On 04/12/12 14:21 -0500, Lon Hohberger wrote: > On 11/27/2012 12:00 PM, Jan Pokorný wrote: >> Reported-by: rpmlint <http://rpmlint.zarb.org/cgi-bin/trac.cgi> >> Signed-off-by: Jan Pokorný <[email protected]> >> --- >> man/Makefile.in | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/man/Makefile.in b/man/Makefile.in >> index 1fdde67..21dda33 100644 >> --- a/man/Makefile.in >> +++ b/man/Makefile.in >> @@ -21,11 +21,11 @@ install: all >> if [ ! -d ${DESTDIR}/${mandir}/man8 ]; then \ >> install -d ${DESTDIR}/${mandir}/man8 ; \ >> fi >> - install -m755 ${MAN8} ${DESTDIR}/${mandir}/man8 >> + install -m655 ${MAN8} ${DESTDIR}/${mandir}/man8 >> if [ ! -d ${DESTDIR}/${mandir}/man5 ]; then \ >> install -d ${DESTDIR}/${mandir}/man5 ; \ >> fi >> - install -m755 ${MAN5} ${DESTDIR}/${mandir}/man5 >> + install -m655 ${MAN5} ${DESTDIR}/${mandir}/man5 >> >> >> clean: >> > > Should be 644, not 655.
Agreed; caused by paying too much attention to rpmlint check as opposed to the case itself :) Looks like the order of checks there would deserve to be reversed. But even in its current form, 655 would be spotted. I was to argue that 444 would be even saner (lvm2 and device-mapper do this in RHEL, for example), but for some reason it seems non-standard to mark something root-non-writeable (Debian:[1], Fedora:N/A). [1] http://www.debian.org/doc/debian-policy/ch-files.html#s-permissions-owners -- Jan
