WebPageProxy's setViewNeedsDisplay() and viewSize() should check page validity. Soon after WebProcess dies there's still a call to those methods, and at the time the PageClient the WebPageProxy holds is invalid.
Created attachment 196002 [details] Patch
Comment on attachment 196002 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=196002&action=review > Source/WebKit2/ChangeLog:4 > + WebPageProxy's setViewNeedsDisplay() and viewSize() should check page validity. > + https://bugs.webkit.org/show_bug.cgi?id=113713 No they shouldn't. You must explain _why_ the change is needed. And add a test unless you have a good reason.
(In reply to comment #2) > (From update of attachment 196002 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=196002&action=review > > > Source/WebKit2/ChangeLog:4 > > + WebPageProxy's setViewNeedsDisplay() and viewSize() should check page validity. > > + https://bugs.webkit.org/show_bug.cgi?id=113713 > > No they shouldn't. > > You must explain _why_ the change is needed. And add a test unless you have a good reason. Since you r- this, so why they shouldn't check the validity then?
(In reply to comment #3) > (In reply to comment #2) > > (From update of attachment 196002 [details] [details]) > > View in context: https://bugs.webkit.org/attachment.cgi?id=196002&action=review > > > > > Source/WebKit2/ChangeLog:4 > > > + WebPageProxy's setViewNeedsDisplay() and viewSize() should check page validity. > > > + https://bugs.webkit.org/show_bug.cgi?id=113713 > > > > No they shouldn't. > > > > You must explain _why_ the change is needed. And add a test unless you have a good reason. > > > Since you r- this, so why they shouldn't check the validity then? Other ports rely on being able to call setViewNeedsDisplay even if the page is invalid. You mustn’t break this contract.
(In reply to comment #4) > (In reply to comment #3) > > (In reply to comment #2) > > > (From update of attachment 196002 [details] [details] [details]) > > > View in context: https://bugs.webkit.org/attachment.cgi?id=196002&action=review > > > > > > > Source/WebKit2/ChangeLog:4 > > > > + WebPageProxy's setViewNeedsDisplay() and viewSize() should check page validity. > > > > + https://bugs.webkit.org/show_bug.cgi?id=113713 > > > > > > No they shouldn't. > > > > > > You must explain _why_ the change is needed. And add a test unless you have a good reason. > > > > > > Since you r- this, so why they shouldn't check the validity then? > > Other ports rely on being able to call setViewNeedsDisplay even if the page is invalid. You mustn’t break this contract. Oh ok, thanks for the explanation! So I guess perhaps this is not the right way to check what Marcelo is looking for... Would checking if (m_pageClient) be acceptable here, since it is a raw pointer ?