On Sat, Feb 28, 2009 at 9:48 AM, Ondrej Certik <ond...@certik.cz> wrote: >
<SNIP> Hi Ondrej, >> If building openmx really benefits from "make -j" you might consider >> putting something clever in spkg-install to detect "available cores" and >> maybe build using them... > > Yes, but I need this for all packages, so I think I'll rather > customize my own derived sage to do this automatically. For Michael: > we'll it's not sage. :) Is this a jedi mind trick? ;) <SNIP> >> This reminds of how Debian is a Linux distribution that dozens >> (hundreds?) of other custom Linux distributions derive from. Sage >> could be the same. That would be cool. > > Ok. I did some progress yesterday: > > http://code.google.com/p/qsnake/wiki/qsnake_all > > Basically you download 3.2MB tarball and then install everything else > using "sage -i". That's what I want! There is not even python in my > qsnake_all. If you type ./sage, it will enter an infinite loop, see my > wiki above for instructions how to fix it. > > However, there is a big problem with dependencies, as you need to > install the packages in correct order. So I will create some meta > packages that will do that for me, e.g. scipy-all-0.1, that would > install all packages needed for scipy in correct order. Or > sage-all-0.1, that would install Sage+packages. Or mayavi2-all for > mayavi. Well, I would not consider the name sage-all-0.1 elegant in this context. I would suggest you call it qsnake-base or something. By the way: You don't need * boehm_gc-7.1.p1.spkg * palp-1.1.p1.spkg * graphs-20070722.spkg You can probably skip * blas-20070724.spkg since it is only there for the LAPACK testers. cvxopt might depend on it, but I will probably fix that since I plan to remove the blas.pskg from Sage unless I find another issue. Since you use an external Fortran just dump all the Linux binaries from the spkg - you won't need them. The fortran.spkg will also go away form Sage completely in the 4.0 time frame. We will have a special OSX quad universal gfortran.spkg that matches with the current XCode releases. So in any case you can strip down * fortran-20071120.p5.spkg substantially. > Secretly I hope that it will spark interest of other people, so that > we can maintain this stripped sage (michael: not sage) together, as I > think it can do the job for all the people who wanted something like > python(x,y) for linux. Well, here you have it. Well, python(x,y) is getting ported to Linux AFAIK. If that will make a large difference or is even practical we will see, but the main issue will plainly be that doing binaries on Windows is a lot easier than on Linux. > When Sage is ported to windows, I'll try to do the same, even though, > I don't have any windows here, so maybe someone else will do it for > me. :) > > Ondrej Cheers, Michael > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
