gh-124703: Change back to raising bdb.BdbQuit when exiting pdb in 'inline' mode in a REPL session

[ADThomas-astro]

This change should mean it's again possible to call breakpoint() in a REPL session and return to the REPL session with q/quit/exit/EOF.

The test has a slight workaround/hack because I couldn't figure out how to call breakpoint() and simulate using pdb interactively within a simulated REPL session in a subprocess. After set_trace() was called, pdb would receive an EOF and exit before it could receive further commands.

If there's interest, I could also add the code from PR129768 to shorten the BdbQuit traceback by one frame.

• Issue: Do not raise an Exception when exiting pdb #124703

[bedevere-app]

Most changes to Python require a NEWS entry. Add one using the blurb_it web app or the blurb command-line tool.

If this change has little impact on Python users, wait for a maintainer to apply the skip news label instead.

[gaogaotiantian]

Can we import the spawn_repl from test_repl here? So we don't need to keep a copy.

[ADThomas-astro]

Yeah, I wasn't sure about this. I see three options:

1. Keep a copy of spawn_repl here - it duplicates some non-obvious code and risks things getting out of sync
2. Do from test.test_repl import spawn_repl. I tried it and it works. But this introduces a dependency/import between test files, which seems discouraged and rare from grepping through the test directory.
3. Move spawn_repl from test_repl.py to somewhere in test/support and import from test.support in both test_repl and test_pdb. This is more invasive and would involve touching at least two other files, but probably more maintainable and keeps test files independent/isolated.

Which do you think?

[gaogaotiantian]

For 2, we can do that inside the test function, so there won't be any dependencies. It's not that common to import a util function from other tests, that's true, but in this case I think it's fine. It's a single specific test using repl.