On Fri, 24 Aug 2007, Kevin B. McCarty wrote:
On 8/24/07, Frederic Lehobey <[EMAIL PROTECTED]> wrote:
I have gathered this information (and more) into a temptative
debian-science source package building several science-* packages
(with cdd-dev machinery).
Great!
I am unsure if it is ready to be uploaded to unstable. But I would be
happy to hear your feedback and suggestions on dependencies to have
(or not) and take them into account.
http://lehobey.net/debian-science
I'll have a look at this in the end of next week. I would like to
ask you not to upload to unstable for one reason: I'm in the middle
of heavy changes for cdd-dev and thus I would like to upload (and
surely test before an upload) after the release of cdd-dev 0.4.
I think I will need about one week for finishing and than I will
test your work on science.
1) Should these metapackages Recommend instead of Depending on the
real ones? I see that is what the education-* metapackages do and
presumably they have some reason for it.
This will be done automagically by the new cdd-dev package. The
reason is that Debian-Edu decided to use only Recommends to be able
to remove the meta package (they think it might be needed in some
cases). It is no real harm since apt-get now by default installs
all recommends and thus all Depends in the tasks are turned to
recommends.
2) Maybe these metapackages could Suggest the corresponding
education-* metapackages.
I see no real advantage in doing so but it also does no harm.
3) science-physics should definitely recommend root-system once that's
out of experimental.
It can be added to the lÃist of Recommends also now. The cdd-dev build
system checks for available packages anyway and it is tunred to a
suggests if it is not found in the apt-cache list. Once it is available
the next package build will recognise this.
Thanks for working on this
Andreas.
--
http://fam-tille.de