You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After which a very large remote pull takes place, sometimes. This can lead to the run taking 2 minutes at the short version and up to 40 minutes in our case for the complete pull.
The thought is that perhaps this should be using --depth=1 in some cases (rather than --unshallow)
I think I might have been using --unshallow unnecessarily for some fetches. I've made change and added it to the v6 release candidate. Please give it a try and let me know if it's an improvement. You can use it like this:
- uses: peter-evans/create-pull-request@v6-rc
I'll be releasing v6 at some point in the near future.
Subject of the issue
My workflow is straightforward, containing this step:
This workflow can run multiple times and force pushes; so it may update an existing branch/pr. When run, I see this line executed:
After which a very large remote pull takes place, sometimes. This can lead to the run taking 2 minutes at the short version and up to 40 minutes in our case for the complete pull.
The thought is that perhaps this should be using
--depth=1
in some cases (rather than--unshallow
)Looking at
create-pull-request/src/git-command-manager.ts
Line 119 in f5ba48a
Which is called from
https://github.com/peter-evans/create-pull-request/blob/main/src/create-or-update-branch.ts#L189
The code in there is nuanced enough that it's not clear why we need to do a full fetch in the case that we are pushing to an existing PR.
Steps to reproduce
The text was updated successfully, but these errors were encountered: