Re: Some questions about dependencies
On Fri, May 02, 2003 at 09:27:33AM +0100, Colin Watson wrote:
> On Fri, May 02, 2003 at 09:52:05AM +0200, Sven Luther wrote:
> > On Thu, May 01, 2003 at 08:31:09PM +0100, Colin Watson wrote:
> > > Uh? The 110 packages Anthony mentions are the packages *which have just
> > > been accepted into testing*, not ones that are blocking anything. You
> > > can find them all by grepping for "accepted:".
> > Ok, sorry, i did misunderstand. Still, when i was investigating the
> > ocaml situation, update_output told me there where 35 or so package that
> > caused problems but around 30 or so only where listed :
> > 2003/03/12 :
> > skipped: ocaml (1221+11)
> > got: 35+0: a-35
> > * alpha: libcamlimages-ocaml, libcamlimages-ocaml-dev,
> > * libconfigwin-ocaml-dev, libgdome2-ocaml, libgdome2-ocaml-dev,
> > * libgdome2-xslt-ocaml, libgdome2-xslt-ocaml-dev,
> > * liblablgl-ocaml,
> > * liblablgl-ocaml-dev, liblablgtk-ocaml, liblablgtk-ocaml-dev,
> > * liblablgtkmathview-ocaml, liblablgtkmathview-ocaml-dev,
> > * libmlgtk-ocaml, libmlgtk-ocaml-dev, libnetclient-ocaml-dev,
> > * libocamlnet-ocaml-dev, libpcre-ocaml, libpcre-ocaml-dev,
> > * libpgsql-ocaml-dev, libpxp-ocaml-dev, libsdl-ocaml,
> > * libsdl-ocaml-dev, libshell-ocaml, libshell-ocaml-dev,
> > * libxstr-ocaml-dev, libzip-ocaml, libzip-ocaml-dev, ocamltk,
> > * zoggy
> > This is 5 missing packages, at first i thought these where the binary
> > package of the ocaml source package, but there are only 4 of them.
> "a-35" means that after experimentally adding ocaml to testing there are
> 35 uninstallable packages on alpha. However, testing isn't in a perfect
> state, so there are usually some uninstallable packages to begin with.
> On that day, it seems that testing started out with 5 uninstallable
> packages on alpha, hence the difference.
Could you (or someone) add this to the testing faq, it had me (and
probably others) wondering about it for days.
> You can find out what this original state is by looking for "ori:" lines
> in update_output.txt.
Ok. We still need a proper recursive tool that lists the true reason for
a package being hold back for the PTS though. Having to hand check the
dependent packages is rather time consuming.