WebKit2 needs layoutTestController.setDeferMainResourceDataLoad
Created attachment 286339 [details]
Note that only one test actually called setDeferMainResourceDataLoad, and even in DumpRenderTree, this test would pass if the function was not called.
Comment on attachment 286339 [details]
This does not seem right. We should fix the test so it tests what it intended to, rather than removing it. I see no rationale here for removing the test other than “it never worked”, but that is not sufficient.
One path here is to come up for a better rationale for why it's OK to remove this test.
Unfortunately, the fact that the test passes when setDeferMainResourceDataLoad does nothing does *not* mean the test had no value. It's entirely possible that it can detect bugs that other tests cannot. It's also possible, though, that it is true that the test no longer has value.
Another possibility is to keep the tests but remove calls to setDeferMainResourceDataLoad, but keep the test. Looking at the test, it seems that deferring load is the default behavior.
Created attachment 286639 [details]
Comment on attachment 286639 [details]
I looked back at when we originally added the test:
I read it carefully, and I see that this was a test specific to a peculiar way that WebKit was used by the now-long-obsolete Qt port. Given that, I changed my mind and I think that we *can* remove the test.
Created attachment 286864 [details]
Comment on attachment 286864 [details]
Clearing flags on attachment: 286864
Committed r204918: <http://trac.webkit.org/changeset/204918>
All reviewed patches have been landed. Closing bug.
Attempted Windows build fix in http://trac.webkit.org/projects/webkit/changeset/204933