gh-80744: do not read .pdbrc twice when cwd == $home #136816
Conversation
Lib/pdb.py
Outdated
| except OSError: | ||
| pass | ||
| if os.path.join( | ||
| os.path.abspath(os.curdir), ".pdbrc" |
There was a problem hiding this comment.
I think the equivalent check should be os.path.abspath(".pdbrc") right?
There was a problem hiding this comment.
@gaogaotiantian yeah i don't know what i was thinking there, updated
btw would it be ok to send a separate pr updating the pdb module & its tests to start using pathlib or preferring to keep os.path ?
Lib/test/test_pdb.py
Outdated
| rc_path = os.path.join(cwd, ".pdbrc") | ||
| with open(rc_path, "w") as f: | ||
| f.write("invalid") | ||
| self.assertEqual(pdb.Pdb().rcLines[0], "invalid") |
There was a problem hiding this comment.
I think you should combine this line and the line below. Just check pdb.Pdb().rcLines directly.
There was a problem hiding this comment.
I've changed it self.assertEqual(pdb.Pdb().rcLines[-1], "invalid"), this test is not patching a .pdbrc at $HOME so potentially there can be other lines first in rcLines, so now it only checks that the last one comes from the file written during the test
incidentally, i just found there are 5 other tests that fail if you do happen to have a non-empty .pdbrc at $home, it's ok in CI because there is none, but locally a few unrelated things fail when the file is present
Lib/test/test_pdb.py
Outdated
| self.assertEqual(pdb.Pdb().rcLines[0], "invalid") | ||
| self.assertEqual(len(pdb.Pdb().rcLines), 1) | ||
|
|
||
| def test_readrc_home_twice(self): |
There was a problem hiding this comment.
Let's rename the function name to be something more descriptive. test_readrc_cwd_is_home maybe?
This is for #80744
There is a previous PR fixing this issue #12731 but it's been abandoned for years as the author is not active in github anymore