Add some better error handling with named pid receive() and receive_forever()

[mjsir911]

This is all kind of assuming that there can only be one pid registered to a name, but I think that's correct?

doesn't add a named process receive_forever test but I can add that in as well pretty easily.

[mjsir911]

this currently doesn't handle the possibility of:

1. main process is registered to a name
2. main process gets sent some data
3. the name gets re-registered to a different process
4. the main process then tries receive()ing on the named subject, to get the data that's existed on the mailbox
5. previous behavior: would succeed, and allow any number of receive()s on names that the process is not registered to.
   current behavior: fails because the process is not registered to the name, even if there is some data for it in the mailbox. maybe this warrants a receive_unchecked()?

[lpil]

Hello!

That's an interesting point about messages in the inbox from a previously registered name. I wonder what would be best to do about that.

[mjsir911]

That's an interesting point about messages in the inbox from a previously registered name. I wonder what would be best to do about that.

maybe just verify that the subject is owned by the current process on subject creation, store it within the namedsubject & then assume it's receivable on?

could also just give out a warning? or something? if receive is run on a non-owned named subject

[lpil]

That would mean that only the subject owner could make the subject, which would defeat the purpose of names!

How about we document that a selector can be used if they want to avoid this check.