Repository navigation
fix: Shape comparable should include log mode to avoid collisions - #3486
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #3486 +/- ##
==========================================
+ Coverage 67.74% 67.78% +0.04%
==========================================
Files 184 184
Lines 9943 9943
Branches 408 404 -4
==========================================
+ Hits 6736 6740 +4
+ Misses 3205 3201 -4
Partials 2 2
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Wat no test? Could we have some property test that mutates all the keys and asserts the comparable-ness?
I'm also really nervous about this need to keep updating @backup_version... At least some scary comment in Shape.comparable about the dependency (or could we automate it? Calculate the version from the comparable fields or something?)
|
Tests are already broken so they are doing their work :D. Claude-generated fix incoming |
commit: |
✅ Deploy Preview for electric-next ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
I've pushed a fix for the tests. When deleting a shape we were always passing |
Sorry yeah opened the PR last thing before leaving haha just for reference of what the fix is - property test was exactly what I was thinking as well
I'm updating it even without releasing electric just for the sake of consistency of the commits and the need to update it. I agree that automating it somehow would be better. Either way recovery right now is broken anyway (see #3482) I have an idea to not rely on shape comparable in general but will elaborate when back |
d91adfe to
b6a3ff5
Compare
Fixes #3484
What I was worried about happened - the comparable was not updated as we updated the shape. I think we should have a better more solid way to define the comparable.
Also updating this will kind of invalidate existing stored shapes if using shape status backup, but not quite sure what the best way to do that is.
I've invalidated it by bumping the shape status backup version so this works now.