This bug is to track removal of the temporary workaround for Gears compatibility in WebKit (see https://bugs.webkit.org/show_bug.cgi?id=24215). Closing this bug should involve removing the workaround in favor of a long term solution. The underlying problem is that current users of gears use a standard gears_init.js that relies on a plugin loading if it's display attribute is set to hidden.
What's the status here? This code is still in the codebase.
The plan is to make plugins not go away when an element has display:none. Not sure if we have a bugzilla bug on that.
See bug 27775.
Created attachment 75005 [details] Patch
Comment on attachment 75005 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=75005&action=review I think we can safely remove the entire rendererIsNeeded() call. I don't see any parent implementations preventing creation of renderers when a document doesn't have a frame, but I believe that to be the case. Frameless documents won't have a FrameView, thus won't be rendered is my understanding. I think this whole override can be removed unless we have a test which proves otherwise. > WebCore/html/HTMLObjectElement.cpp:289 > if (!frame) > return false; Are we sure this is still needed?
I'd rather see you commit this by hand and remove the whole function (or post a new patch to do so). I don't believe the !frame check has any utility.
I'm not familiar with the surrounding code, I'm just removing the workaround I put in. I'd prefer for more ambitious cleanups to be done by other people who have better knowledge of the area. Does that sound OK?
Comment on attachment 75005 [details] Patch Clearing flags on attachment: 75005 Committed r72799: <http://trac.webkit.org/changeset/72799>
All reviewed patches have been landed. Closing bug.
I'm confused. Bug 27775 is not fixed. Was Gears changed to avoid this issue?