A `readonly attribute` is unlikely to have any effect on the object being traced by CallTracer. Therefore, CallTracer probably doesn't need to instrument those getters, as it would most likely be unneeded information.
Created attachment 319289 [details] Patch
Comment on attachment 319289 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=319289&action=review r=me > Source/WebCore/ChangeLog:8 > + Updates existing bindings test with more cases. It seems like the thing to note is that the CallTracingCallback is not invoked for readonly attributes. Aren't the test cases just auto-generated? > Source/WebCore/ChangeLog:12 > + If the attribute is makred `readonly` do not add CallTracer logic. makred -> marked
Comment on attachment 319289 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=319289&action=review >> Source/WebCore/ChangeLog:8 >> + Updates existing bindings test with more cases. > > It seems like the thing to note is that the CallTracingCallback is not invoked for readonly attributes. Aren't the test cases just auto-generated? The IDL file for the test case is not autogenerated, but the corresponding cpp/h files are.
Created attachment 319293 [details] Patch
Created attachment 319302 [details] Patch
Comment on attachment 319302 [details] Patch Clearing flags on attachment: 319302 Committed r221336: <http://trac.webkit.org/changeset/221336>
All reviewed patches have been landed. Closing bug.
<rdar://problem/34150808>