On Fri, 23 Dec 2011 09:34:01 -0500
Jeff Layton <jlayton@xxxxxxxxxx> wrote:
> On Thu, 22 Dec 2011 11:20:11 -0500
> Mike Frysinger <vapier@xxxxxxxxxx> wrote:
>
> > On Thursday 22 December 2011 09:38:55 Jeff Layton wrote:
> > > On Fri, 2 Sep 2011 15:16:55 -0400 Mike Frysinger wrote:
> > > > --- a/aclocal/libnfsidmap.m4
> > > > +++ b/aclocal/libnfsidmap.m4
> > > >
> > > > - AC_CHECK_LIB([nfsidmap], [nfs4_owner_to_uid], [enable_nfsidmap=yes],
> > > > - [enable_nfsidmap=no])
> > > > + AC_CHECK_LIB([nfsidmap], [nfs4_owner_to_uid])
> > >
> > > The above delta in this patch has caused a build-time regression.
> > >
> > > When you eliminated the 3rd argument to AC_CHECK_LIB, this started
> > > adding -lnfsidmap to $LIBS. That means that that library gets linked
> > > into every binary that's being built.
> > >
> > > I haven't determined the right fix for this yet, but I think
> > > libnfsidmap.m4 could use some cleanup.
> >
> > AC_CHECK_LIB([nfsidmap], [nfs4_owner_to_uid], [])
> > -mike
>
> Uhhh, no. That turns the result of this test into a noop. We might as
> well remove it at that point.
>
> Perhaps I should phrase this question differently: What is the above
> test supposed to achieve? IOW, what should happen if it passes or fails?
>
Ahh nm, I think I see now...
We just need to ensure that $ac_cv_lib_nfsidmap_nfs4_owner_to_uid gets
set properly. If that's the case, then we probably want something like
this:
AC_CHECK_LIB([nfsidmap], [nfs4_owner_to_uid], [:])
...autoconf treats bare brackets as if the argument weren't specified.
I have a patch that I'll pass along once I've done a bit of testing
with it.
--
Jeff Layton <jlayton@xxxxxxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
