On Mon, Mar 20, 2000 at 08:19:05AM -0700, Jim Franklin wrote: Thanks. egcs ought to work for the entire process - I chose to use the most recent compiler for 2 reasons:
1) There were some significant improvements in the cross-compiler stuff in the 2.95 series. This is not particularily important when staying on an i386 platform, but very important when cross-compiling from other platforms like UltraSparc. 2) Bigger, better, sexier optimizer. My system just 'feels' a touch faster when compiled with 2.95.2 - I haven't substantiated this with benchmarks. I need to go through the effort of figuring out how to setup OSKit && GNUMach. Once that is done, I will change the instructions so that egcs isn't used at all. (Contributions welcome!! <hint, hint - for that and any other 'help documentation that people want posted'>). Tks, Jeff Bailey > Hi Jeff > I really like your page on cross compilation. Is there any reason to > not use egcs 1.1.2 for the entire process? > > Jim > > Jeff Bailey wrote: > > > > On Sun, Mar 19, 2000 at 10:11:43PM -0500, Colin Walters wrote: > > > > I have documented what works for me at http://hurd.zugzug.com/ - The only > > change I recommend from those is to update to Glibc 2.1.3, and that the > > latest Hurd CVS can now be compiled with gcc 2.95.2. > > > > I will update the instructions this week sometime. > > > > > What versions of the various GNU tools do people use for a cross > > > compilation environment from GNU/Linux? > > > > > > i.e. should I try using the latest GCC? Or would it be better to > > > stick with 2.7.2.3? Are there any problems with the latest binutils? > > > Knowing what is a currently working set will probably save me lots of > > > time. > > > > > > -- > > > Colin Walters <[EMAIL PROTECTED]> > > > http://web.verbum.org/levanti > > > (1024D/C207843A) A580 5AA1 0887 2032 7EFB 19F4 9776 6282 C207 843A > > > > -- > > There is no sin except stupidity. > > - Oscar Wilde -- There is no sin except stupidity. - Oscar Wilde
