Can you add a kill primitive that kills all the sessions and closes the program.
I think that the exit primitive does not suffice since when I use the exit primitive it waits for the end of the active sessions.
Adding a forceful service termination may indeed be useful.
The only doubt I have is if it should go in as an operation of the Runtime standard service or as a language primitive; probably the first. Something like kill@Runtime or die@Runtime.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am not so sure that a kill primitive for aborting an entire service could be safe. I cannot manage a correct ending of the current sessions. It is probably safe to introduce a specific fault (raised when an exit primtive is executed in a session) whose handler could be installed into the main scope in order to allow the programmer for introducing some safe termination activities.
Furthermore, I suggest to introduce a primitive for exiting from a single session without affecting the other ones.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I like the idea of making other sessions handling a fault. How could we implement that semantically? Something along the line of "the next instruction executed by every session will throw a fault" ?
Forcefully terminating the current session without affecting the other ones can already be done by throwing a fault. This, however, cannot interrupt pending Solicit-Response calls (which will wait for the response nevertheless, iirc Jacopo, the original requester, raised this as a problem) which, by the way, could be a sane behaviour considering that even if you are forcefully exiting you probably want your parallel activities in the same session to execute their termination handlers.
If we go this way we have to decide a name for this fault. Exit? ExitFault? Mmh.. :-)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Adding a forceful service termination may indeed be useful.
The only doubt I have is if it should go in as an operation of the Runtime standard service or as a language primitive; probably the first. Something like kill@Runtime or die@Runtime.
I am not so sure that a kill primitive for aborting an entire service could be safe. I cannot manage a correct ending of the current sessions. It is probably safe to introduce a specific fault (raised when an exit primtive is executed in a session) whose handler could be installed into the main scope in order to allow the programmer for introducing some safe termination activities.
Furthermore, I suggest to introduce a primitive for exiting from a single session without affecting the other ones.
I like the idea of making other sessions handling a fault. How could we implement that semantically? Something along the line of "the next instruction executed by every session will throw a fault" ?
Forcefully terminating the current session without affecting the other ones can already be done by throwing a fault. This, however, cannot interrupt pending Solicit-Response calls (which will wait for the response nevertheless, iirc Jacopo, the original requester, raised this as a problem) which, by the way, could be a sane behaviour considering that even if you are forcefully exiting you probably want your parallel activities in the same session to execute their termination handlers.
If we go this way we have to decide a name for this fault. Exit? ExitFault? Mmh.. :-)
Last comment was mine, I forgot to login. I'm putting the feature tracker on compulsory login.
Also this seems pretty old to me...maybe we should close?
This has been fixed by introducing the halt@Runtime operation in runtime.iol.