when a link looks like <a href="about.html">link</a> it implies that the "target" of this link is empty. In IE and firefox, this link will be opened inside current frame. And in webkit, the real target was reset to current document's target name. If this target is "_blank", this link will be opened in a new window( which is not correct).
Thank you for the report! Do you have a test case or an example URL showing the problem? does this affect any live Web sites?
try this: http://club3.kdnet.net/dispbbs.asp?boardid=1&id=6659651 (note it is in Chinese!) Click the page number (at the bottom of the page), it will open a new tab while IE&firefox will open in current tab. I've fix this problem, code as below: @@ -288,8 +288,11 @@ void FrameLoader::urlSelected(const ResourceRequest& request, const String& pass return; String target = passedTarget; - if (target.isEmpty()) + if (target.isEmpty()) { target = m_frame->document()->baseTarget(); + if (target == "_blank") + target = "_self"; + } FrameLoadRequest frameRequest(request, target);
sorry, the patch need a fix @@ -288,8 +288,11 @@ void FrameLoader::urlSelected(const ResourceRequest& request, const String& pass return; String target = passedTarget; - if (target.isEmpty()) + if (target.isEmpty()) { target = m_frame->document()->baseTarget(); + if (target.isEmpty()) + target = "_self"; + } FrameLoadRequest frameRequest(request, target); (In reply to comment #2) > try this: > http://club3.kdnet.net/dispbbs.asp?boardid=1&id=6659651 > (note it is in Chinese!) > > Click the page number (at the bottom of the page), it will open a new tab while IE&firefox will open in current tab. > > I've fix this problem, code as below: > > @@ -288,8 +288,11 @@ void FrameLoader::urlSelected(const ResourceRequest& request, const String& pass > return; > > String target = passedTarget; > - if (target.isEmpty()) > + if (target.isEmpty()) { > target = m_frame->document()->baseTarget(); > + if (target == "_blank") > + target = "_self"; > + } > > FrameLoadRequest frameRequest(request, target);
Would you be willing to submit a patch with a test case, as described in <http://webkit.org/coding/contributing.html>?