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

Re: ocaml 3.04 in testing?



On Tue, Feb 19, 2002 at 06:25:16PM +0100, Judicaël Courant wrote:
> le jeu 07-02-2002 à 12:33, Sven a écrit :
> 
> > > I also looked at update_excuses.html for ocaml, but here it's reported
> > > as valid candidate so I can't understand why it currently isn't in
> > > testing.
> > 
> > I also don't understand, but anyway, we will have to wait at least 2 more
> > weeks for ocaml 3.04-6 going into testing, after it has been rebuilt for all
> > arches. If there is no major problem with it, i will wait until 3.04-6 enters
> > testing for future changes.
> > 
> 
> >From http://ftp-master.debian.org/testing/update_output.txt :
> --------------------------
> 
> skipped: ocaml (0) (120+0)
>     got: 18+0: a-18
>     * alpha: camlp4, camlzip, lablgl, mlgtk, ocamltk
> --------------------------
> 
> So, as far as I understand, installing ocaml would break the binary
> packages camlp4, camlzip, lablgl, mlgtk and ocamltk. Indeed, they depend
> on ocaml <= 3.03... and no "testing" version exists for them.

camlp4 is to be removed anyway, since ocaml now provides it and conflicts with
lder versions.

camlzip is the work of Stefano, and i think he will look at it.

I will personnally look at my packages lablgl, mlgtk and ocamltk (between
my other packages).

So there should be no major problem, especially once ocaml 3.04-9 with the
propper powerpc fix hits unstable.

Anyway, the new ocaml package did enter unstable at least.

> >From http://people.debian.org/~jules/testingfaq.html:
> -----------------------------------
>    5. The operation of installing the package into testing must not
> break any packages currently in testing. (See below for more
> information)
> 
> [...]
> 
> * How could installing a package into testing possibly break other
> packages?
> 

Erm, ...

As i understood it, once there is a version of all these still conflicting
packages in unstable that are ready to enter testing, the whole lot of them
(the ocaml pool as you would say) will enter testing in one whole bunch.

This was the whole point in the testing stuff, i only hope it will work as
advertized, if not we should probably investigate and/or fill bugs.

Anyway, i still think that the dependencies are the right thing to do, as the
behavior we are facing now is the right one, as it will prevent ocaml to go
into testing together with incompatible libraries. It would be preferable if
those libraries were removed altogether, but this is not possible with the
current testing suite.

(Anyway, the problem with ocaml right now, is that it is broken on powerpc,
and was otherwise uninstallable, it will be fixed in the -9 version though.)

Thanks for the report though, it clarified a lot of things for me.

Friendly,

Sven Luther



Reply to: