Jeevan Mathew(deleted)
|
Re: Queston about "MsgDeliverEvent"
|
Jeevan Mathew(deleted)
02/15/2010 12:58 PM
post47322
|
Re: Queston about "MsgDeliverEvent"
Hope I imagined/guessed your doubts the right way..
>
> Isn't it much easier to use "SignalWaitinfo" insteand of "MsgReceivePulse"?
> What is difference between "SignalWaitinfo" and "MsgReceivePulse"
> and what is the relation between "signal" and "pulse" in QNX neutrino? Does
> pulse is implemented by using "signal" or vice vesa in QNX neutrino?
>
Perhaps this clears you doubts:
http://www.qnx.com/developers/docs/6.4.1/neutrino/lib_ref/s/sigevent.html
> 2.
> Once server get "rcvid" for a client, could this "rcvid" be used to call "
> MsgDeliverEvent" multi-times ?
Yes.
>Or does it become invalid once "MsgDeliverEvent
> " is called?
No. except by programmer's fault ?
> I means, could Server keep use this "rcvid" to trigger MsgSend
> from client?
If the client does have a I-let-me-trigger-routine for doing so, yes.
> Does "rcvid" value of client A is fixed as long as client A
> alive?
I think that rcvid is client-unique, but depends on the channel the client called ConnectAttach() on.
(a client could have done :
coid1=ConnectAttach(chid1..
coid2=ConnectAttach(chid2..
MsgSend(coid1...)
MsgSend(coid2...)
results in rcvid1=3 and rcvid2=4 on server side)
>or Is it changed whenever client A send a message ?
no.
>
> 3. Is thereany issue if I use MsgSendPulse/MsgSendReceive instead of
> MsgDeliverEvent/SignalWaitinfo to trigger MsgSend from client ?
>
???
The issue is mostly the design of the msg-passing interface, where imho signals should be used in trad-unix way, this
doesn't mean that you cannot do it the other way around.
> Thanks,
> Dooeui
Hope this helps.
-Jeevan
|
|
|