RESOLVED FIXED 97919
[chromium] Add fields to WebRenderingStats to track frequency of impl-thread vs. main-thread scrolling 
https://bugs.webkit.org/show_bug.cgi?id=97919
Summary [chromium] Add fields to WebRenderingStats to track frequency of impl-thread ...
Terry Anderson
Reported 2012-09-28 08:59:43 PDT
Add |bool didImplThreadScroll| to WebRenderingStats.h which will be true if and only if we have impl-thread scrolled.
Attachments
Patch (2.44 KB, patch)
2012-09-28 09:01 PDT, Terry Anderson 		no flags
DetailsFormatted DiffDiff
Patch (2.29 KB, patch)
2012-10-01 12:28 PDT, Terry Anderson 		no flags
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Terry Anderson
Comment 1 2012-09-28 09:01:11 PDT
Created attachment 166261 [details] Patch
WebKit Review Bot
Comment 2 2012-09-28 09:03:58 PDT
Please wait for approval from abarth@webkit.org, dglazkov@chromium.org, fishd@chromium.org, jamesr@chromium.org or tkent@chromium.org before submitting, as this patch contains changes to the Chromium public API. See also https://trac.webkit.org/wiki/ChromiumWebKitAPI.
Terry Anderson
Comment 3 2012-09-28 09:20:08 PDT
Corresponding chromium patch: https://codereview.chromium.org/10985082/
James Robinson
Comment 4 2012-09-28 12:30:36 PDT
Comment on attachment 166261 [details] Patch A bool isn't a good way to capture this. What's the use case? What about keeping counts of how many scroll-type input events resulted in scrolling vs ignoring vs sending to the main thread?
Terry Anderson
Comment 5 2012-10-01 12:28:32 PDT
Created attachment 166526 [details] Patch
Terry Anderson
Comment 6 2012-10-01 12:31:28 PDT
(In reply to comment #4) > (From update of attachment 166261 [details]) > A bool isn't a good way to capture this. What's the use case? What about keeping counts of how many scroll-type input events resulted in scrolling vs ignoring vs sending to the main thread? We want to use this data to be able to track the frequency of impl-thread vs. main-thread scrolling, so I have added two counters instead of using a bool. The cumulative total of both counters will be reported each time renderingStats() is called.
James Robinson
Comment 7 2012-10-01 15:49:10 PDT
Comment on attachment 166526 [details] Patch R=me We should figure out a way to route stats like this without involving WebKit patches
WebKit Review Bot
Comment 8 2012-10-02 09:49:36 PDT
Comment on attachment 166526 [details] Patch Clearing flags on attachment: 166526 Committed r130180: <http://trac.webkit.org/changeset/130180>
WebKit Review Bot
Comment 9 2012-10-02 09:49:39 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.