| Protocol message specifications [message #1797514] |
Wed, 31 October 2018 21:57  |
Eclipse User |
|
|
|
As Ernesto Posse explains in "PapyrusRT Modelling and Code Generation" there are certain limitations in the implementation.
"A port in the outer boundary of a capsule can be connected to an internal
part, in which case it is called a relay port, as it simply relays messages to and
from the internal part. A boundary port not connected to an internal part is
called an external port, and messages arriving there are handled by the capsule's state machine. Similarly, it is through external ports that the capsule's state machine can send messages outside. Non-boundary ports are called internal, and are used by the capsule's state machine to communicate with the capsule's internal parts."
I want to understand if I can design a protocol that contains both external messages that determine the state machine in he capsule and then also messages that are relayed to internal sub-states that may have sub-state machine, or do I need to have separate protocols for external messages and internal messages relayed to sub-states (and eventually sub-state machines).
|
|
|
| Re: Protocol message specifications [message #1797515 is a reply to message #1797514] |
Wed, 31 October 2018 22:42  |
| Eclipse User |
|
|
|
Hi Dimo,
What you describe is a concept that was often requested in legacy tooling: the concept of a cable the can contain multiple connectors, each with their own protocol. It was discussed as a potential addition to the language, but it was never assigned to an MVP - mostly a matter of prioritization where parity with legacy tooling was primordial.
However, you could define a protocol that would do what you want by, for example, have a "forward" property that would specify on which port the message would be forwarded. This means that, on reception, it would process this (or these) elements to be forwarded to their final destination. You may still have to handle synchronisation in some way, as the semantics of the state machines will still process one message, to completion, at the time.
|
|
|
|
| Re: Protocol message specifications [message #1797562 is a reply to message #1797547] |
Thu, 01 November 2018 13:56 |
| Eclipse User |
|
|
|
You're welcome,
By the way, Papyrus for Real Time has its own forum where you have a better chance of interacting with the Papyrus-RT experts out there. Not every one 06 them monitors the newcomers (this) forum.
|
|
|
Powered by
FUDForum. Page generated in 0.08364 seconds
Home