Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!apple!uokmax!servalan!epmooch!ben From: ben@epmooch.UUCP (Rev. Ben A. Mesander) Newsgroups: comp.sys.amiga.programmer Subject: Re: gcc hunger Message-ID: Date: 14 Apr 91 11:00:50 GMT References: <9104150213.aa27911@muse.compulink.co.uk> <14APR91.22254747@uc780.umd.edu> Lines: 39 >In article <14APR91.22254747@uc780.umd.edu> cs450a03@uc780.umd.edu writes: >Richard Harrison writes: >>You can't compile GCC with SAS/C mainly due to limitations in the SAS >>pre-processor. Using GCC to compile G++ I need 4MB (ie. 3MB free). > >Maybe this is a stupid question, but: >(1) can't gnu-cpp be compiled? If not, isn't there a gnu version of >cpp for use in porting? (Lattice wasn't the first to introduce a weak >preprocessor, I think.) One thing you can do is compile on a Sun 3 system (or with any GCC host machine with a Sun 3 GCC cross-compiler) and use SobjA to convert the Sun .o files to Amiga .o files. Link and go! You might also look at the DECUS cpp to see if it can compile under SAS/C and if it can handle the GCC source. >(2) can't you just preprocess before handing off to lc? (with >includes turned off, hopefully.) Maybe. GCC code is hairy, I'm not sure I'd trust lc. I know that Sun's cc doesn't work quite right on it. Remember that GCC requires a special library of routines to be written (GNULIB) before it can be used. I'd think the point is moot! After all, now you can use GCC to compile whatever changes you like to GCC. >Anyways, I guess that I can accept things as they are... 4MB doesn't >seem that bad [I've heard numbers like 32MB memory required on other >systems -- rather scary.] I've not seen a system that required this much memory to compile GCC. >Raul Rockwell -- | ben@epmooch.UUCP (Ben Mesander) | "Cash is more important than | | ben%servalan.UUCP@uokmax.ecn.uoknor.edu | your mother." - Al Shugart, | | !chinet!uokmax!servalan!epmooch!ben | CEO, Seagate Technologies |