| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\
| |
| | |
Add python 3.9
|
| | |
|
| | |
|
| | |
|
|/ |
|
|\
| |
| | |
add a changelog entry for #1124
|
|/
|
|
|
| |
- add a changelog entry for #1124
- correct duplicate entry for 3.1.12 -> 3.1.13
|
|\
| |
| | |
add replace method to git.Commit
|
|/
|
|
|
|
|
|
|
|
|
| |
This adds a replace method to git.Commit. The replace method returns a
copy of the Commit object with attributes replaced from keyword
arguments. For example:
>>> old = repo.head.commit
>>> new = old.replace(message='This is a test')
closes #1123
|
| |
|
| |
|
|\
| |
| | |
Fix inheritance issue at commit.iter_items
|
|/
|
|
|
|
|
| |
The iterator used to yield Commit() objects, which does not play well
with inheritance. Yield cls() instead.
Signed-off-by: Yuri Volchkov <yuri@volch.org>
|
|
|
|
| |
Fixes #1116
|
|
|
| |
Add __next__ method to git.cmd.Git.CatFileContentStream, so it can actually be used as an iterable
|
|
|
|
|
|
|
|
| |
FETCH_HEAD is one of the symbolic references local to the current
worktree and as such should _not_ be looked up in the 'common_dir'. But
instead of just hard coding the "right thing" (git_dir) lets defer this
to the SymbolicReference class which already contains this knowledge in
its 'abspath' property.
|
|
|
|
| |
Resolves #1106
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
However, the test asserts on the provided context to be correct,
which is hard to do in this branch while it's easy to doubt the value
of this.
Lastly, there seems to be no way to ignore errors in `git` without
muting all output, which is in fact parsed.
Maybe it's possible to ignore errors while parsing the new kind of
error message.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Authored based on
https://github.com/gitpython-developers/GitPython/issues/1099#issuecomment-754606044
Fixes #1099
|
| |
|
|
|
|
|
|
| |
There is a small typo in git/repo/base.py.
Should read `repository` rather than `repostory`.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
In c96476b, the new default_name nested function does not contain a
retun statement. This leads to an issue when the environment variables
are not present, where the actor name would not be set.
Signed-off-by: Athos Ribeiro <athos@redhat.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Calling getpass.getuser may lead to breakage in environments where there
is no entries in the /etc/passwd file for the current user.
Setting the environment variables for the git user configurations should
prevents GitPython from using values from /etc/passwd. However, doing so
will not prevent reading /etc/passwd and looking for an entry with the
current user UID.
This patch changes the behavior described above so GitPython will
perform a lazy evaluation of /etc/passwd, only doing so when the
environment variables for the git user configuration are not available.
Signed-off-by: Athos Ribeiro <athos@redhat.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's possible to specify a pathspec (eg :!foo) to git diff/status/...
but it currently fails with:
git.exc.GitCommandError: Cmd('/usr/bin/git') failed due to: exit code(128)
cmdline: /usr/bin/git diff --abbrev=40 --full-index --raw :!foo
stderr: 'fatal: ambiguous argument ':!foo': unknown revision or path not in the working tree.
Add missing '--' to the arguments to fix this ambiguity
Signed-off-by: Arnaud Patard <apatard@hupstream.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After a recent 'cleanup' operation that attempted to simplify my
GPG key workflow with Yubikeys, it looks like my GPG installation
has 'forgotten' how to interact with the key I typically used
to sign GitPython releases.
Since I never managed to establish a chain of trust with my only
remaining 'good' key, for you this means you cannot trust new
GitPython releases anymore.
There is nothing I can do about except to apologize for the hassle.
If you want to make constructive suggestions on how to fix this, I am
happy to work with you on that.
|
| |
|
|\
| |
| | |
Read conditional include
|
| | |
|
| | |
|
| | |
|