Fix password automation hooks to get pyauto test testSavePassword working.

Fix password automation hooks to get pyauto test testSavePassword working.

Previously, all passwords tests were diabled in pyauto.  This CL
re-enables passwords.PasswordTest.testSavePassword (by explicitly
disabling all other passwords tests).

BUG=73078
TEST=None

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=86154

[dennis_jeffrey, 2011-05-10 01:31:37]

The description for patch set 2 says "Disabling a single pyauto passwords test".
 It should be "Enabling" instead of "Disabling".

[Nirnimesh, 2011-05-10 19:25:51]

Automation portions LGTM

Please get review from someone for content/browser/*

http://codereview.chromium.org/6994005/diff/2001/chrome/browser/automation/au...
File chrome/browser/automation/automation_provider_observers.cc (right):

http://codereview.chromium.org/6994005/diff/2001/chrome/browser/automation/au...
chrome/browser/automation/automation_provider_observers.cc:1689: const
std::string result_key)
std::string&

http://codereview.chromium.org/6994005/diff/2001/chrome/test/functional/passw...
File chrome/test/functional/passwords.py (right):

http://codereview.chromium.org/6994005/diff/2001/chrome/test/functional/passw...
chrome/test/functional/passwords.py:57: # There may be multiple logins for the
same 'signon_realm', since other
Wouldn't tests begin with a clean profile?

[dennis_jeffrey, 2011-05-17 01:40:21]

I addressed the two comments and synced with trunk.

The changes in content/browser/* have been removed because they are no longer
needed (due to http://codereview.chromium.org/7001015/).

I also enabled the passwords tests on all platforms except Mac and Linux (they
were previously disabled in lots of places, but never enabled).  I verified that
all passwords tests currently pass on ChromeOS (with --repeat=20).  I did not
check on Windows.

http://codereview.chromium.org/6994005/diff/2001/chrome/browser/automation/au...
File chrome/browser/automation/automation_provider_observers.cc (right):

http://codereview.chromium.org/6994005/diff/2001/chrome/browser/automation/au...
chrome/browser/automation/automation_provider_observers.cc:1689: const
std::string result_key)
On 2011/05/10 19:25:52, Nirnimesh wrote:
> std::string&

Done.

http://codereview.chromium.org/6994005/diff/2001/chrome/test/functional/passw...
File chrome/test/functional/passwords.py (right):

http://codereview.chromium.org/6994005/diff/2001/chrome/test/functional/passw...
chrome/test/functional/passwords.py:57: # There may be multiple logins for the
same 'signon_realm', since other
On 2011/05/10 19:25:52, Nirnimesh wrote:
> Wouldn't tests begin with a clean profile?

Yes, this is true in Windows and ChromeOS (the only platforms for which the
passwords tests are enabled).  I added a setUp() function to assert this on the
start of each test in this file.

[Nirnimesh, 2011-05-17 20:23:25]

LGTM

http://codereview.chromium.org/6994005/diff/15001/chrome/test/functional/pass...
File chrome/test/functional/passwords.py (right):

http://codereview.chromium.org/6994005/diff/15001/chrome/test/functional/pass...
chrome/test/functional/passwords.py:30:
self.assertEqual(self.GetSavedPasswords(), [])
self.assertFalse(self.GetSavedPasswords())