gh-65052: Prevent pdb from crashing when trying to display objects #110578
Conversation
Lib/test/test_pdb.py
Outdated
| (Pdb) args | ||
| self = *** AttributeError: 'A' object has no attribute 'a' *** | ||
| (Pdb) display self | ||
| display self: *** AttributeError: 'A' object has no attribute 'a' *** |
There was a problem hiding this comment.
I think it would be easier for a user to understand what happened if this printed sometime like what traceback._safe_string prints when str or repr fail (the fact that repr failed is more informative in this context than the precise exception. This could be read as "self is an AttributeError").
There was a problem hiding this comment.
Yeah that makes sense. I added an error message repr({expr}) failed: before the exception. Still think it might be helpful for users to know what actually failed - and now it's less ambiguous with the extra error message (hopefully).
| try: | ||
| return repr(obj) | ||
| except Exception as e: | ||
| return _rstr(f"*** repr({expr}) failed: {self._format_exc(e)} ***") |
There was a problem hiding this comment.
Do we use f-strings or % in this script? Or a mixture of the two?
There was a problem hiding this comment.
I'm using f-strings for all new code. There are plenty of usages with % as pdb is old. Do you think it would be helpful to have a PR to update all strings to f-strings? It would eliminate the mixture.
There was a problem hiding this comment.
Probably not worth the risk of breaking something.
|
Should this be backported? In working cases it does not change the output of |
|
Yes, good point. |
|
Thanks @gaogaotiantian for the PR, and @iritkatriel for merging it 🌮🎉.. I'm working now to backport this PR to: 3.11. |
|
Thanks @gaogaotiantian for the PR, and @iritkatriel for merging it 🌮🎉.. I'm working now to backport this PR to: 3.12. |
|
Sorry, @gaogaotiantian and @iritkatriel, I could not cleanly backport this to |
…cts (pythonGH-110578) (cherry picked from commit c523ce0) Co-authored-by: Tian Gao <gaogaotiantian@hotmail.com>
|
GH-110734 is a backport of this pull request to the 3.12 branch. |
|
@iritkatriel , do you need me to cp this to 3.11? |
|
If you think it should be backported there then yes, it would require a manual PR. |
…ay objects (pythonGH-110578) (cherry picked from commit c523ce0) Co-authored-by: Tian Gao <gaogaotiantian@hotmail.com>
|
GH-111002 is a backport of this pull request to the 3.11 branch. |
Currently in
pdbwe have a couple of explicit/implicit usage ofreprwhen the object is arbitrary - which could potentially cause apdbcrash when the object is not displayable. Even though it's not super common, we should try out best makingpdbnot crash. Hence_safe_repris introduced to print objects for: