Enable ruff's flake8-return rule

[CoolCat467]

In this pull request, we enable ruff's flake8-return rule and fix all the new issues from doing this.

[codecov]

Codecov Report

Attention: Patch coverage is 97.19626% with 3 lines in your changes missing coverage. Please review.

Project coverage is 99.63%. Comparing base (e3cfb23) to head (b3ac062).
Report is 1 commits behind head on main.

Files	Patch %	Lines
src/trio/_core/_io_kqueue.py	0.00%	1 Missing ⚠️
src/trio/_dtls.py	85.71%	1 Missing ⚠️
src/trio/_tests/test_subprocess.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3047      +/-   ##
==========================================
+ Coverage   99.51%   99.63%   +0.12%     
==========================================
  Files         121      121              
  Lines       17841    17832       -9     
  Branches     3206     3206              
==========================================
+ Hits        17754    17767      +13     
+ Misses         67       46      -21     
+ Partials       20       19       -1     

Files	Coverage Δ
src/trio/_core/_ki.py	100.00% <100.00%> (ø)
src/trio/_core/_mock_clock.py	100.00% <100.00%> (ø)
src/trio/_core/_run.py	99.38% <100.00%> (ø)
src/trio/_core/_tests/test_guest_mode.py	100.00% <100.00%> (ø)
src/trio/_core/_tests/test_run.py	100.00% <ø> (ø)
src/trio/_core/_tests/test_thread_cache.py	100.00% <100.00%> (ø)
src/trio/_core/_tests/test_windows.py	100.00% <100.00%> (ø)
src/trio/_core/_unbounded_queue.py	100.00% <100.00%> (ø)
src/trio/_file_io.py	100.00% <100.00%> (ø)
src/trio/_highlevel_generic.py	100.00% <100.00%> (ø)
... and 23 more

... and 4 files with indirect coverage changes

[TeamSpen210]

Not sure if this is worth it, it just causes a lot of churn indenting/dedenting things for no real reason. The elif technically is redundant if the previous if always returns, but it makes it more clear that they're mutually exclusive, and might cause bugs in future if someone removes one of the returns. This also causes symmetric if-else checks to become non-symmetric now, which might be confusing to the reader.

[A5rocks]

I agree with @TeamSpen210 that this doesn't really help anything. It doesn't make things clearer (unlike flake8-commas) and removes semantics. Looking through the error codes listed on https://pypi.org/project/flake8-return/ (I'm sure ruff implements them smarter, ofc):

• R501 conflicts with mypy nvm, I was confusing this with implicit None return w/ a return type like None | int. I guess this is useful for standardization reasons, but I don't really like it.
• R502/R503 is just type checking
• R504 is potentially useful, but in general assigning to a variable is useful because it allows us to show a nicer name in tracebacks. (maybe we could just enable this? I'm not too sure)
• R505/R506/R507/R508 remove semantics