More work towards <rdar://problem/44807048>.
I think we can just call it label() since the class is called UndoStep.
Created attachment 359856 [details] Patch
(In reply to Ryosuke Niwa from comment #1) > I think we can just call it label() since the class is called UndoStep. Oops, I just saw this now :/
Created attachment 359857 [details] Patch
Comment on attachment 359857 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359857&action=review > Source/WebKit/ChangeLog:45 > + Adjust this to take Ref<UndoStep>&&. Why are we making this change? > Source/WebKit/UIProcess/WebEditCommandProxy.h:42 > - static Ref<WebEditCommandProxy> create(WebUndoStepID commandID, WebCore::EditAction editAction, WebPageProxy* page) > + static Ref<WebEditCommandProxy> create(WebUndoStepID commandID, const String& label, WebPageProxy& page) Hm... I would have kept WebCore::EditAction for this one but okay.
Comment on attachment 359857 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359857&action=review >> Source/WebKit/ChangeLog:45 >> + Adjust this to take Ref<UndoStep>&&. > > Why are we making this change? I thought it would be preferable to UndoStep&, but I can revert this part. >> Source/WebKit/UIProcess/WebEditCommandProxy.h:42 >> + static Ref<WebEditCommandProxy> create(WebUndoStepID commandID, const String& label, WebPageProxy& page) > > Hm... I would have kept WebCore::EditAction for this one but okay. I think it should be okay to remove editAction from WebEditCommandProxy, since the only thing we used it for was the undo/redo label.
Created attachment 359882 [details] Review feedback & fix GTK build
Comment on attachment 359882 [details] Review feedback & fix GTK build Clearing flags on attachment: 359882 Committed r240342: <https://trac.webkit.org/changeset/240342>
Is the undo label controllable by the web content?
(In reply to Simon Fraser (smfr) from comment #9) > Is the undo label controllable by the web content? Yes. That being said, the UndoManager API isn't web-exposed (at least, not currently).
(In reply to Wenson Hsieh from comment #10) > (In reply to Simon Fraser (smfr) from comment #9) > > Is the undo label controllable by the web content? > > Yes. That being said, the UndoManager API isn't web-exposed (at least, not > currently). We probably need to do some sanitization if we decide to expose this API to the Web.