Re: Patch for alicia: better handling source + binary overrides
On Wed, Dec 21, 2005 at 10:27:06AM +0000, Colin Watson wrote:
> On Tue, Dec 20, 2005 at 05:13:05PM -0800, Jeroen van Wolffelaar wrote:
> > Please find below (or in spohr:~jeroen/katie/alicia{,.diff,.changelog})
> > a reasonable smallish patch for alicia, to get her to work properly with
> > binary vs source overrides, instead of bailing out if an override exists
> > for both (could happen in the past, happens much more now that cindy
> > will add missing source overrides). Without this patch, override
> > handling is harder to do, natalie isn't very convenient, in usage nor in
> > logging.
>
> I wrote a somewhat different patch for Ubuntu's dak installation a while
> back, which lets you use alicia to work on udeb overrides too. (I don't
> know if it'll apply cleanly to current CVS.) With this you use '-t dsc'
> to work on source packages.
I specifically considered udeb's, and also with my patch it just works,
on the original assumption, currently in alicia, that the same package
name will not occur in multiple types, something which should surely
should hold for udeb & deb -- but indeed much less so for dsc vs
deb/udeb. So I choose for the easy common case, also because deb & udeb
both are binary packages. But yeah, technically, with -t would be more
analog to the database & the real layout of the overrides, and
corresponding better to natalie's behaviour. But madison otoh uses more
like the -S vs not- -S, so it isn't consistent either way.
My patch does have provisions against changing the priority of source
packages though (which is not allowed).
Anyway, I don't mind which of the two patches get applied, either would
do.
--Jeroen
--
Jeroen van Wolffelaar
Jeroen@wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)
http://Jeroen.A-Eskwadraat.nl
Reply to: