gh-108791: Fix pdb CLI invalid argument handling

[chgnrdv]

Fixes #108791.
pdb module, if invoked with invalid command line arguments, produces large traceback and/or tries to run debugger on errorneus target, such as directory. This patch improves error handling in pdb CLI, making error messages more concise.

• Issue: pdb CLI doesn't handle incorrect arguments properly #108791

[chgnrdv]

I had to make changes in test_module_without_a_main, as it was taking error message from context exception.

Traceback (most recent call last):
  File "/home/radislav/projects/cpython/Lib/runpy.py", line 148, in _get_module_details
    return _get_module_details(pkg_main_name, error)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/radislav/projects/cpython/Lib/runpy.py", line 142, in _get_module_details
    raise error("No module named %s" % mod_name)
ImportError: No module named t_main.__main__

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/radislav/projects/cpython/Lib/pdb.py", line 166, in check
    self._details
  File "/home/radislav/projects/cpython/Lib/functools.py", line 1014, in __get__
    val = self.func(instance)
          ^^^^^^^^^^^^^^^^^^^
  File "/home/radislav/projects/cpython/Lib/pdb.py", line 174, in _details
    return runpy._get_module_details(self)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/radislav/projects/cpython/Lib/runpy.py", line 152, in _get_module_details
    raise error(("%s; %r is a package and cannot " +
ImportError: No module named t_main.__main__; 't_main' is a package and cannot be directly executed

[sunmy2019]

Can you make it something like

        try:
            self._details
        except ... : # Something you expect
            print(f"{type(e).__name__}: {e}")
        except Exception:
            traceback.print_exc()

to leave the general case untouched.

[chgnrdv]

Looking at the runpy._get_module_details source, ImportError seems to be the only type of exception that can be raised here.

[gaogaotiantian]

I think the improvement is fine. The argument parser piece is a bit hacky, but I'm planning to convert that part to argparse anyway (getopt has not been updated for like 10 years) so it's okay.

[gaogaotiantian]

Hi @chgnrdv , as I updated pdb to use argparse instead of getopt, the argument parsing piece might be fixed (more importantly, the code in this PR was definitely outdated). Could you resolve the conflicts in this PR, hopefully keep the relevant tests? I'll try to find the core dev to review this once it's fixed. Thanks.

[chgnrdv]

@gaogaotiantian, thanks for notice, I'll take a look tomorrow.

[gaogaotiantian]

LGTM, @iritkatriel could you take a look? Thanks!

[iritkatriel]

LGTM.

Should this be backported?

[gaogaotiantian]

Rare case, but safe and backward compatible. I think it should be back-ported.

[sunmy2019]

Agreed

[miss-islington]

Thanks @chgnrdv for the PR, and @iritkatriel for merging it 🌮🎉.. I'm working now to backport this PR to: 3.11, 3.12.
🐍🍒⛏🤖

[bedevere-app]

GH-110915 is a backport of this pull request to the 3.12 branch.

[bedevere-app]

GH-110916 is a backport of this pull request to the 3.11 branch.

[bedevere-app]

GH-111063 is a backport of this pull request to the 3.11 branch.

[bedevere-app]

GH-111064 is a backport of this pull request to the 3.12 branch.