> > And there's base.repos.doSetup(), so I don't know why it touches
> > every repo.repoXML.  The following works in HEAD:

> doSetup() just sets up the callbacks, it doesn't touch repo.repoXML
> which needs to be touched to trigger the repomd.xml/etc. downloads

Right, I was confused because doSetup() calls prereposetup, and 
the langpack plugin touches repoXML.. but can't rely on that.

> > Also 'makecache' sets mdpolicy to "group:all", so there seems
> > to be no need to explicitly load metadata/filelists/otherdata.
> 
>  Yeh, there is ... read the comment ...
> > -            # These convert the downloaded data into usable data,
> > -            # we can't remove them until *LoadRepo() can do:
> > -            # 1. Download a .sqlite.bz2 and convert to .sqlite
> > -            # 2. Download a .xml.gz and convert to .xml.gz.sqlite

I still think it's not necessary, as _getRepoXML() 
calls _commonRetrieveDataMD() and there's unconditional
_groupCheckDataMDValid() which does the decompression.
_______________________________________________
Yum-devel mailing list
[email protected]
http://lists.baseurl.org/mailman/listinfo/yum-devel

Reply via email to