Patch to follow.
Created attachment 122946 [details] Patch
Comment on attachment 122946 [details] Patch Can we have a test for this?
Comment on attachment 122946 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=122946&action=review > Source/WebCore/inspector/front-end/DOMAgent.js:-1144 > - var doc = this._domAgent._documentURLToDocument[resource.url]; Do you still need the _documentURLToDocument map? I feel that map is flawed in the case where multiple frames with the same URL exist. And this patch fixes that issue by using frameId. Who still needs _documentURLToDocument?
(In reply to comment #3) > (From update of attachment 122946 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=122946&action=review > > > Source/WebCore/inspector/front-end/DOMAgent.js:-1144 > > - var doc = this._domAgent._documentURLToDocument[resource.url]; > > Do you still need the _documentURLToDocument map? I feel that map is flawed in the case where multiple frames with the same URL exist. And this patch fixes that issue by using frameId. Who still needs _documentURLToDocument? Good point. Will nuke it prior to landing.
Great!
Committed r105288: <http://trac.webkit.org/changeset/105288>
Seems like it makes Qt regress again: http://build.webkit.sed.hu/results/x86-64%20Linux%20Qt%20Release/r105288%20(32590)/inspector/elements/set-html-via-resource-pretty-diff.html
(In reply to comment #7) > Seems like it makes Qt regress again: http://build.webkit.sed.hu/results/x86-64%20Linux%20Qt%20Release/r105288%20(32590)/inspector/elements/set-html-via-resource-pretty-diff.html Fixed it, sorry about that.