On Sat, 13 Apr 2024 at 09:12, Bob Bolch <Bob@...> wrote:
Hi Tony,
I? wrote entirely new code in the APL supervisor to support 3270s. I replaced the old supervisor code to run 1 APL?user in 1 CMS Virtual Machine.
I'm very impressed - that can't have been trivial. It wasn't clear to me if you'd repurposed the "standalone" supervisor to run in a VM (and that's why I asked what gets IPL'd in the VM). But your approach sounds like the best one to get a single-user APL under CMS.
?
All that new code is included in the download from this site,along with build instructions. I complied with the license agreement so you have to download the old IBM code from the Computer museum, and then build a MODULE using the provided instructions.
Ya, that makes sense. IIRC that's what Juergen's APL for MVT (or MVT for APL...) does too.
Well one of these centuries I hope to get the goo out of my old 2741 and get it talking to Hercules and hopefully APL. I do have an APL typeball.
Tony H.
On Fri, Apr 12, 2024 at 8:38?PM Tony Harminc <tharminc@...> wrote:
On Fri, 12 Apr 2024 at 18:07, Bob Bolch <Bob@...> wrote:
The code I started with is available here:
That code ran in a special environment in OS/360.
The runtime environment from OS/360 was not included.
You have to agree not to redistribute the source, to get a copy.
I've seen the code (both 50 years ago on fiche and more recently), but I didn't agree to? anything.
I built a runtime environment on VM/370 to use that code. I made it work on a VM370 3270 with the APL support available for that device.?
So you wrote entirely new code in the APL supervisor to support 3270s? Wow - I missed that somehow. And also to allow the supervisor to run in a VM? Or did you write en entirely new Supervisor to host the Interpreter??
What gets IPL'd in the VM? APL\360 came with three options: OS/360, DOS. and "standalone". But the standalone version was just the DOS one except that after initialization it overlaid the DOS supervisor with I/O buffers.
That you were able to do either says a lot about the original APL implementors and their rigid distinction between the APL Interpreter and Supervisor at a time when such modularity and layering was little known.
?
I don't remember any linemode 2741 support being available in VM/370. If you know where it is, please show me.
Mark has beaten me to it (DMKCNS), but other than the channel-attached 1052/3215 console, line-mode on 2741/1050/etc via async 270x-style lines was the only kind of VM/370 terminal device to start with. Supporting such devices in a VM might require Diagnose help with the dynamic channel programs used by APL. I remember writing that to allow MVT TCAM to work in a VM.
The Adesse VM course manual, although aimed at VM/SP, still has a great deal of useful info on how VM/370 works.
I took that course in person in 1985, and much later tried to track down the copyright owner with no success. I very much doubt anyone will be going after bitsavers (or me!) over it.