Bug 196394 – Add more tests for ios iframe scrolling with fixed/sticky layers

• Component: UI Events
• Status: NEW

[Frédéric Wang (:fredw)]

I'll import the remaining tests I wrote for ios frame scrolling. Not sure why commits fixed the issues I had, but it seems they all pass now https://bugs.webkit.org/show_bug.cgi?id=194433#c14

[Frédéric Wang (:fredw)]

Created attachment 366262 [details]
Patch

[Antti Koivisto]

Comment on attachment 366262 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=366262&action=review

> LayoutTests/fast/scrolling/ios/scroll-iframe-005.html:29
> +          // This checks scrolling to the location of the green square.
> +          var c = centerOf("positionFixed");
> +          await touchAndDragFromPointToPoint(c.x, c.y, c.x - 150, c.y - 150);
> +          await liftUpAtPoint(c.x - 150, c.y - 150);
> +
> +          // Wait for scrolling to stabilize and for scrollbars to disappear.
> +          setTimeout(() => {testRunner.notifyDone(); }, 1000);

Can you make these use UIHelper.immediateScrollElementAtContentPointToOffset and remove the timeouts? See

LayoutTests/fast/scrolling/ios/overflow-scroll-overlap.html (and LayoutTests/fast/scrolling/resources/overflow-scroll-overlap.js)

for an example.

[Frédéric Wang (:fredw)]

Created attachment 366274 [details]
Additional patch to use immediateScrollElementAtContentPointToOffset

[Frédéric Wang (:fredw)]

Comment on attachment 366262 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=366262&action=review

>> LayoutTests/fast/scrolling/ios/scroll-iframe-005.html:29
>> +          setTimeout(() => {testRunner.notifyDone(); }, 1000);
> 
> Can you make these use UIHelper.immediateScrollElementAtContentPointToOffset and remove the timeouts? See
> 
> LayoutTests/fast/scrolling/ios/overflow-scroll-overlap.html (and LayoutTests/fast/scrolling/resources/overflow-scroll-overlap.js)
> 
> for an example.

Mmh, I tried it (attachment 366274 [details]) but that does not seem to work. Not sure which mistake I made...

[Simon Fraser (smfr)]

Comment on attachment 366262 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=366262&action=review

>>> LayoutTests/fast/scrolling/ios/scroll-iframe-005.html:29
>>> +          setTimeout(() => {testRunner.notifyDone(); }, 1000);
>> 
>> Can you make these use UIHelper.immediateScrollElementAtContentPointToOffset and remove the timeouts? See
>> 
>> LayoutTests/fast/scrolling/ios/overflow-scroll-overlap.html (and LayoutTests/fast/scrolling/resources/overflow-scroll-overlap.js)
>> 
>> for an example.
> 
> Mmh, I tried it (attachment 366274 [details]) but that does not seem to work. Not sure which mistake I made...

I think it's worth trying. You're adding 2s of tests, and that's not a sustainable rate.

[Frédéric Wang (:fredw)]

(In reply to Simon Fraser (smfr) from comment #5)
> >> LayoutTests/fast/scrolling/ios/overflow-scroll-overlap.html (and LayoutTests/fast/scrolling/resources/overflow-scroll-overlap.js)
> >> 
> >> for an example.
> > 
> > Mmh, I tried it (attachment 366274 [details]) but that does not seem to work. Not sure which mistake I made...
> 
> I think it's worth trying. You're adding 2s of tests, and that's not a
> sustainable rate.

Yes, I agree with that. The thing is that I was not sure whether I made something wrong in my attempt attachment 366274 [details] as the change makes the test then fails when I tried the other day. Not sure why they pass with the other API.

[Frédéric Wang (:fredw)]

(In reply to Simon Fraser (smfr) from comment #5)
> Comment on attachment 366262 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=366262&action=review
> 
> >>> LayoutTests/fast/scrolling/ios/scroll-iframe-005.html:29
> >>> +          setTimeout(() => {testRunner.notifyDone(); }, 1000);
> >> 
> >> Can you make these use UIHelper.immediateScrollElementAtContentPointToOffset and remove the timeouts? See
> >> 
> >> LayoutTests/fast/scrolling/ios/overflow-scroll-overlap.html (and LayoutTests/fast/scrolling/resources/overflow-scroll-overlap.js)
> >> 
> >> for an example.
> > 
> > Mmh, I tried it (attachment 366274 [details]) but that does not seem to work. Not sure which mistake I made...
> 
> I think it's worth trying. You're adding 2s of tests, and that's not a
> sustainable rate.

I'm still not able to make them pass with the internal APIs so I'll move this to a separate bug and handle the other tests here.

[Frédéric Wang (:fredw)]

Created attachment 368235 [details]
Patch

Rebasing on top of bug 197280.

[Frédéric Wang (:fredw)]

Created attachment 368236 [details]
Patch (alternative version using immediateScrollElementAtContentPointToOffset)

With that version, the tests fail for me.

[Frédéric Wang (:fredw)]

Unassigning myself since I'm no longer working on this.