🚀 go-pugleaf

There have been a number of requests for an undump program for the
Sequent systems (both Balance and Symmetry) so I thought I'd take
a moment to point one or two things out about undump.

Firstly, undump was created by Donald Knuth to get over the fact
that PASCAL was *incredibly* bad at loading the format strings from
the *.fmt files generated. Thus he had the program go through the
pain of loading the *.fmt files once and then essentially core 
dumping. Then, the running program is "pasted" back on to the core
image to make an executable that does not have to re-load the
*.fmt file.

With the introduction of ctex (and other C versions of TeX) there
is not the same penalty for loading in real-time. The net result?
Don't bother to undump. Instead, set up a shell script that looks
something like this:

#!/bin/csh
setenv TEXFORMATS /usr/lib/tex/inputs
setenv TEXINPUTS .:/usr/lib/tex/inputs
setenv TEXFONTS /usr/lib/tex/fonts
setenv TEXPOOL /usr/lib/tex

switch($0)
	case latex:
		virtex '&lplain' $1
		breaksw
	case tex:
		virtex '&plain' $1
		breaksw
	case slitex:
		virtex '&splain' $1
		breaksw
endsw

and call it tex. Then link latex, and slitex to it. This will provide you
with a command for tex, latex and slitex that loads the correct *.fmt file
and at the same time allows for all instantiations of tex, latex and slitex
to share the same text segment --- which is much healthier in a busy 
environment.

-david-

Disclaimer: Sequent does not offer nor support tex, latex, slitex or ctex
in any way or manner. These comments are my own and may not reflect Sequent
policies or practices. Your mileage may vary.

--
David Haynes	Sequent Computer Systems (Canada) Ltd.	david@torsqnt.uucp