= numpy =
== Description ==
This package adds numerical linear algebra and other numerical computing
capabilities to python.
== Maintainers ==
* Josh Kantor
== Upstream Contact ==
* Travis Oliphant
* Fernando Perez
* Brian Granger
== Dependencies ==
* GNU patch
* Python
* Lapack
* Blas
* Atlas
* Fortran
== Special Update/Build Instructions ==
* Scipy uses numpy's distutils to control its compilation of fortran code.
Whenever numpy is updated it is necessary to make sure that scipy still builds ok.
== Changelog ==
=== numpy-1.7.0 (François Bissey 19 Mar 2013) ==
* #11334 trivial bump to numpy 1.7.0. Sources are identical to 1.7.0.rc2.
=== numpy-1.7.0.rc2 (François Bissey 14 Jan 2013) ==
* #11334 upgrade to numpy-1.7.0rc2.
* remove darwin_x86_64.patch as it is included upstream
* drop sage_fortran but be careful to set -fpic in Fortran flags like in scipy
=== numpy-1.5.1.p1 (Jeroen Demeyer, 10 February 2012) ==
* #12423: add patch darwin_x86_64.patch from upstream such that the
"x86_64" platform gets recognized as Intel (and therefore, hopefully
the correct CFLAGS are used).
* Use patch for patching.
=== numpy-1.5.1 (Francois Bissey, 18 Feb 2011) ==
* Update to numpy-1.5.1
* Upstream tickets #1399, #1605 and #1625 have been fixed in this version,
so the related patches are no longer needed.
* remove unneeded patches and useless bin folder.
=== numpy-1.5.0 (Karl-Dieter Crisman, 13 October 2010) ==
* Fix a bug in architecture detection on Mac OS X where too many
architectures were detected on 10.6 for fat binaries. See
Numpy ticket #1399 (http://projects.scipy.org/numpy/ticket/1399).
We fix it by patching src/numpy/distutils/fcompiler/gnu.py to
remove the ppc64 option in that specific instance. When
numpy 1.5.1 is released, check to see if this patch is still
necessary.
=== numpy-1.5.0 (John Palmieri, 5 October 2010) ==
* Fix a bug when building on Solaris: numpy was not built with c99,
but used isfinite, which is only defined in c99 on Solaris. See
Sage trac ticket #9808 and numpy trac ticket #1625
(http://projects.scipy.org/numpy/ticket/1625). To fix, apply
the patch from numpy trac ticket #1625, patching
src/numpy/core/setup.py and src/numpy/distutils/command/config.py.
When numpy 1.5.1 is released, check to see if these patches are
still necessary.
* In spkg-install, export F77, F90, and F95 to make sure that
packages which use numpy (like scipy and scipy_sandbox) use the
correct Fortran.
=== numpy-1.5.0 (Stefan Reiterer, 4 September 2010) ==
* Update to numpy-1.5.0
* numpy/core/src/multiarray/buffer.c has to be patched
(see on numpy trac ticket #1605)
=== numpy-1.4.1 (Stefan Reiterer, 4 September 2010) ==
* Update to numpy-1.4.1
* Remove the file gcc_fake from the Numpy package, and instead
set the variable CC to consist of both the compiler and the flag
needed for 64-bit builds on platforms where 64-bits is not the default.
So typically, CC will be set to "gcc -m64", though the solution is more
portable
* The function numpy.core.defmatrix moved to numpy.matrixlib.defmatrix.
This has to be changed in SAGE_HOME/local/lib/python2.6/networks/convert.py
in the package networkx-1.0.1. For newer versions of networkx this is obsolete!!
* After upgrading from numpy.1.3.x to 1.4.1 runtime warnings will appear.
Actually they don't do any harm. If one wants to get rid of them
one has to run sage -ba [branch].
* 2 doctests in sage had to be changed due some output changes.
* cygwin-core-setup.py had to be removed from the patches.
=== numpy-1.3.0.p3 (Mike Hansen, 27 October 2009) ===
* Make sage_fortran the default compiler for cygwin. See #7321.
=== numpy-1.3.0.p2 (Ondrej Certik, 20 September 2009) ===
* fix the build system to work in cygwin
=== numpy-1.3.0.p0 (William Stein, 8 July 2009) ===
* Comment a line out in gnu.py to get distutils to work when building scipy on 64-bit OS X.
See trac http://trac.sagemath.org/sage_trac/ticket/6493 and http://projects.scipy.org/numpy/ticket/1087
=== numpy-1.3.0 (Jason Grout, 27 May 2009) ===
* Update to numpy 1.3.0. Since an spkg for scipy 0.7 is now available, I deleted the deprecation warning fixes mentioned below.
=== numpy-1.2.0.p0 (Michael Abshoff, September 27th, 2008) ===
* quiet a couple deprecation warnings in arrayobject.c and numpytest.py. These changes need to be reverted once we update to scipy 0.7. We also need to fix some API issues in our code once those fixes are reverted - see #4205.
=== numpy-1.2.0 (Jason Grout, September 25th, 2008) ===
* update to numpy 1.2.0, rebase and update patches
=== numpy-1.1.0.p0 (Michael Abshoff, August 27th, 2008) ===
* clean up patches directory
=== numpy-1.1.0 (Jason Grout, July 10th, 2008) ===
* update to numpy 1.1.0
=== numpy-20080104-1.0.4.p4 (Michael Abshoff, May 18th, 2008) ===
* add 64 bit OSX build support
=== numpy-20080104-1.0.4.p3 (William Stein, May 16, 2008) ===
* (Stupid) patch to numpy/linalg/setup.py to support cygwin
=== numpy-20080104-1.0.4.p2(Josh Kantor) ===
* update gnu.py to fix issue with inline fortran on osx
=== numpy-20080104-1.0.4.p1 (Josh Kantor & Michael Abshoff) ====
* update SPK.txt (Josh Kantor)
* remove *~
* update .hgignore
=== numpy-20080104-1.0.4.p0 (Michael Abshoff) ====
* add SPKG.txt
* create hg repo
=== numpy-20080104-1.0.4 (Josh Kantor) ====
* update to numpy 1.0.4 (Josh Kantor)
