A lot of git operations were being repeated, for example, to build a url
you would ask if the path was a Tree, which would call a recursive routine
in Gitlab::Git::Tree#where, then ask if the path was a Blob, which would
call a recursive routine at Gitlab::Git::Blob#find, making reference to
the same git objects several times. Now we call Rugged::Tree#path, which
allows us to determine the type of the path in one pass.
Some other minor improvement added, like saving commonly used references
instead of calculating them each time.
There are several changes to this module:
1. The use of an explicit stack in Participable#participants
2. Proc behaviour has been changed
3. Batch permissions checking
== Explicit Stack
Participable#participants no longer uses recursion to process "self" and
all child objects, instead it uses an Array and processes objects in
breadth-first order. This allows us to for example create a single
Gitlab::ReferenceExtractor instance and pass this to any Procs. Re-using
a ReferenceExtractor removes the need for running potentially many SQL
queries every time a Proc is called on a new object.
== Proc Behaviour Changed
Previously a Proc in Participable was expected to return an Array of
User instances. This has been changed and instead it's now expected that
a Proc modifies the Gitlab::ReferenceExtractor passed to it. The return
value of the Proc is ignored.
== Permissions Checking
The method Participable#participants uses
Ability.users_that_can_read_project to check if the returned users have
access to the project of "self" _without_ running multiple SQL queries
for every user.
Any mention of Issues, MergeRequests, or Commits via GitLab-flavored markdown
references in descriptions, titles, or attached Notes creates a back-reference
Note that links to the original referencer. Furthermore, pushing commits with
commit messages that match a (configurable) regexp to a project's default
branch will close any issues mentioned by GFM in the matched closing phrase.
If accepting a merge request would close any Issues in this way, a banner is
appended to the merge request's main panel to indicate this.
The good:
- You can do a merge request for a forked commit and it will merge properly (i.e. it does work).
- Push events take into account merge requests on forked projects
- Tests around merge_actions now present, spinach, and other rspec tests
- Satellites now clean themselves up rather then recreate
The questionable:
- Events only know about target projects
- Project's merge requests only hold on to MR's where they are the target
- All operations performed in the satellite
The bad:
- Duplication between project's repositories and satellites (e.g. commits_between)
(for reference: http://feedback.gitlab.com/forums/176466-general/suggestions/3456722-merge-requests-between-projects-repos)
Fixes:
Make test repos/satellites only create when needed
-Spinach/Rspec now only initialize test directory, and setup stubs (things that are relatively cheap)
-project_with_code, source_project_with_code, and target_project_with_code now create/destroy their repos individually
-fixed remote removal
-How to merge renders properly
-Update emails to show project/branches
-Edit MR doesn't set target branch
-Fix some failures on editing/creating merge requests, added a test
-Added back a test around merge request observer
-Clean up project_transfer_spec, Remove duplicate enable/disable observers
-Ensure satellite lock files are cleaned up, Attempted to add some testing around these as well
-Signifant speed ups for tests
-Update formatting ordering in notes_on_merge_requests
-Remove wiki schema update
Fixes for search/search results
-Search results was using by_project for a list of projects, updated this to use in_projects
-updated search results to reference the correct (target) project
-udpated search results to print both sides of the merge request
Change-Id: I19407990a0950945cc95d62089cbcc6262dab1a8