findutils-4.1 compilation error

Edward Pinski pinskie at
Thu Jan 9 20:08:54 PST 2003

> On Thursday 09 January 2003 18:53, you wrote:
> > On Thu, Jan 09, 2003 at 05:57:25PM -0700, Mark Lee wrote:
> > > > This isn't exactly true.  I just attempted to compile findutils on my
> > > > IBM RS/6000 43P-133, which is of PPC arch.  I have Debian 2.2r7 on
> > > > there for now since I have not yet done the LFS on it, but without the
> > > > -D_GNU_SOURCE it compiled just fine.  So as far as it fixing for
> > > > non-x86, this is not entirely true at all, Maybe tomorrow or something,
> > > > I'll try it on my Sun Sparcsation 10, plain ol' sparc arch there.
> > >
> > > Try the patch I posted yesterday (or the day before). I think that should
> > > make it work.
> >
> > The point wasn't that it worked without -D_GNU_SOURCE
> > but that it worked without that flag on something non-x86.
> > Pinski didn't say anything about -D_GNU_SOURCE breaking
> > compilation.
> Ok, I see what you mean. I think this is because the -D_GNU_SOURCE exposes all 
> of the gnu extensions, so name conflicts are more likely. But, if you get the 
> package to work with the _GNU_SOURCE, than it will be more likely to work on 
> non-x86 systems (which I think is why it was added). 
> My patch should do the second part (getting it to work with -D_GNU_SOURCE). I 
> think in general, the _GNU_SOURCE is always good practice (in gnu software), 
> but you have to make sure it doesn't break anything (which in this case it 
> did).

Okay let me say this, if -D_GNU_SOURCE was added to "aid" compliation
on non-x86 platforms, yet non-x86 platforms(PPC at least) do not
need it, there should be no point in having the -D_GNU_SOURCE at all.
That is, if the reason why it says in the Changelog is the exact
reason why it was added to the book, because if it was, it's not entirely
true, at least for me on PPC.  That being, I was suggesting it to be
dropped from the book, since there is really no reason to have it in
the book in the first place.

Edward Pinski
Unsubscribe: send email to listar at
and put 'unsubscribe lfs-dev' in the subject header of the message

More information about the lfs-dev mailing list