Have XHR.getResponseHeader() return null in initial ready states.

Have XHR.getResponseHeader() return null in initial ready states.

Follow the XMLHttpRequest specification for getResponseHeader() and
return null if the object is in ready states {UNSENT, OPENED}.

Similarly for getAllResponseHeaders() and returning "" (the empty
string) when in these initial two ready states.

This replaces the older spec behavior of throwing InvalidStateError,
something that no other engines currently do, hence cross-browser
compatibility risk is low.

When reconciling the relevant tests, reworked these to use
js-test-{pre,post}.js instead.

R=tyoshino@chromium.org
BUG=311274
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=160679

[sof, 2013-10-26 17:55:37]

please have a look (or suggest someone else that might want to :) )

[tyoshino (SeeGerritForStatus), 2013-10-26 19:05:13]

thanks

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:38: else {
join l37 and l38 please

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:48:
shouldBeGreaterThanOrEqual("XMLHttpRequest.OPENED", "xhr.readyState");
seems it's recommended to pass actual value as the first argument and expected
and 2nd

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:58:
shouldNotThrow('xhr.open("GET","resources/1251.html", true);');
space. i.e.

"GET", "

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:12: var
headerName="Content-Type";
please insert a space on the right and left of =

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:14: "sEt-coOkIE2",
use spaces plz

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:27:
testPassed("headerValue is " + expected);
this will be shorter if shouldBe is used though we need to set headerValue to
window to make it visible from shouldBe

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:39:
testPassed("header value is unchanged.");
ditto

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:56:
shouldNotThrow('xhr.open("GET","resources/1251.html", true);');
space after "GET",

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:64: <script
src="../resources/js-test-post.js"></script>
put before </head>?

[sof, 2013-10-26 20:02:26]

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:48:
shouldBeGreaterThanOrEqual("XMLHttpRequest.OPENED", "xhr.readyState");
On 2013/10/26 19:05:13, tyoshino wrote:
> seems it's recommended to pass actual value as the first argument and expected
> and 2nd

Yes, but there's no *LessThanOrEqual predicate :) I've tried to side-step this
concern now.

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:27:
testPassed("headerValue is " + expected);
On 2013/10/26 19:05:13, tyoshino wrote:
> this will be shorter if shouldBe is used though we need to set headerValue to
> window to make it visible from shouldBe

Done, but I don't particularly like this way of expressing it.

[tyoshino (SeeGerritForStatus), 2013-10-26 20:22:46]

lgtm

Needs core/ OWNER's approval. Please wait for abarth's review.

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:48:
shouldBeGreaterThanOrEqual("XMLHttpRequest.OPENED", "xhr.readyState");
On 2013/10/26 20:02:26, sof wrote:
> On 2013/10/26 19:05:13, tyoshino wrote:
> > seems it's recommended to pass actual value as the first argument and
expected
> > and 2nd
> 
> Yes, but there's no *LessThanOrEqual predicate :) I've tried to side-step this
> concern now.

Oh... I see.

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html (right):

https://codereview.chromium.org/46913002/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/getResponseHeader.html:27:
testPassed("headerValue is " + expected);
On 2013/10/26 20:02:26, sof wrote:
> On 2013/10/26 19:05:13, tyoshino wrote:
> > this will be shorter if shouldBe is used though we need to set headerValue
to
> > window to make it visible from shouldBe
> 
> Done, but I don't particularly like this way of expressing it.

Yes. I agree it's weird... If you strongly prefer that, feel free to rollback
but maybe after abarth@'s comment.

https://codereview.chromium.org/46913002/diff/70001/LayoutTests/http/tests/xm...
File LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html (right):

https://codereview.chromium.org/46913002/diff/70001/LayoutTests/http/tests/xm...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:33:
testPassed("header values are unchanged.");
do the same as getResponseHeaders.html for consistency?

https://codereview.chromium.org/46913002/diff/70001/LayoutTests/http/tests/xm...
LayoutTests/http/tests/xmlhttprequest/getAllResponseHeaders.html:65: <script
src="../resources/js-test-post.js"></script>
inside </head>

[sof, 2013-10-26 20:57:57]

On 2013/10/26 20:22:46, tyoshino wrote:
> lgtm
> 
> Needs core/ OWNER's approval. Please wait for abarth's review.
> 

Will do, thanks very much for the review.

[abarth-chromium, 2013-10-26 23:08:19]

lgtm

[commit-bot: I haz the power, 2013-10-26 23:08:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/46913002/180001

[commit-bot: I haz the power, 2013-10-26 23:47:21]

Retried try job too often on win_blink_rel for step(s) webkit_unit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_blink_...

[commit-bot: I haz the power, 2013-10-27 02:57:57]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/46913002/180001

[commit-bot: I haz the power, 2013-10-27 15:04:21]

Change committed as 160679