[net.unix-wizards] Query on shells

gregg%okstate.csnet@CSNET-RELAY.ARPA (Gregg Wonderly) (05/27/85)

	Here at OSU, a small number of people have become interested in shells.
We have put together a shell that uses a small subset of VI as a front end.
There are many features that are present that were derived from CSH and SH.
We are interested in comments as to what features of these and other shells
(KSH, VSH, etc) that you as programmers/users deem beneficial.  Also of 
interest, would be ideas of things not present in the shell(s) that you 
have used that you have needed on occasion.  The shell we have developed 
so far, does not support a language in the sense that SH, CSH, and others do.  
Our main interest here, is to provide a working environment, whereby a 
programmer, or a user involved in heavy interaction with the system, can be 
more productive.  Things to consider stem from the available resources.
Memory requirements, system resources (I/O, and computer execution times),
as well as external hardware (terminals, network requirements, and maintenance
on such things), are all non-common variables.  The flavor of UN*X running
is the biggest problem.  Some have certain features (SIGSTOP, SIGCHLD, 
FIONREAD, etc...) that are not available on older flavors.  These features
provide means for easily implementing certain things.  Some features/bugs have 
disappeared in later flavors.  These are just some of the things that have 
been thrown around here.

	Any ideas, big or small, that you have are welcome.  Please reply by direct 
mail if possible.


Gregg Wonderly
Department of Computing and Information Sciences
Oklahoma State University

            1
           /|\  UUCP: {cbosgd, ea, ihnp4, isucs1, mcvax, uokvax}!okstate!gregg
           |o|
   _______//|\\_______
  |_|_|_|_||s||_|_|_|_|
  |_|_|_|_||_||_|_|_|_|
  |_|_|_|_||u||_|_|_|_|  ARPA:  gregg%okstate.csnet@csnet-relay.arpa  

tower@inmet.UUCP (06/04/85)

re: new features for shells

I had oftened wish for an alternative form of pipe syntax that would
indicate that I wanted a temporary file used instead of an in-core pipe.
This would be use for scripts dealing with a large amount of data, and
also on heavily loaded systems.

-len tower	    UUCP:     {bellcore,ima,ihnp4}!inmet!tower
 Intermetrics, Inc. INTERNET: ima!inmet!tower@CCA-UNIX.ARPA
		    USPS:     733 Concord Ave., Cambridge, MA  02138, USA
		    PHONE:    +1 (617) 661-1840

mike@peregrine.UUCP (Mike Wexler) (06/11/85)

> I had oftened wish for an alternative form of pipe syntax that would
> indicate that I wanted a temporary file used instead of an in-core pipe.
> This would be use for scripts dealing with a large amount of data, and
> also on heavily loaded systems.
> 

The whole idea of pipes is that the both ends of the pipe are run in
parellel.  This allows large amounts of data to pass through the
pipes without having a large amount of space to buffer it.  
For instance if I type in the command:
	cat xyz | more
It doesn't not more the whole contents of xyz into memory and then
feed it into more, but it moves the data in 512 byte(system dependent)
chunks into more. More then outputs the stuff to the screen.  Instead
of adding another operator for situations that need large buffers,
the | operator ought to write the stuff out to disk when it runs out
of space(sounds like demand paging to me).


-- 
--------------------------------------------------------------------------------
Mike Wexler(trwrb!pertec!peregrine!mike) | Send all flames to:
15530 Rockfield, Building C              |	trwrb!pertec!peregrine!nobody
Irvine, Ca 92718                         | They will then be given the 
(714)855-3923                            | consideration they are due.

ignatz@aicchi.UUCP (Ihnat) (06/12/85)

In <129@peregrine.UUCP>,  Mike Wexler writes:

>	For instance if I type in the command:
>		cat xyz | more
>	It doesn't not more the whole contents of xyz into memory and then
>	feed it into more, but it moves the data in 512 byte(system dependent)
>	chunks into more. More then outputs the stuff to the screen.

It's not horribly important, I suppose, but don't make the poor, tired
kernel appear more inefficient than it deserves.  A pipe will actually
accept up to 4K of data without any reads from the other end before it
wedges....
-- 
	Dave Ihnat
	Analysts International Corporation
	(312) 882-4673
	ihnp4!aicchi!ignatz