info
Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Compiling Multics PL/I Compiler using the Prime PL/I Compiler
152 views
Skip to first unread message
Daiyu Hurst
Jul 13, 2021, 1:12:25 PM7/13/21
to
Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
Robin Vowels
Jul 13, 2021, 1:30:07 PM7/13/21
to
On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu...@gmail.com wrote:
> Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
>
> Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
.
> Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
>
> Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
Are you saying that the Pr1me PL/I compiler is available somewhere?
Daiyu Hurst
Aug 30, 2021, 12:01:13 AM8/30/21
to
On Tuesday, July 13, 2021 at 1:30:07 PM UTC-4, Robin Vowels wrote:
> On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu wrote:
> > Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
> >
> > Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
> .
> Are you saying that the Pr1me PL/I compiler is available somewhere?
Yes, under Primos, Revs 19 through 24. I usually run it on Rev23.
> > Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
> >
> > Most of the Multics PL/I compiler will compile on Primos using the Prime PL/I (full language) compiler. I'm not terribly surprised by this, since I think Bob Freiburghouse was behind both of them.
> .
> Are you saying that the Pr1me PL/I compiler is available somewhere?
Robin Vowels
Aug 30, 2021, 12:55:14 PM8/30/21
to
Thanks.
.
You have a Pr1me system, or have access to one?
Or you are running an emulator for PRIMOS?
Daiyu Hurst
Sep 3, 2021, 11:37:47 PM9/3/21
to
On Monday, August 30, 2021 at 12:55:14 PM UTC-4, Robin Vowels wrote:
I had a 2455, but I'm pretty sure it ended up in a landfill after the bank took my home from me.
But my drivers got imaged first, and yes, there is an emulator for the 50-Series Prime machines.
You can get it here:
https://github.com/prirun/p50em
A package of demo systems, from Rev18 through Rev24, is also available, link(s) are on that page.
The Rev20 and Rev23.4.Y2K were from my system. One of the Rev19 systems was from Los Angeles Valley College, courtesy of a friend of mine. Another friend provided the Rev 22 system, and I had a partial source tree for Rev18 that seems to be working now, but I haven't been able to get that to build, though the demo binary works.
So anyway, I still don't know which way would be the best way to go: build a 50-Series code generator backend for the Multics PL/I compiler and run it on Multics as a cross-compiler, or port it directly to compile and build on the Prime using the Prime compiler.
A third goal would be to get it to build under IBM PL/I for Windows; didn't I read in old posts, that you have this, too?
5724-B67 IBM(R) PL/I for Windows 6.0 (Built:20060113) 2014.02.04 01:14:21 Page 1
Options Specified
Environment: EXIT
Command: e:pl1_error_print.pl1 (margins(1 200) EXIT DFT(RETURNS(BYADDR))
source stmt list nest attributes(full) xref aggregate limits(fixedbin(31,63)
fixeddec(15,31))
5724-B67 IBM(R) PL/I for Windows * *********************** 2014.02.04 01:14:21 Page 2
Compiler Source
Line.File Stmt LV NT
1.0 /* ***********************************************************
2.0 * *
3.0 * Copyright, (C) Honeywell Information Systems Inc., 1982 *
4.0 * *
5.0 * Copyright (c) 1972 by Massachusetts Institute of *
6.0 * Technology and Honeywell Information Systems, Inc. *
7.0 * *
8.0 *********************************************************** */
9.0
10.0
11.0 /* Print PL/I error messages.
12.0
13.0 Initial Version: 14 June, 1968 by BLW
14.0 Modified: 30 July 1969 by JDM for pl1
15.0 Modified: 10 January 1970 by CTC for new error message format
16.0 Rewritten: 17 August 1970 by P. Green for Version II
17.0 Modified: 16 January 1971 by BLW to print text of bad statement
18.0 Modified: 13 February 1977 by RAB to fix 1577
19.0 Modified: 770712 by PG to add pl1_severity_
20.0 Modified 771020 by PG to fix 1677
21.0 Modified 780807 by PG to permit char(256) tokens to be substituted for '$'.
22.0 Modified 23 October 1980 by M. N. Davidoff to fix 1977, an unitialized variable causes tokens to be quoted when
23.0 they shouldn't be.
24.0 */
25.0 /* format: style3 */
26.0 1 pl1_error_print:
27.0 procedure;
But my drivers got imaged first, and yes, there is an emulator for the 50-Series Prime machines.
You can get it here:
https://github.com/prirun/p50em
A package of demo systems, from Rev18 through Rev24, is also available, link(s) are on that page.
The Rev20 and Rev23.4.Y2K were from my system. One of the Rev19 systems was from Los Angeles Valley College, courtesy of a friend of mine. Another friend provided the Rev 22 system, and I had a partial source tree for Rev18 that seems to be working now, but I haven't been able to get that to build, though the demo binary works.
So anyway, I still don't know which way would be the best way to go: build a 50-Series code generator backend for the Multics PL/I compiler and run it on Multics as a cross-compiler, or port it directly to compile and build on the Prime using the Prime compiler.
A third goal would be to get it to build under IBM PL/I for Windows; didn't I read in old posts, that you have this, too?
5724-B67 IBM(R) PL/I for Windows 6.0 (Built:20060113) 2014.02.04 01:14:21 Page 1
Options Specified
Environment: EXIT
Command: e:pl1_error_print.pl1 (margins(1 200) EXIT DFT(RETURNS(BYADDR))
source stmt list nest attributes(full) xref aggregate limits(fixedbin(31,63)
fixeddec(15,31))
5724-B67 IBM(R) PL/I for Windows * *********************** 2014.02.04 01:14:21 Page 2
Compiler Source
Line.File Stmt LV NT
1.0 /* ***********************************************************
2.0 * *
3.0 * Copyright, (C) Honeywell Information Systems Inc., 1982 *
4.0 * *
5.0 * Copyright (c) 1972 by Massachusetts Institute of *
6.0 * Technology and Honeywell Information Systems, Inc. *
7.0 * *
8.0 *********************************************************** */
9.0
10.0
11.0 /* Print PL/I error messages.
12.0
13.0 Initial Version: 14 June, 1968 by BLW
14.0 Modified: 30 July 1969 by JDM for pl1
15.0 Modified: 10 January 1970 by CTC for new error message format
16.0 Rewritten: 17 August 1970 by P. Green for Version II
17.0 Modified: 16 January 1971 by BLW to print text of bad statement
18.0 Modified: 13 February 1977 by RAB to fix 1577
19.0 Modified: 770712 by PG to add pl1_severity_
20.0 Modified 771020 by PG to fix 1677
21.0 Modified 780807 by PG to permit char(256) tokens to be substituted for '$'.
22.0 Modified 23 October 1980 by M. N. Davidoff to fix 1977, an unitialized variable causes tokens to be quoted when
23.0 they shouldn't be.
24.0 */
25.0 /* format: style3 */
26.0 1 pl1_error_print:
27.0 procedure;
Daiyu Hurst
Sep 4, 2021, 12:19:22 AM9/4/21
to
On Friday, September 3, 2021 at 11:37:47 PM UTC-4, Daiyu Hurst wrote:
> On Monday, August 30, 2021 at 12:55:14 PM UTC-4, Robin Vowels wrote:
> > On Monday, August 30, 2021 at 2:01:13 PM UTC+10, daiyu wrote:
> > > On Tuesday, July 13, 2021 at 1:30:07 PM UTC-4, Robin Vowels wrote:
> > > > On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu wrote:
> > > > > Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
It occurs to me, I should just cast bread on the waters:
> On Monday, August 30, 2021 at 12:55:14 PM UTC-4, Robin Vowels wrote:
> > On Monday, August 30, 2021 at 2:01:13 PM UTC+10, daiyu wrote:
> > > On Tuesday, July 13, 2021 at 1:30:07 PM UTC-4, Robin Vowels wrote:
> > > > On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu wrote:
> > > > > Just tossing this out there. If I did already, let me know; I've had a stroke, and there are definitely some parity errors.
https://www.dropbox.com/s/a7ht8nvbbusd57i/mxPL1.zip?dl=0
If you want this as a Primos MAGSAV tape image, let me know, and I'll cook that up for you. Don't waste a lot of time uploading each file to Primos.
-Dai
Peter Flass
Sep 4, 2021, 7:48:30 PM9/4/21
to
compile Multics stuff with Iron Spring PL/I, but I can’t imaging getting
anywhere without massive source changes.
--
Pete
Daiyu Hurst
Sep 4, 2021, 7:55:42 PM9/4/21
to
On Saturday, September 4, 2021 at 7:48:30 PM UTC-4, wrote:
-Dai
> Daiyu Hurst wrote:
> > On Monday, August 30, 2021 at 12:55:14 PM UTC-4, Robin Vowels wrote:
> >> On Monday, August 30, 2021 at 2:01:13 PM UTC+10, daiyu wrote:
> >>> On Tuesday, July 13, 2021 at 1:30:07 PM UTC-4, Robin Vowels wrote:
> >>>> On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu wrote:
> >>>>> Just tossing this out there. If I did already, let me know; I've had
> >>>>> a stroke, and there are definitely some parity errors.
> >>>>>
> >>>>> Most of the Multics PL/I compiler will compile on Primos using the
> >>>>> Prime PL/I (full language) compiler. I'm not terribly surprised by
> >>>>> this, since I think Bob Freiburghouse was behind both of them.
> >>>> .
> >>>> Are you saying that the Pr1me PL/I compiler is available somewhere?
..snip..
> > On Monday, August 30, 2021 at 12:55:14 PM UTC-4, Robin Vowels wrote:
> >> On Monday, August 30, 2021 at 2:01:13 PM UTC+10, daiyu wrote:
> >>> On Tuesday, July 13, 2021 at 1:30:07 PM UTC-4, Robin Vowels wrote:
> >>>> On Wednesday, July 14, 2021 at 3:12:25 AM UTC+10, daiyu wrote:
> >>>>> Just tossing this out there. If I did already, let me know; I've had
> >>>>> a stroke, and there are definitely some parity errors.
> >>>>>
> >>>>> Most of the Multics PL/I compiler will compile on Primos using the
> >>>>> Prime PL/I (full language) compiler. I'm not terribly surprised by
> >>>>> this, since I think Bob Freiburghouse was behind both of them.
> >>>> .
> >>>> Are you saying that the Pr1me PL/I compiler is available somewhere?
> A major problem is the 36-bit word length. I’m playing around trying to
> compile Multics stuff with Iron Spring PL/I, but I can’t imaging getting
> anywhere without massive source changes.
It's NEVER as easy as simply re-compiling. But even trying to get it to compile on the IBM compiler as well as I have on the Prime compiler is like a night and day difference. What a picky compiler! But the IBM compiler does produce binaries that run on Windows 10. So it's compelling to attempt.
> compile Multics stuff with Iron Spring PL/I, but I can’t imaging getting
> anywhere without massive source changes.
-Dai
Markus Loew
Oct 2, 2021, 7:52:36 AM10/2/21
to
Two Questions:
1st Where can I get the PL/I source text, you are speaking about, you are trying to compile ?
2nd How can I make work " ./em.linux-i386 "
archives@archives-X71Sr:~/Desktop/collect/embinaries$ ll
total 1020
drwxr-xr-x 4 archives archives 4096 oct. 2 13:29 ./
drwxrwxr-x 18 archives archives 4096 sept. 19 06:39 ../
-rw-r--r-- 1 archives archives 742 mai 6 2020 00README
-rw-rw-r-- 1 archives archives 0 oct. 2 13:31 console.log
drwxr-xr-x 2 archives archives 4096 mai 6 2020 embinaries/
-rwxr-xr-x 1 archives archives 224248 mai 5 2020 em.freebsd-amd64*
-rwxr-xr-x 1 archives archives 195088 mai 5 2020 em.linux-amd64*
-rwxr-xr-x 1 archives archives 189112 mai 6 2020 em.linux-armhf*
-rwxr-xr-x 1 archives archives 196324 mai 5 2020 em.linux-i386*
-rwxr-xr-x 1 archives archives 205736 mai 5 2020 em.solaris-amd64*
-rw-r--r-- 1 archives archives 297 oct. 2 13:31 error.log
-rw-r--r-- 1 archives archives 0 sept. 2 05:23 .lock
drwxr-xr-x 2 archives archives 4096 sept. 2 05:36 m2214repack/
archives@archives-X71Sr:~/Desktop/collect/embinaries$ ./em.linux-i386
[Prime Emulator ver ebf5511 May 5 2020]
[Copyright (C) 2005-2019 Jim Wilcoxson pri...@gmail.com]
Can't find disk boot device file disk26u0.*, or multiple files match this device.
To boot from tape (mt0), use -boot 10005
Fatal error
instruction #0 at 0/0 ^000000^
A='0/0 B='0/0 L='0/0 X='0/0 K=0 [16SF000/0 ]
owner=0 , modals=0 [I0N]
Check error.log for more information
archives@archives-X71Sr:~/Desktop/collect/embinaries$
1st Where can I get the PL/I source text, you are speaking about, you are trying to compile ?
2nd How can I make work " ./em.linux-i386 "
archives@archives-X71Sr:~/Desktop/collect/embinaries$ ll
total 1020
drwxr-xr-x 4 archives archives 4096 oct. 2 13:29 ./
drwxrwxr-x 18 archives archives 4096 sept. 19 06:39 ../
-rw-r--r-- 1 archives archives 742 mai 6 2020 00README
-rw-rw-r-- 1 archives archives 0 oct. 2 13:31 console.log
drwxr-xr-x 2 archives archives 4096 mai 6 2020 embinaries/
-rwxr-xr-x 1 archives archives 224248 mai 5 2020 em.freebsd-amd64*
-rwxr-xr-x 1 archives archives 195088 mai 5 2020 em.linux-amd64*
-rwxr-xr-x 1 archives archives 189112 mai 6 2020 em.linux-armhf*
-rwxr-xr-x 1 archives archives 196324 mai 5 2020 em.linux-i386*
-rwxr-xr-x 1 archives archives 205736 mai 5 2020 em.solaris-amd64*
-rw-r--r-- 1 archives archives 297 oct. 2 13:31 error.log
-rw-r--r-- 1 archives archives 0 sept. 2 05:23 .lock
drwxr-xr-x 2 archives archives 4096 sept. 2 05:36 m2214repack/
archives@archives-X71Sr:~/Desktop/collect/embinaries$ ./em.linux-i386
[Prime Emulator ver ebf5511 May 5 2020]
[Copyright (C) 2005-2019 Jim Wilcoxson pri...@gmail.com]
Can't find disk boot device file disk26u0.*, or multiple files match this device.
To boot from tape (mt0), use -boot 10005
Fatal error
instruction #0 at 0/0 ^000000^
A='0/0 B='0/0 L='0/0 X='0/0 K=0 [16SF000/0 ]
owner=0 , modals=0 [I0N]
Check error.log for more information
archives@archives-X71Sr:~/Desktop/collect/embinaries$
Markus Loew
Oct 9, 2021, 6:17:01 AM10/9/21
to
0 new messages