Throttle rendering of detached FrameViews |
||||
Issue descriptionDetached frames are frames whose owner node (<frame> or <iframe>) has display:none or other styles making the owner node not rendered. For now we update almost the whole document cycle for detached FrameViews. Can we use the frame throttling mechanism for detached FrameViews? We also refer FrameViews with null layoutView() as detached. However, I see the detached FrameViews have null frame().ownerLayoutObject() but non-null layoutView(). What's the case of null layoutView()?
,
Apr 8 2016
Saw some layout tests and unit tests testing <iframe style="display: none">. Will run a cluster-telemetry task to see how common they are in 10k sites.
,
Apr 8 2016
,
Apr 9 2016
https://ct.skia.org/chromium_perf_runs/ run 736 For desktop top 10k sites: Total number of frames: 38963 Number of throttled frames: 13622 Number of detached frames: 14823 Number of detached frames that are also throttled: 4661 Based on this data, throttling detached frames will benefit us. Besides performance, we'll also simplify some code handling detached frames, e.g. FrameView::invalidateTreeIfNeededRecursive().
,
Apr 11 2016
Thanks for putting the data together. Seems like an even better idea now :)
,
Apr 11 2017
This issue has been available for more than 365 days, and should be re-evaluated. Please re-triage this issue. The Hotlist-Recharge-Cold label is applied for tracking purposes, and should not be removed after re-triaging the issue. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Apr 11 2017
Note that we accidentally shipped throttling display:none frames in M56 (I think) and had to roll back because it broke some sites.
,
Aug 7 2017
I believe that we can't throttle display:none frames because it will cause major breakage. Sami, am I right? If not, please reopen the bug.
,
Aug 9 2017
That's right, although there's some ongoing spec work to allow that which should ultimately let us do this. |
||||
►
Sign in to add a comment |
||||
Comment 1 by skyos...@chromium.org
, Apr 8 2016