15 Jul
2021
15 Jul
'21
7:01 p.m.
Nelson H. F. Beebe:
P.S. Jay was the first to get Steve Johnson's Portable C Compiler,
pcc, to run on the 36-bit PDP-10, and once we had pcc, we began the
move from writing utilities in Pascal and PDP-10 assembly language to
doing them in C.
======
How did that C implementation handle ASCII text on the DEC-10?
Were it a from-scratch UNIX port it might make sense to store
four eight- or nine-bit bytes to a word, but if (as I sense it
was) it was C running on TOPS-10 or TOPS-20, it would have had
to work comfortably with DEC's convention of five 7-bit characters
(plus a spare bit used by some programs as a flag).
Norman Wilson
Toronto ON
15 Jul
15 Jul
7:27 p.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
The 'second' C compiler was a PDP-10 and Honeywell (36-bit) target Alan
Synder did for his MIT Thesis.
It was originally targeted to ITS for the PDP-10, but it ran on Tops-20
also.
My >>memory<< is he used a 7-bit Character, ala SAIL, with 5 chars stored
in a word with a bit leftover.
You can check it out: https://github.com/PDP-10/Snyder-C-compiler
I believe that C compiler Nelson is talking about I believe is actually
Synder's that Jay either ported from ITS or WAITS.
We had some form of the Synder compiler on the PDP-10's at CMU in the late
1970s.
It was either Mike Accetta or Fil Aleva that wrote a program to read PDP-10
backup tapes, that I updated to deal with TOPS-20/TENEX 'dumper' format
which was similar/only different.
ᐧ
On Thu, Jul 15, 2021 at 3:03 PM Norman Wilson <norman(a)oclsc.org> wrote:
Nelson H. F. Beebe:
P.S. Jay was the first to get Steve Johnson's Portable C Compiler,
pcc, to run on the 36-bit PDP-10, and once we had pcc, we began the
move from writing utilities in Pascal and PDP-10 assembly language to
doing them in C.
======
How did that C implementation handle ASCII text on the DEC-10?
Were it a from-scratch UNIX port it might make sense to store
four eight- or nine-bit bytes to a word, but if (as I sense it
was) it was C running on TOPS-10 or TOPS-20, it would have had
to work comfortably with DEC's convention of five 7-bit characters
(plus a spare bit used by some programs as a flag).
Norman Wilson
Toronto ON
7:28 p.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
g/Synder/s//Snyder/ -- sigh....
ᐧ
On Thu, Jul 15, 2021 at 3:27 PM Clem Cole <clemc(a)ccc.com> wrote:
The 'second' C compiler was a PDP-10 and
Honeywell (36-bit) target Alan
Synder did for his MIT Thesis.
It was originally targeted to ITS for the PDP-10, but it ran on Tops-20
also.
My >>memory<< is he used a 7-bit Character, ala SAIL, with 5 chars stored
in a word with a bit leftover.
You can check it out: https://github.com/PDP-10/Snyder-C-compiler
I believe that C compiler Nelson is talking about I believe is actually
Synder's that Jay either ported from ITS or WAITS.
We had some form of the Synder compiler on the PDP-10's at CMU in the late
1970s.
It was either Mike Accetta or Fil Aleva that wrote a program to read
PDP-10 backup tapes, that I updated to deal with TOPS-20/TENEX 'dumper'
format which was similar/only different.
ᐧ
On Thu, Jul 15, 2021 at 3:03 PM Norman Wilson <norman(a)oclsc.org> wrote:
Nelson H. F. Beebe:
P.S. Jay was the first to get Steve Johnson's Portable C Compiler,
pcc, to run on the 36-bit PDP-10, and once we had pcc, we began the
move from writing utilities in Pascal and PDP-10 assembly language to
doing them in C.
======
How did that C implementation handle ASCII text on the DEC-10?
Were it a from-scratch UNIX port it might make sense to store
four eight- or nine-bit bytes to a word, but if (as I sense it
was) it was C running on TOPS-10 or TOPS-20, it would have had
to work comfortably with DEC's convention of five 7-bit characters
(plus a spare bit used by some programs as a flag).
Norman Wilson
Toronto ON
7:34 p.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
The C compiler we had at NMT that Greg Titus wrote/rewrote allowed one to
pick a number
of different choices for character size (5, 6, 7 or 8). It defaulted to 7
or 8. I recall that the
defaults produced OK results for student work, but that was a bit slow for
pushing the
envelope without some very careful choices. But it was good enough for me
to write
my OS group project running under 'ZAYEF' a DecSystem-20 emulator running
on the
DecSystem 20 under TOPS-20... My first exposure to virtual machines... It
was a total
trip to have 18 bit pointers and weird interrupt semantics....
I really rather working on the VAX 11/750 in 'C' and later on the Sun3/50s
more, though.
In part because the debugger was better (or at least more approachable by my
poor undergraduate mind).
Warner
On Thu, Jul 15, 2021 at 1:28 PM Clem Cole <clemc(a)ccc.com> wrote:
The 'second' C compiler was a PDP-10 and
Honeywell (36-bit) target Alan
Synder did for his MIT Thesis.
It was originally targeted to ITS for the PDP-10, but it ran on Tops-20
also.
My >>memory<< is he used a 7-bit Character, ala SAIL, with 5 chars stored
in a word with a bit leftover.
You can check it out: https://github.com/PDP-10/Snyder-C-compiler
I believe that C compiler Nelson is talking about I believe is actually
Synder's that Jay either ported from ITS or WAITS.
We had some form of the Synder compiler on the PDP-10's at CMU in the late
1970s.
It was either Mike Accetta or Fil Aleva that wrote a program to read
PDP-10 backup tapes, that I updated to deal with TOPS-20/TENEX 'dumper'
format which was similar/only different.
ᐧ
On Thu, Jul 15, 2021 at 3:03 PM Norman Wilson <norman(a)oclsc.org> wrote:
Nelson H. F. Beebe:
P.S. Jay was the first to get Steve Johnson's Portable C Compiler,
pcc, to run on the 36-bit PDP-10, and once we had pcc, we began the
move from writing utilities in Pascal and PDP-10 assembly language to
doing them in C.
======
How did that C implementation handle ASCII text on the DEC-10?
Were it a from-scratch UNIX port it might make sense to store
four eight- or nine-bit bytes to a word, but if (as I sense it
was) it was C running on TOPS-10 or TOPS-20, it would have had
to work comfortably with DEC's convention of five 7-bit characters
(plus a spare bit used by some programs as a flag).
Norman Wilson
Toronto ON
16 Jul
16 Jul
7:38 a.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
Warner Losh <imp(a)bsdimp.com> wrote:
... But it was good enough for me to write my OS group
project running
under 'ZAYEF' a DecSystem-20 emulator running ...
Fascinating. That's the Hebrew word used for "forgery" or
"fakery".
Appropriate for an emulator. Whoever named it both knew Hebrew and
had a sense of humor. :-)
Arnold
8:05 a.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
Clem Cole wrote:
The 'second' C compiler was a PDP-10 and
Honeywell (36-bit) target
Alan Synder did for his MIT Thesis. It was originally targeted to ITS
for the PDP-10, but it ran on Tops-20 also. My >>memory<< is he used
a 7-bit Character, ala SAIL, with 5 chars stored in a word with a bit
leftover.
On ITS it only ever stored characters as full 36-bit words! So sizeof
char == 1 == sizeof int. This is allowed per the C standard. (Maybe it
was updated somewhere else, I dunno.)
KCC does support 6/7/8/9 bits per character. I think 9 is the default,
or else things like memcpy doesn't work.
I believe that C compiler Nelson is talking about I
believe is
actually Synder's that Jay either ported from ITS or WAITS.
I think it's a different compiler based on pcc. But I also think code
was moved between various PDP-10 C compilers and libraries, so it's
sometimes hard to tell one from another.
There was also "Sargasso C", but I don't know much about that one.
Maybe its claim to fame is as the original implementation language for
the VT100 test program vttest still in use today.
There was even an attempt to port GCC, and maybe it's still in use today
somewhere around the Seattle area.
2:19 p.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
On Fri, Jul 16, 2021 at 4:05 AM Lars Brinkhoff <lars(a)nocrew.org> wrote:
Clem Cole wrote:
Ah - that makes sense. I never programmed the Honeywell in anything but
Dartmouth BASIC (mostly) and any early FORTRAN (very little) and the whole
idea of storage size was somewhat oblivious to me at the point as I was a
youngster when I did that. Any idea did the Honeywell treat chars as
36-bit entities also? Steve, maybe you remember?
Also, please remember that the standard does not yet exist for a good 10
years ;-) At this point, the 'standard' was the Ritchie Compiler for the
PDP-11.
At the time, we to run wanted the program on all of the UNIX/v6 systems and
CMU's version of TOPS-10 and later TOPS-20 as an interchange format. Thus,
I have memories of having to use the "c =& 0177" idiom in the
backup/dumper program in a number of places [remember tar does not yet
exist, and tp/stp was a binary program]. Beyond that, I don't remember
much about the running C on the 10s. I think we started trying to move
Harvard's stp to TOPS-10, but ran into an issue [maybe the directory size]
and stopped. Since backup (dumper) was heavily used, we were trying to get
IUS and SUS to be able to be backed up and handled the same way the
operators did the backup for the 10's. In my own case, I had learned SAIL
(and BLISS) on the 10s before C on the PDP-11, plus this was an early C
program for me, maybe my second or third non-trivial one after I worked
with Ted on fsck, so coming from the PDP-10/SAIL/BLISS *et al *world, 7-bit
chars certainly seemed normal. I also remember having an early 'ah-ha'
moment, when the difference between a 7-bit and 8-bit char started to
become important.
Clem
ᐧ
The 'second' C compiler was a PDP-10
and Honeywell (36-bit) target
Alan Synder did for his MIT Thesis. It was originally targeted to ITS
for the PDP-10, but it ran on Tops-20 also. My >>memory<< is he used
a 7-bit Character, ala SAIL, with 5 chars stored in a word with a bit
leftover.
On ITS it only ever stored characters as full 36-bit words! So sizeof
char == 1 == sizeof int. This is allowed per the C standard. (Maybe it
was updated somewhere else, I dunno.)
4:09 p.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
On Fri, Jul 16, 2021, 1:38 AM <arnold(a)skeeve.com> wrote:
Warner Losh <imp(a)bsdimp.com> wrote:
It was billed as not really a DECsystem 20, but a really good fake. Clearly
a nod to that meaning.
Warner
Arnold
... But it was good enough for me to write my OS
group project running
under 'ZAYEF' a DecSystem-20 emulator running ...
Fascinating. That's the Hebrew word used for "forgery" or
"fakery".
Appropriate for an emulator. Whoever named it both knew Hebrew and
had a sense of humor. :-)
17 Jul
17 Jul
12:34 a.m.
New subject: head/sed/tail (was The Unix shell: a 50-year view)
On Fri, Jul 16, 2021 at 7:20 AM Clem Cole <clemc(a)ccc.com> wrote:
On Fri, Jul 16, 2021 at 4:05 AM Lars Brinkhoff <lars(a)nocrew.org> wrote:
The Honeywell 6000 machines ran GCOS; the system standard was six six-bit
characters per word.
The Honeywell 6100 machines ran Multics; the system standard was four
nine-bit characters per word.
For Multics C, sizeof (*) != sizeof (int) and NULL != 0, so a lot of
"portable" C code wasn't.
-- Charles
Clem Cole wrote:
Ah - that makes sense. I never programmed the Honeywell in anything but
Dartmouth BASIC (mostly) and any early FORTRAN (very little) and the whole
idea of storage size was somewhat oblivious to me at the point as I was a
youngster when I did that. Any idea did the Honeywell treat chars as
36-bit entities also? Steve, maybe you remember?
The 'second' C compiler was a PDP-10
and Honeywell (36-bit) target
Alan Synder did for his MIT Thesis. It was originally targeted to ITS
for the PDP-10, but it ran on Tops-20 also. My >>memory<< is he used
a 7-bit Character, ala SAIL, with 5 chars stored in a word with a bit
leftover.
On ITS it only ever stored characters as full 36-bit words! So sizeof
char == 1 == sizeof int. This is allowed per the C standard. (Maybe it
was updated somewhere else, I dunno.)
1015
days inactive
1017
days old
8 comments
6 participants
participants (6)
-
arnold＠skeeve.com -
Charles Anthony -
Clem Cole -
Lars Brinkhoff -
Norman Wilson -
Warner Losh