Ville Skyttä

> Please don't do that :) PR's including test suite additions welcome :)

The make completion used to be in a state that nobody was willing to touch it. It has been completely redone between 2.11 and now, and even though people might...

Thanks for the report! I would rather add some arch specific expectations than exclude those tests altogether on non-x86. For -mtune, perhaps even do something like this ```diff --- a/test/t/test_gcc.py...

Sorry about the huuuuuuuuge delay on my part. I agree there's no rush to switch away from the bash_completion.d based approach for now, or for a long time. Then again,...

Closing due to lack of further info regarding the reproducer. Let's reopen if we receive that later.

One could say a similar thing about pretty much all our completions. We do however choose to restrict filename completions based on common "suffixes" for the context almost everywhere, including...

I have not tested the behavior, but I'm inclined to agree with @akinomyoga here, unless some more feedback confirming that deassociating these would be the right thing to do.

Closing due to lack of feedback.

Reproducer: ```shellsession $ docker run --rm -dp 3128:3128 ubuntu/squid e0c6faf........ $ cat bug914.py import vcr import requests with vcr.use_cassette("bug914.yaml"): response = requests.get("https://vcrpy.readthedocs.io/en/latest/") print(response) $ env HTTPS_PROXY=http://localhost:3128 python3 bug914.py $ grep...

Does not seem fixed. For example, running the musl build on GH ubuntu-latest: https://github.com/scop/home-assistant-jatekukko/actions/runs/13101294499/job/36549863591#step:5:38 > mise [email protected] extract dprint-x86_64-unknown-linux-musl.zip https://github.com/scop/home-assistant-jatekukko/actions/runs/13101294499/job/36549863591#step:6:24 > libunwind: __unw_add_dynamic_fde: bad fde: FDE is really a CIE