Add support for codepen to form_based_credentials_background

Add support for codepen to form_based_credentials_background

Fix google credentials support.

Migrate to action_runner when possible.

BUG=

Committed: https://crrev.com/c3ce3fff3f0f9fd365be29721e9911fbfb1fac47
Cr-Commit-Position: refs/heads/master@{#291705}

[sullivan, 2014-08-18 16:33:40]

I'm having two problems getting credentials to work when recording pagesets in
telemetry: 

1) I am getting logged out pages for codepen.io. For the pages I'm loading,
/debug does a 301 redirect to /full when logged out. When I run
tools/perf/record_wpr, I see the login page briefly, but then the logged out
"/full" pages are recorded. No errors on the console. Any ideas how to debug?

2) For the gmail page with Google credentials, it goes this login page and waits
~30s:

https://accounts.google.com/ServiceLogin?passive=1209600&continue=https%3A%2F...

I get the following output:
[ RUN      ] https://mail.google.com/mail/
The connection to Chrome was lost to the Inspector UI.
Telemetry is waiting for the inspector to be closed...

Unhandled exception while running https://mail.google.com/mail/

Traceback (most recent call last):
  _RunPage at tools/telemetry/telemetry/page/page_runner.py:521
    page_state.PreparePage(test)
  PreparePage at tools/telemetry/telemetry/page/page_runner.py:159
    self.tab, self.page.credentials):
  LoginNeeded at tools/telemetry/telemetry/core/browser_credentials.py:61
    tab, self._credentials[credentials_type])
  LoginNeeded at
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:105
    _SubmitFormAndWait(self.login_form_id, tab)
  _SubmitFormAndWait at
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:27
    util.WaitFor(FinishedLoading, 60)
  WaitFor at tools/telemetry/telemetry/core/util.py:76
    res = condition()
  FinishedLoading at
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:24
    'document.querySelector("#%s") !== null' % form_id)
  EvaluateJavaScript at tools/telemetry/telemetry/core/web_contents.py:117
    expr, context_id=None, timeout=timeout)
  EvaluateJavaScriptInContext at
tools/telemetry/telemetry/core/web_contents.py:133
    expr, context_id=context_id, timeout=timeout)
  EvaluateJavaScript at
tools/telemetry/telemetry/core/backends/chrome/inspector_backend.py:176
    return self._runtime.Evaluate(expr, context_id, timeout)
  Evaluate at
tools/telemetry/telemetry/core/backends/chrome/inspector_runtime.py:40
    res = self._inspector_backend.SyncRequest(request, timeout)
  SyncRequest at
tools/telemetry/telemetry/core/backends/chrome/inspector_websocket.py:66
    res = self._Receive(timeout)
  _Receive at
tools/telemetry/telemetry/core/backends/chrome/inspector_websocket.py:107
    if 'method' in res and self._notification_handler(res):
  _HandleNotification at
tools/telemetry/telemetry/core/backends/chrome/inspector_backend.py:227
    self._WaitForInspectorToGoAwayAndReconnect()
  _WaitForInspectorToGoAwayAndReconnect at
tools/telemetry/telemetry/core/backends/chrome/inspector_backend.py:258
    self._socket.close()
AttributeError: 'NoneType' object has no attribute 'close'

[tonyg, 2014-08-18 16:38:57]

[+ari]

[ariblue, 2014-08-18 19:24:57]

On 2014/08/18 16:38:57, tonyg wrote:
> [+ari]

For the gmail part, I don't think this is a recording specific error. I'm seeing
the same issue when running via the ./run_benchmark command. More specifically,
it looks like we're hanging during the _SubmitFormAndWait part of the
form_based_credentials_backend.py. When increasing the verbosity, I am seeing
this loop over and over:

DEBUG:root:sent [{"params": {"returnByValue": true, "expression":
"document.querySelector(\"#gaia_loginform\") !== null"}, "method":
"Runtime.evaluate", "id": 49}]
DEBUG:root:got
[{"result":{"result":{"type":"boolean","value":true},"wasThrown":false},"id":49}]

I haven't really looked into it much, but when submitting the form with valid
credentials, it just seems to reload the page sometimes. Other times it seems to
work fine.

[ariblue, 2014-08-18 21:08:56]

On 2014/08/18 16:38:57, tonyg wrote:
> [+ari]

For the codepen part, I took a quick look: 

(1) IsAlreadyLoggedIn was evaluating to true in codepen_credentials_backend.py,
so we were going down the wrong codepath. Once changed to something like
document.querySelector(".login-area") === null, we correctly try and fill out
the form.

(2) Filling out the form failed, since line 98 in
form_based_credentials_backend.py had issues with strings that included the '-'
character:
email_id = 'document.querySelector("#%s").%s.value = "%s"; ' % (
changing this to something like:
document.querySelector("#%s #%s").value = "%s"; 
might help

(3) For some reason, the codepen.io pages are refreshing in a loop. I don't know
why this would be occurring, but it is causing issues with trying to fill out
the form.

[sullivan, 2014-08-20 15:51:09]

Thanks so much for your help, Ari!

I managed to get codepen working by fixing the IsAlreadyLoggedIn implementation,
making the querySelector change you suggested, and then changing the code to
click the "log in" button. Clicking the button only worked on a timeout,
otherwise I get a permissions error from codepen.

I still haven't figured out the issue with gmail. It correctly types the email
and password and submits the form, but the login page just reloads. When I copy
and paste the JavaScript it's running onto that same login page in the developer
tools, the log in works correctly.

Before I move forward with actually submitting this page set, Tony, can you take
a look at the trace I generated in crbug.com/405558 ? Are these actually good
energy cases? They show places where we do and do not break out of the
background timer throttling, but right now the cases that do break out run so
much that I had to increase the timeout for gathering the trace on the 30s blank
page, and when I did that I ended up with a 400MB trace file which sad tabs
about:tracing. Should I split the codepen credentials part of this CL out while
we discuss good energy cases?

[tonyg, 2014-08-20 17:37:18]

Regarding the large amount of activity. I think that's a good thing for a
"tough" page set :)

We can drastically cut down on the trace volume by specifying trace categories.
For instance, specifying only the "toplevel" category should keep us informed of
wakeups without overflowing the buffer.

https://codereview.chromium.org/485743002/diff/40001/tools/perf/page_sets/dat...
File tools/perf/page_sets/data/tough_energy_cases.json (right):

https://codereview.chromium.org/485743002/diff/40001/tools/perf/page_sets/dat...
tools/perf/page_sets/data/tough_energy_cases.json:29:
"http://codepen.io/anniesullie/debug/qclzA"
We want testificate364 here, no?

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:1: #
Copyright 2014 The Chromium Authors. All rights reserved.
Let's split the tools/telemetry/* changes into a new CL and get it solid first.
Then we'll land the page set changes separately.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:16: def
credentials_type(self):
Not for this CL, but just thinking aloud that it would be nice if these
properties were specified in the credentials file itself instead of hardcoded in
python like this.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
File
tools/telemetry/telemetry/core/backends/codepen_credentials_backend_unittest.py
(right):

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend_unittest.py:17:
'https://codepen.io/login', 'login-email-field', 'login-password')
I can't tell, does this test hit the live site? If so, we should disable it
before checking in or else change it to use a recording.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:11:
return (tab.EvaluateJavaScript(
This file pre-dates the ActionRunner. Check out
tools/telemetry/telemetry/page/actions/action_runner.py. There are some useful
methods like WaitForElement that might help with any edge cases or races
throughout this file.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:22:
'}, 1000);' % button_id)
Is the timeout necessary in addition to the submit() vs. click() change? Or
would just click() work?

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:101:
if self.IsAlreadyLoggedIn(tab):
Just a wild guess, but should this be checked after we
WaitForDocumentReadyStateToBeComplete()?

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:105:
tab.WaitForDocumentReadyStateToBeInteractiveOrBetter()
Does changing this to Complete instead of Interactive help with the race?

[jeremy, 2014-08-21 08:49:56]

I think the gmail issue is not new and if you can figure out how to fix it
that would be incredibly awesome - crbug.com/386152 .


On Wed, Aug 20, 2014 at 8:37 PM, <tonyg@chromium.org> wrote:

> Regarding the large amount of activity. I think that's a good thing for a
> "tough" page set :)
>
> We can drastically cut down on the trace volume by specifying trace
> categories.
> For instance, specifying only the "toplevel" category should keep us
> informed of
> wakeups without overflowing the buffer.
>
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/perf/page_sets/data/tough_energy_cases.json
> File tools/perf/page_sets/data/tough_energy_cases.json (right):
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/perf/page_sets/data/tough_energy_cases.json#newcode29
> tools/perf/page_sets/data/tough_energy_cases.json:29:
> "http://codepen.io/anniesullie/debug/qclzA"
> We want testificate364 here, no?
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py
> File
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py
> (right):
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py#
> newcode1
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:1:
> # Copyright 2014 The Chromium Authors. All rights reserved.
> Let's split the tools/telemetry/* changes into a new CL and get it solid
> first. Then we'll land the page set changes separately.
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py#
> newcode16
> tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:16:
> def credentials_type(self):
> Not for this CL, but just thinking aloud that it would be nice if these
> properties were specified in the credentials file itself instead of
> hardcoded in python like this.
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/codepen_
> credentials_backend_unittest.py
> File
> tools/telemetry/telemetry/core/backends/codepen_
> credentials_backend_unittest.py
> (right):
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/codepen_
> credentials_backend_unittest.py#newcode17
> tools/telemetry/telemetry/core/backends/codepen_
> credentials_backend_unittest.py:17:
> 'https://codepen.io/login', 'login-email-field', 'login-password')
> I can't tell, does this test hit the live site? If so, we should disable
> it before checking in or else change it to use a recording.
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
> File
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
> (right):
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py#
> newcode11
> tools/telemetry/telemetry/core/backends/form_based_
> credentials_backend.py:11:
> return (tab.EvaluateJavaScript(
> This file pre-dates the ActionRunner. Check out
> tools/telemetry/telemetry/page/actions/action_runner.py. There are some
> useful methods like WaitForElement that might help with any edge cases
> or races throughout this file.
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py#
> newcode22
> tools/telemetry/telemetry/core/backends/form_based_
> credentials_backend.py:22:
> '}, 1000);' % button_id)
> Is the timeout necessary in addition to the submit() vs. click() change?
> Or would just click() work?
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py#
> newcode101
> tools/telemetry/telemetry/core/backends/form_based_
> credentials_backend.py:101:
> if self.IsAlreadyLoggedIn(tab):
> Just a wild guess, but should this be checked after we
> WaitForDocumentReadyStateToBeComplete()?
>
> https://codereview.chromium.org/485743002/diff/40001/
> tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py#
> newcode105
> tools/telemetry/telemetry/core/backends/form_based_
> credentials_backend.py:105:
> tab.WaitForDocumentReadyStateToBeInteractiveOrBetter()
> Does changing this to Complete instead of Interactive help with the
> race?
>
> https://codereview.chromium.org/485743002/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[sullivan, 2014-08-21 21:55:23]

I haven't addressed all the review comments yet, but I found a hack that ensures
both Gmail and Codepen work correctly--I wait 1s after submitting the form. Is
there a less hacky way to do it?

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
File
tools/telemetry/telemetry/core/backends/codepen_credentials_backend_unittest.py
(right):

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend_unittest.py:17:
'https://codepen.io/login', 'login-email-field', 'login-password')
On 2014/08/20 17:37:17, tonyg wrote:
> I can't tell, does this test hit the live site? If so, we should disable it
> before checking in or else change it to use a recording.

FormBasedCredentialsBackendUnitTestBase.testLoginUsingMock uses a mock for the
navigation:
tab.ExpectCall('Navigate', login_page_url)

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:11:
return (tab.EvaluateJavaScript(
On 2014/08/20 17:37:18, tonyg wrote:
> This file pre-dates the ActionRunner. Check out
> tools/telemetry/telemetry/page/actions/action_runner.py. There are some useful
> methods like WaitForElement that might help with any edge cases or races
> throughout this file.

Done.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:22:
'}, 1000);' % button_id)
On 2014/08/20 17:37:17, tonyg wrote:
> Is the timeout necessary in addition to the submit() vs. click() change? Or
> would just click() work?

Yes, it's necessary. I get an error message with an immediate click() :/

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:101:
if self.IsAlreadyLoggedIn(tab):
On 2014/08/20 17:37:18, tonyg wrote:
> Just a wild guess, but should this be checked after we
> WaitForDocumentReadyStateToBeComplete()?

It's kind of hard to read, but _WaitForLoginFormToLoad() gets the page loaded to
the point where there is either a login form, or an element we know indicates
the user is already logged in. So this code should be okay.

https://codereview.chromium.org/485743002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:105:
tab.WaitForDocumentReadyStateToBeInteractiveOrBetter()
On 2014/08/20 17:37:17, tonyg wrote:
> Does changing this to Complete instead of Interactive help with the race?

No, but I changed it anyway, since it seems like in general waiting a bit longer
makes things go a little smoother.

[sullivan, 2014-08-21 21:57:05]

https://codereview.chromium.org/485743002/diff/60001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/60001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:81:
ar.Wait(1) # TODO(sullivan): no hack!!!
This is the hack I really want to get rid of, is there some better way to figure
out that the form is submitted fully? The code to wait until the form is gone
above does not work correctly for gmail.

[tonyg, 2014-08-21 22:34:06]

https://codereview.chromium.org/485743002/diff/60001/tools/perf/page_sets/tou...
File tools/perf/page_sets/tough_energy_cases.py (right):

https://codereview.chromium.org/485743002/diff/60001/tools/perf/page_sets/tou...
tools/perf/page_sets/tough_energy_cases.py:50: # TODO(sullivan): uncomment this
when Google login works properly.
This comment can go away now

https://codereview.chromium.org/485743002/diff/60001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/60001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:55:
def _WaitForLoginState(self, ar):
s/ar/action_runner/ throughout.

https://codereview.chromium.org/485743002/diff/60001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:81:
ar.Wait(1) # TODO(sullivan): no hack!!!
On 2014/08/21 21:57:04, sullivan wrote:
> This is the hack I really want to get rid of, is there some better way to
figure
> out that the form is submitted fully? The code to wait until the form is gone
> above does not work correctly for gmail.

Have you tried replacing this and the WaitForJavaScriptCondition above with a
WaitForNavigate()?

[sullivan, 2014-08-22 21:18:59]

This is ready for review. I will move tough_energy_cases to a follow-on CL.

[tonyg, 2014-08-22 21:35:10]

lgtm

Excellent :)

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:30:
return ('if (window.LoginSettings) {'
I think we should get rid of the conditional. I just had it there for when the
code path was shared with other login pages.

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:33:
'document.getElementById("log-in-button").click();')
Formatting nit, I think this is more readable:

return """
    foo();
    bar();
    """

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:30:
"""Some sites (like codepen) have custom JS to log in."""
Let's avoid mentioning codepen here.

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/browser_credentials.py (right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/browser_credentials.py:61: ar =
action_runner.ActionRunner(tab)
s/ar/action_runner/

[sullivan, 2014-08-22 21:46:54]

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:30:
return ('if (window.LoginSettings) {'
On 2014/08/22 21:35:09, tonyg wrote:
> I think we should get rid of the conditional. I just had it there for when the
> code path was shared with other login pages.

Done.

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/codepen_credentials_backend.py:33:
'document.getElementById("log-in-button").click();')
On 2014/08/22 21:35:09, tonyg wrote:
> Formatting nit, I think this is more readable:
> 
> return """
>     foo();
>     bar();
>     """

Done.

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py
(right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/backends/form_based_credentials_backend.py:30:
"""Some sites (like codepen) have custom JS to log in."""
On 2014/08/22 21:35:09, tonyg wrote:
> Let's avoid mentioning codepen here.

Done.

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
File tools/telemetry/telemetry/core/browser_credentials.py (right):

https://codereview.chromium.org/485743002/diff/120001/tools/telemetry/telemet...
tools/telemetry/telemetry/core/browser_credentials.py:61: ar =
action_runner.ActionRunner(tab)
On 2014/08/22 21:35:10, tonyg wrote:
> s/ar/action_runner/

Done.

[sullivan, 2014-08-22 21:47:01]

The CQ bit was checked by sullivan@chromium.org

[commit-bot: I haz the power, 2014-08-22 21:50:52]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sullivan@chromium.org/485743002/140001

[commit-bot: I haz the power, 2014-08-22 23:15:33]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  linux_chromium_rel_swarming on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  mac_chromium_rel_swarming on tryserver.chromium.mac
(http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)
  win_chromium_rel_swarming on tryserver.chromium.win
(http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)
  win_chromium_x64_rel_swarming on tryserver.chromium.win
(http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[commit-bot: I haz the power, 2014-08-22 23:54:34]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-08-22 23:54:35]

Try jobs failed on following builders:
  mac_chromium_rel_swarming on tryserver.chromium.mac
(http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[tonyg, 2014-08-22 23:57:27]

On 2014/08/22 23:54:35, I haz the power (commit-bot) wrote:
> Try jobs failed on following builders:
>   mac_chromium_rel_swarming on tryserver.chromium.mac
>
(http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

Looks legit:

browser_credentials_unittest.TestBrowserCredentials:

Traceback (most recent call last):
run at
/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/unittest/case.py:331
testMethod()
testCredentialsInfrastructure at
tools/telemetry/telemetry/core/browser_credentials_unittest.py:59
ret = browser_cred.LoginNeeded(tab, 'google')
LoginNeeded at tools/telemetry/telemetry/core/browser_credentials.py:63
tab, runner, self._credentials[credentials_type])
TypeError: LoginNeeded() takes exactly 3 arguments (4 given)

Locals:
credentials_type : 'google'
runner           : <telemetry.page.actions.action_runner.ActionRunner object at
0x107257b50>
tab              : {}

[sullivan, 2014-08-25 16:35:59]

The CQ bit was checked by sullivan@chromium.org

[commit-bot: I haz the power, 2014-08-25 16:36:52]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sullivan@chromium.org/485743002/160001

[commit-bot: I haz the power, 2014-08-25 17:35:33]

Committed patchset #9 (160001) as a64e885edb85f66035f35a848c0add3726a663af

[jam, 2014-08-26 17:20:45]

A revert of this CL (patchset #9) has been created in
https://codereview.chromium.org/502533003/ by jam@chromium.org.

The reason for reverting is: The test change is causing flakiness on trybots
even with retries, see

telemetry/core/backends/google_credentials_backend_unittest/TestGoogleCredentialsBackend/testLoginUsingMock:

http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_....

[tonyg, 2014-08-26 17:27:14]

John, thanks for the fix. BTW, is there any easy way we can notice things
ourselves like this in the future w/o manually polling the flakiness dashboard?

Annie, here's the stack:

[----------] 1 test
[ RUN      ]
backends.google_credentials_backend_unittest.TestGoogleCredentialsBackend.testLoginUsingMock

Traceback (most recent call last):
  run at E:\b\depot_tools\python276_bin\lib\unittest\case.py:331
    testMethod()
  testLoginUsingMock at
tools\telemetry\telemetry\core\backends\google_credentials_backend_unittest.py:18
    'document.getElementById("gb")!== null')
  _LoginUsingMock at
tools\telemetry\telemetry\core\backends\form_based_credentials_backend_unittest_base.py:132
    backend.LoginNeeded(tab, ar, config)
  LoginNeeded at
tools\telemetry\telemetry\core\backends\form_based_credentials_backend.py:102
    tab.Navigate(url)
  handler at tools\telemetry\telemetry\unittest\simple_mock.py:92
    expected_call.VerifyEquals(got_call)
  VerifyEquals at tools\telemetry\telemetry\unittest\simple_mock.py:37
    raise Exception('Self %s, got %s' % (repr(self), repr(got)))
Exception: Self Navigate('https://accounts.google.com/'), got
Navigate('https://accounts.google.com/ServiceLogin?continue=https%3A%2F%2Faccounts.google.com%2FManageAccount')

Locals:
  got    :
Navigate('https://accounts.google.com/ServiceLogin?continue=https%3A%2F%2Faccounts.google.com%2FManageAccount')
  got_a  :
'https://accounts.google.com/ServiceLogin?continue=https%3A%2F%2Faccounts.google.com%2FManageAccount'
  i      : 0
  self_a : 'https://accounts.google.com/'

[  FAILED  ]
backends.google_credentials_backend_unittest.TestGoogleCredentialsBackend.testLoginUsingMock
(1 ms)
[----------] 1 test (0 ms total)

[jam, 2014-08-26 17:29:25]

On 2014/08/26 17:27:14, tonyg wrote:
> John, thanks for the fix. BTW, is there any easy way we can notice things
> ourselves like this in the future w/o manually polling the flakiness
dashboard?

The automated way is when
https://code.google.com/p/chromium/issues/detail?id=403906 gets fixed,

until then, I saw this myself when looking at the various builder's stats, i.e.
build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel?numbuilds=200
(yes, painful even for me)

[commit-bot: I haz the power, 2014-09-10 02:35:21]

Patchset 9 (id:??) landed as
https://crrev.com/c3ce3fff3f0f9fd365be29721e9911fbfb1fac47
Cr-Commit-Position: refs/heads/master@{#291705}