![]() process types) of the same executable image so that different types of processes don't trample each others' prefetch behavior. We reserve 8 to mean "whatever", and this will ultimately lead to processes with /prefetch:8 having inconsistent behavior thus disabling prefetch in practice. TODO(rockot): Make it possible for embedders to override this argument on a per-service basis. ↪įlag indicating that a resource provider must be set up to provide cast receiver with resources. Apps cannot start until provided resources. This flag implies -alsa-check-close-timeout=0. ↪Īdd additional capabilities to the AppContainer sandbox on the GPU process. M tools/metrics/histograms/histograms.↪Īdd additional capabilities to the AppContainer sandbox used for XR compositing. M chrome/browser/chromeos/arc/session/arc_session_manager.h ![]() M chrome/browser/chromeos/arc/session/arc_session_ M chrome/browser/chromeos/arc/auth/arc_auth_service.h ![]() M chrome/browser/chromeos/arc/auth/arc_auth_ M chrome/browser/chromeos/arc/arc_optin_uma.h TEST=deploy to dut, disable playstore, re-enable playstore, successfullyĬhange-Id: I8bb83f7366260f1a09223e2791e5cfab7e94ca68Ĭr-Commit-Position: chrome/browser/chromeos/arc/arc_optin_uma.cc Introducing cloud dpc error enum values to be used for showing helpful OnAuthorizationComplete(Status status, bool initial_signin, string account_name)Īrc: Add support for authorization failure reason After the refactor (in a separate CL) the method may look like ```ĪrcSignInSuccess success // success and already provisioned enumĪrcSignInError error // union of all different kinds of subcode enums and general error enum initial signin vs account_name by putting them in a union would not be technically correct. ![]() if it is not an initial signin, it could still be primary device thus making it mutually exclusive i.e. I'll add a TODO note here and address this signature change after separating out the error and success codes in a separate enum and updating the android and chrome side in different shorter refactoring-only CLs.Īlso it is not clear that account_name must be non-null when initial_signin is false i.e. Mixing functional changes with large refactoring can easily cause bugs to be introduced and makes it hard to understand what went wrong when looking back at git history. I agree that more explicit contract is safer and I'd like to add that the refactoring proposed here is non-trivial. OK, I took a closer look but I think we should still just add a new method. Patch Set #2, Line 335: ArcSignInError? error)
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |