[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: debian-cd BoF at DebConf


directing this mail to -dpkg@, which hopefully makes sense.

Steve McIntyre <steve@einval.com> (21/07/2012):
> On Wed, Jul 18, 2012 at 09:01:21PM +0200, Guillem Jover wrote:
> >Hi!
> >
> >On Wed, 2012-07-18 at 17:19:40 +0100, Steve McIntyre wrote:
> >> CD sizing problems
> >> ==================
> >
> >> There has been much discussion about switching packages over to using
> >> xz compression instead of gzip by default, including Hideki Yamane's
> >> excellent session "Let's shrink Debian package archive!" [3]. Ansgar
> >> has been looking into the possibilities here of re-building a subset
> >> of the core packages using xz, and I think it's clear that this is the
> >> solution for Wheezy at least. In discussion after Hideki's xz talk, I
> >> think there was broad agreement that we should just switch to xz by
> >> default, *but* with the option to use a different (or even null)
> >> compressor where it makes sense (e.g. in packages full of
> >> already-compressed files such as open-clipart). There has been a
> >> suggestion that we should leave base packages using gzip for the sake
> >> of foreign users of debootstrap, but I firmly believe we should just
> >> tell them they'll need xz in future. Let's not hold ourself back
> >> here...
> >
> >I cooked a patch for dpkg several weeks ago, when this got discussed
> >in debian-devel, to add configurable default compressor at dpkg build
> >time and to switch it to xz for Debian, including updated documentation,
> >etc. So if there's agreement, and the release team would accept this
> >change, then I can quickly prepare such dpkg upload.
> Let's ask the RT then...

dpkg's current diff between testing and unstable, once *.po and *.gmo
stripped is:
 323 files changed, 7307 insertions(+), 4626 deletions(-)

There's #681332 about that, which was left unanswered.

Also, I didn't see a diff for that compressor thingy.


Attachment: signature.asc
Description: Digital signature

Reply to: