-
Notifications
You must be signed in to change notification settings - Fork 32
AutoMerge: switch to RegistryInstances #643
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Conversation
Codecov Report❌ Patch coverage is ❌ Your project check has failed because the head coverage (51.20%) is below the target coverage (70.00%). You can increase the head coverage or adjust the target coverage.
Additional details and impacted files@@ Coverage Diff @@
## master #643 +/- ##
===========================================
- Coverage 78.40% 51.20% -27.21%
===========================================
Files 29 29
Lines 1936 1914 -22
===========================================
- Hits 1518 980 -538
- Misses 418 934 +516 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
So, if I understand correctly, you're only using RegistryInstances (and the packed registry) for the upstream registry head, right? For the checked-out PR branch, you're still using the unpacked registry that we have checked out on disk? Do I have that correct? |
|
no, I'm using RegistryInstances for both. It can handle both packed and unpacked registries |
|
@GunnarFarneback I know you might not have time for a full review, but do you have any thoughts on this approach? |
|
I think it's generally fine and simplifies some of the registry analysis code. |
this ends up kind of undoing #614 because it seemed easier to use RegistryInstances everywhere. Now we can use packed registries everywhere except the RegistryCI.jl-based check which needs to support older Julia's. In particular, this should make it more feasible to set up a "local" version that users can use without needing a unpacked registry.
written by claude-code but with a very annoying amount of iteration, probably it would have been faster to do it myself.
The most gnarly checks here are the deps and julia compat, and I tested that we got the same results on the latest release of all 11k packages (this did reveal a bug in the initial implementation, in which there was only a julia dep and no other deps). I also included some tests with the old version on a few packages to show they are the same, but we could delete those tests before merging if we want.