LFS 6.0 BootCD - A call for testers

Matthew Burgess matthew at linuxfromscratch.org
Wed Sep 15 15:54:15 PDT 2004


On Wed, 15 Sep 2004 17:42:45 -0500
Bruce Dubbs <bdubbs at cis.sac.accd.edu> wrote:

> Ian Molton wrote:
> > On Wed, 15 Sep 2004 10:38:47 -0500
> > Bruce Dubbs <bdubbs at swbell.net> wrote:
> > 
> > 
> >>Note:  The make check will always give the error "ufs2_stage1_5 is
> >too
> >>
> >>big."  This is due to a compiler issue, but can be ignored unless
> >you 
> >>plan to *boot* from ufs partition that is normally only used by Sun 
> >>workstations.
> > 
> > 
> > Thats not fair. its not a compiler issue its a 'the code isnt
> > written
> > small enough' issue.
> 
> 
> Really?  I thought it compiled small enough on earlier versions of
> gcc.  Is 
> that wrong?

I thought it did too - however it's a pretty big assumption to make
really.  What if someone overrides the CFLAGS with -O0 for example. 
*Relying* on compiler optimisation passes to generate a small-enough
binary seems pretty dangerous to me.

Matt.



More information about the lfs-dev mailing list