I ported libchrome, libmojo and the
mojo service manager from cros project.
A question maybe about mojo idiom.
One interesting design of mojo is the remote is considered bound once it is bound to the message pipe in the local process, and the process can start using this remote to make mojo IPC call. If the receiver end is not yet online/running, the IPC call is cached locally until the receiver end is available.
What's the general/best practice here when I want to design/think in the mojo way?
If a process A is connected to mojo service manager and has requested service Foo. Should A start making Foo service IPC calls, and then patiently wait for the response to trigger the next step work flow on A side? Does that mean some sort of timer based IPC call is anti-mojo-idiom, because it may have the risk of caching too many pending IPC calls? Or a more high level question, how should I decide when to start making mojo IPC call if I don't know if there is a remote receiver handling this call.
Quite a newbie to mojo, and appreciate your insights and sharing !
Tao