web-platform-tests / interop

web-platform-tests Interop project
https://wpt.fyi/interop
280 stars 28 forks source link

Pointer Events: replace the coalesced event test in Interop 2024 #690

Closed mustaqahmed closed 2 weeks ago

mustaqahmed commented 2 weeks ago

Test List

pointerevents/coalesced_events_attributes_under_load.https.html?mouse

We are renaming the test to .optional.html to correctly reflect the test's scope.

Rationale

This test enforces an optional requirement in the spec. When we included the test in Interop 2024, we missed that one assert wrongly fails the test when the user agent chooses to skip event coalescing. (FYI when run in the WPT server, Chrome can't be forced to coalesce!)

The good news is that we have a different test already to cover the required asserts in the above test. I missed proposing this one originally, sorry:

Possible ways forward for Interop 2024:

FYI @smaug---- @nt1m @nairnandu

EDIT: updated the FYI with correct usernames.

zcorpan commented 2 weeks ago

Also see https://bugzilla.mozilla.org/show_bug.cgi?id=1885287

cc @masayuki-nakano

nt1m commented 2 weeks ago

After checking with @aprotyas, I support swapping the tests.

aprotyas commented 2 weeks ago

+1, let's swap the tests.

smaug---- commented 2 weeks ago

Could we rather just fix the test. There seems to be > 1 check, that should be >=

smaug---- commented 2 weeks ago

...but I can live with renaming

mustaqahmed commented 2 weeks ago

The optional test has been replaced in Interop dashboard. I will close this issue now.

Could we rather just fix the test. There seems to be > 1 check, that should be >=

Sec 10.1 says "As a result, the coalesced events lists for these events always contain at least one event". If this looks questionable, let's discuss in the PEWG.

nt1m commented 2 weeks ago

The optional test has been replaced in Interop dashboard. I will close this issue now.

Could we rather just fix the test. There seems to be > 1 check, that should be >=

Sec 10.1 says "As a result, the coalesced events lists for these events always contain at least one event". If this looks questionable, let's discuss in the PEWG.

@mustaqahmed "at least one event" reads to me as >= 1, not > 1 to me. To me it means "one or more", not "strictly greater than one".

mustaqahmed commented 1 week ago

Sorry my last post was unclear: the WPT seems to follow the spec: assert_greater_than_equal(coalesced_events.length, 1, ...).

Please send a PR to correct whatever I missed.