On 01/05/2008 10:38 PM, Kevin Kofler wrote:
Manuel Wolfshant <wolfy <at> nobugconsulting.ro> writes:
I have also included a patch which removes from the sage install
script the parts which I think that are already available in rawhide.
Still, a monolithic solution like this is not going to fly, every single of
these separate projects needs to be packaged separately,
I absolutely agree with you. I have mentioned that I agree with
having all prereqs as separate packages. But I wanted to see things
started and to isolate the BRs
usually directly from
upstream, not from the bundled often outdated version in SAGE.
agree again
Unfortunately,
SAGE tries to be yet another program which tries to be a distro, this sucks.
they just want top be sure that what they ship, works. I can understand
that. That does not mean that I agree.
Out of spkg/standard/*.spkg, as a first guess, only
spkg/standard/*-2.9.1.1.spkg makes sense to package as part of SAGE, that would
be doc, examples, extcode, sage and sage_scripts. And these should be 5
separate packages or subpackages (e.g. sagemath, sagemath-doc,
sagemath-examples, sagemath-extcode, sagemath-scripts).
Moreover, at least in your build.log, it appears also to still build stuff like
ATLAS
I have not noticed it in the repo...
and gnutls which are already in Fedora (and the gnutls build failed).
no, it's python-gnutls that failed
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
