUALENV) or WIN, reason="Flaky...

None that I can think of, but most likely just historical baggage, `shutil.which` was added in `3.3` and `asv` used to support `2.7` for a long time. A PR would...

Thanks for the issue and checking in! These I think would be more in scope of https://github.com/airspeed-velocity/asv_spyglass In particular after https://github.com/airspeed-velocity/asv_spyglass/pull/7 it should be easy to manipulate the resulting polars...

Yes but the problem as you'll find is that the internals of `asv` are pretty brittle and make this hard. The read only structures in `asv_spyglass` make this a better...