[Chrome OS] Respect periods in google apps account names

[Chrome OS] Respect periods in google apps account names

When logging in to @gmail.com accounts via the web, . characters are
ignored -- which we do on chrome OS.  When logging into google apps
accounts for OTHER domains, . characters should NOT be ignored.  We
were canonicalizing them out on Chrome OS, breaking login for apps
users who have . embedded in their usernames.

BUG=chromium-os:10397
TEST=unit tests, install on device and try cmasone@gmail.com, c.masone@gmail.com, why.you@mydomain.org and whyyou@mydomain.org.  The first three work, the last one does not, as expected.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=70318

[whywhat, 2010-12-30 17:37:39]

LGTM

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator.cc (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator.cc:35: // account servers do not
tolerate them, so we don't need to either.
TODO or actually remove the code? :)

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator.h (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator.h:26: static const char
kSpecialCase[];
Could be more meaningful, i.e. kTolerantDomain. A comment about what it is is
welcome too.

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator_unittest.cc (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator_unittest.cc:38:
first.append(Authenticator::kSpecialCase);
I'd just use "gmail.com" here to improve readability a little bit.

[Chris Masone, 2010-12-30 18:52:50]

will push when the tree reopens

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator.cc (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator.cc:35: // account servers do not
tolerate them, so we don't need to either.
On 2010/12/30 17:37:39, whywhat wrote:
> TODO or actually remove the code? :)

Added a TODO instead of pulling it out, as I want to keep the change as small as
possible, as we need to backport it to 552d and 597

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator.h (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator.h:26: static const char
kSpecialCase[];
On 2010/12/30 17:37:39, whywhat wrote:
> Could be more meaningful, i.e. kTolerantDomain. A comment about what it is is
> welcome too.

I went with kSpecialCaseDomain, and added a comment.

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/authenticator_unittest.cc (right):

http://codereview.chromium.org/6089004/diff/2001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/authenticator_unittest.cc:38:
first.append(Authenticator::kSpecialCase);
On 2010/12/30 17:37:39, whywhat wrote:
> I'd just use "gmail.com" here to improve readability a little bit.

ok.