[Info-vax] About $ run /mailbox DCL command

[A famous IT technical writer]

On Jan 4, 6:10 pm, moro... at world.std.spaamtrap.com (Michael Moroney)
wrote:
> hb <becker.... at gmail.com> writes:
> >On Jan 4, 4:53=A0pm, IanMiller <g... at uk2.net> wrote:
> >> I don't think the mailbox driver supports a timeout directly.
> >> In a program you can set a timer and abort the I/O.
> >The mailbox driver supports IO$M_NOW. That's not a timeout, or only a
> >very short one :-) It looks like the DCL read doesn't set this I/O
> >modifier for mailboxes; neither by default nor on request. Coding an
> >own read utility with this looks like a SMOP.
>
> If I recall correctly, READ/TIME_OUT=0 does set IO$M_NOW on a mailbox
> read.  No other values for /TIME_OUT are supported or work.

Me too, I have no working pure DCL solution to offer Tristan. As I
tested it, the DCL read/time_out just hanged my VMS session even not
responding to CTRL-Y. I did propose Tristan some C code which does
what he wants. However he said he has already one but was forward
looking a pure DCL solution. If someone is interested with my C code,
let it me know by email you'll find at http://vouters.dyndns.org/ If
it happens there is demand for, I may especially write an article
describing everything I know on this subject. This will describe the
read/time_out non DCL solution and the corresponding workaround which
interfaces with DCL.

As another topic related subject, I misunderstood what the OpenVMS
manual say about lib$set_symbol. Jim's f$cvui() lexical function made
me more rework this subject. I thus wrote a special article upon at
http://vouters.dyndns.org/tima/OpenVMS-lib$set_symbol-lib$set_symbol_only_set_a_string_symbol_type.html
If I did not clearly understood what the VMS manual states, this may
be as well unclear to other readers. Hence this special article
attempting to make things clear for everyone about what one can expect
from this VMS routine.

Philippe