summaryrefslogtreecommitdiffstats
path: root/net/websockets
diff options
context:
space:
mode:
authordarin@chromium.org <darin@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2009-12-07 04:15:39 +0000
committerdarin@chromium.org <darin@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2009-12-07 04:15:39 +0000
commitb36d9baa98e397fd194716a6041083ed989524c8 (patch)
tree335b3e294a41f795af59f72b1fc1cc0c6f171752 /net/websockets
parent6e41eae8b3ea421a67104fe1849d5f261a4f2b20 (diff)
downloadchromium_src-b36d9baa98e397fd194716a6041083ed989524c8.zip
chromium_src-b36d9baa98e397fd194716a6041083ed989524c8.tar.gz
chromium_src-b36d9baa98e397fd194716a6041083ed989524c8.tar.bz2
Coalesce damage rects that share an edge.
This fixes the Bloat HTTP page cycler regression, and allows me to remove the hack I checked in to disable multiple-paints due to an observed DHTML page cycler regression. I added a new histogram, RW_IntermediatePaintRectCount, that would have shown the problem clearly had it been there before. I included some cleanup in PaintAggregator: 1- rename "r" to "existing_rect" for clarity 2- check for contained (i.e., redundant) invalidates up front I also changed the opacity of the paint rects used when --show-paint-rects is specified. I find myself dogfooding with this command line option enabled, and I want it to be a little less annoying but still just as useful. R=brettw BUG=29477 TEST=none Review URL: http://codereview.chromium.org/464057 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@33949 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'net/websockets')
0 files changed, 0 insertions, 0 deletions