[net.sources.bugs] csh on 3b2's?

rhealey@umd-cpe.UUCP (Rob Healey) (07/01/86)

In article <17600017@smu>, rab@smu writes:
> 
> 
> anybody ever heard of csh for the 3b2?
> 
> i have been looking until i am blue in the face.  the shell script that
> was written to look like csh is interesting, but painfully slow.
> 
> if ANYONE has ANY idea about this, please let me know.
> 
> thanks.
> 
> rick barrett
> convex!smu!rab

	Rick,

	 You can try to use older versions of the ucb csh if you have the
      license for the source and if your license says it's ok to do that.

	I found a 1979 version of the csh on our PDP-11 and sort of got it
      to work. The PDP-11 version has no job control and the sources say
      that the version I used isn't even supported by Berkeley. I rewrote
      parts of the code because of PDP-11isms but everything eventually
      worked, sort of. The shell is unstable and likes to crash when one
      sources a file with errors in it. I use the shell a little but since
      I don't know if it'll crash when I do something I mostly use the
      Bourne shell. The shell works for about 60-90 commands before it starts
      acting weird and I have to exit and restart the shell. I heard AT&T was
      going to release a version of the C shell for the 3b2 but that was just a
      rumor. 

	I think the 3b2 C compiler is brain-damaged somewhere though.
      The code generated by the compiler would use registers for important
      pointers, call a function without saving the registers, and wham,
      instant core dump when you try to use the pointer. What's really a shame
      is that none of the pointers used were declared register, the compiler
      did that on it's own but forgot to save/restore the registers.
      I couldn't release ANY memory space used by the shell or it would dump
      core.

	The lack of a functioning C shell is a pain on the 3b2, it's not a
      bad machine for 3-6 users it just lacks the BSD "extras" that make
      BSD systems more pleasant to use.

	I didn't really answer your question I suppose but it is an idea.
      although a very shakey one.


			-Rob Healey
			 CpE Dept.
			 University of Minnesota, Duluth

		  ..!ihnp4!umn-cs!umd-cs!umd-ua!umd-cpe!rhealey


#include <disclaimers.h>

david@ukma.UUCP (David Herron, NPR Lover) (07/06/86)

Why do you want the csh there?  

I can think of two reasons only:

1)  Aliases -- but the SysVr2 shell's shell procedures are MUCH MUCH
    better.
2)  History -- But there was some patches posted by somebody at GaTech
    about a year ago (Was it Arnold Robbins???  I don't recall, if so
    he's at Emory now).  The patches are for SysVr2 shell on SysV, same
    shell on 4.2BSD and Bezerkely's Bourne shell on 4.2BSD.  The history
    is different from the csh's mechanism so I haven't gone through
    the trouble of learning it completely but that's mainly because
    I still use csh from time to time...


I find the csh to be badly thought out in general -- I don't understand
why people think it's so neat..  

There *are* a few problems with gsh (that's what I'm calling the sVr2
shell with those patches) but they're really because Arnold (?) doesn't
have time to really debug the thing (and neither do I for that matter).

Fr'instance, occasionally it crashes on me in a similar way to the bug
people have been talking about (Re: memory allocation) when I use history.
Also on BSD systems it doesn't do the same things the csh does with
process groups -- The controlling shell stays in the same process
group as the processes it execs which occasionally causes odd behavior.


-- 
David Herron,  cbosgd!ukma!david, david@UKMA.BITNET, david@uky.csnet

mikel@codas.ATT.UUCP (Mikel Manitius) (07/07/86)

> Why do you want the csh there?  
> 
> I can think of two reasons only:
> 
> 1)  Aliases -- but the SysVr2 shell's shell procedures are MUCH MUCH
>     better.
> 2)  History -- But there was some patches posted by somebody at GaTech
>     about a year ago (Was it Arnold Robbins???  I don't recall, if so
>     he's at Emory now).  The patches are for SysVr2 shell on SysV, same
>     shell on 4.2BSD and Bezerkely's Bourne shell on 4.2BSD.  The history
>     is different from the csh's mechanism so I haven't gone through
>     the trouble of learning it completely but that's mainly because
>     I still use csh from time to time...
>
> David Herron,  cbosgd!ukma!david, david@UKMA.BITNET, david@uky.csnet

Get ksh (Korn Shell), it will give you aliases, history, and much
more, like command editing and re-entry ala vi style.
-- 
    ___
   /   \		Mikel Manitius @ AT&T-IS Altamonte Springs, FL
  | RPI |		...{seismo!akgua|ihnp4|cbosgd|mcnc}!codas!mikel
  |  .  | 
\\-------//