Sebastian Thiel comments

Results 519 comments of


Sebastian Thiel

3.1.20: pytest (6.2.4) error

Additionally, `./init-tests-after-clone.sh` should be executed to get the clone ready for running the tests.

submodule.update with init=True does not work if the submodule does not have a 'master' branch

Thanks for taking the time for this fantastic issue description! Unfortunately, nothing I could write here would do it any justice because I wouldn't know how to fix that. I...

index.add('.') adds even .git directory into the index, and adds files also with ./ prefix

Thanks for letting me know ! As I see the `IndexFile` implementation, it's definitely located in the _plumbing_ area within the git ecosystem. Therefore I would expect it to do...

index.add('.') adds even .git directory into the index, and adds files also with ./ prefix

I think you would have to filter this path out by yourself. It is valid to pass in iterators, and a filter could look like `(p for p in paths...

index.add('.') adds even .git directory into the index, and adds files also with ./ prefix

Just for completeness, [here is the link to the comment describing the workaround](https://github.com/gitpython-developers/GitPython/issues/292#issuecomment-107800965).

Possible new feature (a new commit query) if not implemented yet

Having looked at it just now, I can only agree with @dongs0104 . However, as the code above calls git directly, this also means that git itself is unable to...

Possible new feature (a new commit query) if not implemented yet

I wouldn't mind seeing a PR for this. The method would probably be called `ref_exists(…)`.

Library leaves dangling processes after use

Thanks for all the research done to investigate, here are my thoughts. > The class comment indicates that the process should be killed when the object goes out of scope....

Not all of diff types documented and used in iter_change_type

Absolutely right, thanks a lot for taking the time to point this out. It would be great to see a PR to improve this, such that * the comment is...

Is gitdb an implementation detail?

Thanks for posting! I believe this is indeed an issue, and it's worth fixing for better usability and less surprises. Historically, GitDB was meant as standalone low-level library anyone could...