On Sat, Jul 18, 2009 at 4:27 PM, Franco Saliola<sali...@gmail.com> wrote: > > On Wed, Jul 15, 2009 at 8:39 PM, William Stein<wst...@gmail.com> wrote: >> >> On Tue, Jul 14, 2009 at 5:45 PM, Tim Lahey<tim.la...@gmail.com> wrote: >>> >>> It looks like the unladen-swallow Python branch >>> has been making good progress. They now can >>> pass their test suite of third party packages >>> (including NumPy and SymPy). >>> >>> http://code.google.com/p/unladen-swallow/wiki/Release2009Q2 >>> >>> For those who don't know, it's an attempt to compile >>> Python down to LLVM to improve the Python speed. >>> >>> Cheers, >> >> Can somebody make an spkg of this? Could one build Sage on top of it >> (instead of Python-2.6.2?). >> >> I tried typing >> >> svn checkout >> http://unladen-swallow.googlecode.com/svn/branches/release-2009Q2-maint >> unladen-2009Q2 >> >> and got >> >> teragon-2:unladen-2009Q2 wstein$ ls >> Demo Include Mac Modules PCbuild >> README Util install-sh >> Doc LICENSE Makefile.pre.in Objects Parser >> RISCOS configure pyconfig.h.in >> Grammar Lib Misc PC Python >> Tools configure.in setup.py >> >> teragon-2:... wstein$ more README >> This is Python version 2.6.1 >> ============================ >> ... >> >> So maybe it works just like Python at this point? >> >> (To make an spkg of this, just take the python-2.6.2.spkg from >> sage-4.1.tar, and replace src by the above unladen-2009Q2 directory, >> then try ty build all of Sage from scratch by typing "make".) > > I couldn't resist testing this out. Here is a report. > > > PACKAGE CREATION: > > As suggested, I > > * extracted the python-2.6.2.spkg from sage-4.1; > * replaced src by the unladen-2009Q2 source; > * repackaged it and named it python-unladen-2009Q2.spkg; > * replaced python-2.6.2.spkg with with python-unladen-2009Q2.spkg > * compiled > > After having a few build errors with this spkg, I think I tracked the > problem down to building unladen in parallel. I editted the spkg-install > script to avoid this. You can find the spkg here: > > > http://sage.math.washington.edu/home/saliola/sage-4.1+unladen-2009Q2/python-unladen-2009Q2.spkg > > > BUILD RESULTS: > > * On sage.math: success! > * On my Ubuntu-8.10 64-bit Macbook: success! > * On a 64-bit Gentoo Linux machine: success! > > > DOCTEST RESULTS: > > On all three systems mentioned above, 'make test' and 'make testlong' > returned doctest failures in the files: > > devel/sage/sage/structure/sage_object.pyx > devel/sage/sage/misc/explain_pickle.py > > All these errors are related to pickling. For sage_object.pyx, > len(dumps(2/3)) has changed. The output of the 21 doctests in > explain_pickle.py vary slightly from the expected output. > > > You can find the test logs here: > > > http://sage.math.washington.edu/home/saliola/sage-4.1+unladen-2009Q2/sage-4.1/ > > > For those with access to sage.math, you can use the above test it out on a > few computations. > > Take care, > Franco >
Thanks for doing this! I'm really glad this comes so close to working. It will be interesting to compare some timings, and I also posted your spkg to the official "experimental" spkg list (so it appears when people do "sage -experimental"). William --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
