Xref: utzoo comp.text.tex:5579 comp.os.msdos.apps:1127 Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!unido!sbsvax!fb14vax!arno From: arno@fb14vax.cs.uni-sb.de (Arno Schmitt) Newsgroups: comp.text.tex,comp.os.msdos.apps Subject: tex3.1 on MSDOS machine Keywords: tex3.1, gcc, 386 instructions Message-ID: <9505@sbsvax.cs.uni-sb.de> Date: 21 Feb 91 13:18:47 GMT Sender: news@sbsvax.cs.uni-sb.de Followup-To: comp.text.tex Lines: 38 Hello, I'm trying to compile tex3.1 on an IBM compatible machine using Keisuke Yoshida's gcc C-compiler for MSDOS. I've managed to compile the program, but, unfortunately, it will not run correctly. To be more precise: 1. I obtain a running version of INITEX. INITEX is able to read e.g. the file lplain.tex and produces a file lplain.fmt. 2. I obtain a version of VIRTEX. But when starting it with the lplain.fmt file, it won't give me the LaTeX version number, and so on. Moreover, it doesn't recognize any LaTeX command. I have compiled the version I used for my PC on a Sun and both programs worked without any trouble. Now my questions: - Is there anyone who got tex3.1 working on a 386-based MSDOS machine such that tex uses the protected mode and the 386 machine instructions? - Does anyone know about bugs of the gcc mentioned above? - Does tex depend on byte ordering? - Are there any functions that have to be changed in order to make tex work under MSDOS? (Note that I have already changed the code in such a way that reading and writing of files takes place in binary mode) My machine is a 386-IBM clone with 4.7 Meg ram and an ATI-EGA wonder graphics card, if this is of any interest. I know that there are a lot of tex-versions for PC's, but I want to have one using the capabilities of the 386 processor. I've tested the gcc mentioned above with several applications using a lot of floating point operations and the test programs were 2-3 times faster than the same program compiled with Turbo C. Thank you very much for your help arno (arno@cs.uni-sb.de)