[Webkit-unassigned] [Bug 191001] REGRESSION(r237218): [GTK] Broke imported/w3c/web-platform-tests/intersection-observer/same-document-root.html

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=191001

--- Comment #1 from Ali Juma <ajuma at chromium.org> ---
Looking through the failures, the more common failure mode seems to be:
--- /home/slave/webkitgtk/gtk-linux-64-debug-tests/build/layout-test-results/imported/w3c/web-platform-tests/intersection-observer/same-document-root-expected.txt
+++ /home/slave/webkitgtk/gtk-linux-64-debug-tests/build/layout-test-results/imported/w3c/web-platform-tests/intersection-observer/same-document-root-actual.txt
@@ -4,6 +4,6 @@
 PASS document.scrollingElement.scrollTop = window.innerHeight. 
 PASS root.scrollTop = 150 with root scrolled into view. 
 PASS document.scrollingElement.scrollTop = 0. 
-PASS root.scrollTop = 0 
-PASS root.scrollTop = 150 with root scrolled out of view. 
+FAIL root.scrollTop = 0 assert_equals: entries.length expected 3 but got 2
+FAIL root.scrollTop = 150 with root scrolled out of view. assert_equals: entries.length expected 4 but got 2

In this case, |root| is an overflow scroller. What's happening is that after we scroll |root|, we're not getting the expected intersection observation by the time the next rAF happens. So either overflow scrolls don't trigger flushes on GTK, or they trigger flushes too late.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20181029/974a3ade/attachment-0001.html>