GitLab merge requestshttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests2016-10-06T19:36:26Zhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/326Validate with Git Hooks before accepting a merge request2016-10-06T19:36:26ZGabriel MazettoValidate with Git Hooks before accepting a merge requestWe made possible to add some Git Hook based validation to things like commit message or commit author's email.
When we accept a merge request, it actually goes by the same validations an reject the merge if the merge commit does not f...We made possible to add some Git Hook based validation to things like commit message or commit author's email.
When we accept a merge request, it actually goes by the same validations an reject the merge if the merge commit does not follow Git Hook's predefined validations.
The problem is that it executes these validations when doing the very hard work of manipulating git repository, and cannot throw a usable and readable error to the user.
We are now executing these validations before scheduling git hard-work operations, and saving the error to display for the end-user.
Fixes #468, #3058.7Robert SpeicherRobert Speicherhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/528CE to EE for 8.10.0-rc12016-10-06T19:40:04Zusername-removed-128633CE to EE for 8.10.0-rc1Unresolved conflicts:
- `app/views/projects/_home_panel.html.haml`
- `app/views/shared/_clone_panel.html.haml`
- `app/helpers/projects_helper.rb`
@iamphill Could you please resolve the conflicts introduced by gitlab-org/gitlab-ce...Unresolved conflicts:
- `app/views/projects/_home_panel.html.haml`
- `app/views/shared/_clone_panel.html.haml`
- `app/helpers/projects_helper.rb`
@iamphill Could you please resolve the conflicts introduced by gitlab-org/gitlab-ce!4989 in `app/views/projects/_home_panel.html.haml` and check that project headers look good? Thanks in advance!
@patricio Could you please resolve the conflicts introduced by gitlab-org/gitlab-ce!4696 in `app/helpers/projects_helper.rb` and `app/views/shared/_clone_panel.html.haml`? Thanks in advance!
I need this to be able to release 8.10.0-rc1!8.10username-removed-128633username-removed-128633https://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/531Prevent stale data in LDAP group sync last owner check2016-10-06T19:40:10ZDrew BlessingPrevent stale data in LDAP group sync last owner checkRepresenting https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/529
Fixes gitlab-org/gitlab-ce#17764
Turns out that switching to an AR association has resolved the issue. However, a merge problem from CE to EE caused the owners...Representing https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/529
Fixes gitlab-org/gitlab-ce#17764
Turns out that switching to an AR association has resolved the issue. However, a merge problem from CE to EE caused the owners method to not be removed so it took precedence over the association. This removes that method and adds a test to ensure we don't have stale owner data in the future. 8.10Robert SpeicherRobert Speicherhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/556Bump missing gitlab-license version2016-10-06T19:40:23ZStan HuBump missing gitlab-license versionhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/563WIP: Fix of 'Commits being passed to custom hooks are already reachable when ...2016-10-06T19:40:55ZValery SizovWIP: Fix of 'Commits being passed to custom hooks are already reachable when using the UI'Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/19771
This MR changes the way how we create commit, we create it directly in the branch and we don't use temporary ref anymore. This also works similar to how `git-receive-pack` works...Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/19771
This MR changes the way how we create commit, we create it directly in the branch and we don't use temporary ref anymore. This also works similar to how `git-receive-pack` works so it's a natural way to create commit.
- [x] Replace reference to gitlab_git gem in Gemfile when one released with our change (https://gitlab.com/gitlab-org/gitlab_git/merge_requests/102)
- [x] Specs
- [x] Fix CreateBranchService
- [x] Back-port to CE (https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/5337)8.11Valery SizovValery Sizovhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/580Display the JSON payload in settings page for EE usage ping2016-10-06T19:40:57ZStan HuDisplay the JSON payload in settings page for EE usage pingScreenshot:
![image](/uploads/b85639c029c2caa065aa27aa8220ffb3/image.png)
Screenshot:
![image](/uploads/b85639c029c2caa065aa27aa8220ffb3/image.png)
8.10Douwe MaanDouwe Maanhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/587Refactor Slack service documentation2016-10-06T19:41:01ZAchilleas PipinellisRefactor Slack service documentation8.10Achilleas PipinellisAchilleas Pipinellishttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/592Fix pagination on search result page when ES search is enabled2016-10-06T19:41:06ZValery SizovFix pagination on search result page when ES search is enabledFixes https://gitlab.com/gitlab-org/gitlab-ee/issues/811Fixes https://gitlab.com/gitlab-org/gitlab-ee/issues/8118.10username-removed-128633username-removed-128633https://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/599Fix regression in Git Annex permission check2016-10-06T19:41:09ZStan HuFix regression in Git Annex permission checkThe `changes` in git-annex-shell mean a different thing than git-receive-pack. Revert to the previous code and use a simple check whether Git Annex is enabled.The `changes` in git-annex-shell mean a different thing than git-receive-pack. Revert to the previous code and use a simple check whether Git Annex is enabled.8.10Stan HuStan Huhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/601Change `LdapGroupSync` worker to use new Group Sync classes2016-10-06T19:41:10ZDrew BlessingChange `LdapGroupSync` worker to use new Group Sync classesThis was missed before - somehow :disappointed: The old `group_sync` class was left and the worker still pointed to it. Unfortunately, this means that 8.11 shipped with the new modular code but it is not in use. In the interest of not in...This was missed before - somehow :disappointed: The old `group_sync` class was left and the worker still pointed to it. Unfortunately, this means that 8.11 shipped with the new modular code but it is not in use. In the interest of not introducing a possible regression, we may as well wait until 8.11.
Douwe MaanDouwe Maanhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/605Fix of `Commit search breaks for some URLs on gitlab-ce project`2016-10-06T19:41:14ZValery SizovFix of `Commit search breaks for some URLs on gitlab-ce project`Fixes https://gitlab.com/gitlab-org/gitlab-ee/issues/835Fixes https://gitlab.com/gitlab-org/gitlab-ee/issues/8358.10username-removed-128633username-removed-128633https://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/606Add migration guide for 8.112016-10-06T19:41:16ZRubén DávilaAdd migration guide for 8.118.11Robert SpeicherRobert Speicherhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/598ES common index2016-10-06T19:41:16ZValery SizovES common indexThis will allow as to use parent/child relationship in ES. Related to https://gitlab.com/gitlab-org/gitlab-ee/issues/375
- [x] Fix settings overlapping
- [x] Specs
- [x] Test everything
- [x] Change rake tasks
- [x] Change index ...This will allow as to use parent/child relationship in ES. Related to https://gitlab.com/gitlab-org/gitlab-ee/issues/375
- [x] Fix settings overlapping
- [x] Specs
- [x] Test everything
- [x] Change rake tasks
- [x] Change index settings for all models
- [x] Update the install doc
- [x] Update the update doc (I created https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/5545)
- [x] Prepare text for release blog post
Proposed addition to releasepost:
We totally changed the structure of Elasticsearch index this is why we need to remove whole index and build new one. To be able to use parent/child relationship in ES we decided to move everything in a single index. This solution has some drawback - if we decide to change the type of some existing field in ES we will need to rebuild whole index for every entity. So it makes it harder to support but we expect to have more performance advantages with parent/child relationships.
**Migration**
Remove old indexes
```
curl -XDELETE 'http://localhost:9200/_all/'
```
Build new indexes as described in [Elasticsearch integration](../integration/elasticsearch.md#add-gitlabs-data-to-the-elasticsearch-index)
8.11Valery SizovValery Sizovhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/608Hide SSH repository mirroring2016-10-06T19:41:20ZNick ThomasHide SSH repository mirroring## What does this MR do?
Removes reference to `ssh://` URLs from the repository mirror settings page
## Are there points in the code the reviewer needs to double check?
- [ ] Is the screenshot now used in the documentation an appr...## What does this MR do?
Removes reference to `ssh://` URLs from the repository mirror settings page
## Are there points in the code the reviewer needs to double check?
- [ ] Is the screenshot now used in the documentation an appropriate resolution?
- [ ] Do the remaining screenshots need re-doing to match?
Repository mirroring is also referenced from the 'Import project by URL' functionality, where much of the text of `_instructions.html.haml` is duplicated in a separate partial. It doesn't mention `ssh://`, though.
## Why was this MR needed?
Mirroring repositories over SSH needs more work before it's generally useful.
Currently, these URLs are interpreted with the implicit context of the SSH client configuration for the GitLab installation's `git` user, particularly `known_hosts` and any SSH private keys it has access to. Unless the user manually alters these details - especially `known_hosts` - then `ssh://` URLs just don't work.
We don't wish to support and document manual SSH configuration management, and the naive approach (a single `id_rsa` file used when mirroring all repositories) is insecure - especially for GitLab.com or similar deployments.
Future MRs will re-add documentation for `ssh://` URLs for both password and public-key authentication once `known_hosts` and unique SSH keys for repository mirroring are handled automatically.
Disabling support for `ssh://` URLs entirely would break existing customer setups.
## What are the relevant issue numbers?
#98 #621
## Screenshots (if relevant)
![Screen_Shot_2016-08-02_at_10.58.23](/uploads/940bce89b6e8cc109f5df766cd287c7d/Screen_Shot_2016-08-02_at_10.58.23.png)Douwe MaanDouwe Maanhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/629CE upstream2016-10-06T19:41:32ZValery SizovCE upstream8.11Valery SizovValery Sizovhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/637Used cached value of project count to reduce DB load2016-10-06T19:41:36ZStan HuUsed cached value of project count to reduce DB loadRequires gitlab-org/gitlab-ce!5746
Possible fix for gitlab-com/infrastructure#303Requires gitlab-org/gitlab-ce!5746
Possible fix for gitlab-com/infrastructure#3038.10username-removed-128633username-removed-128633https://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/639Removed geo_node_key_id index2016-10-06T19:41:37Zyorickpeterse-stagingRemoved geo_node_key_id indexThis index does not appear to be used, see https://gitlab.com/gitlab-org/gitlab-ce/issues/20767 for more information.
cc @brodock This index does not appear to be used, see https://gitlab.com/gitlab-org/gitlab-ce/issues/20767 for more information.
cc @brodock 8.11Valery SizovValery Sizovhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/652Run Sidekiq in inline mode for Elastic Search related specs2016-10-06T19:41:51ZRubén DávilaRun Sidekiq in inline mode for Elastic Search related specsStan HuStan Huhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/641Fix race condition with UpdateMirrorWorker Lease2016-10-06T19:41:52ZPaco GuzmanFix race condition with UpdateMirrorWorker LeaseAny time we modify the project import state to an in progress we're going to execute the sidekiq job not lease to try to obtain in this workflow. This fixes the UI problem when an user try to update a mirror just inside a lease timeout.
...Any time we modify the project import state to an in progress we're going to execute the sidekiq job not lease to try to obtain in this workflow. This fixes the UI problem when an user try to update a mirror just inside a lease timeout.
Closes #748
/cc @jacobvosmaer-gitlab @stanhu Robert SpeicherRobert Speicherhttps://staging.gitlab.com/gitlab-org/gitlab/-/merge_requests/647Optimize commit and diff changes access check to reduce git operations2016-10-06T19:41:59Zusername-removed-367626Optimize commit and diff changes access check to reduce git operationsGit operations are costly. Before, if file locks and all file locks were
enabled we would iterate over each commit twice and over each diff as much
as 4 times. This updates uses lambdas to go through all neccessary validations
in only...Git operations are costly. Before, if file locks and all file locks were
enabled we would iterate over each commit twice and over each diff as much
as 4 times. This updates uses lambdas to go through all neccessary validations
in only one iteration per commit and per diff. It also removes some unused
code in lib/gitlab/git_access.rb and adds examples to ensure the code keeps
working as intended.
Related to https://gitlab.com/gitlab-org/gitlab-ce/issues/14280
Valery: Closes https://gitlab.com/gitlab-org/gitlab-ee/issues/8808.11yorickpeterse-stagingyorickpeterse-staging