RESOLVED FIXED 99540
Web Inspector: NMI instrument ImageDecoder. It owns a buffer that could be quite big. 
https://bugs.webkit.org/show_bug.cgi?id=99540
Summary Web Inspector: NMI instrument ImageDecoder. It owns a buffer that could be qu...
Ilya Tikhonovsky
Reported 2012-10-16 22:30:01 PDT
EOM
Attachments
Patch (5.71 KB, patch)
2012-10-16 22:33 PDT, Ilya Tikhonovsky 		yurys: review+
buildbot: commit-queue-
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Ilya Tikhonovsky
Comment 1 2012-10-16 22:33:18 PDT
Created attachment 169089 [details] Patch
Build Bot
Comment 2 2012-10-16 22:59:18 PDT
Comment on attachment 169089 [details] Patch Attachment 169089 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/14387436
Build Bot
Comment 3 2012-10-16 23:04:19 PDT
Comment on attachment 169089 [details] Patch Attachment 169089 [details] did not pass win-ews (win): Output: http://queues.webkit.org/results/14393466
Yury Semikhatsky
Comment 4 2012-10-17 00:46:35 PDT
Comment on attachment 169089 [details] Patch Can we add a test for this? Now that there is a framework for writing pure protocol inspector tests it should be quite easy.
Ilya Tikhonovsky
Comment 5 2012-10-17 05:58:23 PDT
(In reply to comment #4) > (From update of attachment 169089 [details]) > Can we add a test for this? Now that there is a framework for writing pure protocol inspector tests it should be quite easy. Looks like it is a tricky task because: a) we have only aggregated statistic and other patches could increase the reported size and thus hide the regression; b) it is not clear for me how I can reproduce the situation with cached bitmap in image decoder; c) a test with both, low and upper limit in LayoutTests will hurt other developers.
Ilya Tikhonovsky
Comment 6 2012-10-17 07:26:19 PDT
Committed r131593: <http://trac.webkit.org/changeset/131593>
Note You need to log in before you can comment on or make changes to this bug.