Hi David,

I noticed some issues with net::CookieStoreIOS::AddCallbackForCookie method and since we don't have anyone on iOS team who understands this method, I was wandering if you have someone in mind who can help.

There are 2 ways how net::CookieStoreIOS is used:
1.) Backed up by CookieMonster::PersistentCookieStore and not synchronized with NSHTTPCookieStore (created via CookieStoreIOS constructor)
2.) Synchronized with NSHTTPCookieStore but not backed up by any storage (created via CreateCookieStore factory method).

We had unit tests which tested non-existing configuration (backed up by CookieMonster::PersistentCookieStore and synchronized with NSHTTPCookieStore), so I changed them to use configuration #2 (created via CreateCookieStore). But after the change tests actual result changed and I don't know if that was a bug or expected behavior: https://codereview.chromium.org/2601973002/diff/1/ios/net/cookies/cookie_store_ios_unittest.mm#newcode576

Thanks!

I know absolutely nothing about cookie storage stuff and am at a conference for the next week. :-)

+mmenke might know [someone who knows] more about that stuff.

This is not urgent at all. If this is a bug then we had this bug pretty much since M48.

[+droger, +marq] The iOS CookieStore isn't owned by the Chrome network stack team.  If it turns out this looks like a CookieMonster bug, I'm happy to help investigate.

Matt, if you are familiar with net::CookieStore interface, could you please tell if the fact that actual results were changes is a bug or not? This is not a CookieMonster bug, because behavior has changed for net::CookieStoreIOS which is not backed up by CookieMonster. So if it's a bug, then it's ios specific bug.

Unfortunately, the iOS cookie store is backed by the iOS platform cookie store.  There are plenty of cases where the iOS cookie store behaves differently from ours.  I have no idea if these are those cases (Nor exactly what those cases are).

You could just run the tests with CookieMonster, and see what happens.  For the 1U -> 2U changes, the old code looks correct.  Otherwise, the code is being notified of a cookie change that happened before it set the cookie change callback.

I can't make any sense of the other one.

Thank you, Matt.

CCing Benoit, because CookieStoreIO created for SignIn via CreateCookieStore has this bug.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot