Changes to .tentative tests
Closed this issue · 5 comments
In 15c5ae8 @dizhang168 with review from @tkent-google made quite major changes to a tentative test contributed by WebKit.
The rationale given was that the behavior wasn't standardized. But that's what .tentative means.
Could you please revert this change?
cc @foolip
The original test was only for Selection.direction, which is a new not fully standardized test that Webkit added. I agree it should be kept as tentative.
55dd5fe
However, the part I removed was for an addendum that Gecko made to test multi click (single, double and triple clicks).
3af40b4
04e0f3e
I looked into this and the click behavior is not specified. Because of OS differences, it is also hard to have this standardized:
w3c/selection-api#70 (comment)
w3c/selection-api#177 (comment)
When this patch was landed, I did cc the OP to make sure they are aware and so they can correct me:
#47043 (comment)
If we still feel strong about keeping this part of the test in WPT, I can add it as a separate new file called selection-direction-on-clicks.tentative.html and make it clear that shipping selection.direction doesn't mean this test should become non-tentative.
If the expectation is that it can become non-tentative modulo a set of subtests, putting those subtests in a separate file seems like the best way forward indeed. Unless you have actual confirmation from the writer of the tests that removing them is okay. (Probably best done as a PR against WPT itself with them as reviewer.)
Thanks for the feedback. I have created the PR for it and asked @sefeng211's review: #48267
Note: I agree with you it is not good to remove a test without the original writer's approval. I did cc and gotten their approval in the original PR: #47043 (comment)
Closing this as I have re-added the subtests as "selection/Selection-direction-on-clicks.tentative.html".
Thanks @dizhang168 and apologies for not entirely appreciating the sequence of events here. Glad it was a misunderstanding.