rpmbuild doesn't produce debuginfos

[Ralf Corsepius]

On Fri, 2006-05-19 at 16:34 +0200, Matthias Saou wrote:
> Caolan McNamara wrote :
> 
> > On Fri, 2006-05-19 at 16:23 +0200, Ralf Corsepius wrote:
> > > Hi,
> > > 
> > > I am facing a bizarre issue on rpmbuilding *some* rpms on FC5:
> > > 
> > > rpmbuild doesn't build debuginfos, despite the package contains
> > > binaries[1].
> > > 
> > > Also, on packages exposing these symptoms, executables don't seem get
> > > stripped:
> > > $ rpmlint ~/src/rpms/RPMS/i386/xxx-0.0.2-0.fc5.i386.rpm
> > > ...
> > > W: xxx unstripped-binary-or-object /usr/bin/xyz
> > 
> > The created binaries might not have permissions which allow them to be
> > written by the rpmbuild-ing owner.

> ...or the binary might have had setuid or setgid bits set.
#8()

>  Not sure why,
> but binaries don't get stripped properly in that case.
Not in the cases I trip over this issue. These are normal application,
with nothing special about them. Actually, as it seems, the more trivial
a package is, the more likely it seems to be hit by the problem.

Even a package containing "hello world" and building in mock exposes
this problem ;).

> No idea about the debuginfo packages not being created, though, since in
> my experience, even if there are no binaries in the package, you'll get
> one (although empty) as long as the package isn't explicitly "noarch".
I had wanted to send an example src.rpm of a package exposing the issue
to this list, but the FESCo/RH God reigning this list has proven to
insist on ridiculously unreasonable low limits on postings ... which had
forced me to file this PR to ...
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=192422

Ralf