🚀 go-pugleaf

In article <1990Dec11.191456.10816@ioe.lon.ac.uk> teexnma@ioe.lon.ac.uk (Nino Margetic) writes:
>If my memory still serves me, the problem with the original AWF as
>distributed by Henry Spencer (henry@zoo.toronto.edu), when running under 
>MSDOS, was with nested loops in the AWF script, and the person who later 
>claimed that he/she has done the "port" to MSDOS, said that he/she only 
>changed AWF to reduce the number of levels in nested loops.
>To be honest, I haven't tried running the AWF script since the time it
>was originally released (after my first trial has failed).
>N.B. If I rememeber correct, Henry Spencer himself was very skeptic
>about running his script on 16bit machines (see README file in the original
>distribution of AWF - comp.sources.unix, v23, issue 27; released
>Sept 7th 1990).

I did not port awf to DOS but I got AWF running last night under OS/2
with GAWK 2.11.1. I did only modify the awf script for path names etc.
and a bug in my SH with "(cmd; cmd ...) >file" but nothing in the awk
passes. Maybe it's just a problem of available memory? It worked well,
the regression test was 100% ok and it formatted another man page
correctly. That means, AWF *works* with 16bit machines if they only
have enough (process) address space. Besides GAWK, AWF also needs SH
(and SED, CAT, RM) which may cause memory shortage under DOS if you run
GAWK from within a SH script.

Kai Uwe Rommel

--
/* Kai Uwe Rommel
 * Munich
 * rommel@lan.informatik.tu-muenchen.dbp.de
 */