Re: ^Z signal

[pete@minster.york.ac.uk]

> 	Did you strip the binary?  On my machine, that cut it down from 624K
> 	to 192K.  And that is with editline linked in.  Compiling with
> 	assertions off might make it smaller yet.  On our machine, tcsh and
> 	bash are considerably larger (though bash is statically linked, so the
> 	comparison is not quite fair).
> 
> 	- Harald

Yes, to both stripping and assertions. Here's some sizes: (on Sparcs, under
sunos 4.1.1)

dcs118% ls -l $b/sh $b/csh $b/rc $b/ksh $s/es $s/wash ~spqr/bin/bash
-rwxr-xr-x   3  root     staff     68312 Feb 13  1990 /bin/sh
-rwxr-xr-x   1  root     staff    147456 Oct 11  1990 /bin/csh
-rwxr-xr-x   1  root     staff     90112 Apr  7  1992 /bin/rc
-rwxr-xr-x   3  root     staff     68312 Feb 13  1990 /bin/ksh
-rwxr-xr-x   1  pete     staff    172032 May  2 17:24 /usr/pete/bin/sun4/es
-rwxr-xr-x   1  pete     staff    159224 Oct  1  1992 /usr/pete/bin/sun4/wash
-rwxrwxr-x   1  spqr     staff    483328 Feb 24  1992 /usr/spqr/bin/bash

(hmmm... looks like ksh is a link to sh here :-))

Admittedly (like your system) bash is statically linked, and I know we've
got an out-of-date version of rc...

Now, on a fast-ish machine with large-ish memory I agree that image size is
not vitally important, but try telling that to some poor soul on a 3/50
with only 4Mb!

This is precisely why I want es to stay small and neat -- so that _everyone_
can get the benefits of relatively clean syntax and easy customisability. 
Heading into gnu-ish bloat is the last thing I want to see happen. Commnand
line editing and job control are things which can, and should, be handled
elsewhere. 

pete
--
 *Peter Fenelon -- Research Associate -- Software Safety Assessment Procedures*
 Dept. of Computer Science, University of York, York, Y01 5DD (+44 0904 433388)
 EMAIL: pete@minster.york.ac.uk ``You are what you drink and I'm a bitter man''