I reached out to Tim Sneddon, but haven't heard back from him yet, so I
forged ahead. I have released MMK V5.1-1 that will run on X86_64, in
addition to VAX, Alpha, and I64.

The X86_64 version was compiled using the native C compiler under
OpenVMS V9.2-1:

VSI C X7.4-785 (GEM 50X65) on OpenVMS x86_64 V9.2-1

I made modifications to a number of the files in the MMK github
repository. Those changes have not been fed back to github yet, as I
want to coordinate that with Tim.

You can download MMK V5.1-1 from the Process Software freeware archive:

https://vms.process.com/scripts/fileserv/fileserv.com?MMK

I supply both VMSINSTAL and PCSI kits. The current sources are included
in the .F saveset, if you want my modified sources.

These kits will install and run on VAX V6.1 or higher, Alpha V6.2 or
higher, I64 V8.2 or higher, and X86_64 V9.2-1.

I've made a few of the freeware packages available for X86_64, so far. I
hope to add more in the coming months.

Search the FILESERV for packages that run on x86_64

https://www.process.com/resources/openvms/

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 7/13/23 3:42 PM, Hunter Goatley wrote:
> I reached out to Tim Sneddon, but haven't heard back from him yet, so I
> forged ahead. I have released MMK V5.1-1 that will run on X86_64, in
> addition to VAX, Alpha, and I64.
>
> The X86_64 version was compiled using the native C compiler under
> OpenVMS V9.2-1:
>
>    VSI C X7.4-785 (GEM 50X65) on OpenVMS x86_64 V9.2-1
>
> I made modifications to a number of the files in the MMK github
> repository. Those changes have not been fed back to github yet, as I
> want to coordinate that with Tim.

The usual way to coordinate is to send a pull request as I already did
some months ago:

https://github.com/endlesssoftware/mmk/pull/101

You may have noticed things I didn't (haven't looked yet) but it's a
shame to duplicate efforts.

> You can download MMK V5.1-1 from the Process Software freeware archive:
>
> https://vms.process.com/scripts/fileserv/fileserv.com?MMK
>
> I supply both VMSINSTAL and PCSI kits. The current sources are included
> in the .F saveset, if you want my modified sources.
>
> These kits will install and run on VAX V6.1 or higher, Alpha V6.2 or
> higher, I64 V8.2 or higher, and X86_64 V9.2-1.
>
> I've made a few of the freeware packages available for X86_64, so far. I
> hope to add more in the coming months.
>
> Search the FILESERV for packages that run on x86_64
>
> https://www.process.com/resources/openvms/
>

On 7/13/2023 5:00 PM, Craig A. Berry wrote:
>
> The usual way to coordinate is to send a pull request as I already did
> some months ago:
>
> https://github.com/endlesssoftware/mmk/pull/101
>
Thanks, Craig. I have never used github for anything but downloading
kits, so I knew nothing about this.

> You may have noticed things I didn't (haven't looked yet) but it's a
> shame to duplicate efforts.

Agreed, it is. But the source changes weren't bad. I spent more time
fixing problems in the KITINSTAL.COM and making it consistent with the
PCSI kits.

It looks like I did everything you did except for the bit about defining
MMSVAX, MMSALPHA, etc. That's a nice change, and one that I had
overlooked (those are also symbols I didn't know about and have never
used ;-) ). We also took different approaches to the MACRO compiler
warning and checking for not-VAX, but the end results are the same.

Tim replied to me overnight, and I have sent my changes to him. I'll let
him know about your pull request, too, as I'm assuming he hasn't seen
it, as he's had no time for MMK for a long time.

Thanks!

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 7/14/2023 2:06 PM, Hunter Goatley wrote:
>
> It looks like I did everything you did except for the bit about defining
> MMSVAX, MMSALPHA, etc. That's a nice change, and one that I had
> overlooked (those are also symbols I didn't know about and have never
> used ;-) ).

That bit me yesterday, when I found that MMSIA64 was defined on X86. I
took Craig's approach to fix it, though I rewrote the code because VAX
doesn't have snprintf().

I've released MMK V5.1-2 for VAX, Alpha, I64, and X86_64.

You can download MMK V5.1-2 from the Process Software freeware archive:

https://vms.process.com/scripts/fileserv/fileserv.com?MMK

I've released a few other packages for X86_64. You can see them all from
a link on this page:

https://www.process.com/resources/openvms/

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 9/6/23 7:40 AM, Hunter Goatley wrote:
> On 7/14/2023 2:06 PM, Hunter Goatley wrote:
>>
>> It looks like I did everything you did except for the bit about
>> defining MMSVAX, MMSALPHA, etc. That's a nice change, and one that I
>> had overlooked (those are also symbols I didn't know about and have
>> never used ;-) ).
>
> That bit me yesterday, when I found that MMSIA64 was defined on X86. I
> took Craig's approach to fix it, though I rewrote the code because VAX
> doesn't have snprintf().
>
> I've released MMK V5.1-2 for VAX, Alpha, I64, and X86_64.
>
> You can download MMK V5.1-2 from the Process Software freeware archive:
>
> https://vms.process.com/scripts/fileserv/fileserv.com?MMK

Thanks, Hunter. I've been meaning to update my pull request with your
changes but I don't see the source code in your kits -- are they buried
inside one of those savesets?

On 9/6/2023 10:18 AM, Craig A. Berry wrote:
>
> Thanks, Hunter.  I've been meaning to update my pull request with your
> changes but I don't see the source code in your kits -- are they buried
> inside one of those savesets?
>
Yes, MMK051.F contains MMK051_SOURCE.ZIP, which has all of the sources.

Thanks. I've had no time to look at how to use GitHub, and I've heard
nothing more from Tim since I told him about my first release.

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 9/6/2023 8:40 AM, Hunter Goatley wrote:
>
> I've released MMK V5.1-2 for VAX, Alpha, I64, and X86_64.
>
Now it's MMK V5.1-3. After releasing V5.1-2, I discovered that MMK on
VAX did not support command lines longer than 255 characters. There was
code to turn those into continuation lines, but a miscalculation in the
line lengths prevented that from working. I have corrected that code, so
now long DCL lines are supported on VAX, too.

I've released MMK V5.1-3 for VAX, Alpha, I64, and X86_64.

You can download MMK V5.1-3 from the Process Software freeware archive:

https://vms.process.com/scripts/fileserv/fileserv.com?MMK

I've released a few other packages for X86_64. You can see them all from
a link on this page:

https://www.process.com/resources/openvms/

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 9/6/23 2:19 PM, Hunter Goatley wrote:
> On 9/6/2023 8:40 AM, Hunter Goatley wrote:
>>
>> I've released MMK V5.1-2 for VAX, Alpha, I64, and X86_64.
>>
> Now it's MMK V5.1-3. After releasing V5.1-2, I discovered that MMK on
> VAX did not support command lines longer than 255 characters. There was
> code to turn those into continuation lines, but a miscalculation in the
> line lengths prevented that from working. I have corrected that code, so
> now long DCL lines are supported on VAX, too.

I was about to tell you that in 5.1-2 you'd regressed something you
fixed 7 years ago:

https://github.com/endlesssoftware/mmk/commit/d669df76b69bd273b22f5ea159d4c92d1df7e229

I'm working on merging 5.1.3 with my previous changes and the code on
GitHub, and I wondered about a couple things. There are multiple places
where you do:

#if defined(__ALPHA) || defined(__ia64__) || defined(__x86_64__)

instead of:

#ifndef __VAX

Is there any reason not to prefer the simpler and more future-proof
version? I don't have docs going back far enough to know if VAXC defines
__VAX or only VAX, but I don't see a reason to target that compiler when
it was superseded by DEC C for OpenVMS VAX so long ago.

In mmk_default_rules_x86_64.mms, some of the comments still said IA64
rather than x86. I'm keeping the ones that say x86.

On 9/6/2023 5:19 PM, Craig A. Berry wrote:
>
> I was about to tell you that in 5.1-2 you'd regressed something you
> fixed 7 years ago:
>
I don't know how that got lost. That change was not in the files I used
from 5.1-1, which was, I think, the last kit I got from Tim. It's on
GitHub, but he must have checked that in after releasing V5.1-1? It
doesn't matter now. I've fixed it again. 😎

> #if defined(__ALPHA) || defined(__ia64__) || defined(__x86_64__)
>
> instead of:
>
> #ifndef __VAX
>
> Is there any reason not to prefer the simpler and more future-proof
> version?

No, there's not, and I had actually intended to go back and change
those, but forgot to do so. When I first did it, I just used my REPLACE
TPU script to change them all in batch because it was the faster/lazier
thing to do. I meant to go back and replace them, but never did.

I don't have docs going back far enough to know if VAXC defines
> __VAX or only VAX, but I don't see a reason to target that compiler when
> it was superseded by DEC C for OpenVMS VAX so long ago.

Agreed.

> In mmk_default_rules_x86_64.mms, some of the comments still said IA64
> rather than x86. I'm keeping the ones that say x86.

Yes, I forgot to update the IA64 reference at the top.

Thanks!

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/

On 9/8/23 3:07 PM, Hunter Goatley wrote:
> On 9/6/2023 5:19 PM, Craig A. Berry wrote:

> >    #if defined(__ALPHA) || defined(__ia64__) || defined(__x86_64__)
> >
> > instead of:
> >
> >    #ifndef __VAX
> >
> > Is there any reason not to prefer the simpler and more future-proof
> > version?
>
> No, there's not, and I had actually intended to go back and change
> those, but forgot to do so. When I first did it, I just used my REPLACE
> TPU script to change them all in batch because it was the faster/lazier
> thing to do. I meant to go back and replace them, but never did.

Sounds good. That was the only remaining difference I had from your
changes so I kept the #ifndef __VAX and revised the pull request:

https://github.com/endlesssoftware/mmk/pull/101

Tim now just needs to click one button and the sources on GitHub will be
up-to-date.

On 9/8/2023 6:17 PM, Craig A. Berry wrote:
>
> Sounds good. That was the only remaining difference I had from your
> changes so I kept the #ifndef __VAX and revised the pull request:
>
> https://github.com/endlesssoftware/mmk/pull/101
>
> Tim now just needs to click one button and the sources on GitHub will be
> up-to-date.

Cool. Thanks. I'll try emailing Tim again....

--
Hunter
------
Hunter Goatley, Process Software, http://www.process.com/
goathunter@goatley.com http://hunter.goatley.com/