wrt bug #425092 --- Comment #3 from Paweł Hajdan, Jr. <phajdan.jr@gentoo.org> --- Bumped, thanks. I only switched to EAPI 4, I don't see the reason for other updates in your patch. Please file a separate bug and explain more if needed. Further explanations: -DESCRIPTION="Mails anomalies in the system logfiles to the administrator." +DESCRIPTION="Mails anomalies in the system logfiles to the administrator" > It is not common to have a period in the description. -DEPEND="" RDEPEND="!app-admin/logsentry app-misc/lockfile-progs dev-lang/perl dev-perl/mime-construct + virtual/mailx" - virtual/mailx - ${DEPEND}" > DEPEND is a variable, that is optional. So it can be dropped. + doexe "${FILESDIR}"/${PN}.cron + + fowners -R logcheck:logcheck /etc/logcheck /var/{lib,lock}/logcheck - doexe "${FILESDIR}/${PN}.cron" } pkg_postinst() { - chown -R logcheck:logcheck /etc/logcheck /var/{lib,lock}/logcheck || die > If you could do the same in ${D}, is there a reason to change the users etc-dir unasked? If it is related to a previous logsentry install that uses the same dir, then an elog would be better imho because there could be file collisions, too. + elog "Please read the guide at http://www.gentoo.org/doc/en/logcheck.xml" - elog "Please read the guide ad http://www.gentoo.org/doc/en/logcheck.xml" > Typo fix.
(In reply to comment #0) > > DEPEND is a variable, that is optional. So it can be dropped. Not really. Do you have something more to support that point? "It works now" does not count. > + doexe "${FILESDIR}"/${PN}.cron > + > + fowners -R logcheck:logcheck /etc/logcheck /var/{lib,lock}/logcheck > - doexe "${FILESDIR}/${PN}.cron" > } > > pkg_postinst() { > - chown -R logcheck:logcheck /etc/logcheck /var/{lib,lock}/logcheck || > die > > > If you could do the same in ${D}, is there a reason to change the users etc-dir unasked? If it is related to a previous logsentry install that uses the same dir, then an elog would be better imho because there could be file collisions, too. Collisions in /etc are handled differently. I don't remember now, but unless there is a serious, real problem with the current code, I'd not like to change it.
(In reply to comment #1) > (In reply to comment #0) > > > DEPEND is a variable, that is optional. So it can be dropped. > > Not really. Do you have something more to support that point? "It works now" > does not count. > http://dev.gentoo.org/~ulm/pms/head/pms.html#x1-660007.2
(In reply to comment #2) > (In reply to comment #1) > > (In reply to comment #0) > > > > DEPEND is a variable, that is optional. So it can be dropped. > > > > Not really. Do you have something more to support that point? "It works now" > > does not count. > > > > http://dev.gentoo.org/~ulm/pms/head/pms.html#x1-660007.2 Thank you. I didn't know that. Still, the ebuild may define the variable, and I prefer to define it, even if it's empty.