search

NOAA-FIMS / FIMS

The repository for development of FIMS
https://noaa-fims.github.io/FIMS/
GNU General Public License v3.0
12 stars 8 forks source link

[Developer Issue]: Use GHA to create a pull request to modify pull requests #660

Open Bai-Li-NOAA opened 1 month ago

Bai-Li-NOAA commented 1 month ago

Description

Comments from @kellijohnson-NOAA: should the documentation GitHub action be run during the PR phase rather than after merging into main. We thought it was not worth implementing at this current point in time but if anyone has ideas for how to ensure that clean up is part of the PR rather than after merging into main that would be amazing ⭐!

Bai-Li-NOAA commented 1 month ago

@iantaylor-NOAA thinks that running doc-and-style on PRs makes a lot of sense. It gives PR authors better feedback on their code styling and gives the reviewers a chance to look over changes to the .Rd files and other documentation updated by doc-and-style. He just implemented that for r4ss (via https://github.com/r4ss/r4ss/commit/5be028c1825610d3c61be85b2691afd1735b75f9) and posted an issue about it here: https://github.com/nmfs-fish-tools/ghactions4r/issues/135. There may be some redundant runs of the workflows but he doesn't think that's a big deal for something that runs pretty fast.

I found an use case in the create-pull-request repository that might work for us.

iantaylor-NOAA commented 1 month ago

My naive implementation in r4ss failed (here) with the following error

Checking the base repository state
  /usr/bin/git symbolic-ref HEAD --short
  fatal: ref HEAD is not a symbolic ref
  /usr/bin/git rev-parse HEAD
  7b34b3d68667987f1623a38a[45](https://github.com/r4ss/r4ss/actions/runs/10116442340/job/27979299598#step:19:49)a96207bff0d4ca
  Working base is commit '7b34b3d68667987f1623a38a45a96207bff0d4ca'
  Error: When the repository is checked out on a commit instead of a branch, the 'base' input must be supplied.

I'm guessing that this was related to the workflow needing to create a pull request to the originating branch rather than main but not having the specifications to do that right. The approach used in the use case that @Bai-Li-NOAA linked above seems more promising.