Hi all, Since I am back to speed I think it is time I posted an update on the state of sage-on-gentoo. I released sage-4.6.1 ebuilds in the hour after the release was announced. That was made possible by the fact that I have started to make ebuilds for alpha and rc releases. 4.6.1 is in very good shape! it offers mpc as an option and the sage ebuild itself is down to 6 patches compared with 12 for 4.6. The main reason for this reduction is that sage has now caught up with sage-on-gentoo in terms of numpy/scipy/matplotlib/sphinx/ecl/maxima. We are using python-2.6.5 or 2.6.6 while sage is at 2.6.4, we also can use a combination of BLAS/LAPACK libraries not just ATLAS. We are still stuck with gmp for technical reasons. There are other differences between a vanilla build and a gentoo build but those are the one that stand out. List of differences on demand.
We may have an increase in the patch number we carry as I had request for getting mpmath to version 0.16 (ticket #9969) and I am interested in including ppl (ticket #10039 and the necessary #10233) as well. I may do that after I do an initial lot of 4.6.2.alpha2 ebuilds tomorrow. The big news however is that during the last month I had an iMac running 10.5 and with the help of Georg S. Weber we got sage-4.6.1 running on a Gentoo prefix on OS X. We have an ugly message at start up but sage -testall -long gives very good results. ie 4 more failures than for a regular sage-on-gentoo build. 2 are related to the tachyon bug and I can reproduce them with a vanilla sage on the same machine, one is an ordering difference and the last one is related to weird error message at start up. Details here: https://github.com/cschwan/sage-on-gentoo/issues#issue/35 It has been very instructive and Georg was very helpful (he really got us building polybori on his own - no small feat). Of all the things I have seen during the month I will only mention the strange things about singular. There are 5 libraries that ship with the singular executable (before you build libsingular): dbmsr.so p_Procs_FieldGeneral.so p_Procs_FieldIndep.so p_Procs_FieldQ.so p_Procs_FieldZp.so On OS X you are supposed to have .dylib or .bundle not .so. In any case we were always wondering about these libraries as their structure is quite strange. I discovered that they were missing from my vanilla install of sage on OS X. spkg_install wants to copy .dylib libraries but we have .so. I doesn't have any adverse effects on sage's tests making me think that they are in fact useless and we shouldn't bother with them at all. Next in line: I have a nokia N900 so I could look in sage on arm and contribute to the corresponding bug. I have access to a power5 cluster running AIX so I could help David on getting things running on AIX, there is also a ppc-aix gentoo-prefix... I also have access to a BlueGene/L but I am not sure I'll try to get sage to run on that just yet. Cheers, Francois -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
