On 2023-09-11, gah4 <gah4@u.washington.edu> wrote:
>
> It was a C program that I wrote, and that had exit(8); in it.
>
> There was no access violation.
>

I am surprised nobody told you the error message could be suppressed
if you set the right bit in the exit status you return to DCL:

$ exit 8
%SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000000000008, PC=000000007AF0D9F0, PS=00000000
$ write sys$output $status
%X00000008

$ exit %x10000008
$ write sys$output $status
%X10000008

Note that it's that specific bit which needs setting as you still get
the message in this case:

$ exit %x40000008
%SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000040000008, PC=000000007AF0D9F0, PS=00000000
$ write sys$output $status
%X40000008

You can mask off the top 4 bits (the internal control flags) to
extract out the actual number your program returned to your DCL script.

Not sure where I found this out because it's not in the manuals
I just checked.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On Tuesday, September 12, 2023 at 1:34:40 PM UTC+1, Simon Clubley wrote:
> On 2023-09-11, gah4 <ga...@u.washington.edu> wrote:
> >
> > It was a C program that I wrote, and that had exit(8); in it.
> >
> > There was no access violation.
> >
>
> I am surprised nobody told you the error message could be suppressed
> if you set the right bit in the exit status you return to DCL:
>
> $ exit 8
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000000000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X00000008
>
> $ exit %x10000008
> $ write sys$output $status
> %X10000008
>
> Note that it's that specific bit which needs setting as you still get
> the message in this case:
>
> $ exit %x40000008
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000040000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X40000008
>
> You can mask off the top 4 bits (the internal control flags) to
> extract out the actual number your program returned to your DCL script.
>
> Not sure where I found this out because it's not in the manuals
> I just checked.
>
> Simon.
>
> --
> Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
> Walking destinations on a map are further away than they appear.

See the OpenVMS Calling Standard chapter on OpenVMS Conditions where all this is documented https://docs.vmssoftware.com/vsi-openvms-calling-standard/#VMSCONDITIONS_CHAP

> See the OpenVMS Calling Standard chapter on OpenVMS Conditions [...]

Also: HELP CRTL exit

its $ type exit8.c
#include <stdlib.h>

int main( int argc, char **argv)
{ exit( 8);
} its $

its $ cc exit8
its $ link exit8
its $ run exit8
%SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=000000000000
0008, PC=FFFFFFFFFF561663, PS=0000000F

its $ cc exit8 /define = _POSIX_EXIT /obj=exit8p
its $ link exit8p
its $ run exit8p
its $ write sys$output $status
%X0035A041
its $ write sys$output f$message( %X0035A041)
%C-S-EXIT8, exit value 8

> [...] it's not in the manuals I just checked.

Valuable info for the psychics who know which "the manuals" those
might be.

On 2023-09-12 12:34:37 +0000, Simon Clubley said:

> On 2023-09-11, gah4 <gah4@u.washington.edu> wrote:
>>
>> It was a C program that I wrote, and that had exit(8); in it.
>>
>> There was no access violation.
>>
>
> I am surprised nobody told you the error message could be suppressed
> if you set the right bit in the exit status you return to DCL:
>
> $ exit 8
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual
> address=0000000000000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X00000008
>
> $ exit %x10000008
> $ write sys$output $status
> %X10000008
>
> Note that it's that specific bit which needs setting as you still get
> the message in this case:
>
> $ exit %x40000008
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual
> address=0000000040000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X40000008
>
> You can mask off the top 4 bits (the internal control flags) to
> extract out the actual number your program returned to your DCL script.
>
> Not sure where I found this out because it's not in the manuals
> I just checked.
>
> Simon.

See $STSDEF ;)

— VAXman

On 9/12/2023 8:34 AM, Simon Clubley wrote:
> On 2023-09-11, gah4 <gah4@u.washington.edu> wrote:
>> It was a C program that I wrote, and that had exit(8); in it.
>>
>> There was no access violation.
>
> I am surprised nobody told you the error message could be suppressed
> if you set the right bit in the exit status you return to DCL:
>
> $ exit 8
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000000000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X00000008
>
> $ exit %x10000008
> $ write sys$output $status
> %X10000008
>
> Note that it's that specific bit which needs setting as you still get
> the message in this case:
>
> $ exit %x40000008
> %SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual address=0000000040000008, PC=000000007AF0D9F0, PS=00000000
> $ write sys$output $status
> %X40000008
>
> You can mask off the top 4 bits (the internal control flags) to
> extract out the actual number your program returned to your DCL script.
>
> Not sure where I found this out because it's not in the manuals
> I just checked.

It is documented and has been as long as I can remember.

$ type simon.c
#include <string.h>

#include <stsdef.h>
#include <starlet.h>

int main(int argc, char *argv[])
{ if(argc > 1 && strcasecmp(argv[1], "simonsaysquiet") == 0)
{
sys$exit(STS$M_INHIB_MSG | 8);
}
else
{
sys$exit(8);
}
} $ cc simon
$ link simon
$ simon :== $sys$disk:[]simon.exe
$ simon
%SYSTEM-W-ACCVIO, access violation, reason mask=00, virtual
address=0000000000000008, PC=0000000000000012, PS=00000003
$ sh symb $status
$STATUS == "%X00000008"
$ simon SimonSaysQuiet
$ sh symb $status
$STATUS == "%X10000008"

Arne

On 2023-09-12, Arne Vajhøj <arne@vajhoej.dk> wrote:
>
> It is documented and has been as long as I can remember.
>

Yes, but I couldn't remember where. It's one of things you read about,
file the knowledge away for later use and then forget many years later
where you read it.

I thought it was in the message manual, but the status breakdown in
that manual didn't cover the top 4 bits when I looked. Ian ended up
posting the answer. Thanks Ian.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On Wednesday, September 13, 2023 at 1:13:21 PM UTC+1, Simon Clubley wrote:
> On 2023-09-12, Arne Vajhøj <ar...@vajhoej.dk> wrote:
> >
> > It is documented and has been as long as I can remember.
> >
> Yes, but I couldn't remember where. It's one of things you read about,
> file the knowledge away for later use and then forget many years later
> where you read it.
>
> I thought it was in the message manual, but the status breakdown in
> that manual didn't cover the top 4 bits when I looked. Ian ended up
> posting the answer. Thanks Ian.
> Simon.
>
> --
> Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
> Walking destinations on a map are further away than they appear.

The key thing to remember is that they are called Condition Values then one can search the manuals for the details :-)