On Sun, Jun 22, 2014 at 8:11 AM, Tim Lauridsen <tim.lauridsen(a)gmail.com> wrote:
On Sat, Jun 21, 2014 at 8:28 PM, Jon <jdisnard(a)gmail.com> wrote:
>
> Would it be too much trouble to use the sqlite data in DNF?
> I suppose it would be a step backwards to have our primary (future)
> tool using gzip metadata.
the information in the sqlite files is orded i a way the yum is using the
data, dnf parses the .xml files into another format used by libsolv.
the sqlite was invented because the yum .xml to sqlite format was slow and
has to happen on every client out there, so it was moved serverside to speed
up things.
Well can't we change create repo to create xml.xz files (in addition
to .gz to preserve backwards compatibility) ?