Unflake http/tests/xmlhttprequest/event-listener-gc.html

Unflake http/tests/xmlhttprequest/event-listener-gc.html

Using a zero-timeout and timer task to set up the test condition is not
as precise as we would like. Rephrase to use an immediate message
dispatch instead.

R=
BUG=503842
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197724

[sof, 2015-06-24 07:47:26]

sigbjornf@opera.com changed reviewers:
+ tyoshino@chromium.org, yhirano@chromium.org

[sof, 2015-06-24 07:47:26]

please take a look.

[tyoshino (SeeGerritForStatus), 2015-06-24 09:42:57]

On 2015/06/24 07:47:26, sof wrote:
> please take a look.

postMessage() is the best way to ensure collect() is run async to test() but
before onreadystatechange dispatch?

[sof, 2015-06-24 09:45:11]

On 2015/06/24 09:42:57, tyoshino wrote:
> On 2015/06/24 07:47:26, sof wrote:
> > please take a look.
> 
> postMessage() is the best way to ensure collect() is run async to test() but
> before onreadystatechange dispatch?

Better than was, yes.

[tyoshino (SeeGerritForStatus), 2015-06-24 09:46:46]

On 2015/06/24 09:45:11, sof wrote:
> On 2015/06/24 09:42:57, tyoshino wrote:
> > On 2015/06/24 07:47:26, sof wrote:
> > > please take a look.
> > 
> > postMessage() is the best way to ensure collect() is run async to test() but
> > before onreadystatechange dispatch?
> 
> Better than was, yes.

I understand more about the background, but lgtm on the patch.

[sof, 2015-06-24 09:47:17]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-06-24 09:47:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1201393003/1

[tyoshino (SeeGerritForStatus), 2015-06-24 09:48:49]

Promise.resolve().then(collect); should work too?

[tyoshino (SeeGerritForStatus), 2015-06-24 09:49:37]

On 2015/06/24 09:48:49, tyoshino wrote:
> Promise.resolve().then(collect); should work too?

Hmm, microtask is not so reliable for this purpose for now?

[commit-bot: I haz the power, 2015-06-24 09:51:12]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197724

[sof, 2015-06-24 09:56:25]

On 2015/06/24 09:49:37, tyoshino wrote:
> On 2015/06/24 09:48:49, tyoshino wrote:
> > Promise.resolve().then(collect); should work too?
> 
> Hmm, microtask is not so reliable for this purpose for now?

Or requestAnimationFrame(collect) (or some setImmediate() thing, should that
transpire.)

I don't think it is possible to guarantee expected ordering 100% here -- i.e.,
triggering the GC after OPENED (but without having the XHR object on the stack
while doing that GC), but before other readystate transitions.