Crash with unicode path

Hi, I ran into a problem that this library doesn't support unicode in path

I use:

Debian 9
Python 3.5.3
GitPython==2.1.10 (installed by pip3 install GitPython command)

i found old issue, similar to mine, but it doesn't helped me:
#532

Everything works just fine if i will use path without cyrillic letters.
Code what doesn't work:

#!/usr/bin/python3
import git
def main():

    repo = git.Repo.clone_from("https://github.com/gitpython-developers/GitPython.git",
        "/home/vlad/AltSTU-Submodules/Диплом/Microcontroller-Unit-Testing/scripts/")

if __name__ == '__main__':
    main()

Traceback (most recent call last):
  File "gitl.py", line 13, in <module>
    main()
  File "gitl.py", line 9, in main
    "/home/vlad/AltSTU-Submodules/Диплом/Microcontroller-Unit-Testing/scripts/")
  File "/home/vlad/.local/lib/python3.5/site-packages/git/repo/base.py", line 982, in clone_from
    return cls._clone(git, url, to_path, GitCmdObjectDB, progress, **kwargs)
  File "/home/vlad/.local/lib/python3.5/site-packages/git/repo/base.py", line 933, in _clone
    finalize_process(proc, stderr=stderr)
  File "/home/vlad/.local/lib/python3.5/site-packages/git/util.py", line 333, in finalize_process
    proc.wait(**kwargs)
  File "/home/vlad/.local/lib/python3.5/site-packages/git/cmd.py", line 402, in wait
    stderr = force_bytes(stderr)
  File "/home/vlad/.local/lib/python3.5/site-packages/gitdb/utils/encoding.py", line 16, in force_bytes
    return data.encode(encoding)
UnicodeEncodeError: 'ascii' codec can't encode characters in position 54-59: ordinal not in range(128)

Back in the days when this library was originally conceived, it was very easy to get text encodings wrong, unfortunately.
Given the current state of affairs, is there any good way to deal with this in what I would call 'modern' python?
After all, the actual OS encoding can only be guessed if it is not provided from the outside, and today one might argue that utf-8 is a better default than ascii.

Well, as far as i know, python 3 uses UTF-8 by default, so adding

# -*- coding: utf-8 -*-

to my code didn't worked out. I don't know where to look for this bug: in system locale configuration or in library. Can you confirm that, this actually a bug in library?

It is a problem with GitPython as it tries to decode something that is not ascii encoded as ascii.
The coding declaration you provided tells python how to decode source code files while making sense of them, and thus not affect the runtime.

I have the same error in GitPython 2.1.11. Is there any solutions?

@brizjin

@brizjin I don't think there is unless someone contributes a patch.

This should be fixed with ebf4656 (#955) in v3.0.5. See #920.

Navigation Menu

Repository navigation

Crash with unicode path #761

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Crash with unicode path #761

Description

Activity

Byron commented on Jun 5, 2018

vtsisyk commented on Jun 7, 2018

Byron commented on Jun 10, 2018

brizjin commented on Aug 13, 2018

Byron commented on Oct 14, 2018

Harmon758 commented on Feb 18, 2020

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions