On Tuesday, May 26 2009, Seth Vidal said: > On Fri, 22 May 2009, James Antill wrote: >> Some thoughts about group DB (having groups as real objects). I mainly >> wrote this for me, to see if "group DB" was actually possible. I think >> so :). [snip] >> ...a couple of things that fall out of this: >> >> i. You can have all the packages for a group installed, but the group >> not installed.
This has in the past been a contention from a UI perspective >> ii. You can have a group installed, but no packages from that group >> installed. And this definitely feels weird. What does having the group "installed" mean then? Doing this via some yum-only-database still feels like we're going to get into trouble for all the reasons we had talked about back in January > and all of the above means we're doing it via a yum "groupdb" of some > kind. I'm not necessarily opposed to it but if: > > 1. we could construct rpm hdr objects in memory from python > 2. we could insert these objects into the rpmdb in a transaction > > would that be enough to step around the weird extra db? > > specifically it would allow us to make meta-pkg "Group" packages on the > fly and have the rpmdb keep things consistent - including something like > a "version". There's still some amount of fiddles and bits perhaps required due to repo enabling/disabling. But it feels a little bit better if we can get the information into the rpmdb rather than more and more dbs in more and more places... Jeremy _______________________________________________ Yum-devel mailing list Yum-devel@lists.baseurl.org http://lists.baseurl.org/mailman/listinfo/yum-devel
