Bug 208285 - Remove initial layout throttle
Summary: Remove initial layout throttle
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: zalan
URL:
Keywords: InRadar
: 204031 (view as bug list)
Depends on: 208536
Blocks:
  Show dependency tree / graph
 
Reported: 2020-02-26 21:27 PST by Ben Nham
Modified: 2020-03-05 10:15 PST (History)
13 users (show)

See Also:

Attachments
Patch (17.80 KB, patch)
2020-02-26 21:35 PST, Ben Nham 		no flags Details | Formatted Diff | Diff
Patch (19.30 KB, patch)
2020-02-27 23:48 PST, Ben Nham 		no flags Details | Formatted Diff | Diff
Patch (27.17 KB, patch)
2020-03-03 10:18 PST, zalan 		no flags Details | Formatted Diff | Diff
Patch (19.26 KB, patch)
2020-03-04 08:55 PST, zalan 		no flags Details | Formatted Diff | Diff
Patch (19.75 KB, patch)
2020-03-04 09:02 PST, zalan 		no flags Details | Formatted Diff | Diff
Show Obsolete (4) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Ben Nham 2020-02-26 21:27:14 PST 
Remove initial layout throttler
Comment 1 Radar WebKit Bug Importer 2020-02-26 21:34:44 PST 
<rdar://problem/59835321>
Comment 2 Ben Nham 2020-02-26 21:35:17 PST 
Created attachment 391837 [details]
Patch
Comment 3 Antti Koivisto 2020-02-26 22:41:07 PST 
Comment on attachment 391837 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=391837&action=review

> Source/WebCore/page/SettingsBase.cpp:-67
> -// This amount of time must have elapsed before we will even consider scheduling a layout without a delay.
> -// FIXME: For faster machines this value can really be lowered to 200. 250 is adequate, but a little high
> -// for dual G5s. :)
> -static const Seconds layoutScheduleThreshold = 250_ms;

My dual G5s are going to fly!
Comment 4 Ben Nham 2020-02-27 09:51:29 PST 
*** Bug 204031 has been marked as a duplicate of this bug. ***
Comment 5 Ben Nham 2020-02-27 23:48:34 PST 
Created attachment 391966 [details]
Patch
Comment 6 Ben Nham 2020-02-28 09:40:51 PST 
Alan said he could take a look at the layout test issues here, so assigning to him. I already put in a couple workarounds for LayoutTests in the latest patch for ios-wk2:

 - tiled-drawing-async-frame-scrolling.html: the third iframe wasn't in the render tree, which was causing this to fail; worked around this by waiting for the entire window and all iframes to load before starting the test
 - autoscroll-in-iframe-body.html: I think this was issuing a click event in an iframe before the iframe had rendered

There are also a couple of mac-wk2 layout tests that are failing now that are somewhat similar (they attempt to move/click the mouse after an onload event).

As Alan said, we should probably remove the workarounds and make sure we fully understand the implications of removing the layout timer.
Comment 7 zalan 2020-03-02 21:28:18 PST 
LayoutTests/compositing/tiling/tiled-drawing-async-frame-scrolling.html is fixed at r257752
Comment 8 zalan 2020-03-03 10:18:07 PST 
animations/cross-fade-list-style-image.html is fixed at r257764
Comment 9 zalan 2020-03-03 10:18:40 PST 
Created attachment 392291 [details]
Patch
Comment 10 zalan 2020-03-03 19:41:49 PST 
fast/images/animated-gif-no-layout.html is fixed at r257811
Comment 11 zalan 2020-03-03 20:27:45 PST 
fast/events/autoscroll-in-iframe.html fails on trunk (run locally)
Comment 12 zalan 2020-03-04 08:55:01 PST 
Created attachment 392417 [details]
Patch
Comment 13 zalan 2020-03-04 09:02:04 PST 
Created attachment 392418 [details]
Patch
Comment 14 WebKit Commit Bot 2020-03-04 11:05:02 PST 
Comment on attachment 392418 [details]
Patch

Clearing flags on attachment: 392418

Committed r257862: <https://trac.webkit.org/changeset/257862>
Comment 15 WebKit Commit Bot 2020-03-04 11:05:04 PST 
All reviewed patches have been landed.  Closing bug.
Comment 16 Jacob Uphoff 2020-03-05 10:15:55 PST 
The changes made here caused https://bugs.webkit.org/show_bug.cgi?id=208653