test: prefer expected failures to skips in WPT status files - #65311
test: prefer expected failures to skips in WPT status files#65311avivkeller wants to merge 1 commit into
Conversation
Signed-off-by: avivkeller <me@aviv.sh>
|
Review requested:
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #65311 +/- ##
==========================================
- Coverage 90.13% 90.13% -0.01%
==========================================
Files 752 752
Lines 251568 251568
Branches 47270 47286 +16
==========================================
- Hits 226759 226757 -2
+ Misses 16168 16154 -14
- Partials 8641 8657 +16 🚀 New features to boost your workflow:
|
|
I don't see the point in unskipping
In general I trust the maintainers of those systems to know better and to curate their WPT expectations and wouldn't cause this churn which is also a PITA when backporting. |
|
I'm happy to close, I figured this would be easier for the respective maintainers to identify unexpected successes or failures instead of skips |
|
Bump on the above @nodejs/web-standards :-) |
panva
left a comment
There was a problem hiding this comment.
doing this for tentative tests that are undergoing their spec's development is going to cause hassle when it comes to WPT update automation. I'm -1 on just unskipping all of this.
As titled. By avoiding
skiping web platform tests (only when possible), and instead expecting failures, we can better detect when our WPT compatibility changes. For instance, in this PR, several previously skipped WPT files actually pass, and are now runcc @panva