Sage 2.8.11:
Release team: Michael Abshoff (chair), William Stein, Carl Witty
While there were many bug fixes, performance improvements and features added
the main goal of this release was to get Sage building out of the box in 32
bit mode on OSX 10.5. We didn't make it 100% to the finish line, but more
about that below. The main participants in the port were William Stein,
Ralf-Philipp Weinmann and Michael Abshoff. If you plan to participate in
Sage Bug Day 5 you should build this release since it will be the base we
build upon.
This release contains patches and updated spkgs from (in ascending order
of ticket number):
* Jaap Spies
* Carl Witty
* Tom Boothby
* William Stein
* Martin Albrecht
* Josh Kantor
* Ondrej Certik
* Michael Abshoff
* Joel Mohler
* Mike Hansen
* Jennifer S. Balakrishnan
* Jason Grout
* Robert Bradshaw
* Ralf-Philipp Weinmann
The following 33 tickets have been closed during the 2.8.11 release cycle:
#217: optimize matrix permanents (Jaap Spies)
#389: bug in mpfi C library (Carl Witty)
#431: dsage jobs get lost by server (Yi Qiang, fixed by an earlier patch)
#750: permutation group element (dict method, acting on lists) (Tom Boothby)
#762: Elliptic curve L-series bug (William Stein)
#830: SAGE does not install 'gphelp' for extended PARI/GP help (Carl Witty)
#845: can't pass boolean as parameter to Magma (Martin Albrecht)
#948: very slow factorization over a numberfield in a 2-variable ring
(Martin Albrecht, Hannes Schönemann)
#959: errors building sage because singular gets confused by system-wide boost
(Martin Albrecht)
#969: cvxopt miscompiled on OSX ppc (Josh Kantor, fixes by Michael Abshoff)
#971: update Sympy.spkg to 0.5.6, add some minimal doctests (Carl Witty,
Ondrej Certik)
#973: Unhandled SIGSEGV: A segmentation fault occured running dance(10
(Michael Abshoff, Carl Witty)
#993: Pari's gp interpreter has built-in library search path (Carl Witty)
#1011: MagmaElement.__nonzero__ (Martin Albrecht)
#1026: memleak in linbox's gmp++_int_io.C (Michael Abshoff)
#1030: MPolynomial_libsingular mutates with call to factor (Joel Mohler)
#1032: Latex'ing variable names is more robust and consistent (Joel Mohler)
#1033: updated optional spkg: biopython 1.44 (Mike Hansen, fixes by William
Stein, Michael Abshoff)
#1034: clean up 'revlex' term ordering mess (Martin Albrecht)
#1037: arithmetic with Schubert polynomials (Mike Hansen)
#1038: bad interaction between numpy and in-place operations (Robert Bradshaw)
#1039: Dokchitser L-series of number field (Jennifer S. Balakrishnan, referred
by William Stein)
#1044: segfault apply morphism to field element (Carl Witty)
#1045: add eulerian testing and circuit-finding (Jason Grout, referred by
Robert Miller)
#1049: graphs: transitive reduction function (Jason Grout, referred by
Robert Miller)
#1053: updated Cython to the 0.9.6.8b release (Robert Bradshaw)
#1056: Fix Givaro 3.2.6 build problem on MacOS X 10.5 (Ralf-Philipp Weinmann,
integrated by Michael Abshoff)
#1059: fix lcalc installation on OSX 10.5 (William Stein, integrated by
Michael Abshoff)
#1051: pari/gp extended help stops working when sage tree is moved (Carl Witty)
#1057: Order elements do not have Z as a (proper) basering (Robert Bradshaw)
#1060: fix flintqs compile on OSX 10.5 (William Stein, integrated by Michael
Abshoff)
#1061: updated python spkg on 10.5 (William Stein)
#1063: "sage -sh" should run $SHELL with Sage environment variables set
(Carl Witty)
Build instruction for OSX 10.5
Install gfortran (e.g., binaries from http://sagemath.org/packages/binary/gfortran/
or http://hpc.sourceforge.net/ and export SAGE_FORTRAN=`which gfortran`.
Then the build will work.
There is one more doctest to fix and I am certain that this will be done
quickly. If you do not install gfortran Sage will use the prebuilt
g95 which will fail to compile scipy on 10.5. Please report any issues you
encounter.
Apologies to anybody I might have forgotten.
Cheers,
Michael