UNCONFIRMED 66615
panlex.org loads slower in WK2 than in WK1 (a page with lots of buttons) 
https://bugs.webkit.org/show_bug.cgi?id=66615
Summary panlex.org loads slower in WK2 than in WK1 (a page with lots of buttons)
Jonathan Pool
Reported 2011-08-19 18:02:07 PDT
Created attachment 104603 [details] web page with 5000-button form WebKit 93384 loads about 70 button elements per second, so a page containing a form with 5000 button elements requires about 75 seconds to load and be ready for navigation. By comparison, Firefox requires under 2 seconds for the same page. WebKits in Safari up through version 5.0 were as fast as other browsers on such pages. Attachment is a simplified example of such a page. The real-world page that made this bug evident to me is at the above URL.
Attachments
web page with 5000-button form (677.81 KB, text/html)
2011-08-19 18:02 PDT, Jonathan Pool 		no flags
Details
View All Add attachment proposed patch, testcase, etc.
Jonathan Pool
Comment 1 2011-08-19 18:03:16 PDT
Bug verified to exist also in Mac OS X 10.7.
Alexey Proskuryakov
Comment 2 2011-08-19 21:17:20 PDT
<rdar://problem/9990319>
Jonathan Pool
Comment 3 2011-10-14 10:29:07 PDT
Bug continues to be present in Safari 5.1.1.
Jonathan Pool
Comment 4 2011-12-02 15:32:07 PST
Bug remains in Safari 5.1.2 and in webkit r101843.
Jonathan Pool
Comment 5 2012-02-11 21:38:00 PST
Further testing has revealed that this bug exists only in Multi-process Windows, not in Single-process Windows.
Jonathan Pool
Comment 6 2012-03-17 12:17:15 PDT
On 17 March 2012 further testing failed to replicate this bug on Webkit r101843. The bug was also not exhibited by several later Webkit builds, or by Safari 5.1.4 on either OS X 10.6.8 or OS X 10.7.3. The test pages appear to load slightly faster in single-process windows than in multi-process windows, but in either case about 20 times as fast as when the bug was in effect. I am not aware of any change in the environment to which the disappearance of this bug could be attributed.
Alexey Proskuryakov
Comment 7 2012-03-18 20:37:29 PDT
Thank you. Safari 5.1.4 improves this case, however the remaining performance difference is still something we should track as a bug. For Apple employees looking at this: Safari 5.1.4 improvement is <rdar://problem/9937149>.
Alexey Proskuryakov
Comment 8 2012-07-23 12:51:15 PDT
<rdar://problem/9837023> is what tracks the remaining regression now internally at Apple.
Note You need to log in before you can comment on or make changes to this bug.