Recategorize ARC auth sign in errors.

components/arc/common/auth.mojom

Index: components/arc/common/auth.mojom
diff --git a/components/arc/common/auth.mojom b/components/arc/common/auth.mojom
index 503cb3615f08997216ca32f62a90c683828612db..679e70bf7d3d5ee03818a7b06740e02f48c67f5e 100644
--- a/components/arc/common/auth.mojom
+++ b/components/arc/common/auth.mojom
@@ -6,12 +6,57 @@ module arc.mojom;
 
 // These values describe failure reason of sign-in.
 enum ArcSignInFailureReason {

[Luis Héctor Chávez, 2016/07/28 17:16:40]

This is going to fail validation (and thus cause a connection closure) for the
brief period of time Android is newer than Chrome. We might need to do this in
two steps, first adding the [Extensible] attribute to the enum and once both
sides are in sync, we actually introduce this change.

Alternatively, we can bump the host version and use that bump to detect if we
can send anything with value > 5 from Android if you don't want to do this in
two steps.

[hidehiko, 2016/07/28 17:41:37]

Thank you for pointing this out.
How about adding [Extensible] in this CL?

I was planning to wait for this roll, then to land the ARC side change, to avoid
the error.
Using [Extensible] makes more sense to me, but I think we do not need to split
CLs in this case.
Indeed, conceptually those are independent, but IMHO it could be acceptable
size. Also, anyway, we need to wait for the roll, even with [Extensible].

WDYT?

[Luis Héctor Chávez, 2016/07/28 17:45:41]

That should be done regardless.
Yeah, any way of ensuring that there are no Chrome OS builds with newer Android
than Chrome works :)

[hidehiko, 2016/07/29 04:25:42]

Done.

+  // Negative values are reserved for internal use.
+  // The values are shuffled to keep the backward compatibility.
+  // Next value: 15.
   UNKNOWN_ERROR = 0,
-  NETWORK_ERROR = 1,
-  SERVICE_UNAVAILABLE = 2,
-  BAD_AUTHENTICATION = 3,
-  GMS_CORE_NOT_AVAILABLE = 4,
-  CLOUD_PROVISION_FLOW_FAIL = 5,
+
+  // Mojo errors:
+  // MOJO_VERSION_MISMATCH: is sent when an API is not supported
+  //   due to Host/Instance version mismatch.
+  // MOJO_CALL_TIMEOUT: is sent when a Mojo invocation is started
+  //   but not completed with time out.
+  MOJO_VERSION_MISMATCH = 6,
+  MOJO_CALL_TIMEOUT = 7,
+
+  // Device check-in errors:
+  // DEVICE_CHECK_IN_FAILED: is sent when "check-in" procedure is completed
+  //   but is actually failed.
+  // DEVICE_CHECK_IN_TIMEOUT: is sent when "check-in" procedure started
+  //   but is not completed.
+  // DEVICE_CHECK_IN_INTERNAL_ERROR: is sent when "check-in" procedure is
+  //   started, but some unexpected error situation happens so that it could
+  //   not completed.
+  DEVICE_CHECK_IN_FAILED = 4,
+  DEVICE_CHECK_IN_TIMEOUT = 8,
+  DEVICE_CHECK_IN_INTERNAL_ERROR = 9,
+
+  // GMS errors:
+  // GMS_NETWORK_ERROR: is sent when GMS sign-in procedure is started, but
+  //   it reports NETWORK_ERROR.
+  // GMS_SERVICE_UNAVAILABLE: is sent when GMS sign-in procedure is started
+  //   but it reports SERVICE_UNAVAILABLE. Note that this is not generic
+  //   service unavailable error code.
+  // GMS_BAD_AUTHENTICATION: is sent when GMS sign-in procedure is started
+  //   but it reports BAD_AUTHENTICATION.
+  // GMS_SIGN_IN_FAILED: is sent when GMS sign in procedure is started but
+  //   it reports an error other than above NETWORK_ERROR, SERVICE_UNAVAILABLE
+  //   or BAD_AUTHENTICATION. In general, we do not expect this kind of cases.
+  // GMS_SIGN_IN_TIMEOUT: is sent when GMS sign in procedure is started but
+  //   not returned.
+  // GMS_SIGN_IN_INTERNAL_ERROR: is sent when GMS sign in procedure is started
+  //   but could not completed.
+  GMS_NETWORK_ERROR = 1,
+  GMS_SERVICE_UNAVAILABLE = 2,
+  GMS_BAD_AUTHENTICATION = 3,
+  GMS_SIGN_IN_FAILED = 10,
+  GMS_SIGN_IN_TIMEOUT = 11,
+  GMS_SIGN_IN_INTERNAL_ERROR = 12,
+
+  // Cloud provisioning errors.
+  CLOUD_PROVISION_FLOW_FAILED = 5,
+  CLOUD_PROVISION_FLOW_TIMEOUT = 13,
+  CLOUD_PROVISION_FLOW_INTERNAL_ERROR = 14,
 };
 
 interface AuthHost {