Add credential passing API for Chrome OS SAML login

Add credential passing API for Chrome OS SAML login

This CL adds an API that SAML IdPs can use to pass credentials to
Chrome OS during login instead of relying on password scraping.

BUG=127095
TEST=New browser test + manual

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=245480

[bartfab (slow), 2014-01-14 19:41:10]

Hi Xiyuan,

Could you please review?

[xiyuan, 2014-01-14 21:21:29]

Awesome. LGTM.

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
File chrome/browser/resources/gaia_auth/background.js (right):

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
chrome/browser/resources/gaia_auth/background.js:23: *       of the scraped
passwords by sending the |getScrapedPasswords| message.
Thank you for the better documentation.

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
File chrome/browser/resources/gaia_auth/saml_injected.js (right):

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
chrome/browser/resources/gaia_auth/saml_injected.js:40: if (event.source !=
window || event.data.type != 'gaia_saml_api')
nit: I might be a bit paranoid but would put more check before accessing
event.data.type just in case saml idp page uses postMessage to itself as well.

Something like
if (event.source != window ||
    typeof event.data != 'object' ||
    !event.data.hasOwnProperty('type') ||
    event.data.type != 'gaia_saml_api')
...

[bartfab (slow), 2014-01-15 11:00:29]

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
File chrome/browser/resources/gaia_auth/saml_injected.js (right):

https://codereview.chromium.org/138133006/diff/1/chrome/browser/resources/gai...
chrome/browser/resources/gaia_auth/saml_injected.js:40: if (event.source !=
window || event.data.type != 'gaia_saml_api')
On 2014/01/14 21:21:30, xiyuan wrote:
> nit: I might be a bit paranoid but would put more check before accessing
> event.data.type just in case saml idp page uses postMessage to itself as well.
> 
> Something like
> if (event.source != window ||
>     typeof event.data != 'object' ||
>     !event.data.hasOwnProperty('type') ||
>     event.data.type != 'gaia_saml_api')
> ...
> 

You are not paranoid at all. I agree that a conflict with messages used
internally by the SAML provider is a danger. I picked the signature
'gaia_saml_api' in the hopes that it will not clash with anything SAML providers
are likely to use. I am wondering whether a message signature with 'google' in
it would be even better, as third-party companies are unlikely to use our
company name in their implementation.

And the extra checks you suggested make very good sense as well of course. Done.

[xiyuan, 2014-01-15 17:08:40]

LGTM++

[commit-bot: I haz the power, 2014-01-16 17:43:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bartfab@chromium.org/138133006/70001

[commit-bot: I haz the power, 2014-01-16 20:29:52]

Retried try job too often on win_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2014-01-16 23:24:30]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bartfab@chromium.org/138133006/70001

[commit-bot: I haz the power, 2014-01-17 10:12:02]

Change committed as 245480