- 10 Sep, 2019 2 commits
-
-
Georges Racinet authored
It should be the most reliable way to tear down the test projects and seems to be a bit faster.
-
Georges Racinet authored
Hence, for clarity and consistence, we also rename the existing method as `webdriver_destroy()` This new method should be more reliable for teardowns and more efficient.
-
- 09 Sep, 2019 4 commits
-
-
Georges Racinet authored
Preliminary tests indicate it makes tests about 1/3 faster to use it in project fixtures rather than the webdriver based one, BUT to do so, we would need to add a dedicated test using the webdriver method, because the Web UI is critical and has to be tested.
-
Georges Racinet authored
This avoids importing the project module, and in particular the masking niceties that can endeavour.
-
Georges Racinet authored
also, stop hardcoding 'root' for the group part of the project path.
-
Georges Racinet authored
This is a bit simpler for callers, but the fixtures need it to become idempotent (hence skip_missing) for tests that call destroy() directly. In turn those tests (currently test_destroy only) don't need to depend on many fixtures nor to import 'project'
-
- 06 Sep, 2019 1 commit
-
-
Georges Racinet authored
-
- 29 Aug, 2019 1 commit
-
-
Georges Racinet authored
-
- 30 Aug, 2019 3 commits
-
-
Georges Racinet authored
This second test demonstrates that it works in Heptapod 0.5.2
-
Georges Racinet authored
As usual, we use the RESTful API, and assume that's the same as from the Web UI. This new test passes right away on Heptapod 0.5.2
-
Georges Racinet authored
It's possible to get an intermediate page telling us that the page will refresh once the project is created. That should not be a reason to fail.
-
- 06 Sep, 2019 1 commit
-
-
Georges Racinet authored
-
- 29 Aug, 2019 1 commit
-
-
Georges Racinet authored
This will be useful to control various flags directly from the functional tests, without having to wait for proper exposition (if ever)
-
- 28 Aug, 2019 5 commits
-
-
Georges Racinet authored
Pushing public changesets must be restricted to project masters.
-
Georges Racinet authored
Now `owner_api_get()`, `owner_api_post()` etc all go through the central `owner_api_request()`.
-
Georges Racinet authored
-
Georges Racinet authored
-
Georges Racinet authored
This is necessary for some API calls, such as adding to project members
-
- 13 Aug, 2019 1 commit
-
-
Georges Racinet authored
The two new tests do fail on Heptapod 0.5.1
-
- 10 Aug, 2019 4 commits
-
-
Georges Racinet authored
-
Georges Racinet authored
-
Georges Racinet authored
In case of complex or hard to reproduce states, this is a really useful shortcut.
-
Georges Racinet authored
-
- 12 Aug, 2019 5 commits
-
-
Georges Racinet authored
Here, the user name that can be problematic is Gitlab's, not the one that can be seen in a previously commit made client-side.
-
Georges Racinet authored
We didn't support full name until now, but it's important because of the non-ASCII tests.
-
Georges Racinet authored
This is much more direct, especially for non-root users. The webdriver method currently works for root only, even though we fixed the first impediment: next page is the one for choosing the branches and cannot be skipped by validation of a default, as root can do.
-
Georges Racinet authored
This was missing, and of course now we'll need to deduplicate
-
Georges Racinet authored
Useful for various API calls.
-
- 01 Aug, 2019 1 commit
-
-
Georges Racinet authored
Preliminary results where that non-ASCII in user name doesn't harm, but we have a failure in production of the merge commit in test_mergerequest_api_explicit.
-
- 31 Jul, 2019 2 commits
-
-
Georges Racinet authored
as explained in the comments/docstring actually, if hg-git fails in the calls from GitLab (potentially not exactly covered by other tests), this test fails.
-
Georges Racinet authored
This goes with the latest fix in hg-git so that the topic Git branches get forwarded even if publication occurred in the same push.
-
- 30 Jul, 2019 6 commits
-
-
Georges Racinet authored
This paves some of the way for tests where several merge requests would be closed in a single push. Somewhat a long shot: our MR creation is not able yet to select the source topic, but it was easy to spot and fix right now.
-
Georges Racinet authored
This one was in my list of stuff that could go wrong - but it passes with the current fix for heptapod#47 Had to improve the rebasing logic so that it can handle a topic with several changesets.
-
Georges Racinet authored
and also stop hardcoding the MR id, we can read it from the redirected URL after creation.
-
Georges Racinet authored
as the docstring explains, it's not obvious that this would be detected, and we'd have a risk about that if we start tweaking the tripler inner push system for better performance or maintainability.
-
Georges Racinet authored
it's indeed easier to fix the scenario by accidentally disabling that removal.
-
Georges Racinet authored
The separate robustness test is not meaningful any more, we simply include that check in the rebase_push scenario
-
- 05 Aug, 2019 2 commits
-
-
Georges Racinet authored
these can be amended later, but it's occurring to me that the interdependency with changes in Heptapod has to be clarified, if only as a reference for myself.
-
Georges Racinet authored
This happened for me in about 5% of tests in tear down phase. It has value to test project destruction through the regular UI, hence not converting to API call right away. An API call would probably be more efficient for teardowns though, and would help us make the difference between tear down and testing of the delete feature.
-
- 16 Jul, 2019 1 commit
-
-
Georges Racinet authored
With this change, a persistent unprivileged user `test_basic` is created, with its API token. The Heptapod fixture retrieves the token and starts a web session as this user, the project fixtures have it own their projects. It felt reasonible to have this generic user persist across test sessions, notably because user destruction is partly asynchronous in GitLab, because removing all objects authored by this user can be very long. We'll probably have this user own further test projects and create auxiliary, transient additional users to interact with the projects.
-