[webkit-reviews] review requested: [Bug 32437] [Qt] [LayoutTestController]Output of document.write() is vanished by waitUntilDone()/notifyDone() : [Attachment 44761] Patch

[bugzilla-daemon at webkit.org]

robert <robert at roberthogan.net> has asked  for review:
Bug 32437: [Qt] [LayoutTestController]Output of document.write() is vanished by
waitUntilDone()/notifyDone()
https://bugs.webkit.org/show_bug.cgi?id=32437

Attachment 44761: Patch
https://bugs.webkit.org/attachment.cgi?id=44761&action=review

------- Additional Comments from robert <robert at roberthogan.net>
This fixes it for me. It also fixes the following from the skipped list on my
machine:

 fast/forms/textarea-linewrap-dynamic.html
 fast/forms/textarea-setvalue-submit.html
 fast/forms/textarea-hard-linewrap-empty.html
 fast/forms/submit-to-url-fragment.html
 http/tests/misc/percent-sign-in-form-field-name.html
 http/tests/security/escape-form-data-field-names.html

The 'but' is that these 2 need to be added to the skipped list:

 http/tests/xmlhttprequest/access-control-basic-denied-preflight-cache.html
 svg/custom/use-instanceRoot-as-event-target.xhtml

I think the bug with the drt was masking deeper problems with these two. In the
first case, the test returns a 'SUCCESS' *and* an exception. I believe up to
now it wasn't catching this exception because notifyDone shut down the page
before it had finished loading.

In the second case, notifyDone seems to get called three times by WebCore,
hence the three sets of identical results for 'test 10'. So in a sense, the
test is actually three times more successful than before. ;-) Again, the
behaviour of the DRT would have been to shut down the still-loading page before
receiving these two extra bits of output.

Csaba, can you see if you can reproduce my success with the above tests with
this patch?