Hi Anton,

the dependencies are (probably incompletely) modelled in 
$SAGE_ROOT/spkg/standard/deps. If you receive a failure for a package you 
believe that is not essential you can simply 'touch' the file 

$SAGE_ROOT/spkg/installed/package_name_without_spkg_ending

However, many packages are actually essential because lots of Sage's 
functionality is not written in Python but only exposed using Python. Take 
Singular for example (I work with that one the most). It itself requires gmp 
and ntl. Singular is also built as a library and a stand alone application. 
This libSingular is linked in from SAGE to provide multivariate polynomials 
(also the stand alone application is used to provide some highlevel 
functionality) The multivariate polynomials are used all over (I might 
exaggerate here) the Sage Python library because they are quite essential 
building blocks. libSingular is also a component of Sage we never got to 
compile under Cygwin. If you want to take a shot at that I am very confident 
every Sage developer who ever tried this will very happily provide you with 
as much information on this as he/she has (e.g., ask mabshoff or me on IRC) 
Also I am sure the Singular team will be very supportive.

The problem is that Cygwin is not a perfect bug-for-bug POSIX replacement for 
your standard Linux desktop and many little and big tweaks are necessary to 
get each component of Sage to compile. Also, Cygwin is so damn slow.

> Therefore I ask you for help. Tell me which packages are
> really necessary for building minimal working version of sage. Without
> notebooks, just command prompt. I tried to build 'sage-2.8.9.spkg'.
> Ok, it builds but then? 

sage-2.8.9.spkg builds? I doubt that (i.e. this might be a bug if it claims it 
built, I am a overlooking something very obvious here) because that would 
mean e.g. libsingular built.

> To Jaap:
>
> I decided forking python is not a good idea :).
>
> > SAGE has it all, and is certainly not a download of 2
> > GB!http://www.sagemath.org/SAGEbin/vmware/says 614 MB
>
> Yes, but when you install it you get some more GBs.
>
> Sage supports pari a little. But not fully. Take this line from the
> documentation:
> gp("a = intnum(x=0,6,sin(x))")
> or this
> x = gp("znlog( Mod(2,7), Mod(3,7))")
> or this
> gp('[1,2;3,4]')
> You see, you have to pass strings to the gp interpreter. This means
> you cannot fully work with pari using python language.

Yes, it is not complete yet but it is very easy to add the functions 
NotImplemented yet. See:

http://www.sagemath.org/doc/html/prog/node66.html

Note that we have two interfaces to pari: one is via the pseudy-tty interface 
and gp and one is using the C interface directly. So far no-one sat down to 
wrap every single pari function yet, though this would be awesome. 

What are you using to link Python and Pari's C?

Martin

PS: Just to make sure, I didn't come across wrong. I (and probably every other 
Sage developer) am very happy to help out and receive criticism about Sage. 
The crucial point is to have specific information one can relate to. So feel 
very welcome to ask questions about the build process, criticize it if you 
think something could/should be done better and (even better) get involved 
and fix stuff ;-)

-- 
name: Martin Albrecht
_pgp: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x8EF0DC99
_www: http://www.informatik.uni-bremen.de/~malb
_jab: [EMAIL PROTECTED]


--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-forum
URLs: http://sage.math.washington.edu/sage/ and http://sage.scipy.org/sage/
-~----------~----~----~----~------~----~------~--~---

Reply via email to