I have successfully installed VMS C, C++ and Fortran on AXP, running
OpenVMS V8.4-L2. When I try to install VMS Pascal, it crashes out with
this error:

Installing VSI Pascal compiler...
%VMSINSTAL-I-RESTORE, Restoring product save set B ...
Installing VSI Pascal definition files...
%VMSINSTAL-I-RESTORE, Restoring product save set C ...
Generating PASCAL$C_ROUTINES.PAS/PEN...
Generating STARLET.PAS/PEN...
Generating PASCAL$ACLEDIT_ROUTINES.PAS/PEN...
Generating PASCAL$CLI_ROUTINES.PAS/PEN...
Generating PASCAL$CMA_ROUTINES.PAS/PEN...
Generating PASCAL$CONV_ROUTINES.PAS/PEN...
Generating PASCAL$CVT_ROUTINES.PAS/PEN...
Generating PASCAL$DCX_ROUTINES.PAS/PEN...
Generating PASCAL$DTK_ROUTINES.PAS/PEN...
Generating PASCAL$EDT_ROUTINES.PAS/PEN...
Generating PASCAL$FDL_ROUTINES.PAS/PEN...
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
address=FFFFFFFFFFFF
FFB8, PC=000000007AFA3570, PS=0000001B
%TRACE-F-TRACEBACK, symbolic stack dump follows
image module routine line rel PC abs
PC
DCL 0 00000000000AB570
000000007AFA3570
DCL 0 00000000000AA784
000000007AFA2784
0 FFFFFFFF8008A0D8
FFFFFFFF8008A0D8
DCL 0 00000000000AB564
000000007AFA3564
0 FFFFFFFF8008A0D8
FFFFFFFF8008A0D8
PASCAL 0 000000000044EBAC
000000000044EBAC
PASCAL GEM_CP_VMS PARSE_BOOL_QUAL 4810 0000000000004634
0000000000468ED4
PASCAL GEM_CP_VMS DO_QUALIFIERS 1922 0000000000000DF0
0000000000465690
PASCAL GEM_CP_VMS GEM_CP_MAIN 2532 000000000000169C
0000000000465F3C
0 FFFFFFFF8038FC04
FFFFFFFF8038FC04
%TRACE-I-END, end of TRACE stack dump
VSI Pascal Installation aborted
%VMSINSTAL-F-UNEXPECTED, Installation terminated due to unexpected
event.

VMSINSTAL procedure done at 15:20

Any ideas?
--
Tactical Nuclear Kittens

On Thursday, July 28, 2022 at 11:34:35 AM UTC-4, Single Stage to Orbit wrote:
> I have successfully installed VMS C, C++ and Fortran on AXP, running
> OpenVMS V8.4-L2. When I try to install VMS Pascal, it crashes out with
> this error:
>
> Installing VSI Pascal compiler...
> %VMSINSTAL-I-RESTORE, Restoring product save set B ...
>
> Installing VSI Pascal definition files...
> %VMSINSTAL-I-RESTORE, Restoring product save set C ...
> Generating PASCAL$C_ROUTINES.PAS/PEN...
> Generating STARLET.PAS/PEN...
> Generating PASCAL$ACLEDIT_ROUTINES.PAS/PEN...
> Generating PASCAL$CLI_ROUTINES.PAS/PEN...
> Generating PASCAL$CMA_ROUTINES.PAS/PEN...
> Generating PASCAL$CONV_ROUTINES.PAS/PEN...
> Generating PASCAL$CVT_ROUTINES.PAS/PEN...
> Generating PASCAL$DCX_ROUTINES.PAS/PEN...
> Generating PASCAL$DTK_ROUTINES.PAS/PEN...
> Generating PASCAL$EDT_ROUTINES.PAS/PEN...
> Generating PASCAL$FDL_ROUTINES.PAS/PEN...
> %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
> address=FFFFFFFFFFFF
> FFB8, PC=000000007AFA3570, PS=0000001B
> %TRACE-F-TRACEBACK, symbolic stack dump follows
> image module routine line rel PC abs
> PC
> DCL 0 00000000000AB570
> 000000007AFA3570
> DCL 0 00000000000AA784
> 000000007AFA2784
> 0 FFFFFFFF8008A0D8
> FFFFFFFF8008A0D8
> DCL 0 00000000000AB564
> 000000007AFA3564
> 0 FFFFFFFF8008A0D8
> FFFFFFFF8008A0D8
> PASCAL 0 000000000044EBAC
> 000000000044EBAC
> PASCAL GEM_CP_VMS PARSE_BOOL_QUAL 4810 0000000000004634
> 0000000000468ED4
> PASCAL GEM_CP_VMS DO_QUALIFIERS 1922 0000000000000DF0
> 0000000000465690
> PASCAL GEM_CP_VMS GEM_CP_MAIN 2532 000000000000169C
> 0000000000465F3C
> 0 FFFFFFFF8038FC04
> FFFFFFFF8038FC04
> %TRACE-I-END, end of TRACE stack dump
> VSI Pascal Installation aborted
> %VMSINSTAL-F-UNEXPECTED, Installation terminated due to unexpected
> event.
>
> VMSINSTAL procedure done at 15:20
>
> Any ideas?
> --
> Tactical Nuclear Kittens
That's odd. It didn't do that for me. Do you have older Pascal compilers on your system?
(I don't think so or the installation would have said so). The crash deep in DCL from PARSE_BOOL_QUAL
is certainly inside of CLI$GET_VALUE or CLI$PRESENT. The installation script would have
done a SET COMMAND on the PASCAL.CLD in the kit. We need to compile those PAS files into
PEN files and if there isn't a compiler already on the system, we would use the compiler that is
currently being installed. The fact that we compiled several PAS files into PEN files before the
crash makes it feel like some sort of quota issue maybe?

Try $ SET PROC/DUMP before invoking @VMSINSTAL, that should get you a process dump file (.DMP), which you could look at.

Volker.

On 2022-07-28 16:07:45 +0000, John Reagan said:

> That's odd. It didn't do that for me. ... The fact that we compiled
> several PAS files into PEN files before the crash makes it feel like
> some sort of quota issue maybe?

Entirely FWIW...

I have fixed myriad weird app crashes by adding explicit app startup
quota checks enforcing the minimal process quotas within the tooling.

If the current process settings don't meet the minimums, and don't
approximate the expected ratios (for cases where ratios might exist
among the quotas), the app prints a message and exits.

With a few of these cases, the tooling went further and detected
newly-added users with insufficient values and automatically fixed the
incorrect quotas overnight. But that's not an option here.

Caveat: There was a case where this check-and-fix design went awry many
years ago, with the DECwindows startup. IIRC, there were two sources of
quota truth in that case (oops), and only one had been updated. (nb:
not my bug.)

Defensive programming on OpenVMS usually involves quotas checks, system
parameter checks, the expected and required settings for the DECC$
logical names, having had the app-required startup previously run, and
other such.

These checks have fixed massive numbers of bug reports, and have made
the apps much easier to maintain.

--
Pure Personal Opinion | HoffmanLabs LLC

On Thu, 2022-07-28 at 09:07 -0700, John Reagan wrote:
> > %TRACE-I-END, end of TRACE stack dump
> > VSI Pascal Installation aborted
> > %VMSINSTAL-F-UNEXPECTED, Installation terminated due to unexpected
> > event.
> >
> > VMSINSTAL procedure done at 15:20
> >
> > Any ideas?

> That's odd.  It didn't do that for me.  Do you have older Pascal
> compilers on your system? (I don't think so or the installation would
> have said so).  The crash deep in DCL from PARSE_BOOL_QUAL is
> certainly inside of CLI$GET_VALUE or CLI$PRESENT.  The installation
> script would have done a SET COMMAND on the PASCAL.CLD in the kit. 
> We need to compile those PAS files into PEN files and if there isn't
> a compiler already on the system, we would use the compiler that is
> currently being installed.  The fact that we compiled several PAS
> files into PEN files before the crash makes it feel like some sort of
> quota issue maybe?

Yes, there are no other Pascal compiler on the system, and I don't
think there is any quotas set yet. I have tried SET PROC/DUMP and re-
running the Pascal installation, just to see where that takes me.
--
Tactical Nuclear Kittens

On Thu, 2022-07-28 at 09:18 -0700, Volker Halle wrote:
> Try $ SET PROC/DUMP before invoking @VMSINSTAL, that should get you a
> process dump file (.DMP), which you could look at.

Trying that, thanks.
--
Tactical Nuclear Kittens

Den 2022-07-28 kl. 22:04, skrev Single Stage to Orbit:
> On Thu, 2022-07-28 at 09:07 -0700, John Reagan wrote:
>>> %TRACE-I-END, end of TRACE stack dump
>>> VSI Pascal Installation aborted
>>> %VMSINSTAL-F-UNEXPECTED, Installation terminated due to unexpected
>>> event.
>>>
>>> VMSINSTAL procedure done at 15:20
>>>
>>> Any ideas?
>
>> That's odd.  It didn't do that for me.  Do you have older Pascal
>> compilers on your system? (I don't think so or the installation would
>> have said so).  The crash deep in DCL from PARSE_BOOL_QUAL is
>> certainly inside of CLI$GET_VALUE or CLI$PRESENT.  The installation
>> script would have done a SET COMMAND on the PASCAL.CLD in the kit.
>> We need to compile those PAS files into PEN files and if there isn't
>> a compiler already on the system, we would use the compiler that is
>> currently being installed.  The fact that we compiled several PAS
>> files into PEN files before the crash makes it feel like some sort of
>> quota issue maybe?
>
> Yes, there are no other Pascal compiler on the system, and I don't
> think there is any quotas set yet.

Quotas are *always* set in one way or another.
I take it granted that you have ead the "Installation
Procedure Requirements" in the install manual.

But since the compiler is actally used by the install script,
I'd also look at the "After Installing VSI Pascal" section.

Anyway, if you still have all your quotas at the defaults,
just jack them up and try again.

I have tried SET PROC/DUMP and re-
> running the Pascal installation, just to see where that takes me.

On Thu, 2022-07-28 at 09:07 -0700, John Reagan wrote:
> That's odd.  It didn't do that for me.  Do you have older Pascal
> compilers on your system? (I don't think so or the installation would
> have said so).  The crash deep in DCL from PARSE_BOOL_QUAL
> is certainly inside of CLI$GET_VALUE or CLI$PRESENT.  The
> installation script would have done a SET COMMAND on the PASCAL.CLD
> in the kit.  We need to compile those PAS files into
> PEN files and if there isn't a compiler already on the system, we
> would use the compiler that is currently being installed.  The fact
> that we compiled several PAS files into PEN files before the
> crash makes it feel like some sort of quota issue maybe?

I have managed to get Pascal installed. I simply told it to install
Pascal without all the other bits (STARLET etc), that was successful. I
have now re-run the installer to get all the other bits installed.
Fingers crossed all is good ...
--
Tactical Nuclear Kittens

On 7/28/2022 12:07 PM, John Reagan wrote:
> On Thursday, July 28, 2022 at 11:34:35 AM UTC-4, Single Stage to Orbit wrote:
>> I have successfully installed VMS C, C++ and Fortran on AXP, running
>> OpenVMS V8.4-L2. When I try to install VMS Pascal, it crashes out with
>> this error:

I'm going to assume OpenVMS V8.4-2L2, the version using EV6 instructions.

> The fact that we compiled several PAS files into PEN files before the
> crash makes it feel like some sort of quota issue maybe?
>

Got to ask, has VSI set new default quotas in the new releases, perhaps avoiding
the quotas from 30-40 years ago?

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 2022-07-28, Single Stage to Orbit <alex.buell@munted.eu> wrote:
> I have successfully installed VMS C, C++ and Fortran on AXP, running
> OpenVMS V8.4-L2. When I try to install VMS Pascal, it crashes out with
> this error:
>

Are you using an emulator to run Alpha VMS and if so, which one ?

Simon.

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

On Fri, 2022-07-29 at 13:00 +0000, Simon Clubley wrote:
> > I have successfully installed VMS C, C++ and Fortran on AXP,
> > running OpenVMS V8.4-L2. When I try to install VMS Pascal, it
> > crashes out with
> > this error:
> >
> Are you using an emulator to run Alpha VMS and if so, which one ?

Originally I was using the ES40 emulator, but I've since switched to
AXPbox as that appears to have more fixes included.

If that's still an issue, I can always wait for the x86_64 VSI release
to become available for hobbyist use. The only issue is that it
probably will only work on my Intel based laptop as my main machine is
a mighty AMD threadripper.
--
Tactical Nuclear Kittens

On 2022-07-29, Single Stage to Orbit <alex.buell@munted.eu> wrote:
> On Fri, 2022-07-29 at 13:00 +0000, Simon Clubley wrote:
>> > I have successfully installed VMS C, C++ and Fortran on AXP,
>> > running OpenVMS V8.4-L2. When I try to install VMS Pascal, it
>> > crashes out with
>> > this error:
>> >
>> Are you using an emulator to run Alpha VMS and if so, which one ?
>
> Originally I was using the ES40 emulator, but I've since switched to
> AXPbox as that appears to have more fixes included.
>

That's what I suspected, thanks.

To everyone else: I think asking if someone is running on an emulator,
and if so, which emulator, should be a standard question from now on
when unusual errors such as the one Alex experienced are reported.

Simon.

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

Simon Clubley schrieb am Montag, 1. August 2022 um 02:04:12 UTC+2:
> On 2022-07-29, Single Stage to Orbit <alex....@munted.eu> wrote:
> > On Fri, 2022-07-29 at 13:00 +0000, Simon Clubley wrote:
> >> > I have successfully installed VMS C, C++ and Fortran on AXP,
> >> > running OpenVMS V8.4-L2. When I try to install VMS Pascal, it
> >> > crashes out with
> >> > this error:
> >> >
> >> Are you using an emulator to run Alpha VMS and if so, which one ?
> >
> > Originally I was using the ES40 emulator, but I've since switched to
> > AXPbox as that appears to have more fixes included.
> >
> That's what I suspected, thanks.
>
> To everyone else: I think asking if someone is running on an emulator,
> and if so, which emulator, should be a standard question from now on
> when unusual errors such as the one Alex experienced are reported.
> Simon.
>

I've had a similar failure with installation of pascal in axpbox, when using the IDE disk interface. After switching to SCSI, it worked without problems.

On Mon, 2022-08-01 at 01:15 -0700, osf1 wrote:
> Simon Clubley schrieb am Montag, 1. August 2022 um 02:04:12 UTC+2:
> > On 2022-07-29, Single Stage to Orbit <alex....@munted.eu> wrote:
> > > On Fri, 2022-07-29 at 13:00 +0000, Simon Clubley wrote: 

[ snip ]

> > > Originally I was using the ES40 emulator, but I've since switched
> > > to AXPbox as that appears to have more fixes included.
> > >
> > That's what I suspected, thanks.
> >
> > To everyone else: I think asking if someone is running on an
> > emulator, and if so, which emulator, should be a standard question
> > from now on when unusual errors such as the one Alex experienced
> > are reported.
>
> I've had a similar failure with installation of pascal in axpbox,
> when using the IDE disk interface. After switching to SCSI, it worked
> without problems.

I'm using the sym53c810 scsi bus if thats of any interest, but I have
this:

pci0.15 = ali_ide
{
}

Should I remove it compelely? No IDE devices used in the system.

Thanks,
Alex
--
Tactical Nuclear Kittens

On Thu, 2022-07-28 at 09:07 -0700, John Reagan wrote:
> That's odd.  It didn't do that for me.  Do you have older Pascal
> compilers on your system? (I don't think so or the installation would
> have said so).  The crash deep in DCL from PARSE_BOOL_QUAL
> is certainly inside of CLI$GET_VALUE or CLI$PRESENT.  The
> installation script would have done a SET COMMAND on the PASCAL.CLD
> in the kit.  We need to compile those PAS files into
> PEN files and if there isn't a compiler already on the system, we
> would use the compiler that is currently being installed.  The fact
> that we compiled several PAS files into PEN files before the
> crash makes it feel like some sort of quota issue maybe?

I have found the STARLETPAS.TLB file in the SYS$LIBRARY directory, I
have extracted its contents, it should be possible to run BUILD.COM to
reconstruct the STARLET.PEN files, but what is VMI$KWD? It is not
defined on my system.

Thanks,
Alex
--
Tactical Nuclear Kittens

On Tuesday, August 2, 2022 at 10:01:41 AM UTC-4, Single Stage to Orbit wrote:
> On Thu, 2022-07-28 at 09:07 -0700, John Reagan wrote:
> > That's odd. It didn't do that for me. Do you have older Pascal
> > compilers on your system? (I don't think so or the installation would
> > have said so). The crash deep in DCL from PARSE_BOOL_QUAL
> > is certainly inside of CLI$GET_VALUE or CLI$PRESENT. The
> > installation script would have done a SET COMMAND on the PASCAL.CLD
> > in the kit. We need to compile those PAS files into
> > PEN files and if there isn't a compiler already on the system, we
> > would use the compiler that is currently being installed. The fact
> > that we compiled several PAS files into PEN files before the
> > crash makes it feel like some sort of quota issue maybe?
> I have found the STARLETPAS.TLB file in the SYS$LIBRARY directory, I
> have extracted its contents, it should be possible to run BUILD.COM to
> reconstruct the STARLET.PEN files, but what is VMI$KWD? It is not
> defined on my system.
>
> Thanks,
> Alex
> --
> Tactical Nuclear Kittens

VMI$KWD is the logical name that VMSINSTAL assigns to the
subdirectory of SYS$UPDATE that it creates to temporarily hold
any files that it uses during the installation process. That
directory is deleted and logical deassigned at the end of the
installation process. I think that the KWD is KitWorkingDirectory
or similar.

Single Stage to Orbit schrieb am Dienstag, 2. August 2022 um 16:01:41 UTC+2:
> On Mon, 2022-08-01 at 01:15 -0700, osf1 wrote:
> > Simon Clubley schrieb am Montag, 1. August 2022 um 02:04:12 UTC+2:
> > > On 2022-07-29, Single Stage to Orbit <alex....@munted.eu> wrote:
> > > > On Fri, 2022-07-29 at 13:00 +0000, Simon Clubley wrote:
> [ snip ]

> > I've had a similar failure with installation of pascal in axpbox,
> > when using the IDE disk interface. After switching to SCSI, it worked
> > without problems.
> I'm using the sym53c810 scsi bus if thats of any interest, but I have
> this:
>
> pci0.15 = ali_ide
> {
> }
>
> Should I remove it compelely? No IDE devices used in the system.

Not sure. I've removed the complete ide entry from the config file. But if the block is empty, it also shouldn't be used anywhere in the emulator, 'though I haven't checked the source code of axpbox, to confirm this.

> Thanks,
> Alex
> --
> Tactical Nuclear Kittens

On Tue, 2022-08-02 at 07:39 -0700, Jim wrote:
> > I have found the STARLETPAS.TLB file in the SYS$LIBRARY directory,
> > I have extracted its contents, it should be possible to run
> > BUILD.COM to reconstruct the STARLET.PEN files, but what is
> > VMI$KWD? It is not defined on my system.
>
> VMI$KWD is the logical name that VMSINSTAL assigns to the
> subdirectory of SYS$UPDATE that it creates to temporarily hold
> any files that it uses during the installation process. That
> directory is deleted and logical deassigned at the end of the
> installation process. I think that the KWD is KitWorkingDirectory
> or similar.

That makes a lot of sense now, thanks.
--
Tactical Nuclear Kittens

On Tue, 2022-08-02 at 08:31 -0700, osf1 wrote:
> > I'm using the sym53c810 scsi bus if thats of any interest, but I
> > have
> > this:
> >
> > pci0.15 = ali_ide
> > {
> > }
> >
> > Should I remove it compelely? No IDE devices used in the system.
>
> Not sure. I've removed the complete ide entry from the config file.
> But if the block is empty, it also shouldn't be used anywhere in the
> emulator, 'though I haven't checked the source code of axpbox, to
> confirm this.

I'll give that a try when I get around to re-running the Pascal
installer.
--
Tactical Nuclear Kittens