- 18 May, 2020 2 commits
-
-
Georges Racinet authored
That also will be useful for all tests involving the CI. In most of them we won't care about what the CI configuration says, we'll only need it to exist and be valid. --HG-- branch : heptapod-0-12
-
Georges Racinet authored
This is going to be a very recurring pattern. --HG-- branch : heptapod-0-12
-
- 11 May, 2020 1 commit
-
-
Georges Racinet authored
More inspection of the behaviour in the Git case indicate that going through the creation endpoint again for an update is incorrect. The proper verb for update is PUT, but the URL is the same one, which is a bit surprising for the creation. --HG-- branch : heptapod-0-12
-
- 07 May, 2020 5 commits
-
-
Georges Racinet authored
As we suspected, but didn't have time to test before release, heptapod#215 seems to be fixed in version 0.12.3. --HG-- branch : heptapod-0-12
-
Georges Racinet authored
At this stage of Heptapod, the API returns Git SHAs, so we have to use it again to get the details. Usually our tests are made so that a commit can be identified by its message (`'title'` key of the returned dict) --HG-- branch : heptapod-0-12
-
Georges Racinet authored
As we suspected, but did not have time to test, this confirms that heptapod#261 is fixed in Heptapod 0.12.3 --HG-- branch : heptapod-0-12
-
Georges Racinet authored
We'll have several tests using two linear topics on top of each other. --HG-- branch : heptapod-0-12
-
Georges Racinet authored
The docstring was actually for an earlier attempt where the first topic got to be rebased. That was more akin to `test_mergerquest_stacked_amend_publish_first()` --HG-- branch : heptapod-0-12
-
- 06 May, 2020 2 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-12
-
Georges Racinet authored
Since this relies on the client having hg-evolve 9.3.1, we mark it explicitely in the CI configuration. --HG-- branch : heptapod-0-12
-
- 01 May, 2020 3 commits
-
-
Georges Racinet authored
This is the reproduction of heptapod#145. In truth, we should parametrize the fast forward as we did in anouther topic (no, de facto and required) --HG-- branch : heptapod-0-12
-
Georges Racinet authored
This demonstrates in particular our modification to GitLab's JobRequest API: `hgsha` and `repo_type` (in truth the latter conflicts semantically with GitLab's RepoType, which is `project` or `wiki`, and we should rename it to `vcs_type` at some point) --HG-- branch : heptapod-0-12
-
Georges Racinet authored
This will be the main facility for CI functional tests, see #7. --HG-- branch : heptapod-0-12
-
- 03 May, 2020 1 commit
-
-
Georges Racinet authored
The current heptapod:testing image is a preliminary Heptapod 0.13 version. The CI will have to run against the 0-12-testing image. --HG-- branch : heptapod-0-12
-
- 01 May, 2020 4 commits
-
-
Georges Racinet authored
In a typicall MR situation (from a topic), the length run that is expected to fail is replaced by a short explanation and a failure. People are expected to relaunch manually once the Docker image is ready and only then, to merge.
-
Georges Racinet authored
The intermediate images are now called `testing` with `latest` being the latest *released* version.
-
Georges Racinet authored
we've had enough startup problems that we didn't look at this obvious until now.
-
Georges Racinet authored
For objects that are relative to different type of GitLab objects, such as Runners, it would be convenient to access the Heptapod instance uniformly.
-
- 29 Apr, 2020 3 commits
-
-
Georges Racinet authored
this time we parametrize our three fast-forward cases in a single test method
-
Georges Racinet authored
for example, will be used for squash params
-
Georges Racinet authored
- the link should be to create a MR for the natural target branch - the prompt should not fail if the natural target named branch doesn't exist – no prompt in that case
-
- 28 Apr, 2020 4 commits
-
-
Georges Racinet authored
Before that, we only tested that the actions were displayed, not that they worked. Also, this is with the default group selected in the "New project" page.
-
Georges Racinet authored
Amazing that it didn't exist before.
-
Georges Racinet authored
This one assumes the webdriver is already on the "New project" page. It will save time in some tests, and also does not require a permanent user (ie registered to the `Heptapod` instance) and can work with a fixture such as `external_user`
-
Georges Racinet authored
This is the counterpart of the fix for #256, we still musn't display that link if the external user has no group
-
- 27 Apr, 2020 5 commits
-
-
Georges Racinet authored
This is heptapod#256 As a reminder, those need to have the Developer rights on the group.
-
Georges Racinet authored
This new function allows to perform a webdriver login with a User instance. We have a bit of inconsistency here: it will work only for just-created Users (e.g by a fixture), not for those that `User.search()` can return -- unless we can assume the pasword is a standard one and we set it separately before calling `login_as_user()`.
-
Georges Racinet authored
api_post() can be used for some operations, including but not limited, to supporting api_add_member().
-
Georges Racinet authored
As the name suggests, this user is set to "external". This is needed to test heptapod#256. As a fixture, it is created on the fly and deleted in teardown.
-
Georges Racinet authored
It's been delayed for way too much time. Should in particular help simplifying tests/test_groups.py.
-
- 23 Apr, 2020 1 commit
-
-
Georges Racinet authored
Problems with stacked topics have been annoying for a long time (see heptapod#43). Here is a slightly different scenario, that shows that with hg-evolve 9.3.1 (see heptapod#242), we don't need to force-push but GitLab is confused. --HG-- branch : heptapod-0-12
-
- 05 May, 2020 1 commit
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-12
-
- 04 May, 2020 1 commit
-
-
Georges Racinet authored
This is based on the scenario of heptapod#43, with further complication that we publish a topic in the same push. --HG-- branch : heptapod-0-12
-
- 01 Mar, 2020 1 commit
-
-
thanks to py-heptapod!13 we now have something better, that doesn't look like a bug, and it's now generalized in current Heptapod branches. Still testing the message a bit, trying not to be too specific, but changes that this breaks if we change the message are still high.
-
- 06 Apr, 2020 1 commit
-
-
Georges Racinet authored
it's need to seed the topic being there, even though we disable the evolve extension.
-
- 30 Mar, 2020 1 commit
-
-
Georges Racinet authored
It seems that the topic aware client must have hg-evolve 9.3.0 in order to push the topic. For the time being, besides asserting that clients without the `topic` extension don't get topics, we check that the `topic` extension *alone* is enough to pull by revision, which is what the runners do.
-
- 26 Apr, 2020 1 commit
-
-
Georges Racinet authored
In future Heptapod 0.13, the check for command existence happens first, leading to a 403 before the 404 is indeed generated. This doesn't break the test in Heptapod 0.12
-
- 03 Dec, 2019 1 commit
-
-
Georges Racinet authored
As expected, once Mercurial does 3-way merges of files, it becomes very difficult to test that we don't have discrepancies with Git (and it would depend potentiallhy on bugs that have a tendency to be fixed). So better test that the in-file merge worked
-
- 05 Apr, 2020 1 commit
-
-
Georges Racinet authored
This is for heptapod#99, everything is explained at length in the docstring.
-
- 04 Apr, 2020 1 commit
-
-
Georges Racinet authored
Like the other ones, these are Docker only, since we don't have a real service manager in the source (GDK) case. And that's why they are additions to the existing test: we still want the basic form of the test to be executed in the source case. All of this is terribly inefficient, and it would also be better to minimize the number of restarts by grouping hashed and legacy storage tests together.
-