Avoid calculation of closes_issues.
## What does this MR do?
Avoid unneeded calls to MR closes issues
## Are there points in the code the reviewer needs to double check?
I'm not sure if calling this method from a view is a good practice, but I cannot see another simple way of avoiding this problem. In case we want to avoid this in the controller we need to specify the action, format and status of the merge request, because in that case we know that the `_open` partial will render. We could add some lazy evaluation but it not a thing I see in use along the app but feedback is welcome
## What are the relevant issue numbers?
#14202 , #19490
## Does this MR meet the acceptance criteria?
- [x] [CHANGELOG](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CHANGELOG) entry added
- ~~[ ] [Documentation created/updated](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/development/doc_styleguide.md)~~
- ~~[ ] API support added~~
- Tests
- ~~[ ] Added for this feature/bug~~
- [x] All builds are passing
- [x] Conform by the [style guides](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CONTRIBUTING.md#style-guides)
- [x] Branch has no merge conflicts with `master` (if you do - rebase it please)
- [x] [Squashed related commits together](https://git-scm.com/book/en/Git-Tools-Rewriting-History#Squashing-Commits)
See merge request !5140
Added blank state to issues
## What does this MR do?
Adds new blank state to issues when no issues exist.
Part of #18519
## Screenshots (if relevant)
See merge request !4908
New diff notes
Fixes#12732, #14731, #19375, #14783
Builds on https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/4110
To do:
- [x] Get it mostly working
- [x] Validate position validity
- [x] Fix: Don’t link to `#`
- [x] Fix: Base ref can be `nil`, potentially, when the MR has an oprhan source branch => Yep, doesn’t work. We need to store a `start_id`
- [x] Optimize: Fewer duplicate `git diff` compares
- [x] Optimize: Pass paths to `PositionTracer#diff` for faster diffs
- [x] Refactor: Use `head_id` in `MergeRequest`/`MergeRequestDiff` instead of `source_sha`
- [x] Refactor: Convert existing array-based diff refs to the DiffRefs model
- [x] Tweak: Use `note_type` in `Autosave` key
- [x] Tweak: Remove `line_code: note.line_code` from `link_to_reply_discussion`
- [x] Update: `SentNotifications` and reply-by-email receiver
- [x] Update: MR diff notification email
- [x] Update: API (MR, Commit note creation and entity)
- [x] Update: GitHub importer
- [x] Address any other TODO comments
- [x] Fix: Suppress "edited 4 minutes ago"
- [x] Write tests
- [x] `LineMapper`
- [x] `PositionTracer`
- [x] `Position`
- [x] `DiffPositionUpdateService`
- [x] `DiffNote`
- [x] `MergeRequests::RefreshService` / `MergeRequest#update_diff_notes_positions`
- [x] Make sure commits with diff notes don't get cleaned up, since this would prevent the diff notes from being rendered (https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/5062)
Future improvements:
- Display unresolved comments on files outside the diff, if the comment was added when that file _was_ part of the diff
- Allow commenting on sections between hunks, when expanding the diff using `...`
- (We'd need to generate line code based on Position if we have it, even if it falls outside bounds of diff)
- `diff_hunk` on diff note API entity
- Show diff hunk in notification email
- Resolved line notes would have a boolean, and be inactive through `notes.any? { !active? || resolved? }`
- Multi line notes would store a number of positions, and do the right thing (™) in grouping and then rendering if the first item is multiline? => true
- Image diff notes could store x,y,width,height instead of old_line,new_line for similar grouping. Does it need a reference to say if it's on old or new? These can't have line_codes, clearly. Rendering would be interesting.
- Show commit line comments in the MR diff
- Comment on specific selected words
- Comment on file header
- Unfold top of discussion diff note
- New diff notes API for commits and MRs
/cc @rspeicher
See merge request !4101
Don't show "request access" button to project owners
This MR fixes an issue where project owners that are not in the project's members list (I believe this is how we handled project owners before, now we seem to create a "Master" member for the project creator) would see the `Request Access` button.
This MR fixes this issue in a clean way by adding a new `:request_access` ability to replace an ugly helper.
It also give project owners the ability to update & destroy a requester that would happen to be themselves (since owners could request access to their own project before this MR).
Related to #18790.
See merge request !5091
Add lib/gitlab/emoji.rb instead of gitlab_emoji gem and upgrade Gemojione
- No reason to split it into a separate gem when the gem barely did anything. We can use gemojione directly, making updating gemojione that much easier. Unless there's a particularly good reason we were using the gem?
- Fixes the Rake task since it broke after all the AwardEmoji changes.
- Update gemojione to 2.6.1.
Spring Update changes! http://emojione.com/releases/#2.2.0
See merge request !4919
- Group / project members cannot request access
- Group members cannot request access to a group's project
This addresses an issue where project owners could request access
to their own project, leading to UI inconsistency where their requester
status would replace their owner status.
Signed-off-by: Rémy Coutable <remy@rymai.me>
Exclude requesters from Project#members, Group#members and User#members
## What does this MR do?
It excludes requesters from the `Project#members`, `Group#members` and `User#members` associations, and adds new `Project#requesters` and `Group#requesters` associations.
## Are there points in the code the reviewer needs to double check?
No.
## Why was this MR needed?
Without this, if you call `project.members`, requesters are included in the results! This is at best misleading, and at worst can lead to security issues. By excluding requesters from the `#members` associations, we avoid introducing security inadvertently since you have to call the `#requesters` association explicitly to get requesters.
## What are the relevant issue numbers?
This is something I realized while fixing the security issue #19102.
## Does this MR meet the acceptance criteria?
- [x] I don't think this needs a CHANGELOG since this is an internal change
- Tests
- [x] Added for this feature/bug
- [ ] All builds are passing
- [x] Conform by the [style guides](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CONTRIBUTING.md#style-guides)
- [x] Branch has no merge conflicts with `master` (if you do - rebase it please)
- [x] [Squashed related commits together](https://git-scm.com/book/en/Git-Tools-Rewriting-History#Squashing-Commits)
See merge request !4946
Memoize the maximum access level for the author of notes
Cache the maximum access level for each user in a map in the controller
In #19273, we saw that retrieving ProjectTeam#human_max_access for each note takes the bulk of the time when rendering certain issues or merge requests. We observe that most of the comments in an issue are typically done by the same users. This MR memoizes the max access level by user ID.
See merge request !4982
Fixed URL on label button when filtering
## What does this MR do?
Gives the filtered labels the correct URL. Previously they tried to link to `labels#show` whereas now it links to the correct filter path.
## What are the relevant issue numbers?
Closes#19005
See merge request !4897
No reason to split it into a separate gem when the gem barely did anything. We can use gemojione directly, making updating gemojione that much easier.
Also fix the Rake task and update gemojione to 2.6.1. This adds the EmojiOne Spring update.
Changelog: https://github.com/jonathanwiesel/gemojione/blob/master/CHANGELOG.md
In #19273, we saw that retrieving ProjectTeam#human_max_access for each
note takes the bulk of the time when rendering certain issues or merge requests.
We observe that most of the comments in an issue are typically done by the
same users. This MR memoizes the max access level by user ID.
Rémy Coutable
Robert Speicher
Robert Speicher
Annabel Dunstone
Jacob Schatz
Douwe Maan
Alfredo Sumaran
Paco Guzman
Douwe Maan
Drew Blessing
Patricio Cano
Phil Hughes
Grzegorz Bizon
Jacob Vosmaer
Alejandro Rodríguez
Fatih Acet
Connor Shea
Stan Hu
Felipe Artur
Luke "Jared" Bennett