Paul J. Dorn

(Linking relevant PR #2938 for anyone in the future stumbling across this)

Thanks for that context - though I am not convinced yet that the locking issue in the requests queue is directly related to the locking issue on thread join (btw,...

As a minimal fix for excluding generated html duplicates (and bundled _static that comes with those) from sdist, appending `recursive-exclude docs/build *` to `MANIFEST.in` seems reasonable.

I believe this is no longer needed since https://github.com/benoitc/gunicorn/pull/2043 - there is no longer any confusion between NameError from looking up the application and during import.

Rebased version contains a doc syntax fix missed in #3217 (rst lists use `- dash` not `* star`) Question to reviewers: What about `PATH_INFO`, should it also be in the...

> Do you plane any other changes? In this PR? No. Before any release is tagged that includes this change? Very likely. The only makeshift-release-candidate I tested beyond running unit...

@washeck I saw yours and would like to leave the maintainer with the option of applying your minimal fix + test, with or without bundling more extensive changes resolving much...

> I agree for this but why for `%s` ? The magic happening for single-arg percent-formatting is not unique to `%r`: ``` Python 3.12.4 on linux >>> "repr %r" %...

Dupes: https://github.com/benoitc/gunicorn/pull/2836 https://github.com/benoitc/gunicorn/pull/3189 Since you only sent one while more are in the same codebase, you might be happy to learn that a rebased version of such patch, canonicalizing *all*...

Design consideration: Whether the * failure of this hook should be full responsibility of the hook author (more flexible, wrapping try-except if so desired), or whether * failure of this...