[TUHS] Origin year of BSD csh?

[TUHS] Origin year of BSD csh?

[Aharon Robbins]

Hi.

Can anyone give a definitive date for when Bill Joy's csh first got out
of Berkeley?  I suspect it's in the 1976 - 1977 time frame, but I don't
know for sure.

Thanks!

Arnold

[TUHS] Origin year of BSD csh?

[Mary Ann Horton]

1BSD did not contain csh, and was officially released March 9 1978. It 
contained something called "a shell" which appears to be a predecessor 
to csh, but was still compiled as sh.  The README on 1BSD from Bill Joy 
states:

    Wed Oct 19, 1977

    This directory contains the source for a shell.
    It requires floating point to do the time command which is built-in
    so you will have to cc it -f on machines without floating point.
    It also requires a version 7 C compiler.

    Accurate documentation is in the file "sh.6" to be nroffed with
    /usr/man/man0/naa and a new "version 7" nroff.

    This shell requires the "htmp" data base also used by the editor "ex".
    If you do not set it up so that the "sethome" command is done by "login"
    then you should use the old "osethome" routine in ../s6 rather than
    "sethome"
    and reenable the execl of this sethome in the file "sh.c" (with the
    correct
    pathname).

2BSD did include csh and was first officially released May 1979. I'm 
sure there were informal advance copies of csh sent out sooner. I recall 
csh already being on the UCB systems when I arrived in September of 1978.

I brought csh with me to Bell Labs in the summer of 1979.  The folks at 
Bell Labs recoiled in horror: they had just gone through a painful 
conversion from the Mashey shell to the Bourne shell, and would never 
consider another conversion.  csh was (mostly) upward compatible with 
the Mashey shell, unlike the Bourne shell.  (This was in the Bell Labs 
Computer Center, where I was a summer employee, not Research or the PWB 
group, which I'm sure felt the same way.)

     Mary Ann

On 06/26/2016 03:14 AM, Aharon Robbins wrote:
> Hi.
>
> Can anyone give a definitive date for when Bill Joy's csh first got out
> of Berkeley?  I suspect it's in the 1976 - 1977 time frame, but I don't
> know for sure.
>
> Thanks!
>
> Arnold

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20160626/fbcbca4e/attachment.html>

[TUHS] Origin year of BSD csh?

[Larry McVoy]

On Sun, Jun 26, 2016 at 09:30:39AM -0700, Mary Ann Horton wrote:
> I brought csh with me to Bell Labs in the summer of 1979.  The folks at Bell
> Labs recoiled in horror: they had just gone through a painful conversion
> from the Mashey shell to the Bourne shell

I used csh for a while before ksh became available.  It was an improvement
over the Bourne shell, IMO, but once ksh came out I went back to Bourne
shell syntax.  And now bash is pretty nice.

--lm

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1009 bytes --]

I detested the CSH syntax.   In order to beat back the CSH proponents at BRL, I added JOB control to the SV  (and later SVR2) Bourne Shell.   Then they beat on me for not having command like editing in (a la TCSH), so I added that.   This shell went out as /bin/sh in the Doug Gwyn SV-on-BSD release so every once and a while over the years I trip across a “Ron shell” usually people who were running Mach-derived things that ran my shell as /bin/sh. 


> On Jun 26, 2016, at 1:14 PM, Larry McVoy <lm at mcvoy.com> wrote:
> 
> On Sun, Jun 26, 2016 at 09:30:39AM -0700, Mary Ann Horton wrote:
>> I brought csh with me to Bell Labs in the summer of 1979.  The folks at Bell
>> Labs recoiled in horror: they had just gone through a painful conversion
>> from the Mashey shell to the Bourne shell
> 
> I used csh for a while before ksh became available.  It was an improvement
> over the Bourne shell, IMO, but once ksh came out I went back to Bourne
> shell syntax.  And now bash is pretty nice.
> 
> --lm

[TUHS] Origin year of BSD csh?

[John Cowan]

Ronald Natalie scripsit:

> I detested the CSH syntax.

Tom Christiansen's 1994 flame "csh Programming Considered Harmful"
can be found at <http://www.perl.com/doc/FMTEYEWTK/versus/csh.whynot>.
Current csh releases have undoubtedly fixed many of the complaints,
but as TC notes, many cannot be fixed.

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
                I am a member of a civilization. --David Brin

[TUHS] Origin year of BSD csh?

[Larry McVoy]

On Sun, Jun 26, 2016 at 04:43:26PM -0400, John Cowan wrote:
> Ronald Natalie scripsit:
> 
> > I detested the CSH syntax.
> 
> Tom Christiansen's 1994 flame "csh Programming Considered Harmful"
> can be found at <http://www.perl.com/doc/FMTEYEWTK/versus/csh.whynot>.
> Current csh releases have undoubtedly fixed many of the complaints,
> but as TC notes, many cannot be fixed.

Tom and I went to undergrad together, we were both in the UW-Madison CS
department.  Sad to say, I've lost touch with him and googling is not
finding him (other than wikipedia).  Anyone know what he is doing these
days?

--lm

[TUHS] Origin year of BSD csh?

[John Cowan]

Larry McVoy scripsit:

> Tom and I went to undergrad together, we were both in the UW-Madison CS
> department.  Sad to say, I've lost touch with him and googling is not
> finding him (other than wikipedia).  Anyone know what he is doing these
> days?

Googling with the date filter set to "last year" gives us a lot of other
Tom Christiansens, but also http://stackoverflow.com/users/471272/tchrist,
updated last August.

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
Objective consideration of contemporary phenomena compel the conclusion
that optimum or inadequate performance in the trend of competitive
activities exhibits no tendency to be commensurate with innate capacity,
but that a considerable element of the unpredictable must invariably be
taken into account. --Ecclesiastes 9:11, Orwell/Brown version

[TUHS] Origin year of BSD csh?

[Sven Mascheck]

On Sun, Jun 26, 2016 at 01:32:23PM -0500, Ronald Natalie wrote:
> I added JOB control to the SV  (and later SVR2) Bourne Shell.
> Then they beat on me for not having command like editing in (a la TCSH),
> so I added that.

How interesting, I will try to bother you (perhaps directly) about
in-depth informations :-)


I've always been intrigued by the fact that traditional Bourne shell and
Almquist shell haven't implemented history or command line editing.

I found two interesting references about this:

- Ash announcement, "A reimplementation of the System V shell":

  "I conclude by listing a few features that I have omitted intentionally.

   3.  History.  It seems to me that the csh history mechanism is
       mostly a response to the deficiencies of UNIX terminal I/O.
       Those of you running 4.2 BSD should try out atty (which I am
       posting to the net at the same time as ash) and see if you
       still want history."

- From an article from D. Korn, "ksh - An Extensible High Level Language":

  "Originally the idea of adding command line editing to ksh was
   rejected in the hope that line editing would move into the terminal
   driver. However, when it became clear that this was not likely to
   happen soon, both line editing modes were integrated into ksh and
   made optional so that they could be disabled on systems that provided
   editing as part of the terminal interface."

I believe it's a real pity that it hasn't been implemented in terminal
drivers in general.

Or do I overlook possible disadvantages?  What could be downsides,
apart from possibly inconsistent behaviour across systems?

Sven

[TUHS] Origin year of BSD csh?

[Steve Nickolas]

On Mon, 27 Jun 2016, Sven Mascheck wrote:

> On Sun, Jun 26, 2016 at 01:32:23PM -0500, Ronald Natalie wrote:
>> I added JOB control to the SV  (and later SVR2) Bourne Shell.
>> Then they beat on me for not having command like editing in (a la TCSH),
>> so I added that.
>
> How interesting, I will try to bother you (perhaps directly) about
> in-depth informations :-)
>
>
> I've always been intrigued by the fact that traditional Bourne shell and
> Almquist shell haven't implemented history or command line editing.

There are a couple variants of the Almquist shell (an old one called 
"cash", and one I think FreeBSD uses) which have readline-like history.

-uso.

[TUHS] Origin year of BSD csh?

[Joerg Schilling]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1233 bytes --]

Sven Mascheck <mascheck at in-ulm.de> wrote:

>   "Originally the idea of adding command line editing to ksh was
>    rejected in the hope that line editing would move into the terminal
>    driver. However, when it became clear that this was not likely to
>    happen soon, both line editing modes were integrated into ksh and
>    made optional so that they could be disabled on systems that provided
>    editing as part of the terminal interface."
>
> I believe it's a real pity that it hasn't been implemented in terminal
> drivers in general.
>
> Or do I overlook possible disadvantages?  What could be downsides,
> apart from possibly inconsistent behaviour across systems?

It was in the terminal driver from VMS ;-)

In Summer 1984, I noticed that this feature worked in a similar way as my test 
implementation from 1982 and then worked on an integrated implementation for 
"bsh" at H. Berthold AG. The person that helped me in 1984 was Peter Teuchert.

Jörg

-- 
 EMail:joerg at schily.net                  (home) Jörg Schilling D-13353 Berlin
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.org/private/ http://sourceforge.net/projects/schilytools/files/'

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 295 bytes --]


> 
> It was in the terminal driver from VMS ;-)

The T in TCSH was TENEX, which indeed had such editing as part of the COMND (command) JSYS (essentially an OS call).    This propagated forward into TOPS-20.    It isn’t really the terminal driver, but more of a command line processor call.

[TUHS] Origin year of BSD csh?

[John Cowan]

Sven Mascheck scripsit:

>    3.  History.  It seems to me that the csh history mechanism is
>        mostly a response to the deficiencies of UNIX terminal I/O.
>        Those of you running 4.2 BSD should try out atty (which I am
>        posting to the net at the same time as ash) and see if you
>        still want history."

I personally could not live without !!, ^foo^bar, !foo, etc.

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
Verbogeny is one of the pleasurettes of a creatific thinkerizer.
        --Peter da Silva

[TUHS] Origin year of BSD csh?

[Steve Nickolas]

On Mon, 27 Jun 2016, John Cowan wrote:

> Sven Mascheck scripsit:
>
>>    3.  History.  It seems to me that the csh history mechanism is
>>        mostly a response to the deficiencies of UNIX terminal I/O.
>>        Those of you running 4.2 BSD should try out atty (which I am
>>        posting to the net at the same time as ash) and see if you
>>        still want history."
>
> I personally could not live without !!, ^foo^bar, !foo, etc.

Same.

That's a big reason why I still use bash even though ksh is faster and 
lighter and otherwise works just as well.

-uso.

[TUHS] Origin year of BSD csh?

[Dave Horsfall]

On Mon, 27 Jun 2016, Sven Mascheck wrote:

> I've always been intrigued by the fact that traditional Bourne shell and 
> Almquist shell haven't implemented history or command line editing.

I hated CSH, and only used it when forced to; I used a utility called 
"screen" for my job-switching needs, and was really happy when KSH came 
along (then ZSH).

-- 
Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1062 bytes --]


> On Jun 27, 2016, at 6:27 AM, Sven Mascheck <mascheck at in-ulm.de> wrote:
> 
> On Sun, Jun 26, 2016 at 01:32:23PM -0500, Ronald Natalie wrote:
>> I added JOB control to the SV  (and later SVR2) Bourne Shell.
>> Then they beat on me for not having command like editing in (a la TCSH),
>> so I added that.
> 
> How interesting, I will try to bother you (perhaps directly) about
> in-depth informations :-)
> 

Sure, it was a long time ago, but I’ll tell you what I remember.   The one thing I do remember is that the SV /bin/sh was
written in these horrendous macros that sort of made it look like algol or something.   When the SVR2 shell came out,
someone (not Bourne obviously) had undone all those in favor of the native C++ if/else/while blocking.


> 
> I've always been intrigued by the fact that traditional Bourne shell and
> Almquist shell haven't implemented history or command line editing.

Command line editing might have been implemented in the driver as enhanced editing in “cooked” mode, but the history
is a bit more context specific.

[TUHS] Origin year of BSD csh?

[Clem Cole]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 438 bytes --]

On Mon, Jun 27, 2016 at 4:33 PM, Ronald Natalie <ron at ronnatalie.com> wrote:

> written in these horrendous macros that sort of made it look like algol or
> something.


​Famously called "Bournegol" -- Steve was a member of the Algol68
definition group IIRC.​
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20160627/a2b31909/attachment-0001.html>

[TUHS] Origin year of BSD csh?

[Steve Nickolas]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1426 bytes --]

On Mon, 27 Jun 2016, Ronald Natalie wrote:

>
>> On Jun 27, 2016, at 6:27 AM, Sven Mascheck <mascheck at in-ulm.de> wrote:
>>
>> On Sun, Jun 26, 2016 at 01:32:23PM -0500, Ronald Natalie wrote:
>>> I added JOB control to the SV  (and later SVR2) Bourne Shell.
>>> Then they beat on me for not having command like editing in (a la TCSH),
>>> so I added that.
>>
>> How interesting, I will try to bother you (perhaps directly) about
>> in-depth informations :-)
>>
>
> Sure, it was a long time ago, but I’ll tell you what I remember.  The 
> one thing I do remember is that the SV /bin/sh was written in these 
> horrendous macros that sort of made it look like algol or something.

That was inherited from the original V7 Bourne shell.  And "horrendous" 
doesn't even begin to describe that disaster of coding.

> When the SVR2 shell came out, someone (not Bourne obviously) had undone 
> all those in favor of the native C++ if/else/while blocking.
>>
>> I've always been intrigued by the fact that traditional Bourne shell and
>> Almquist shell haven't implemented history or command line editing.
>
> Command line editing might have been implemented in the driver as 
> enhanced editing in “cooked” mode, but the history is a bit more context 
> specific.

I kind-of like the MS-DOS 5 approach of having a separate tool that the 
shell can optionally link to that provides those capabilities.

*ducks and runs*

-uso.

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

> 
> I kind-of like the MS-DOS 5 approach of having a separate tool that the shell can optionally link to that provides those capabilities.
> 
> *ducks and runs*

Not unique to MSDOS.   As pointed out the COMND JSYS in Tenex/TOPS-20 provided such a feature. 

[TUHS] Origin year of BSD csh?

[John Cowan]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 856 bytes --]

Ronald Natalie scripsit:

> Sure, it was a long time ago, but I’ll tell you what I remember.
> The one thing I do remember is that the SV /bin/sh was written in
> these horrendous macros that sort of made it look like algol or
> something.

I've always wondered what would have happened if Algol 68 (brought back
from England by Bourne) had out-competed C at Bell Labs, and had become
the dominant programming language of Unix.  Probably the commercial
world would have standardized on Pascal, something that almost happened
(the x86 chip is optimized for Pascal in several ways).

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
"But I am the real Strider, fortunately," he said, looking down at them
with his face softened by a sudden smile.  "I am Aragorn son of Arathorn,
and if by life or death I can save you, I will."

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1367 bytes --]

Hard to believe the 8086 chip was “optimized” for anything.  The instruction set was designed for programming terminals.
The iapx32 was designed to run higher level languages (Ada) but despite how “nicely” it implemented this, it couldn’t counter the fact that it was molasses slow doing anything.   It was easier to build craftier compilers than trying to burn the smarts into silicon.

> On Jun 27, 2016, at 4:20 PM, John Cowan <cowan at mercury.ccil.org> wrote:
> 
> Ronald Natalie scripsit:
> 
>> Sure, it was a long time ago, but I’ll tell you what I remember.
>> The one thing I do remember is that the SV /bin/sh was written in
>> these horrendous macros that sort of made it look like algol or
>> something.
> 
> I've always wondered what would have happened if Algol 68 (brought back
> from England by Bourne) had out-competed C at Bell Labs, and had become
> the dominant programming language of Unix.  Probably the commercial
> world would have standardized on Pascal, something that almost happened
> (the x86 chip is optimized for Pascal in several ways).
> 
> -- 
> John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
> "But I am the real Strider, fortunately," he said, looking down at them
> with his face softened by a sudden smile.  "I am Aragorn son of Arathorn,
> and if by life or death I can save you, I will."

[TUHS] Origin year of BSD csh?

[John Cowan]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 611 bytes --]

Ronald Natalie scripsit:

> Hard to believe the 8086 chip was “optimized” for anything.
> The instruction set was designed for programming terminals.

Well, yes.  But the four separate address spaces work fine for Pascal,
where it is always statically known whether a pointer is to code,
global data, the stack (internal only), or the heap.  For C they were
nothing but a nuisance:  C can handle separate I & D space, but that's all.

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
What asininity could I have uttered that they applaud me thus?
        --Phocion, Greek orator

[TUHS] Origin year of BSD csh?

[Peter Jeremy]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 897 bytes --]

On 2016-Jun-27 16:28:14 -0500, Ronald Natalie <ron at ronnatalie.com> wrote:
>Hard to believe the 8086 chip was “optimized” for anything.  The instruction set was designed for programming terminals.

I think "designed" is being generous.  The closest to "designed" would have been for a calculation but then the 4004 grew warts and the warts grew warts.

>The iapx32 was designed to run higher level languages (Ada)

That was the iapx432.  No relationship to the x86.

And, whilst we're dealing with what-if's, what if the M68K had taken off,
rather than the 8086.  IBM had a M68K box in the same timeframe as the PC.

-- 
Peter Jeremy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 949 bytes
Desc: not available
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20160628/1aa071da/attachment.sig>

[TUHS] Origin year of BSD csh?

[arnold]

Thanks everyone for the answers, esp. to Mary Ann for the definitive dates.

My story is similar to most everyone else's. I was exposed to csh on
4.1 BSD but was so horrified by the syntax that I preferred to do without
job control and use the Bourne shell.

Later on when I was at Georgia Tech we got the BRL dist and I used Ron's
job control shell. I wrote a csh-style history mechanism for it and backported
that and the job control to the V7 sh and posted diffs to USENET so that people
without a SV license could benefit.

From there I went to ksh for many years, and thence to Bash. I abandoned
the csh-history-for-sh stuff as soon as I got ksh with vi editing mode
and have never looked back.  Circa 1990 I banged on the bash/readline code
to make its vi mode more like ksh's.

How well I remember Bournegol and how happy I was when I saw that SVR2 had
gotten rid of it.

With respect to history in the terminal, the Bell Labs guys did that by
making the terminal smarter, with the Blit.  I had one but the load it
put on our poor vax 11/780 was awesome.

Arnold

[TUHS] Origin year of BSD csh?

[Random832]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 493 bytes --]

On Mon, Jun 27, 2016, at 16:33, Ronald Natalie wrote:
> Command line editing might have been implemented in the driver as
> enhanced editing in “cooked” mode, but the history
> is a bit more context specific.

MS Windows does a kind of halfway decent job of this by having a
separate history per program, which I don't think requires any
information the terminal driver doesn't have access to (the process that
is trying to read from the terminal, and the binary running in that
process)

[TUHS] Origin year of BSD csh?

[Tony Finch]

Sven Mascheck <mascheck at in-ulm.de> wrote:
>
> I've always been intrigued by the fact that traditional Bourne shell and
> Almquist shell haven't implemented history or command line editing.

The Almquist shell had history support added for (I think) 4.4BSD
https://svnweb.freebsd.org/csrg/bin/sh/histedit.c?view=log

It's a required feature for POSIX sh, though I don't know how far back
that requirement goes...

Tony.
-- 
f.anthony.n.finch  <dot at dotat.at>  http://dotat.at/  -  I xn--zr8h punycode
Fair Isle, Faeroes, Southeast Iceland: Southwesterly backing easterly or
southeasterly, 4 or 5 increasing 6 at times. Moderate, occasionally rough for
a time. Rain or showers. Good, occasionally poor.

[TUHS] Origin year of BSD csh?

[Clem Cole]

Interesting... Horton's timing sounds right because  I'm pretty sure we had
some sort of Berkeley shell @ before CMU had 2BSD on the v6++ systems in
1978 - (I have to ask him, Klone must have been the one that brought it
over to Mellon Institute ). I state that because I remember trying to play
with it as well as another hacked shell V6 (I think from Harvard) around
that time.  I was fascinated by the idea of being able to change the
default command system, which no other OS I was using I could do same
(TOPS*, VMS, TSS, Exec/8).   But I remember I didn't like some of choices
of the Berkeley shell's syntax and tended to avoid it/could not figure it
out.    Within a year or so V7 showed up there after with Bourne shell and
I was happy with that.

A few years later, I did switch to typing to the csh when I got to UCB, but
that was not until after the MIT job control stuff had been spliced into
the BSD kernel (Horton & Kleckner were probably the ones that convinced me
to learn it).  With job control I became a fan, but never warmed up to the
programming syntax.   I picked up the mantra that I still consider wise --
"type to Joy and program to Bourne."   This is comfortable for the ROMS in
the muscles of my fingers, but my scripts are portable.

Clem

PS To this day (like about a month ago), if I need to hack on my .login
script when I move sites (I have some site specific stuff in .login and
.profiles), I have to grab the cshell man page so I don't screw up the
syntax.

On Sun, Jun 26, 2016 at 2:14 PM, Larry McVoy <lm at mcvoy.com> wrote:

> On Sun, Jun 26, 2016 at 09:30:39AM -0700, Mary Ann Horton wrote:
> > I brought csh with me to Bell Labs in the summer of 1979.  The folks at
> Bell
> > Labs recoiled in horror: they had just gone through a painful conversion
> > from the Mashey shell to the Bourne shell
>
> I used csh for a while before ksh became available.  It was an improvement
> over the Bourne shell, IMO, but once ksh came out I went back to Bourne
> shell syntax.  And now bash is pretty nice.
>
> --lm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20160626/9273e2dc/attachment.html>

[TUHS] Origin year of BSD csh?

[Joerg Schilling]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1330 bytes --]

Clem Cole <clemc at ccc.com> wrote:

> A few years later, I did switch to typing to the csh when I got to UCB, but
> that was not until after the MIT job control stuff had been spliced into
> the BSD kernel (Horton & Kleckner were probably the ones that convinced me
> to learn it).  With job control I became a fan, but never warmed up to the
> programming syntax.   I picked up the mantra that I still consider wise --
> "type to Joy and program to Bourne."   This is comfortable for the ROMS in
> the muscles of my fingers, but my scripts are portable.

Job control of course was an important improvement. I took the idea and 
implemented in my bsh in 1985.

Now looking back, it is interesting, that there are just four shells that 
implement support for vfork():

-	csh - the first

-	bsh since 1985

-	ksh vfork() probably since 1984, jobcontrol apparently since 1982.

-	bosh (my recent Bourne Shell) since 2014

But on a decent OS, vfork() helps a lot to speed up the shell.

On Solaris, fork() is copy-on-write based but still 3x slower than vfork().

Jörg

-- 
 EMail:joerg at schily.net                  (home) Jörg Schilling D-13353 Berlin
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.org/private/ http://sourceforge.net/projects/schilytools/files/'

[TUHS] Origin year of BSD csh?

[Ronald Natalie]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 166 bytes --]

vfork() is of use on non-paged (and poorly implemented paging) systems.    If you implemented the copy-on-write fork() behavior, you’d not need the vfork KLUDGE.

[TUHS] Origin year of BSD csh?

[Steffen Nurpmeso]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 528 bytes --]

Ronald Natalie <ron at ronnatalie.com> wrote:
 |vfork() is of use on non-paged (and poorly implemented paging) system\
 |s.    If you implemented the copy-on-write fork() behavior, you’d not\
 | need the vfork KLUDGE.

I think there is currently going on some (i haven't really a glue)
virtually mapped stack in Linux (thread around [1]), and it seems
vfork() there doesn't even copy the page table.  So that seems to
be a measurable difference.

  [1] http://www.openwall.com/lists/kernel-hardening/2016/06/21/10

--steffen

[TUHS] Origin year of BSD csh?

[Joerg Schilling]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 855 bytes --]

Ronald Natalie <ron at ronnatalie.com> wrote:

> vfork() is of use on non-paged (and poorly implemented paging) systems.    If you implemented the copy-on-write fork() behavior, you???d not need the vfork KLUDGE.

This is what the Linux people believe. As a result, they have a vfork() 
implementation that collects all pitfalls from fork() and vfork() ;-)

The basic difference is:

-	With a copy-on-write fork, you copy an address space description and
	need to set up a set of new MMU PTEs.

-	With vfork, you borrow the address space descrition and the MMU PTEs 
	from the parent.

Jörg

-- 
 EMail:joerg at schily.net                  (home) Jörg Schilling D-13353 Berlin
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.org/private/ http://sourceforge.net/projects/schilytools/files/'

[TUHS] Bizarre job control, was csh

[Warren Toomey]

I wrote a shell quite a while ago, based on a friend's shell and also the shell in Marc Rochkind's book. It was portable across a lot of systems but small enough to fit on Minix.
I used ptrace() to implement job control on Minix. See ftp://minnie.tuhs.org/pub/Wish/wish_internals.pdf

Cheers, Warren
-- 
Sent from my Android phone with K-9 Mail. Please excuse my brevity.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20160627/8468cc94/attachment.html>

[TUHS] Bizarre job control, was csh

[Steve Nickolas]

On Mon, 27 Jun 2016, Warren Toomey wrote:

> I wrote a shell quite a while ago, based on a friend's shell and also the shell in Marc Rochkind's book. It was portable across a lot of systems but small enough to fit on Minix.
> I used ptrace() to implement job control on Minix. See ftp://minnie.tuhs.org/pub/Wish/wish_internals.pdf
>
> Cheers, Warren
>

I tried a couple times to figure out how to implement a Bourne shell, and 
just couldn't figure it out.  But then, I was never much of a programmer.

-uso.

[TUHS] Bizarre job control, was csh

[Joerg Schilling]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 830 bytes --]

Steve Nickolas <usotsuki at buric.co> wrote:

> I tried a couple times to figure out how to implement a Bourne shell, and 
> just couldn't figure it out.  But then, I was never much of a programmer.

It is hard and you cannot do it from just reading the POSIX standard.

This is because the POSIX standard tries to write descriptions in an abstract 
notation that missleads people that did never see a working implementation 
before.

After a few years of maintaining the Bourne Shell, I would know how to do it but 
I am not interested to do it from scratch.

Jörg

-- 
 EMail:joerg at schily.net                  (home) Jörg Schilling D-13353 Berlin
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.org/private/ http://sourceforge.net/projects/schilytools/files/'

[TUHS] Bizarre job control, was csh

[Steve Nickolas]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 942 bytes --]

On Mon, 27 Jun 2016, Joerg Schilling wrote:

> Steve Nickolas <usotsuki at buric.co> wrote:
>
>> I tried a couple times to figure out how to implement a Bourne shell, and
>> just couldn't figure it out.  But then, I was never much of a programmer.
>
> It is hard and you cannot do it from just reading the POSIX standard.
>
> This is because the POSIX standard tries to write descriptions in an abstract
> notation that missleads people that did never see a working implementation
> before.
>
> After a few years of maintaining the Bourne Shell, I would know how to do it but
> I am not interested to do it from scratch.
>
> Jörg
>
>

Eh.  I had a specific reason for doing it that most people would find 
idiotic - I was using OSes (DOS - mainly on a Tandy 1000, which in no way 
could ever hope to handle DJGPP - and Win32) that weren't remotely POSIX 
and I wanted to make them feel more Unixy without outright *emulating 
Unix*.

-uso.

[TUHS] Origin year of BSD csh?

[Steve Nickolas]

On Sun, 26 Jun 2016, Larry McVoy wrote:

> On Sun, Jun 26, 2016 at 09:30:39AM -0700, Mary Ann Horton wrote:
>> I brought csh with me to Bell Labs in the summer of 1979.  The folks at Bell
>> Labs recoiled in horror: they had just gone through a painful conversion
>> from the Mashey shell to the Bourne shell
>
> I used csh for a while before ksh became available.  It was an improvement
> over the Bourne shell, IMO, but once ksh came out I went back to Bourne
> shell syntax.  And now bash is pretty nice.
>
> --lm

For some daft reason my first foray onto the Unix command line was tcsh, 
and later I switched to bash, which is still my primary choice - though I 
don't mind using any other Korn-type shell, long as I got my "emacs 
editing" mode (although the real ksh's tab completion is clunkier than 
bash's).

Using a Bourne shell that doesn't have a line editor is a pain in the 
keester, but if I must, I can deal.

I have no idea how to use csh, and if I'm set up with csh as my default 
shell the first thing I'll do is try to switch it to bash or ksh!

-uso.

[TUHS] Origin year of BSD csh?

[Joerg Schilling]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 4013 bytes --]

Aharon Robbins <arnold at skeeve.com> wrote:

> Can anyone give a definitive date for when Bill Joy's csh first got out
> of Berkeley?  I suspect it's in the 1976 - 1977 time frame, but I don't
> know for sure.

In 1977 (published November 23), there was "ashell" with this "READ_ME":

/*--------------------------------------------------------------------------*/
Wed Oct 19, 1977

This directory contains the source for a shell.
It requires floating point to do the time command which is built-in
so you will have to cc it -f on machines without floating point.
It also requires a version 7 C compiler.

Accurate documentation is in the file "sh.6" to be nroffed with
/usr/man/man0/naa and a new "version 7" nroff.

This shell requires the "htmp" data base also used by the editor "ex".
If you do not set it up so that the "sethome" command is done by "login"
then you should use the old "osethome" routine in ../s6 rather than "sethome"
and reenable the execl of this sethome in the file "sh.c" (with the correct
pathname).

                                Bill Joy
                                CS Division
                                Department of EE and CS
                                UC Berkeley
                                Berkeley, California  94704

                                (415) 524-4510          [HOME]
                                (415) 642-4948          [SCHOOL]
/*--------------------------------------------------------------------------*/

Given that ashell/sh.c contains:

/*
 * Shell
 *
 * Modified by Bill Joy
 * UC Berkeley 1976/1977
 *

it was most likely based on the Thompson shell.

Here is the start of the man page:

/*--------------------------------------------------------------------------*/
SH(VI)                       9/15/77                       SH(VI)



NAME
     sh - a shell (command interpreter)

SYNOPSIS
     sh [ -V ] [ -v ] [ -t ] [ -c ] [ -i ] [ name [ arg ...  ] ]

DESCRIPTION
     Sh is a command interpreter.   It  arranges  and  interprets
     command  lines  and  the contents of command files.  It is a
     modification of the standard shell sh (I), and  almost  com-
     pletely upward compatible therewith.  The intent, in working
     on a new shell, is to provide an environment which  is  more
     easily tailored to the wishes of each individual user.  Most
     new features of this shell, especially  the  alias  feature,
     are  toward  this  end.   Later  versions  of this shell may
     include improvements to the command language  of  the  shell
     and allow more easy repetition of commands.  The intent here
     is to make the command language more resemble  a  high-level
     language  - C being the natural choice for UNIX, and to pro-
     vide some  means  of  repeating  modified  commands  without
     retyping,  perhaps  akin to the INTERLISP redo feature.  The
     eventual goal is a C-shell,  csh  (or  ``seashell''  if  you
     prefer.)
/*--------------------------------------------------------------------------*/

BTW: csh was an improvement for most shells from that time, but it lacks a
decent history editor.

In 1982, I wrote my first experimental history editor that supports cursor keys 
but called the commands via system() and in 1984, I integrated this concept 
into a shell called "bsh" that we had at H. Berthold AG on an OS called 
"VBERTOS" that was based on "UNOS" - the first UNIX clone.

A csh port for UNOS was available around 1982, but with the availability of 
a shell with integrated history editor, other shells seemed to be of no 
real interest. So around September 1984, people at H.Berthold AG stopped using
csh even though bsh had similar problems in the shell command language as
seen with csh.


Jörg

-- 
 EMail:joerg at schily.net                  (home) Jörg Schilling D-13353 Berlin
       joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.org/private/ http://sourceforge.net/projects/schilytools/files/'