[webkit-reviews] review requested: [Bug 37759] [Qt] need an implementation of LayoutTestController::computedStyleIncludingVisitedInfo : [Attachment 55986] patch v6

[bugzilla-daemon at webkit.org]

Antonio Gomes (tonikitoo) <tonikitoo at webkit.org> has asked  for review:
Bug 37759: [Qt] need an implementation of
LayoutTestController::computedStyleIncludingVisitedInfo
https://bugs.webkit.org/show_bug.cgi?id=37759

Attachment 55986: patch v6
https://bugs.webkit.org/attachment.cgi?id=55986&action=review

------- Additional Comments from Antonio Gomes (tonikitoo)
<tonikitoo at webkit.org>
Patch is essentially similar to https://bugs.webkit.org/attachment.cgi?id=53815
but instead of 

Instead of receiving a QVariant as parameter of
computedStyleIncludingVisitedInfo it is now getting a QWebElement . The
parameters of DumpRenderTreeSupportQt::computedStyleIncludingVisitedInfo also
changed accordingly , which makes the code simpler.

> The "crash" only happens when running all four tests at one time, or one test
with multiple times of calling the new interface (for > example),

> var ele = document.getElementById('one');
> window.layoutTestController.computedStyleIncludingVisitedInfo(ele);
> window.layoutTestController.computedStyleIncludingVisitedInfo(ele);
> window.layoutTestController.computedStyleIncludingVisitedInfo(ele);
> window.layoutTestController.computedStyleIncludingVisitedInfo(ele);
> window.layoutTestController.computedStyleIncludingVisitedInfo(ele);

I tested it and it works fine now.

> Also, if I change the parameter type of computedStyleIncludingVisitedInfo(),
from QVariant to others, e.g. QString, everything > works fine -- no crash, no
error.

I think that is the case of QWebElement as well.


ps: Yi is still an author in the ChangeLog, since my work is based on his
original work.