Could you confirm there are no runtime issues with GitLab?
Details
Details
Diff Detail
Diff Detail
- Repository
- rP FreeBSD ports repository
- Lint
No Lint Coverage - Unit
No Test Coverage - Build Status
Buildable 9273 Build 9727: arc lint + arc unit
Event Timeline
Comment Actions
I can confirm this update will break GitLab. :D
If you want to check the easy way do:
cd /usr/ports/www/gitlab && make patch
Then go into WRKSRC and checkout the Gemfile. For rubygem-oj (and sidekiq) there is a pessimistic operator requesting 2.17 versions.
Maybe it will work with the new version. In this case you need to incorporate a patch of the Gemfile. But this will need some more testings and i'm to short on time for this. There are some more typical GitLab-users to add for a test. If you update the patch i will assign them for review.
Comment Actions
Ah yes, I should have just checked the Gemfile before having even considered the run-time tests. I see the latest Gemfile uses version 5.0.0 of sidekiq, but still forces the older version of oj. Let's just wait a bit and see if upstream makes this job easier. Thanks.
Comment Actions
You're welcome! :)
But also note: the latest Gemfile isn't in use. We have 8.16 in ports, while afterwards follows 8.17, 9.0 and 9.1. And next week 9.2...
I'm currently working at 8.17, but the GitLab port sadly lacks behind, nearly always. Its a bunch of work and test.
If you need to update the gems, just create an old repo-copy and use them in GitLab. This is already done multiple times.
Comment Actions
I understood there would be some delays until the code I referenced hit the ports tree. I don't think there is a rush to push the latest sidekiq and oj to the ports tree just yet. The reason I considered the update was for the upcoming net-im/mastodon port, but there are some other things that need to be settled there first.