pmaresch@hawk.ulowell.edu (Pierre Mareschal) (10/22/90)
Hi,
Here is a message I get when building latex on the IBM R6000
systems but I never encountered such a problem!
MF doesn't have the same problem.
When done, latex loops infinitely on a simple file (small.tex)
and never exits.
I compiled tangle and web2c with and without -O flags
and got the sampe problem.
Any suggestions?
------ Begin of Message
% make
../web/tangle tex.web ctex.ch
This is TANGLE, C Version 4
*1*17*25*38*54*72*99*110*115*133
! Value too big: 2550000. (l.3229)
@d
penalty(#) == mem[#+1].int {the added cost of breaking a list here} *162*173*
199*203*207*211*220*256*268*289*297*300*321*332*366*402*464*487*511*539*583*592*
644*680*699*719*768*813*862*891*900*919*942*967*980*1029*1055*1136*1208*1299*133
0*1338*1340*1379*1381
Writing the output file.....500.....1000.....1500.....2000.....2500.....3000....
.3500.....4000.....4500.....5000.....5500....
Done.
1035 strings written to string pool file.
(Pardon me, but I think I spotted something wrong.)
1254-004 The error code from the last failed command is 1.
Make Quitting.
------ End of Message
:- Pierre Mareschal <pmaresch@hawk.ulowell.edu>space@ncc1701.stgt.sub.org (Lars Soltau) (10/27/90)
In article <1363@ul-cs.ulowell.edu> pmaresch@hawk.ulowell.edu (Pierre Mareschal) writes: >Hi, Lo! >Here is a message I get when building latex on the IBM R6000 >systems but I never encountered such a problem! Um, just a comment: you aren't building LaTeX, you are building TeX, since LaTeX is just a macro package. OK, enough nitpicking, let's get to the real issue. >------ Begin of Message >% make > ../web/tangle tex.web ctex.ch >This is TANGLE, C Version 4 >*1*17*25*38*54*72*99*110*115*133 >! Value too big: 2550000. (l.3229) >@d > penalty(#) == mem[#+1].int {the added cost of breaking a list here} *162*173* [...] That line looks exactly the same in my version. I don't have the same problem, though. I have build TeX and mf, both in the big (32bit) version, on our /6000 and everything is running (running indeed!) merrily along. So, here's what I think can have caused your problem: 1.) I compile everything with bsdcc. Read /usr/lpp/bsdport for details about bsdcc, if you haven't already done so. 2.) Look at your changes file (ctex.ch). There should be something like ctex32.ch and ctex16.ch, or at least a bigtex.diff, which you should apply (Sorry, no patch on AIX3.1, you have to do it by hand or port patch first). I can't understand why all of you have this problem. I got a tape with the very same TeX sources that were compiled on our Sun4 at my university and compiled it on the /6000 within about an hour. No problems whatsoever. -- Lars Soltau bang: <insert ridiculously long path> Bix: -- no bucks -- smart: space@ncc1701.stgt.sub.org Stuttgart-Zuffenhausen, where dreams come true: Heaven is Eleven!
jpe@egr.duke.edu (John P. Eisenmenger) (10/31/90)
From article <1363@ul-cs.ulowell.edu>, by pmaresch@hawk.ulowell.edu (Pierre Mareschal): > > Hi, > > Here is a message I get when building latex on the IBM R6000 > systems but I never encountered such a problem! > ../web/tangle tex.web ctex.ch > This is TANGLE, C Version 4 > *1*17*25*38*54*72*99*110*115*133 > ! Value too big: 2550000. (l.3229) This is a simple problem, but one which will stump many people, so I am posting to the net. The problem is that the "char" type is *NOT* signed on the RS6000. Simply change the line #define schar char in your site.h file to be #define schar signed char and the problem should clear up. Oh yes, I am also using the bsdcc compiler as outlined in the "porting BSD" notes. -John Eisenmenger Dept. of Electrical Engineering Duke University