Re: Packages maintained by Debian Science maintainers but missing in tasks files
Hi Frederic-Emmanuel,
On Sun, Dec 15, 2013 at 09:45:36AM +0000, PICCA Frederic-Emmanuel wrote:
> clipper:
> libccp4:
> mmdb:
> ssm:
> -> only in experimental. should we put them anyway ?
Yes, definitely! This is one of the cool features of the Blends
framework: You get the advertising (+QA means) for *all* packages - even
those only in VCS (and so you should add even those in VCS into the
tasks files!!!) but the metapackages will be created on the basis what is
actually contained in testing so the resulting set of dependencies will
create a valid metapackage that can smoothly migrate to testing.
Since you did not mention what task I'll leave this to you and please
pretty please add all your stuff in Vcs as well. This has advantages
now and will save us later such big changes like now since usually
people forget updating the tasks files once a package is really accepted
in Debian.
> guidata: python-guidata
> -> false positiv ?
I'd call this as a right positiv since it is actually not in the tasks
files. What I meant with false positives are packages where at least one
binary package of a source package was mentioned inside the tasks files
but the source was part of my list (even if it should not).
> is a dependency of guiqwt. Should I add it somewhere ? this is a qt
> based toolkit used to build scientific applications
Considering the description of python-guidata
Description-en: dataset manipulation GUI generator
Based on the Qt Python binding module PyQt4, guidata is a Python library
generating graphical user interfaces for easy dataset editing and display.
It also provides helpers and application development tools for PyQt4.
it could be suggested in viewing or I'm even tempted to create a
viewing-dev package since we have several packages which are useful to
develop some applications that are viewing scientific data. What do
you think?
> guiqwt: python-guiqwt
> -> false positiv ?
>
> numericalcomputation:Depends: python-guiqwt
> viewing:Depends: python-guiqwt
Yup, sorry.
> mmtk: python-mmtk
> -> chemistry debian-chem or both ?
> Description-en: molecular modeling toolkit
> The Molecular Modeling Toolkit (MMTK) is a library for molecular
> ...
It's listed in DebiChem tasks. SO I just added (+commited) to chemestry.
> numexpr: python-numexpr, python-numexpr-dbg, python3-numexpr, python3-numexpr-dbg
> -> dependency of pytables
I think the description rectifies something like:
$ git diff
diff --git a/tasks/mathematics-dev b/tasks/mathematics-dev
index 1b1d44f..9594ea9 100644
--- a/tasks/mathematics-dev
+++ b/tasks/mathematics-dev
@@ -16,6 +16,8 @@ Why: non-free
Depends: python-numpy
+Depends: python-numexpr | python3-numexpr
+
Depends: libgsl0-dev
Depends: liblapack-dev, libatlas-base-dev, libblas-dev
> pytables: python-tables, python-tables-dbg, python-tables-doc, python3-tables, python3-tables-dbg
> -> numericalcomputation ?
>
> this package is use to read hdf5 files (python-tables), so it is of
> general interest. should we add it in other task ?
$ git diff
diff --git a/tasks/numericalcomputation b/tasks/numericalcomputation
index a95a97d..9fdf17e 100644
--- a/tasks/numericalcomputation
+++ b/tasks/numericalcomputation
@@ -49,6 +49,8 @@ Depends: python-scitools
Depends: python-deap
+Depends: python-tables | python3-tables
+
Depends: octavede
Homepage: http://octavede.sourceforge.net/
Pkg-URL: http://snapshot.debian.org/package/octavede/0.3-1/
> vitables: vitables
> -> numericalcomputation ?
>
> this package provide a qt GUI on top of pytables to explore the
> containt of hdf5 files.
Added + commited.
> pytango: python-pytango, python3-pytango, python-pytango-doc
> -> false positiv ?
> dataacquisition:Suggests: python-pytango, python-taurus
Yep, false positiv. :-(
> qhull: libqhull5, libqhull-dev, libqhull-doc, libqhull-bin
> -> mathematic-dev ?
> maybe some geometric task but it is not available ?
> Description-en: calculate convex hulls and related structures (development files)
> Qhull computes convex hulls, Delaunay triangulations, halfspace
> intersections about a point, Voronoi diagrams, furthest-site
> Delaunay triangulations, and furthest-site Voronoi diagrams. It
> runs in 2-d, 3-d, 4-d, and higher dimensions.
$ git diff
diff --git a/tasks/mathematics b/tasks/mathematics
index 61d1525..f66344a 100644
--- a/tasks/mathematics
+++ b/tasks/mathematics
@@ -135,6 +135,8 @@ Depends: julia
Suggests: matlab-support
+Depends: qhull-bin
+
X-Removed: Packages removed from Debian
Depends: octaviz
diff --git a/tasks/mathematics-dev b/tasks/mathematics-dev
index 9594ea9..0e08f43 100644
--- a/tasks/mathematics-dev
+++ b/tasks/mathematics-dev
@@ -33,6 +33,8 @@ Why: linear programming
Depends: libcgal-dev
Why: computational geometry
+Depends: libqhull-dev
+
Depends: libmatheval-dev
Depends: python-openopt
In the commit log I asked for volunteers to create a geometry task which
I personally would consider sensible.
> sardana: python-sardana, python-sardana-doc
> -> false positiv ?
> dataacquisition:Depends: python-sardana
Yes, sorry.
> tango: liblog4tango5, liblog4tango5-dbg, liblog4tango5-dev, liblog4tango5-doc, libtango-tools, libtango8, libtango8-dbg, libtango8-dev, libtango8-doc, tango-accesscontrol, tango-accesscontrol-dbg, tango-common, tango-db, tango-db-dbg, tango-starter, tango-starter-dbg, tango-test, tango-test-dbg
> -> false positiv ?
> dataacquisition-dev:Depends: libtango8-dev, liblog4tango5-dev
> dataacquisition:Suggests: tango-db, tango-starter, tango-accesscontrol, python-pytango, python-taurus
Yes, sorry.
> taurus: python-taurus, python-taurus-doc
> -> false positiv ?
> dataacquisition:Suggests: tango-db, tango-starter, tango-accesscontrol, python-pytango, python-taurus
Yes, sorry.
> so it seems that there is a bunch of false positiv. I imagine that
> this is because not all binary packages generated for each source
> package listed are not referenced in the tasks files.
No, the problem was a wrong query I did. I totally rewrote the script
which is now based on the fact that Blends data are now also kept in
UDD. I commited the script to
http://anonscm.debian.org/gitweb/?p=blends/website.git;a=blob;f=misc/sql/0-uncategorised.sh
but your recent suggestions are not yet imported (the cron job runs once
a day). So you can run this script on alioth and see what uncategorised
source packages are remaining (including the resulting binary packages
and the latest uploader of the package). I intend to post this update
to the list tomorrow (or the day after tomorrow depending how many
updates might come today).
Please make sure you commit the packages in the beginning (clipper, etc.)
into the according tasks.
> In my case, I provide also -doc and -dbg packages in a few of my packages.
>
> Maybe we should also provide *-doc ans -dbg taskes files to put these
> binary packages into. What is your opinion about this ?
In the very beginning of metapackage creation we had one med-doc package
which turned out to be not very practical since we tried to collect
every doc there. So we dropped it in Debian Med but if other people
think we should provide some foo-doc, bar-doc package that's fine for
me. I guess if we see such tasks appearing in Git somebody considered
the idea good enough to do the work. ;-)
Thanks you very much for your contribution
Andreas.
--
http://fam-tille.de
Reply to: