Patch to follow. Note that it regresses visual representation a bit: we are not special casing with(element) and with(document) anymore and do not tell user that it is "Event target" and "Event document". However, strictly speaking, we should not have done it for with(element) not necessarily related to an event.
Created attachment 88925 [details] Patch
Comment on attachment 88925 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=88925&action=review > Source/WebCore/inspector/InjectedScriptSource.js:-513 > - scopeObjectProxy.isElement = true; We should be able to figure it out on the front-end side and avoid the regression. > Source/WebCore/inspector/front-end/ScopeChainSidebarPane.js:100 > + if (scope.type !== "global") { This way you won't expand global scope when it is the only one. This is a regression.
Created attachment 88977 [details] Patch
Attachment 88977 [details] did not build on qt: Build output: http://queues.webkit.org/results/8376824
(In reply to comment #4) > Attachment 88977 [details] did not build on qt: > Build output: http://queues.webkit.org/results/8376824 It was a false positive alarm, sorry for the inconvenience. I fixed the Qt EWS, and tried to build this patch manually, and it works for me. ;)
Committed r83434: <http://trac.webkit.org/changeset/83434>