On Tue, Jul 17, 2007 at 11:35:04PM +0200, Ralf S. Engelschall wrote:
> On Tue, Jul 17, 2007, Ralf S. Engelschall wrote:
> > On Tue, Jul 17, 2007, Ralf S. Engelschall wrote:
> > > On Tue, Jul 17, 2007, Jeff Johnson wrote:
> > > [...]
> > > So what? Fix the temporary file handling a better way or at least get
> > > rid of the bogus code or really keep it to further test the hacker index
> > > of developers who hack on RPM?
> > > [...]
> >
> > Oh, sorry, I see. I've compared too much with the code in lib/misc.c
> > but overlooked that there is an "#ifndef NOTYET" (note the "n" and not
> > a "#ifdef NOTYET". Ok, forget my previous reply. I missed the fact that
> > mktemp(3) itself is still needed, of course. Well, then then why arent't
> > why at least use the close(mkstemp()) hack consistenty throughout the
> > RPM code? In build.c there is this workaround, in lib/rpmchecksig.c it
> > isn't, etc. Any reasons for this?
> 
> So, what about this to make the stuff at least consistent and at the
> same time finally get rid of the mktemp(3) linker warnings?

No, please do not apply this change.
The problem is real, so let linker warnings remind us about this issue.


-- 
ldv