| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| |\
| |
| | |
chore: add type-hints to gitlab/cli.py
|
| | | |
|
| |\ \
| | |
| | | |
chore(deps): update dependency docker-compose to v1.28.5
|
| |/ / |
|
| |\ \
| |/
|/| |
fix: remove duplicate class definitions in v4/objects/users.py
|
| |/
|
|
|
| |
The classes UserStatus and UserStatusManager were each declared twice.
Remove the duplicate declarations.
|
| |\
| |
| |
| |
| | |
python-gitlab/renovate/wagoid-commitlint-github-action-3.x
chore(deps): update wagoid/commitlint-github-action action to v3
|
| | | |
|
| |\ \
| | |
| | | |
Add type-hints to gitlab/const.py
|
| | | | |
|
| |\ \ \
| | | |
| | | | |
chore: add type hints to gitlab/utils.py
|
| | |/ / |
|
| |\ \ \
| | | |
| | | | |
chore: add type-hints to gitlab/config.py
|
| | |/ / |
|
| |\ \ \
| |/ /
|/| | |
chore: fix wrong variable name in cli.py
|
| |/ /
| |
| |
| |
| | |
Discovered this when I ran flake8 on the file. Unfortunately I was the
one who introduced this wrong variable name :(
|
| |\ \
| |/
|/| |
chore: remove usage of 'from ... import *'
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In gitlab/v4/objects/*.py remove usage of:
* from gitlab.base import *
* from gitlab.mixins import *
Change them to:
* from gitlab.base import CLASS_NAME
* from gitlab.mixins import CLASS_NAME
Programmatically update code to explicitly import needed classes only.
After the change the output of:
$ flake8 gitlab/v4/objects/*py | grep 'REST\|Mixin'
Is empty. Before many messages about unable to determine if it was a
valid name.
|
| |\ \
| | |
| | | |
feat(projects): add project access token api
|
| | | |
| | |
| | | |
Co-authored-by: Nejc Habjan <hab.nejc@gmail.com>
|
| | | | |
|
| |\ \ \
| |/ /
|/| | |
fix: tox pep8 target, so that it can run
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Previously running the pep8 target would fail as flake8 was not
installed.
Now install flake8 for the pep8 target.
NOTE: Running the pep8 target fails as there are many warnings/errors.
But it does allow us to run it and possibly work on reducing these
warnings/errors in the future.
In addition, add two checks to the ignore list as black takes care of
formatting. The two checks added to the ignore list are:
* E501: line too long
* W503: line break before binary operator
|
| |/ |
|
| |\
| |
| | |
fix: undefined name errors in v4 objects
|
| |/
|
|
| |
Discovered that there were some undefined names.
|
| |\
| |
| | |
chore: remove unused function _construct_url()
|
| |/
|
|
|
|
| |
The function _construct_url() was used by the v3 API. All usage of the
function was removed in commit
fe89b949922c028830dd49095432ba627d330186
|
| |\
| |
| | |
Enable mypy type checking and add type hints to gitlab/base.py
|
| | | |
|
| |/
|
|
| |
Add an initial mypy test to test gitlab/base.py and gitlab/__init__.py
|
| |\
| |
| | |
chore: remove usage of 'from ... import *' in client.py
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In gitlab/client.py remove usage of:
* from gitlab.const import *
* from gitlab.exceptions import *
Change them to:
* import gitlab.const
* import gitlab.exceptions
Update code to explicitly reference things in gitlab.const and
gitlab.exceptions
A flake8 run no longer lists any undefined variables. Before it listed
possible undefined variables.
|
| |\
| |
| | |
chore: explicitly import gitlab.v4.objects/cli
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
As we only support the v4 Gitlab API, explicitly import
gitlab.v4.objects and gitlab.v4.clie instead of dynamically importing
it depending on the API version.
This has the added benefit of mypy being able to type check the Gitlab
__init__() function as currently it will fail if we enable type
checking of __init__() it will fail.
Also, this also helps by not confusing tools like pyinstaller/cx_freeze with
dynamic imports so you don't need hooks for standalone executables. And
according to https://docs.gitlab.com/ee/api/,
"GraphQL co-exists with the current v4 REST API. If we have a v5 API, this
should be a compatibility layer on top of GraphQL."
|
| |/ |
|
| |\
| |
| | |
test: extend wait timeout for test_delete_user()
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Have been seeing intermittent failures of the test_delete_user()
functional test. Have made the following changes to hopefully resolve
the issue and if it still fails to know better why the failure
occurred.
* Extend the wait timeout for test_delete_user() from 30 to 60
tries of 0.5 seconds each.
* Modify wait_for_sidekiq() to return True if sidekiq process
terminated. Return False if the timeout expired.
* Modify wait_for_sidekiq() to loop through all processes instead of
assuming there is only one process. If all processes are not busy
then return.
* Modify wait_for_sidekiq() to sleep at least once before checking
for processes being busy.
* Check for True being returned in test_delete_user() call to
wait_for_sidekiq()
|
| |\
| |
| | |
feat: add release links API support
|
| | | |
|
| | | |
|
| | |
| |
| |
| | |
Helps getting the real state of the project from the server.
|
| | | |
|
| |\ \
| |/
|/| |
fix: test_update_group() dependency on ordering
|
| |/
|
|
|
|
|
|
| |
Since there are two groups we can't depend on the one we changed to
always be the first one returned.
Instead fetch the group we want and then test our assertion against
that group.
|
| |\
| |
| | |
chore(deps): update dependency docker-compose to v1.28.4
|
| | | |
|
