rpm-5.4.10-22.i686 loops forever when installing ntpd-4.2.6p5-5.i686.rpm
Elan Ruusamäe
glen at pld-linux.org
Mon Oct 22 10:07:41 CEST 2012
On 22.10.2012 04:46, Jeffrey Johnson wrote:
> On Oct 21, 2012, at 5:04 PM, Elan Ruusamäe <glen at pld-linux.org> wrote:
>
>> NO
>> D: Conflicts: ntp < 4.2.0-3 NO
>> ^CD: Exiting on signal(0x2) …
> This report does nothing except document the existence.
that's how it starts. i had no other ideas than Conflicting-Itself
causes loop, and included that output into the report
> There are too may possible causes to hazard a guess, including
> a well documented difference in behavior detected by Mancoosi years
> ago, and also identifying precisely what metadata (and
> what rpm version was used to build all the packages).
to supply this information, i must know what to supply, thus somebody
should first tell what is needed!
rpm that was used to install package, was included in report. how do i
figure out rpm version used to build package? guess? or is it in metadata?
rpm -qp --yaml, did not give match on "5.4"
i've put the exact rpm online:
http://carme.pld-linux.org/~glen/ntpd-4.2.6p5-5.i686.rpm
or just rebuild with known version and hope it to be fixed by latter
patches?
ps: i have no idea who is mancoosi, or what he identified, you should
had provided link to that, not assume everybody knows your friends
> Diagnosis starts with adding "debug" to the Conflictname
> and Packages macro stanzas in /usr/lib/rpm/macro configuration
> in orde to see exactly what database requests are being performed.
there's no exact macro named "Conflictname" or such file, closes i find
was this:
/usr/lib/rpm/macros:%_dbi_config_3_Conflictname %{_dbi_btconfig}
%{?_bt_dupsort}
you mean add there word "debug", like this?
/usr/lib/rpm/macros:%_dbi_config_3_Conflictname %{_dbi_btconfig}
%{?_bt_dupsort} debug
> FYI: As a member @rpm5.org you *are* expected to know
> how to diagnose problems with RPM. *shrug*
and you as a maintainer of @rpm5.org should know, i haven't been active
there for some years!
--
glen
More information about the pld-devel-en
mailing list