[scponly] New scponly Snapshot Release

[Frank Fegert]

> > On Saturday, November 20, 2010 08:49:28 am Sven Hoexter wrote:
> Oh wait, I somehow assumed that this problem is part of a spec file shipped
> with scponly but now, actually looking at it, I think I've been wrong with
> that assumption. So IMHO this shouldn't be tackled in scponly but fixed
> somehow in the spec file/rpm build enviroment used. Maybe patch out the
> failling chown call and then somehow explain rpmbuild that this file should be
> installed with root:root ownership. Or avoid the install part altogether and
> manualy include the files in the package. I'm pretty sure that's possible via
> the %files, %attr directives and friends.

Yes of course, permissions and ownership can be fixed within the
spec file with the %defattr/%attr directives and the "-o 0 -g 0"
part can be patched out of the Makefile before installing. What
i was merely trying to ask is if it's really necessary to run the
install command with user/group in the first place? As far as i
can see there are three cases here:
 1) You run the build/install process as root, in which case every-
    thing should be fine even without the "-o 0 -g 0", since the
    newly created files will have uid=0/gid=0.
 2) You run the build/install process as non-root (RPM build env
    or not doesn't matter). The install with "-o 0 -g 0" will fail
    anyhow.
 3) You run build as non-root and install as root. The install with
    "-o 0 -g 0" will be actually useful.

Is case number three really that common? If it's causing too much
of a headache, then don't worry about it. I'll simply patch the
Makefile.in during the RPM build process.

Best regards,

    Frank