Summary: | [Qt] QtTestBrowser: Prevent calling load() directly from loadFinished() in robot mode. | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Jocelyn Turcotte <jturcotte> | ||||||
Component: | New Bugs | Assignee: | Jocelyn Turcotte <jturcotte> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Enhancement | CC: | ademar, kling, webkit-ews | ||||||
Priority: | P2 | Keywords: | Qt, QtTriaged | ||||||
Version: | 528+ (Nightly build) | ||||||||
Hardware: | Other | ||||||||
OS: | OS X 10.5 | ||||||||
Attachments: |
|
Description
Jocelyn Turcotte
2010-10-18 03:35:59 PDT
Created attachment 71018 [details]
Patch
Comment on attachment 71018 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=71018&action=review Patch looks OK, but this makes me wonder: if starting a new load in direct response to loadFinished(bool) may cause crashes, shouldn't we be fixing that or documenting that fact? > WebKitTools/QtTestBrowser/urlloader.cpp:96 > + --m_numFramesLoading; We should ASSERT(m_numFramesLoading > 0); before this. Created attachment 71142 [details]
Patch
Attachment 71142 [details] did not build on qt: Build output: http://queues.webkit.org/results/4446084 Comment on attachment 71142 [details]
Patch
r=me, but don't forget to add the missing include. ;)
Committed r70044: <http://trac.webkit.org/changeset/70044> Blocking 2.1 to get it cherry-picked. This patch only affect QtTestBrowser and picking it should involve no risk. Please block the candidates master bug, not the release critical (fixed). In practice the chances of being cherry-picked are almost the same - it all depends on the complexity and if it conflicts or not. Revision r70044 cherry-picked into qtwebkit-2.1 with commit 4d3698b <http://gitorious.org/webkit/qtwebkit/commit/4d3698b> |