Re: math package maintainer for Debian
Firstly, I am not a Debian developer but have been using debian for some
10 years and have a very rudimentary knowledge of debian package
creation. I am very interested in seeing a SAGE package for debian and
recently made some effort to "debianize" the upstream source.
I learnt quite a lot in doing so and would be very happy to share those
experiences with anyone interested in putting together a debian sage
Very briefly, the key issues I ran into and modifications I had to make
1) Upstream brings in local copies of everything from python, various
standard libraries (e.g. freetype, readline) and basic utilities (such
as bzip2). It is relatively straight forward to simply remove all of
those upstream packages which already exist in debian - firstly, delete
the spkg files (bzip2'd tar archives) and secondly remove the build
calls from the main Makefile.
Note that removing these duplications cuts the size of the final SAGE
deb file from many hundreds of megabytes to about 25Mb.
2) Two of the spkg sources need to be unpacked and patched to force
installation into the appropriate local build debian directory - to
streamline this upstream could add an argument in a few build scripts.
(This was the messy bit where I quickly was out of my depth in terms of
debian package development).
Anyway, I successfully built (and used for quite some time) a debianized
SAGE 126.96.36.199. I recently attempted to build 188.8.131.52 but ran into some
compile errors associated with libpari.
As I said, I'd be most happy to share my (rather patchy) knowledge with
Dr James McCaw
School of Physics
University of Melbourne, VIC 3010, AUSTRALIA
Phone: +61 3 8344 9145
On Sat, 2006-10-14 at 12:18 -0500, Dirk Eddelbuettel wrote:
> On 14 October 2006 at 13:00, David Joyner wrote:
> | Hello:
> | I am part of a development team for a
> | GPL'd math software package called SAGE. We could like
> Which is at http://sage.scipy.org/sage but you didn't say that :)
> | to ask a question of a contact person from the Debian team who
> | deals with math packages. If such a person is reading this list
> | (or if someone on this list knows of such a person), please
> | reply to this email or email William Stein (email@example.com,
> | the SAGE lead developer) or Martin Albrecht
> | (firstname.lastname@example.org, another SAGE developer).
> Is there a reason you can't ask directly here on the list where more eyeballs
> will see it? Given how SAGE comprises deveral dozen packages so you'd have to
> contact several maintainers anyway, several of which are likely to be
> listening here.
> I think one you SAGE guys had contacted me before in regards to Quantian
> (www.quantian.org aka dirk.eddelbuettel.com/quantian.html) but I can't recall
> what became of that...
> Hell, there are no rules here - we're trying to accomplish something.
> -- Thomas A. Edison