rawhide report: 20050304 changes

Jeff Spaleta jspaleta at gmail.com
Fri Mar 4 21:12:30 UTC 2005


On Fri, 04 Mar 2005 15:25:39 -0500, Eric Warnke <eric at snowmoon.com> wrote:
> Frank Schmitt wrote:
> For people that just have to have the oly functionality they can build
> the database with rpm -i --justdb --dbpath /path/to/db /path/to/all/RPMS/*

Assuming of course they have a full cache of rpms available locally
and not on the network.

Speaking from my own perspective most of what I have used a local copy
of the rpmdb for is dependancy information. But when I'm trying to
debug something or confirm a reported problem I have used the rpmdb
for other queries so that I dont have to pull the rpms from the net...
especially in previous rounds of development/test release cycles.  I'm
pretty sure at one time or another I have used queries for:
--scripts 
--triggers 
--obsoletes
--conflicts
in an effort to identify one particular package problem or another. 

If I can get access to the same information via queries of repodata I
won't be impacted at all. Even if I can't recover queries against
something like scripts.. being able to make queries against the
requires/provides for the full set Core+Extras packages will be far
more useful to most of the troubleshooting that i end up looking 
into.

Though i will say the dirty little trick of running an rpm -e --test
against the full rpmdb to see what depends on a particular package was
a mighty useful dirty trick. It would be very useful to have something
as simple in the repoquery worldview to accomplish the same thing. Hmm
now that I think about it.. i think Seth told me that Panu had some
leafnode code already that can do this.

-jef"doesn't speak for himself.. i speak for my puppet and my
puppeteer speaks for me"spaleta




More information about the fedora-devel-list mailing list