* Annex E.5 System.RPC
@ 2018-04-02 19:39 Dmitry A. Kazakov
2018-04-06 13:40 ` Marius Amado-Alves
0 siblings, 1 reply; 2+ messages in thread
From: Dmitry A. Kazakov @ 2018-04-02 19:39 UTC (permalink / raw)
1. Why the name is fixed "System.RPC" instead of being specified in by a
representation clause, e.g. in order to have multiple independent
implementations/transports?
2. It seems that the design of Do_RPC and Do_APC interface requires all
parameters stored in Params_Stream_Type. The contents is copied to the
transport. This is one unnecessary copying when the transport has a
stream interface, and also a restriction to have memory ready to keep
all parameters and results cached. The transport cannot establish a
channel to the callee because Do_RPC/APC are called *after" writing the
stream, when the callee is yet unknown. Was it considered *first*
initiate RPC/APC and *then* serialize the parameters?
--
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: Annex E.5 System.RPC
2018-04-02 19:39 Annex E.5 System.RPC Dmitry A. Kazakov
@ 2018-04-06 13:40 ` Marius Amado-Alves
0 siblings, 0 replies; 2+ messages in thread
From: Marius Amado-Alves @ 2018-04-06 13:40 UTC (permalink / raw)
No anwser, but I too hope this is "fixed" because in my mind Annex E is one of the things that makes Ada superior, as I know of no other language with such level of distribution support.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2018-04-06 13:40 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-04-02 19:39 Annex E.5 System.RPC Dmitry A. Kazakov
2018-04-06 13:40 ` Marius Amado-Alves
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox