[Webkit-unassigned] [Bug 151139] [GTK] Web Process crashes on reparenting a WebView with AC mode on

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=151139

--- Comment #3 from Mario Sanchez Prada <mario at webkit.org> ---
Created attachment 265302
  --> https://bugs.webkit.org/attachment.cgi?id=265302&action=review
Backtraces for the UI and Web processes (Debug build)

Seems like I end up getting the backtraces from a debug build (SVN r192098) sooner than what I expected, so I'm attaching them right now.

I'm still not sure the web process backtrace is useful since it shows the moment the X error is noticed but not when it's generated, so I'll try to get one setting GDK_SYNCHRONIZE=1.

In any case, it shows at least why the UI Process it's dying, which seems to be due to drawingArea being NULL in the following snippet:

    static void webkitWebViewBaseRealize(GtkWidget* widget)
    {
        WebKitWebViewBase* webView = WEBKIT_WEB_VIEW_BASE(widget);
        WebKitWebViewBasePrivate* priv = webView->priv;

    #if USE(REDIRECTED_XCOMPOSITE_WINDOW)
        [...]
            if (priv->redirectedWindow) {
                DrawingAreaProxyImpl* drawingArea = static_cast<DrawingAreaProxyImpl*>(priv->pageProxy->drawingArea());
                drawingArea->setNativeSurfaceHandleForCompositing(priv->redirectedWindow->windowID());
            }
        }
    #endif
       [...]
    }

My theory is that when the X error comes bad things happen to the XRedirectedCompositeWindow and the drawing area dies too, causing the proxy here to die as well. But that's a theory, we need facts :)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20151111/013777e2/attachment-0001.html>