From: Darren Senn (sinster@tioga.ucsc.edu)
Date: 06/22/92


From: sinster@tioga.ucsc.edu (Darren Senn)
Subject: Re: Trouble compiling TeX
Date: 22 Jun 1992 20:02:50 GMT

In article <HEIM.92Jun22201009@frieda.peanuts.informatik.uni-tuebingen.de>, heim@peanuts.informatik.uni-tuebingen.de (Gerald Heim) writes:
> I have had the same problems. The key is: flex doesn't work right.
> You should lex the *.l files at another UNIX box that has another lex.
> Or, my solution, precompile tex on a UNIX box (be aware of the right
> endian definition in common/endian.h which is made at compile-time).

I had no real trouble compiling TeX "right out of the box"... Though
I didn't use any "unixtex.tar.Z": I grabbed web2c and the various web
files from labrea.stanford.edu (the central TeX distribution point).

I especially didn't have any trouble with flex: just grab the code from
prep.ai.mit.edu and compile: no modifications or patches needed.

I only needed to make two changes to get it to work:
        a) The program splitup in web2c/web2c doesn't work: it'll terminate
           input when it sees a blank line: change the call to gets() to
           a call to fscanf() or fgets() or just write your own gets, and
           things work fine.
        b) The ed script in web2c/mf that's used to split up the mf.web
           output is broken: it'll generate a bunch of symbols of the
           form [rl]h{field}+ ("r" or "l" followed by "h" followed by
           one or more occurrences of "field"). These need to be folded
           into [rl]hfield (same symbol, but just one "field").
With those two changes, everything works fine.

-- 
Darren Senn                                            Phone:    (408) 479-1521
sinster@scintilla.capitola.ca.us                       Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo.     Santa Cruz, Ca 95062