cache manual results only for pushes #120555
Conversation
| permissions: read-all | ||
|
|
There was a problem hiding this comment.
Let's keep these: these jobs have no need for more than read.
| permissions: read-all | ||
|
|
| with: | ||
| # explicitly enable sandbox | ||
| extra_nix_config: sandbox = true |
There was a problem hiding this comment.
It's the default value, so it provides no benefit. It would also break macOS in case we'd enable it.
| push: | ||
| pull_request: |
There was a problem hiding this comment.
Is this right? It seems weird that an empty body for an on value would fall through to the next, but I'd believe it if you said so :)
There was a problem hiding this comment.
Yes. This is correct syntax. They are not evaluated in order but any of them can trigger the job. They are empty because they could take further parameter like branch lists.
There was a problem hiding this comment.
Currently this change will start building the manuals on future release branches, would need to be updated to keep the original behaviour of only building master.
push:
branches:
- master
Edit: didn't refresh before posting.
I don't understand what this means or what you're trying to achieve here, can you explain in more detail? |
|
I guess the motivation is that it would upload too much rebuild manual data to cachix? |
| on: | ||
| pull_request_target: | ||
| push: | ||
| pull_request: | ||
| branches: | ||
| - master |
There was a problem hiding this comment.
Ah this only happens for master anyway...
| on: | ||
| pull_request_target: | ||
| push: |
There was a problem hiding this comment.
Maybe we only want to run this on master branch than.
There was a problem hiding this comment.
Why not also for releases/staging?
There was a problem hiding this comment.
Release branch doesn't really see that many changes to it's docs/options after branch off. IMO hydra is sufficient for stable, it's unstable where the docs breakage was reasonably common.
PRs targeting staging would likely timeout frequently with mass rebuilds. Any errors missed in the initial PR on staging are caught by the action in the staging-next -> master PR.
There was a problem hiding this comment.
Yes. Mass-rebuilds is something you want to avoid in github actions. Especially without caching.
I don't see much benefit for pushing to the cache for PRs.