I threw together this CL: https://codereview.chromium.org/2928823003

It compiles, no idea if it runs.

After applying Patch Set 2 on top of 59.0.3071.86 and building with 'enable_supervised_users=false', I got a segmentation fault in browser_sync::ProfileSyncService::IsSignedIn().

I tested moving the following line back into its original place:

    SigninManagerBase* signin = SigninManagerFactory::GetForProfile(profile);

And setting init_params.signin_wrapper as follows (not sure if it's correct):

==================================
    init_params.signin_wrapper =
#if BUILDFLAG(ENABLE_SUPERVISED_USERS)
        base::MakeUnique<SupervisedUserSigninManagerWrapper>(profile, signin);
#else
        base::MakeUnique<SigninManagerWrapper>(signin);
#endif
==================================

After that, opening chrome://settings/ resulted in a segfault in settings::ProfileInfoHandler::IsProfileManagingSupervisedUsers(), so I wrapped that in an #if/#else/#endif using BUILDFLAG(ENABLE_SUPERVISED_USERS).

Then, clicking on "Manage other people" resulted in a CHECK() failure, at which point I gave up.

Supervised users seems to be assumed enabled by code all over the place. I'm starting to question whether proper support for 'enable_supervised_users=false' is feasible, and even if it gets fixed now, how long it would last before it gets out of sync and breaks again. Then, there's also the fact that the 'enable_supervised_users=false' case is not part of the tests, so any build and/or runtime failures will go unnoticed.

Seems this is chromium build related issue which is difficult to triage the issue from te team end.Hence adding TE-NeedsTriageHelp  for further investigation.