Reland c91b178b07b0d - Delete dead signin code (SigninGlobalError)

Reland c91b178b07b0d - Delete dead signin code (SigninGlobalError)

original cl - https://codereview.chromium.org/1299543002/
BUG=none
TBR=anthonyvd@chromium.org,sky@chromium.org,rogerta@chromium.org

Committed: https://crrev.com/33b30bb9c715a13874b73f73c12291bda35c1148
Cr-Commit-Position: refs/heads/master@{#344373}

[Evan Stade, 2015-08-19 16:42:53]

estade@chromium.org changed reviewers:
+ groby@chromium.org

[Evan Stade, 2015-08-19 16:44:46]

Rachel, any ideas what might have caused the failure (see original bug / revert
message)? It's not apparent to me and there doesn't seem to be a 10.9 trybot to
test this on (but I've sent it to all the other Mac bots).

[groby-ooo-7-16, 2015-08-19 17:14:08]

On 2015/08/19 16:44:46, Evan Stade wrote:
> Rachel, any ideas what might have caused the failure (see original bug /
revert
> message)? It's not apparent to me and there doesn't seem to be a 10.9 trybot
to
> test this on (but I've sent it to all the other Mac bots).

The test timed out - that's the typical callstack for OSX timeouts (and it
mumbles about that later in the test).

My guess would be that there's an unrelated change around InfiniteUnload, and
you accidentally ran afoul of it. (The test in questions runs a while(true)
script loop and then closes the browser. That _would_ timeout if the detection
for that is broken)

[Evan Stade, 2015-08-19 17:20:42]

On 2015/08/19 17:14:08, groby wrote:
> On 2015/08/19 16:44:46, Evan Stade wrote:
> > Rachel, any ideas what might have caused the failure (see original bug /
> revert
> > message)? It's not apparent to me and there doesn't seem to be a 10.9 trybot
> to
> > test this on (but I've sent it to all the other Mac bots).
> 
> The test timed out - that's the typical callstack for OSX timeouts (and it
> mumbles about that later in the test).
> 
> My guess would be that there's an unrelated change around InfiniteUnload, and
> you accidentally ran afoul of it. (The test in questions runs a while(true)
> script loop and then closes the browser. That _would_ timeout if the detection
> for that is broken)

well, thanks. I can't say I really follow --- what line in the logs indicates a
timeout? Which line of the test is the while loop? Most importantly, what are my
next steps?

[groby-ooo-7-16, 2015-08-19 17:58:32]

On 2015/08/19 17:20:42, Evan Stade wrote:
> On 2015/08/19 17:14:08, groby wrote:
> > On 2015/08/19 16:44:46, Evan Stade wrote:
> > > Rachel, any ideas what might have caused the failure (see original bug /
> > revert
> > > message)? It's not apparent to me and there doesn't seem to be a 10.9
trybot
> > to
> > > test this on (but I've sent it to all the other Mac bots).
> > 
> > The test timed out - that's the typical callstack for OSX timeouts (and it
> > mumbles about that later in the test).
> > 
> > My guess would be that there's an unrelated change around InfiniteUnload,
and
> > you accidentally ran afoul of it. (The test in questions runs a while(true)
> > script loop and then closes the browser. That _would_ timeout if the
detection
> > for that is broken)
> 
> well, thanks. I can't say I really follow --- what line in the logs indicates
a
> timeout? 
Only in the full stdio, not the per-test log. In full stdio, it's right under
the callstack:
    [411/483] UnloadTest.BrowserCloseInfiniteUnload (TIMED OUT)

> Which line of the test is the while loop?
https://code.google.com/p/chromium/codesearch#chromium/src/chrome/browser/unl...

Specifically, the second line there:
     <script>window.onunload=function(e){while(true){}}</script>

> Most importantly, what are my next steps?
I don't have a good answer, because I suspect this is a flaky test. Note that
crbug.com/462671 indicates that UnloadTest.CrossSiteInfiniteBeforeUnloadSync is
flaky as well, so I wouldn't be surprised if it's just in general a rare flake.
(Also, lots of other tests in that file mention the word flakey...)

I'm trying to get the Mac10.9 browser_test flakiness reports, but so far the URL
hangs:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=...

If it _is_ a flake, your best bet is finding somebody who knows more about the
UnloadTest stuff. Or just disabling it. 

If it's not a flake (i.e. you or a reland can repro on the 10.9 bots), then the
question becomes why it times out on 10.9. Presumably, a trooper could VNC and
take a look. I don't think anybody here has a 10.9 machine, we were
force-upgraded to 10.10

[Evan Stade, 2015-08-19 18:09:23]

+cc others

guess I'll try a reland

[Evan Stade, 2015-08-19 18:09:46]

estade@chromium.org changed reviewers:
+ anthonyvd@chromium.org, rogerta@chromium.org, sky@chromium.org

[sky, 2015-08-19 20:11:07]

You're going with the revert was because of flake?
Ok, LGTM

[Evan Stade, 2015-08-19 23:45:43]

The CQ bit was checked by estade@chromium.org

[Evan Stade, 2015-08-19 23:46:07]

On 2015/08/19 20:11:07, sky wrote:
> You're going with the revert was because of flake?
> Ok, LGTM

this bot shows green 10.9 runs:
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_dbg_...

so yea, hopefully it was flake.

[commit-bot: I haz the power, 2015-08-20 00:01:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1298973005/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1298973005/40001

[commit-bot: I haz the power, 2015-08-20 00:11:11]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-08-20 00:11:39]

Patchset 3 (id:??) landed as
https://crrev.com/33b30bb9c715a13874b73f73c12291bda35c1148
Cr-Commit-Position: refs/heads/master@{#344373}