Summit2008/Sessions/Proposals/Platform/WebPerformanceTools: Difference between revisions

Summit2008/Sessions/Proposals/Platform/WebPerformanceTools (view source)

Revision as of 04:50, 25 July 2008

419 bytes added , 25 July 2008

no edit summary

Dbaron

canmove, Confirmed users, Bureaucrats and Sysops emeriti

1,334

edits

@@ Line 23: / Line 23: @@
 * ability to categorize memory usage by frame/page/site and by type of allocation (images, JS, DOM nodes, etc.) [dbaron]
+** allows developers to get a better picture of what's going on and how we can improve memory usage
+*** a way to get useful data from users who report Firefox using too much memory
+** allows users to "blame" pages in addition to the browser as a whole
+*** can lead to feedback to both page authors (page uses too much memory) and browser developers (what page X does shouldn't use this much memory)
 * standard profiling UI for JS [dbaron]
-* profiling UI that can explain time spent on things like style resolution, layout, etc., relative to what nodes are being laid out (and what JS triggered that work, even if it was asynchronous) [dbaron]
+** venkman UI plus callers/callees?
-* profiling UI that can highlight thrashing caused by functions requiring flushes [dbaron]
+* profiling UI that can point out things that we know are issues [dbaron]
-* profiling UI that can highlight the problems that are causing the event queue not to be serviced for a long time (i.e., show the performance problems in the longest events) [dbaron]
+** explain time spent on things like style resolution, layout, etc., relative to what nodes are being laid out (and what JS triggered that work, even if it was asynchronous)
+** highlight thrashing caused by functions requiring flushes
+** highlight the problems that are causing the event queue not to be serviced for a long time (i.e., show the performance problems in the longest events)