[aur-general] sage-mathematics (was: TU application for Thomas Dziedzic)

[Peter Lewis]

Hi,

Sorry for diverting the thread, but since there was some discussion about this, 
I wanted to ask.

On Wednesday 09 Jun 2010 at 21:15 Thomas Dziedzic wrote:
> > well, sage-mathematics is not a package, it is a distribution (btw, do you
> > know scilab?). Some years ago I asked on this list if it should go to
> > community, and the short discussion ended in the conclusion, that it
> > should not. But I am fine with disussing it again.
> 
> I guess you could call it a distribution which puts a lot of tools
> under a common interface because that's basically what it is.
> Afaik, the programs that come with it are isolated to the package. Too
> bad it wont make it to community then, since I was hoping to spare a
> lot of compilation time for people :P

I installed the sage-mathematics package from the AUR recently, and yes, the 
whole way it's put together seems a mess (though this seemed more to do with 
sage itself than the way it's packaged).

But my question is this - I've had some of the constituent parts which make up 
sage installed for a while, so didn't really want to have another whole version 
of them installed in /opt just so I could use them through the sage interface. 
That seemed like a mess. Is it possible to organise this such that it build of 
the constituent parts, rather than installs two versions of them? (Or is it 
basically an inherent problem caused by the way that sage requires them to be 
distributed?)

Cheers,

Pete.