John Cormie
John Cormie
> This is the _destination_ identity? And this is for cross-user binder calls? I guess the system might do some of that. Correct, it's the destination Android user for cross-user...
> > > We're trying to remove most usages of Attributes. In general, we are replacing them with the same pattern of having a reference-equality Key with a generic for...
I've thought about it and favor the approach in this PR because it: 1) fits with the existing/intended architecture of a single global NRR with one shared NRP of each...
Replaced by https://github.com/grpc/grpc-java/pull/11669
> However, it is a bit misleading that the class is marked as `@ThreadSafe`, as this should imply that the class must safely handle multiple concurrent calls to `handleTransaction()`. I...
> @jdcormie What is the problem with just removing @threadsafe from `BinderTransport` without changing which class implements LeakSafeOneWayBinder.TransactionHandler? Because BinderTransport is *intended* to be thread safe from the perspective of...
Parcelables as message payloads probably has similar issues.
Looks like the check failures above are real: ``` /grpc-java/binder/src/main/java/io/grpc/binder/SecurityPolicies.java:125: warning: [static] static variable should be qualified by type name, PackageManager, instead of by an expression "android", packageManager.GET_SIGNING_CERTIFICATES); ^ /grpc-java/binder/src/main/java/io/grpc/binder/SecurityPolicies.java:133:...
@ejona86 Switching off of java.net.URI seems infeasible. Any ideas on how this could be fixed anyway, even if just for the special case of `intent:#Intent;`?