Fix downloading of patches on public merge requests when user logged out
### What does this MR do?
This MR makes it possible to download a diff patch on a public merge request when a user is logged out.
### Why was this MR needed?
An Error 500 would result when a user attempted to click on the "Email Patches" or "Plain Diff" button:
```
NoMethodError - undefined method `id' for nil:NilClass:
lib/gitlab/backend/shell_env.rb:9:in `set_env'
lib/gitlab/satellite/action.rb:20:in `in_locked_and_timed_satellite'
lib/gitlab/satellite/merge_action.rb:49:in `diff_in_satellite'
app/models/merge_request.rb:219:in `to_diff'
app/controllers/projects/merge_requests_controller.rb:42:in `block (2 levels) in show'
```
### What are the relevant issue numbers?
* Closes#1225
* Closes#1854 (dup)
* Closes#1858 (dup)
See merge request !872
Code tweaks in 45e9150a caused the ldap_person method to not return expected results.
Improved tests to cover the ldap_person method, which was previously stubbed.
Now we can simply loop through all themes, among other things.
This also removes the `dark_theme` / `light_theme` classes and the
`theme_type` helper, since they weren't used anywhere.
There is no need to check if `repository_ref` is present as:
```
@repository_ref = if repository_ref.present?
repository_ref
else
nil
end
```
is as same as doing:
```
@repository_ref = repository_ref
```
Fix upgrader script
This is a fix for upgrader script not guessing the latest version correctly.
Upgrader now handles versions where a version part (major/minor/patch) can have multi-digit number, also ensures that the latest version is chosen from git tags by converting tag to Gitlab::VersionInfo and than selecting the latest/greatest version.
Fixes: #1476
See merge request !695
Ths commit does next:
* When we remove project we move repository to path+deleted.git
* Then we schedule removal of path+deleted with sidekiq
* If repository move failed we abort project removal
This should help us with NFS issue when project get removed but
repository stayed. The full explanation of problem is below:
* rm -rf project.git
* rm -rf removes project.git/objects/foo
* NFS server renames foo to foo.nfsXXXX because some NFS client (think
* Unicorn) still has the file open
* rm -rf exits, but project.git/objects/foo.nfsXXX still exists
* Unicorn closes the file, the NFS client closes the file (foo), and the
* NFS server removes foo.nfsXXX
* the directory project.git/objects/ still exists => problem
So now we move repository and even if repository removal failed
Repository directory is moved so no bugs with project removed but
repository directory taken. User still able to create new project with
same name. From administrator perspective you can easily find stalled
repositories by searching `*+deleted.git`
Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>
Until now, a user needed to first sign in with his LDAP identity and then manually
link his/her account with an omniauth identity from their profile.
Only when this is done can the user authenticate with the omniauth provider and at
the same time benefit from the LDAP integration (HTTPS authentication with LDAP
username/password and in EE: LDAP groups, SSH keys etc.).
This feature automates the process by looking up a corresponding LDAP person when a
user connects with omniauth for the first time and then automatically linking the LDAP
and omniauth identities (of course, like the existing allow_single_sign_on setting,
this is meant to be used with trusted omniauth providers).
The result is identical to a manual account link.
Add config initializers for other omniauth settings.
Fix reference links in dashboard activity and ATOM feeds.
References like `@user` were not linked on the dashboard activity feed and in Atom feeds, because the reference parser depends on `@project` which isn't set in those situations.
This MR passes the project in question to the reference parsers in an option.
cc @rspeicher
See merge request !653
This makes the memory growth-triggered Sidekiq restarts more reliable by
reducing the chance that Sidekiq ends up in a state where it is not
accepting new jobs but also not shutting down: SIGKILL is more likely to
work than SIGTERM.
Batch 1 of text improvements
Batch 1 of changes from my effort at !635 to walk through every piece of text in GitLab and see if it can be improved.
This batch includes:
- Improve text on error pages.
- Improve Git access error messages.
- Improve description of branch protection levels.
- Improve OAuth signup error message.
- Improve OAuth application flash messages.
cc @rspeicher
See merge request !642
Dmitriy Zaporozhets
Stan Hu
Alex Lossent
Robert Speicher
Douwe Maan
themaze75
Jeroen van Baarsen
zenati
Eric Maziade
Martins Polakovs
Jakub Jirutka
Nikita Verkhovin
Jacob Vosmaer