Stray definitions of xmalloc and friends?

Stray definitions of xmalloc and friends?

[Stan Shebs]

While going through and whacking uses of "register", I noticed
that there are three apparently-leftover definitions of xmalloc and
other libiberty functions, in gencheck.c, gengenrtl.c, and
gensupport.c.  No longer any reason to keep these alive, right?

Stan

Re: Stray definitions of xmalloc and friends?

[DJ Delorie]

> While going through and whacking uses of "register", I noticed
> that there are three apparently-leftover definitions of xmalloc and
> other libiberty functions, in gencheck.c, gengenrtl.c, and
> gensupport.c.  No longer any reason to keep these alive, right?

The gen* programs aren't linked with libiberty, because they run on
the build machine and we don't yet build a build-libiberty (see my
proposal about that, though).  Thus, they must continue to provide
their own xmalloc et al functions.

Re: Stray definitions of xmalloc and friends?

[Zack Weinberg]

On Tue, Oct 09, 2001 at 05:44:13PM -0700, Stan Shebs wrote:
> While going through and whacking uses of "register", I noticed
> that there are three apparently-leftover definitions of xmalloc and
> other libiberty functions, in gencheck.c, gengenrtl.c, and
> gensupport.c.  No longer any reason to keep these alive, right?

As we still don't have a build libiberty, no, those are still
required.

Wanna fix it?  There's at least two different build-libiberty patches
floating around, they just need testing under appropriately
pathological conditions.

zw