Re: Why are new package versions depending on libc6 in unstable?
On Wed, Nov 20, 2002 at 09:26:58PM +0100, Andreas Metzler wrote:
> Mike Fedyk <mfedyk@matchmail.com> wrote:
> > On Wed, Nov 20, 2002 at 09:36:27AM +0100, Andreas Metzler wrote:
> >> We want the _libraries_ to be tested, *before* they move to testing.
>
> > The libraries DO get tested, by everyone running the unstable binary
> > version. That does NOT mean that you have to compile against it. You know
> > that it compiles by the fact that the library package itself built, and you
> > know if the library works by running programs with that binary package
> > installed.
>
> This argument only holds for forward compatible changes, and I have
> another one:
>
> We want the development-package of the libraries to be tested, *before*
> they move to testing.
>
> I want to know whether compiling xxx against libblah triggers bugs,
> using libblah-dev from sid on the autobuilders at least gives me a
> chance to do that.
Good point.
Then what if we had a experemental->unstable->testing->stable four level
distribution setup for transitions that we already *know* will be painful.
Libc, perl, samba, whatever.
It is quite painful when unstable gets stuck like it has with libc6. Also,
the experemental dist shouldn't automatically move anything to unstable. It
should be more like a staging area for beta/large version changes like
gnome, kde, samba or anything with a large userbase that would want to work
on the beta version and still have the ease of packaging.
Reply to: