Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!pacbell.com!uop!napa!hughes From: hughes@napa.uop.edu (Ken Hughes) Newsgroups: comp.os.386bsd.apps Subject: Re: GCC 2.3.3 on 386bsd Date: 30 Mar 1993 16:17:31 GMT Organization: Electrical and Computer Engineering, University of the Pacific Lines: 53 Message-ID: <1p9rqr$rb@unix1.cc.uop.edu> References: <1993Mar29.222819.25110@njitgw.njit.edu> <CAZ.93Mar30123709@sqsosk.psq.mei.co.jp> <1993Mar30.071217.14597@sifon.cc.mcgill.ca> NNTP-Posting-Host: napa.eng.uop.edu In article <1993Mar30.071217.14597@sifon.cc.mcgill.ca> storm@cs.mcgill.ca (Marc Wandschneider) writes: >In article <CAZ.93Mar30123709@sqsosk.psq.mei.co.jp> caz@psq.mei.co.jp (Caz Yokoyama) writes: >>In article <1993Mar29.222819.25110@njitgw.njit.edu> kxn3796@hertz.njit.edu (Ken Nakata CIS stnt) writes: >> >> I succeeded to make xgcc without severe problems but xgcc complained >> when it was compiling enquire.c. It claimed that numerical constants >> in enquire.c were out of range. But there are only ZEROs and ONEs. >> I'm sorry not to have exact errorneous line number. >>------------------------------------------- >>The problem that yopu encount is following. >> >>$ configure i386-bsd >>$ make >>..... >>./xgcc -B./ -DIN_GCC -g -DNO_MEM -DNO_STDDEF -DNO_LONG_DOUBLE_IO -O0 -I.-c ./enquire.c >>./enquire.c: In function `fprop': >>./enquire.c:2307: floating constant out of range >>*** Error code 1 >> >>Stop. > > The best solution I have seen to this problem is to just IGNORE the >error generated by enquire.c > > ie: > >gzcat gcc-2.3.3.tar.z | tar xf - >cd gcc-2.3.3 >./configure --target=i386-unknown-bsd >make LANGUAGES=c ># IGNORE ALL enquire.c FLOATING POINT ERROR STUFF >make stage1 >......... This is very interesting. I've been trying to compile GCC-2.3.3 for three days and know that ./xgcc (specifically ./cc1) is HANGING in line 2307 in atof(); the compilation comes to a complete stop with no warnings or error messages and I have to kill the process. I've tried using unpatched and patched kernels and libraries, and changing the constants in float.h and math.h but nothing seems to make a difference. I'm using a 486/33 so I assume it's not math-coprocessor incompatibility problem. It doesn't seem to be a specific problem of GCC-2.3.3; the only reason I'm trying to build my own is some programs I've ported to 386BSD (for my disseratation) hang the same way in the floating point routines. Anyone seen anything like this before? Ken -- Ken Hughes | "I can't believe this is my life; (hughes@napa.eng.uop.edu) | I'm going to have to send my SAT FT-Ph D candidate, PT-ex-sysadm | scores to San Quentin instead of University of South Florida | Stanford..." _Heathers_