XHR.abort(): no event dispatching in completed/non-started states.

XHR.abort(): no event dispatching in completed/non-started states.

Follow the spec and other implementations and only transition to an
UNSENT state if abort() is called on request that either hasn't been
sent or has completed. That is, no dispatching of progress events will
be done.

R=
BUG=315470
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=161515

[sof, 2013-11-06 11:06:42]

Please have a look.

[abarth-chromium, 2013-11-06 18:09:35]

lgtm

[commit-bot: I haz the power, 2013-11-06 18:09:41]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/60203010/1

[abarth-chromium, 2013-11-06 18:09:43]

Thanks for the CL!

[commit-bot: I haz the power, 2013-11-06 20:41:54]

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

[commit-bot: I haz the power, 2013-11-06 22:44:12]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/60203010/1

[commit-bot: I haz the power, 2013-11-07 00:56:04]

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

[commit-bot: I haz the power, 2013-11-07 07:43:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/60203010/1

[commit-bot: I haz the power, 2013-11-07 08:39:02]

Change committed as 161515

[tyoshino (SeeGerritForStatus), 2013-11-07 13:38:05]

Thanks for the fix. lgtm. i'll create patches for some minor issues later.

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/readystatechange-and-abort.html
(right):

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/readystatechange-and-abort.html:104:
abortInHeadersReceivedTest.step(function () {
step() is unnecessary. we're in step_func. i'll send you a fix.

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/timeout/xmlhttprequest-timeout.js
(right):

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/timeout/xmlhttprequest-timeout.js:198: if
(me.request.readyState == XMLHttpRequest.DONE || me.request.readyState >
XMLHttpRequest.OPENED) {
maybe this should be
me.request.readyState < XMLHttpRequest.OPENED
but we don't hit this bug?

this code was hard to read before your change. i'll do some refactoring.

[sof, 2013-11-07 14:34:11]

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
File LayoutTests/http/tests/xmlhttprequest/timeout/xmlhttprequest-timeout.js
(right):

https://codereview.chromium.org/60203010/diff/1/LayoutTests/http/tests/xmlhtt...
LayoutTests/http/tests/xmlhttprequest/timeout/xmlhttprequest-timeout.js:198: if
(me.request.readyState == XMLHttpRequest.DONE || me.request.readyState >
XMLHttpRequest.OPENED) {
On 2013/11/07 13:38:05, tyoshino wrote:
> maybe this should be
> me.request.readyState < XMLHttpRequest.OPENED
> but we don't hit this bug?
> 
> this code was hard to read before your change. i'll do some refactoring.

Much better, that was the intent & close enough (i.e., you could still be in
OPENED, pre-send(), and have abort() not dispatch any progress events.) But none
of the tests here exercises such early calls to abort(), afaict.

[The upstream Mozilla version doesn't have this extra condition & silently
completes the final AbortedRequest() (for "aborted") test without outputting
PASS.]