On request error, always fire events on the XMLHttpRequestUpload before the XMLHttpRequest

On request error, always fire events on the XMLHttpRequestUpload before the XMLHttpRequest

On request error, fire events on the XMLHttpRequestUpload object before the XMLHttpRequest
object as per the latest specification:
http://xhr.spec.whatwg.org/#request-error

IE10 behaves according to specification but blink was firing the events on the XMLHttpRequest
object BEFORE the XMLHttpRequestUpload object in case of 'network error' or 'abort error'.
Blink was however behaving according to specification in case of 'timeout error', which was
inconsistent.

BUG=264609
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=155347

[haraken, 2013-07-26 12:11:25]

What about Firefox?

[do-not-use, 2013-07-26 12:24:16]

On 2013/07/26 12:11:25, haraken wrote:
> What about Firefox?

The test fails on Firefox.

[haraken, 2013-07-26 12:33:09]

> The test fails on Firefox.

Does this mean that the behavior of Firefox is the same as the current Blink?

[do-not-use, 2013-07-26 12:41:18]

On 2013/07/26 12:33:09, haraken wrote:
> > The test fails on Firefox.
> 
> Does this mean that the behavior of Firefox is the same as the current Blink?

Yes it does. Sorry for not being clear enough.

[haraken, 2013-07-26 13:11:49]

On 2013/07/26 12:41:18, Christophe Dumez wrote:
> On 2013/07/26 12:33:09, haraken wrote:
> > > The test fails on Firefox.
> > 
> > Does this mean that the behavior of Firefox is the same as the current
Blink?
> 
> Yes it does. Sorry for not being clear enough.

Hmm, then it might be controversial whether we should change the implementation
of Blink and Firefox or the spec. We might want to hear opinions of XHR experts.

[do-not-use, 2013-07-30 06:02:58]

On 2013/07/26 13:11:49, haraken wrote:
> On 2013/07/26 12:41:18, Christophe Dumez wrote:
> > On 2013/07/26 12:33:09, haraken wrote:
> > > > The test fails on Firefox.
> > > 
> > > Does this mean that the behavior of Firefox is the same as the current
> Blink?
> > 
> > Yes it does. Sorry for not being clear enough.
> 
> Hmm, then it might be controversial whether we should change the
implementation
> of Blink and Firefox or the spec. We might want to hear opinions of XHR
experts.

Should I cc anyone in particular?

[tyoshino (SeeGerritForStatus), 2013-07-30 09:08:48]

Commented on the bug. I think it's ok to make this change. Some webapps might
assume this order but they break only when some network error occurs.

[arv (Not doing code reviews), 2013-07-31 16:11:57]

LGTM

This is such an edge case that I doubt it will break any existing code. It makes
sense to follow the spec here.

[commit-bot: I haz the power, 2013-08-01 06:00:56]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ch.dumez@sisa.samsung.com/20560007/1

[commit-bot: I haz the power, 2013-08-01 09:26:34]

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

[commit-bot: I haz the power, 2013-08-01 11:02:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ch.dumez@sisa.samsung.com/20560007/1

[commit-bot: I haz the power, 2013-08-01 14:36:50]

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

[commit-bot: I haz the power, 2013-08-01 15:48:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ch.dumez@sisa.samsung.com/20560007/1

[commit-bot: I haz the power, 2013-08-01 17:42:06]

Change committed as 155347