Create a "destroyed Milestone" event and keep Milestone events around in the DB for posterity
Also fix issue where destroying a Milestone would cause odd, transient messages like "created milestone" or "imported milestone".
Now if a milestone is destroyed, at least it will indicate in the activity feed even if it's not clear which milestone was destroyed:
Closes#2382
See merge request !1227
for posterity.
Also fix issue where destroying a Milestone would cause odd, transient messages like
"created milestone" or "imported milestone".
Add "in" preposition when creating and destroying milestones
Closes#2382
I also revamped the tests on "closing reference commits" (= "Fixes #xxxx" for example).
Now, there are two different contexts:
- when the commits with "closing reference" are pushed to the default branch,
- when the commits with "closing reference" are pushed to a non-default branch.
Closesgitlab-org/gitlab-ce#2190.
Add gitlab-shell to error message to give user a clue that something may be wrong there.
Ran into this in #2082. User was told that repositories were created when they were
not due to hooks symlink being wrong.
Fix hooks for web based events with external issue references
The creation of cross references for external issues (which would fail)
is now prevented. This fixes a 500 error and the execution of hooks when
creating a merge request or commenting on one.
Fixes#1650, GH-9333.
This regression might have been introduced by
8f8a8ab32b. There is actually a TODO by
@rspeicher mentioning external issues which is lost (unhandled) in a
later commit.
This fix is related to !766 and !804 which fix a similar issue for hooks for
Git based events.
See merge request !794
Fix post-receive errors on a push when an external issue tracker is configured
### What does this MR do?
This MR improves upon !766, fixing errors upon a git push that occur only when an external issue tracker (not JIRA) is used. This MR takes into account that external issue trackers, such as JIRA, may close issues. Disabling the processing of post-receive commits when an external issue tracker is configured seems like the wrong behavior.
### Why was this MR needed?
When a user adds an issue reference, the refactoring in 8f8a8ab and a6defd157 caused `project.get_issue` to be called, causing `ExternalIssue` to be returned when an external issue tracker was configured. This object does not have a `close` method, as needed in `CloseService`. Nor does it make sense to associate a `SystemNote` with this object.
GitLab EE uses a [special case for JIRA](https://gitlab.com/gitlab-org/gitlab-ee/blob/master/app/services/git_push_service.rb#L87). I would recommend moving this logic into `CloseService` too.
### What are the relevant issue numbers?
* Closes#1700
* Closes#1720
See merge request !804
Robert Speicher
Kamil Trzcinski
Valery SizovandKamil Trzcinski
Stan Hu
Dmitriy Zaporozhets
Jared Szechy
Douwe Maan
Maël Valais
Daniel Gerhardt
