I am not sure whether my code reading is wrong. So, please let me know to close this bug if I am wrong. While implementing cursor of WebKit2/Efl, I got below assumption. 1. Widget::setCursor is similar. It just passed to chromeclient. - efl port was totally different, but now it is same as other ports after Bug 88803 is landed. - win port have an addition check, qt port have macro. but others are same. 2. Only EventHandler calls setCursor in WebCore. - some ports call it, but I think that they have a way to call ChromeClient::setCursor. 3. setCursor is called although cursor is not changed. - for example, every mouse move event will generate IPC message to change cursor in case of WebKit2 (although it is same). My prototype looks working in efl port and gtk port. But I don't know how to test cursor (Is it possible?) and I am worried about any regressions.
Created attachment 147832 [details] prototype
Comment on attachment 147832 [details] prototype Attachment 147832 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/12961613
Created attachment 147834 [details] prototype(fix warning)
Created attachment 147838 [details] prototype(fix more)
Created attachment 147840 [details] prototype(fix more)
Comment on attachment 147840 [details] prototype(fix more) Attachment 147840 [details] did not pass win-ews (win): Output: http://queues.webkit.org/results/12964569
Created attachment 147903 [details] prototype(fix more)