Closed Bug 834589 Opened 12 years ago Closed 11 years ago

Scrolling jank on page due to SelectorMatches and RuleHash:Enumerate

Categories

(Core :: Layout, defect)

21 Branch
x86_64
Windows 7
defect
Not set
normal

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: mayankleoboy1, Unassigned)

References

()

Details

(Keywords: perf, Whiteboard: [Snappy])

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:21.0) Gecko/20130124 Firefox/21.0 Build ID: 20130124054158 Steps to reproduce: 1. Create fresh profile 2. Go to http://www.businessweek.com/articles/2013-01-10/can-meg-whitman-reverse-hewlett-packards-free-fall#r=rss 3. Scroll only a little bit from the top. The main jank is in the top/heading of the page. Actual results: jerky scrolling. Profile : http://people.mozilla.com/~bgirard/cleopatra/#report=96f66898bfc3574c6af2dc20f19415320450b31b I have seen RuleHash:enumerate and SelectorMatches in numerous profiles i have takes of janky pages. Expected results: smoothies.
Tentatively marking Snappy, as I have seen RuleHash:enumerate and SelectorMatches in countless profiles of janky pages.
Keywords: perf
Whiteboard: [Snappy]
Component: Untriaged → Layout
Product: Firefox → Core
Summary: Scrolling jank on http://www.businessweek.com/articles/2013-01-10/can-meg-whitman-reverse-hewlett-packards-free-fall#r=rss → Scrolling jank on page due to SelectorMatches and RuleHash:Enumerate
I am experiencing similar issue with TreeStyleTab when switching between tabs (Fx Beta, 19). CodeAnalyst tells me: > Name Address Self Children Total > nsFrameManager::ReResolveStyleContext 0x5fa4e680 30 2864 2894 > nsRuleNode::WalkRuleTree 0x5fa53670 488 2054 2542 > nsStyleSet::GetContext 0x5fa53040 6 2025 2031 > nsContainerFrame::GetChildLists 0x5fa778c0 25 1534 1559 > SelectorMatches 0x5fa789f0 404 932 1336 > > 5 functions, Total: 953 self samples, 9409 child samples The total column, as I understand is time in milliseconds. I am not able to correctly interpret CodeAnalyst results, tho. However on my relatively fast machine the UI hangs for 1 - 2 seconds and CPU goes to its per-core limit. Could the suspected job (function) be rewritten to work incrementally to let spin the main loop? For my case neither e10s nor content thread would help.
works Ok-ish for me now. closing.
Status: UNCONFIRMED → RESOLVED
Closed: 11 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.