[issue6721] Locks in the standard library should be sanitized on fork
Change by Jesse Farnham : -- nosy: +jesse.farnham ___ Python tracker <https://bugs.python.org/issue6721> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue36259] exception text is being sourced from the wrong file
Change by Jesse Farnham : -- nosy: +jesse.farnham ___ Python tracker <https://bugs.python.org/issue36259> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue36259] exception text is being sourced from the wrong file
Jesse Farnham added the comment: I did some digging into this, and the problem seems to be that _Py_FindSourceFile() in traceback.c searches through every directory in sys.path (of which the first entry is the working directory) to find a file with the passed filename. So if there's a file in the working directory with name matching the filename passed in from the traceback object, _Py_FindSourceFile will find that one, resulting in the wrong listing being printed. Does the traceback object contain any other information that would avoid the need to search sys.path to find the right file? If it can know the file name, maybe there's a way to find the file path as well? Apologies if this is not useful -- this is my first attempt to contribute to Python. -- ___ Python tracker <https://bugs.python.org/issue36259> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue36259] exception text is being sourced from the wrong file
Jesse Farnham added the comment: Upon further digging, the filename to search for ultimately comes from the PyCodeObject where the error occurred. So a possible solution could be to populate the PyCodeObject with the absolute path to the file when it’s first created in compilation. Then no searching is needed and the problem doesn’t happen. This seems like a potentially very invasive change, though. -- ___ Python tracker <https://bugs.python.org/issue36259> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue29699] shutil.rmtree should not fail with FileNotFoundError (race condition)
Change by Jesse Farnham : -- nosy: +jesse.farnham ___ Python tracker <https://bugs.python.org/issue29699> ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
