[Webkit-unassigned] [Bug 36721] Return correct load result to NPAPI plugins calling getUrlNotify on target frames

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=36721





--- Comment #19 from Robert Hogan <robert at webkit.org>  2010-11-16 14:55:13 PST ---
(In reply to comment #13)
> Created an attachment (id=74026)
 --> (https://bugs.webkit.org/attachment.cgi?id=74026&action=review) [details]
> Patch

Per Alexey's comments in IRC a bit more context is required for why the patch is needed.

Currently, if an NPAPI plugin sends a getURLNotify() with a target frame specified in the request PluginView passes it off to FrameLoader and immediately returns an NPRES_DONE notification to the plugin without waiting to hear how the request actually got on.

Mac gets around this by implementing platform-specific logic (it doesn't use PluginView for a start) in WebKit/mac/plugins/.

This WebCore behaviour means that this test fails for users of PluginView - including qt, gtk and chromium. Chromium passes the test by changing its platform-specific expected result to match the fake NPRES_DONE currently returned by PluginView::performRequest.

So fixing this early return in WebCore should allow other ports to start passing this test as long as they support all the DRT bits and pieces required to do a getUrlNotify() layout test and haven't implemented similar workarounds in platform code.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.