[Webkit-unassigned] [Bug 130942] REGRESSION (r166144?): fast/preloader/document-write.html is very flaky

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=130942





--- Comment #4 from Darin Adler <darin at apple.com>  2014-03-31 10:18:50 PST ---
(In reply to comment #3)
> If we can't roll out this change

We can, it’s just not necessarily worth it.

> if cannot fix it

No clue on the cause of the different behavior yet.

> we should definitely disable or remove the test.

Maybe.

> It's pure noise now, only confirming that it doesn't crash.

Well, the fact that script1.js appears in the output meaningfully indicates that what the test was designed to test is still working; preloading is occurring.

It’s unfortunate that the test machinery classifies that as a failure. Whether document-write-plaintext.js shows up in the output or not is not relevant to what the test is trying to test.

So the test would still have some value if we could find a way to ignore the irrelevant part.

And also it would be interesting to know why the load of document-write-plaintext.js does not always log. Maybe there’s some previous test that loads it and it is cached?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.