Chris Gilbert wrote:
> On Mon, 27 Apr 1998, Matthew Wilcox wrote:
> 
> > William Gallafent
> > > Is it possible, then, that a system with the bugged SA and a compiler
> > > which shouldn't work with it could compile the whole of qt (shared) 
> > > without problem (as mine did) and then fail on bzip2, or is it more likely
> >                                 ^^^^^^^^^^^^^^^^^^^^^^
> > > that something has happened to my system in between? If so, what? (bearing
> > > in mind that I've tried installing both comp sets) ... shared library
> > > problem - does cc1 use a shared library? 
> > 
> > Ah, the one piece of information I needed which you didn't disclose before.
> > I've tried compiling bzip2 for armlinux with an ARM610 and gcc 2.7.2.2.  It
> > fails to compile bzip2.  However, 2.8.0 worked just fine.  It miscompiles
> > other things, but that's by the by.  I haven't tried 2.8.1 or EGCS.  Perhaps
> > someone would like to try it?
> 
> bzip2 works fine on my 1.3 system (using the comp.tgz from the 1.3
> release and the cc1 bits from the 1.3_alpha release), compiled just fine,
> I might have had to use gmake, can't remember...
> 

It failed to compile on my system (using the same comp.tgz and cc1 bits 
as you did). I edited the makefile and removed the -O3 flag to get rid of 
any optimisation, and it worked then...

> 
> Chris
> 
> 
> 
tim