Summary: | [S60] tabbing navigation skips over text content | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Shaun Keller <shaun.keller> | ||||||||
Component: | Layout and Rendering | Assignee: | Nobody <webkit-unassigned> | ||||||||
Status: | CLOSED FIXED | ||||||||||
Severity: | Normal | CC: | zalan | ||||||||
Priority: | P2 | ||||||||||
Version: | 412 | ||||||||||
Hardware: | S60 Emulator | ||||||||||
OS: | S60 3rd edition | ||||||||||
Attachments: |
|
Description
Shaun Keller
2006-11-22 07:35:05 PST
Created attachment 11608 [details]
tabbing navigation fix
Comment on attachment 11608 [details]
tabbing navigation fix
in getNewTabbedCursorPosition, you use the same variable name (as the parameter of this function) to call the subframe's getNewTabbedCursorPosition. is it intentional? as with this setup your nodeIndex remains -1 even if it is set to something else in the subframe call.
int nodeIndex = -1;
bool foundInFrame = foundFrame->getNewTabbedCursorPosition(lr,tb,nodeIndex);
in tabbedNavigation, it looks to me that nodeIndex can be easily -1 and that crashes on the following line
QRect foundRect = (*_focusableNodeList)[nodeIndex];
and please remove the FIXME comment
Created attachment 11642 [details]
apply corrections based on Zalan's comments
Thanks for catching the nodeIndex errors. The delta is: removed the creation of shadowing variables; added a test for return nodeIndex value of -1 which now skips over changing the cursor position based on a found node; corrected comments and indenting. I rebuilt, tested and reviewed the changes with Sriram.
Created attachment 11643 [details]
retry at usable diff/patch
Comment on attachment 11642 [details]
apply corrections based on Zalan's comments
r=me
Comment on attachment 11643 [details]
retry at usable diff/patch
r=me
3.2 17914 3.1m 17915 Verified and Closed. |