[webkit-reviews] review granted: [Bug 90976] If value for responseType defined as type that not supported, it should not throw an exception in XHR 2 : [Attachment 153253] Patch

[bugzilla-daemon at webkit.org]

Alexey Proskuryakov <ap at webkit.org> has granted Kinuko Yasuda
<kinuko at chromium.org>'s request for review:
Bug 90976: If value for responseType defined as type that not supported, it
should not throw an exception in XHR 2
https://bugs.webkit.org/show_bug.cgi?id=90976

Attachment 153253: Patch
https://bugs.webkit.org/attachment.cgi?id=153253&action=review

------- Additional Comments from Alexey Proskuryakov <ap at webkit.org>
View in context: https://bugs.webkit.org/attachment.cgi?id=153253&action=review


r=me, but please double check why cross-platform results don't need CONSOLE
MESSAGE lines. I think that it's a mistake.

> Source/WebCore/xml/XMLHttpRequest.cpp:357
> +	   logConsoleError(scriptExecutionContext(), "\"" + responseType + "\"
is not a valid XMLHttpRequest.responseType.");

I would have said "supported", not "valid", because new types can be added to
the spec in the future, and lack of support for them in a certain release of a
WebKit based browser doesn't make such invalid. But it's not a big deal.

> LayoutTests/ChangeLog:14
> +	   * fast/xmlhttprequest/xmlhttprequest-responsetype-arraybuffer.html:
Removed a line which tries to set invalid type (since we now have a separate
test)
> +	   *
platform/chromium/fast/xmlhttprequest/xmlhttprequest-responsetype-set-invalidty
pe-expected.txt: Added (since chromium expectation includes CONSOLE messages).

This surprises me - I think that every port does this. Looking at Mac
DumpRenderTree source code, I do see console messages printed.