Age | Commit message (Collapse) | Author |
|
|
|
We usually use it for security fixes that don't need to be ported to EE
and the task seems to hang when run against the security branch.
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
Fix BASH usage in the .gitlab-ci.yml
As discussed on Slack. Even though I created the fix, I was a mere code monkey in this case.
See merge request !7693
|
|
Notify broken master on #development
See merge request !7593
|
|
|
|
|
|
Add node_modules cache to npm related builds
See merge request !7548
|
|
This reverts commit 23fc7875a838b072b8869efada6f68a3a19fc9a6.
|
|
|
|
|
|
|
|
|
|
AKA:
- move all conversations to #core
- beating will continue until morale improves
- Learn by pain
- Pain driven development
|
|
prepare stage
|
|
|
|
|
|
ee_compat_check task
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
[ci skip]
|
|
|
|
Add ESLint
## What does this MR do?
- create `package.json`
- add ESLint dependencies to `package.json`
- add JavaScript linting to CI
- add Rake task `lint:javascript`as alias for `eslint` (which itself is an alias for `npm run eslint`)
## Are there points in the code the reviewer needs to double check?
Probably not.
## Why was this MR needed?
My hovercraft was full of eels.
## What are the relevant issue numbers?
- https://gitlab.com/gitlab-org/gitlab-ce/issues/13224#note_12537431
- https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/5006#note_13255658
See merge request !5445
|
|
We don't need to run `ee_compat_check` against `master`
See merge request !7119
|
|
Add a trigger variable to docs trigger job
See merge request !7089
|
|
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
|
|
|
|
Fix the `migration paths` CI job
## What does this MR do?
It fixes the `migration paths` CI job. This job started failing at the time we added a migration that is actually using Redis.
The issue was that the `config/resque.yml` file we're using is not back-compatible with the `v8.5.9` version we're checking out in the first part of the job, that's why we're using a UNIX socket for this version.
To fix the issue:
* We use 'SETUP_DB: 'false'` in the job variables to speed up the job (it does not make sense to run migration before checking out v8.5.9)
* We just copy config/resque.yml.example to config/resque.yml at is simply has correct structure when checked out from v8.5.9
* When we revert to CI_BUILD_REF, we run prepare script again, which swaps config/resque.yml again, using new version this time.
Fixes #23658.
Note: the `migration paths` CI job is run on `master` only but you can see a successful build here: https://gitlab.com/gitlab-org/gitlab-ce/commit/fd99ec1a8ecc49ef70c4358ca7be2987f2a7505f
See merge request !7092
|
|
|
|
|
|
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
|
|
|
|
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
|
|
|
|
|
|
|
|
Do not run before_script, artifacts, cache in trigger_docs job
Fixes
See merge request !6866
|
|
|
|
|
|
Also add a safeguard for non-CI env.
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
Signed-off-by: Rémy Coutable <remy@rymai.me>
|
|
Keep a copy of gems needed by our releases as CI artifacts
This is meant as insurance in case a gem GitLab needs is ever removed
from rubygems.org. After we merge this, all future GitLab tags will
have a CI artifact archive containing all the gems (for all platforms)
specified in Gemfile.lock for that revision.
Only runs on tags (so we don't constantly create 50MB gem cache bundles).
See merge request !6617
|
|
|
|
Add a bundle check step to ensure dependencies are correct
This should help prevent merge issues in the future, which caused !6814 to be needed.
See merge request !6815
|
|
This reverts merge request !6730
|
|
|
|
This should help prevent merge issues in the future, which caused !6814
to be needed.
|
|
|
|
|
|
|