On Thu, 2011-02-24 at 13:49 +0200, Panu Matilainen wrote: > On 02/15/2011 05:16 PM, seth vidal wrote: > > On Tue, 2011-02-15 at 10:08 -0500, James Antill wrote: > >> On Tue, 2011-02-15 at 13:06 +0200, Panu Matilainen wrote: > >>> Avoids another lowlevel rpm routine usage where rpmsack provides > >>> a nicer interface for the thing, and another step towards consolidating > >>> rpm quirk handling to one spot. > >> > >> This makes me twitch a bit, it just seems wrong to have config depend > >> on rpmsack ... I can't see any problems atm. though, so let me think > >> about it a bit more. > >> Anyone else have an opinion? > >> > > > > feels like a place where we'll trip up later by a stray or circular > > import but there's nothing implicitly worrisome about it. I use the > > RpmDBPackageSack object directly for talking to the rpmdb w/o > > instantiating a yumbase object pretty commonly in little random scripts. > > So ... any decision wrt this - is it ok to use rpmdbsack from config or > not? Like said (IIRC), another way to avoid the current circular import > without shuffling anything else around is putting the import inside > _getsysver(), eg > > --- a/yum/config.py > +++ b/yum/config.py > @@ -1000,10 +1000,10 @@ def _getsysver(installroot, distroverpkg): > @param distroverpkg: The value of the distroverpkg option. > @return: The release version as a string (eg. '4' for FC4) > ''' > - ts = rpmUtils.transaction.initReadOnlyTransaction(root=installroot) > - ts.pushVSFlags(~(rpm._RPMVSF_NOSIGNATURES|rpm._RPMVSF_NODIGESTS)) > + from rpmsack import RPMDBPackageSack > + rpmdb = RPMDBPackageSack(root=installroot) > ... >
Let's try it. ACK - put it in rawhide - we'll see what blows up in mock or in the distro. -sv _______________________________________________ Yum-devel mailing list Yum-devel@lists.baseurl.org http://lists.baseurl.org/mailman/listinfo/yum-devel
