Meta-Information needs update

[pajod]

• Are historically referenced MAINTAINERS still active and posess permissions on GitHub?
• There is no primary contact email mentioned in Security.md - should stay in sync with community.rst
  • Security mailbox seems broken #3160
• Many triaged issues / planned changes need (updated) milestone
  • https://github.com/benoitc/gunicorn/issues?q=is%3Aopen+is%3Aissue+milestone%3A22.0
• many existing PR are stuck with no or outdated "what is missing to move this forward" comment
  • and possibly too few PRs and issues are closed with "this is not moving forward for this reason" clarification
• Typo in the Security.md
  • sorry, my bad! I missed the trailing S: https://devguide.python.org/versions/
  • Typo and email address in Security.md #3123
• Last entry in NOTICE reference is unclear
  • maybe util.py/_waitfor() reference and cpython/Lib/test/support/os_helper.py now (can that be deleted eventually?)
• build infrastructure is not well documented.
  • Should what is partially already done - referencing test dependencies via extras be done everywhere, no requirements*.txt whatsoever?
  • maybe putting it all in in a CI file referencing the Python Packaging Authority github action would both document it and ensure we do not reinvent wheels (pun intended)
  • CI: minimal dependency versions: specify them and try them in CI #3134
• Is the Makefile still of any use?
  • the sphinx one is out of date, it does not understand the dummy target (running sphinx-quickstart --extensions=gunicorn_ext gave one one that does)
• supported OS should be clarified, and preferably strictly based on what is really tested in Github CI
  • Linux works well, Mac sort of works (no pipe2()), cygwin and WSL1 are weird about filesystem & socket operation edge cases, WSL2 probably works as good as Linux (broken by CLOCK_REALTIME fun) but is unavailable on Github CI
  • we can test WSL via appveyor right now, like its not even particularly complex or slow, its more a matter of whether anyone cares to look at failures
• I have not been able to use (==) the mentioned minimum (>=) gevent / eventlet versions.
  • The actual minimum supported version is easy enough to verify in CI to not miss that in the future.
  • CI: minimal dependency versions: specify them and try them in CI #3134
• It is easy to forget make -C docs html
  • if we insist on keeping the duplication in-repo, we can and should CI-test
  • CI: verify docs are up to date #3217
• The main README.. in fact the also the full documentation lacks clear goal / non-goal statements
  • should probably point to https://peps.python.org/pep-3333/ and https://datatracker.ietf.org/doc/html/rfc9112 and promise to implement those reasonably well - and point out where we make our own notable choices.