[Yum-devel] [RFC] dbversion 10

[Jeremy Katz]

On Tue, 2007-04-10 at 16:38 -0400, seth vidal wrote:
> On Tue, 2007-04-10 at 16:31 -0400, James Bowes wrote:
> > seth vidal wrote:
> > >> If we keep epoch a string (as it's been in the past), then that's not a
> > >> problem though, no?
> > > 
> > > agreed.
> > > 
> > > 
> > >> The conversion of the other things seems pretty straight forward as I
> > >> think that their usage is far more constrained
> > > 
> > > also agreed.
> > > 
> > > I just wanted to make sure I made mention of the epoch bit, again, if we
> > > were thinking about it for pre-3.2
> > 
> > So to summarize, are you two are thinking that we apply the patch as
> > emailed, minus anything relating to epoch?
> > 
> > Meaning that API breaks for anyone who treats build_time as a string
> > (whoever that might be), but does not break for epoch related code. Also
> > meaning that epoch is still stored as a string in the db.
> 
> That sounds about right to me, yes.
> 
> Jeremy?

Exactamundo.

Jeremy