Change in `pdb` behavior at exit from post-mortem debugger

[chgnrdv]

Bug report

Bug description:

In 3.12 and earlier, if user enters quit/exit commands or does Ctrl-D, pdb restarts script with corresponding message and exits only on repeated quit/exit/Ctrl-D:

$ python -m pdb -c c pdb_bug.py 
Traceback (most recent call last):
  File "/usr/local/lib/python3.11/pdb.py", line 1775, in main
    pdb._run(target)
  File "/usr/local/lib/python3.11/pdb.py", line 1643, in _run
    self.run(target.code)
  File "/usr/local/lib/python3.11/bdb.py", line 600, in run
    exec(cmd, globals, locals)
  File "<string>", line 1, in <module>
  File "/home/radislav/projects/cpython/pdb_bug.py", line 1, in <module>
    raise ValueError()
ValueError
Uncaught exception. Entering post mortem debugging
Running 'cont' or 'step' will restart the program
> /home/radislav/projects/cpython/pdb_bug.py(1)<module>()
-> raise ValueError()
(Pdb) quit
Post mortem debugger finished. The /home/radislav/projects/cpython/pdb_bug.py will be restarted
> /home/radislav/projects/cpython/pdb_bug.py(1)<module>()
-> raise ValueError()
(Pdb) quit
$

Since fa18b0a only message is printed, but Python exits completely:

$ ./python -m pdb -c c pdb_bug.py 
Traceback (most recent call last):
  ...
  File "/home/radislav/projects/cpython/pdb_bug.py", line 1, in <module>
    raise ValueError()
ValueError
Uncaught exception. Entering post mortem debugging
Running 'cont' or 'step' will restart the program
> /home/radislav/projects/cpython/pdb_bug.py(1)<module>()
-> raise ValueError()
(Pdb) quit
Post mortem debugger finished. The /home/radislav/projects/cpython/pdb_bug.py will be restarted
$

@gaogaotiantian, what's your opinion? Should we restore the old behavior or remove confusing message about "restart"? As I can see, pdb docs doesn't state explicitly what quit/exit commands should do in post-mortem debugger.

CPython versions tested on:

CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-118714: Make the pdb post-mortem restart/quit behavior more reasonable #118725
• [3.13] gh-118714: Make the pdb post-mortem restart/quit behavior more reasonable (GH-118725) #121346

[gaogaotiantian]

I think we should make quit quit and restart restart. It's inconsistent for the post-mortem debugger to have a different behavior. We noticed the user that cont and step will restart the program, which is fine and has been the case for a really long time, but quit should not just quit the post mortem debugger and enter the normal debugger, they should explicitly ask a restart.

So we need to do something about the message - I'll work on that. This is a bug fix so it doesn't have to be merged before beta freeze.

[ambv]

Should this still be open?

[gaogaotiantian]

Fixed in #118725