Commits on Aug 1, 2010
  1. DOC: wiki merge, core.umath

    rgommers committed Aug 1, 2010
  2. DOC: wiki merge, rst docs.

    rgommers committed Aug 1, 2010
  3. BUG: (backport of r8568) remove unnecessary import and reload in test…

    This was causing errors when running the test suite.
    rgommers committed Aug 1, 2010
Commits on Jul 31, 2010
  1. TST: (backport r8567) Add tests for matlib. Closes #1242.

    (From Ralf Gommers.)
    pv committed Jul 31, 2010
  2. BUG: fix numscons build issue introduced in r8541.

    The problem was that if endian.h was not detected, @DEFINE_NPY_HAVE_ENDIAN_H@
    was still present in _numpyconfig.h, causing compilation to fail.
    rgommers committed Jul 31, 2010
  3. BUG: (backport of r8553) fix base_repr for py3k.

    Strings can not be indexed with floats anymore.
    rgommers committed Jul 31, 2010
Commits on Jul 29, 2010
  1. DOC: sphinxext: (backport r8549) replace directive mangling with doma…

    …ins (#1489)
    Only Sphinx >= 1.0 is supported now.
    pv committed Jul 29, 2010
  2. REL: Update paver and release scripts for Python 2.7/3.1 binaries.

    Note that the increase in disk image size from 10Mb to 11Mb does not mean numpy
    is actually that large - more than half that is the included docs. It is the
    case though that the 2.7 binaries are slightly larger than the 2.5/2.6 ones, by
    100kb on Windows and 800kb on OS X.
    rgommers committed Jul 29, 2010
  3. BUG: make sure npy_cpu.h is always included in ndarraytypes.h

    (cherry picked from commit b4d7338de5338daed4e543921cf3cb82eb2c3675)
    cournape committed Jul 29, 2010
  4. BUG: look for endian.h.

    We should use this header in npy_endian.h if available.
    (cherry picked from commit 6f59b03b86868794328a39bd0542cc47957d2c80)
    cournape committed Jul 29, 2010
Commits on Jul 28, 2010
  1. ENH: (backport r8539) testing: add assert_tol_equal for testing array…

    … equality with specified tolerances
    pv committed Jul 28, 2010
  2. BUG: (backport r8535) support assignment x[(0,)] = foo in ndarray sub…

    …types (#1563)
    The problem was a code path trying to assign x[0], which returns a
    scalar -- for actual ndarrays a internal function always returning
    arrays was used.
    pv committed Jul 28, 2010
  3. BUG: (backport r8532) fix usability bugs in lookfor

    - allow numbers in function signature
    - print the items in relevance order (not reverse relevance)
    - include ufunc docstrings
    pv committed Jul 28, 2010
Commits on Jul 27, 2010
  1. BUG (backport r8526): fix alpha build.

    (cherry picked from commit c11d86ee1da5841d92402a474d9c7c2d90873dc5)
    cournape committed Jul 27, 2010
Commits on Jul 26, 2010
Commits on Jul 24, 2010
  1. DOC: update the blurb about the array interface

    (cherry picked from commit r8512)
    pv committed Jul 24, 2010
  2. BUG: quick and ugly fix for long double on linux ppc.

    (cherry picked from commit r8511)
    pv committed Jul 24, 2010
  3. ENH: fix long double detection for linux ppc

    This does not fix build on linux ppc, as the implementation for
    long double functions is missing.
    (cherry picked from commit r8510)
    pv committed Jul 24, 2010
  4. BUG: core: fix argmax and argmin NaN handling to conform with max/min (

    This makes `argmax` and `argmix` treat NaN as a maximal element.
    Effectively, this causes propagation of NaNs, which is consistent with
    the current behavior of amax & amin.
    (cherry picked from commit r8509)
    pv committed Jul 24, 2010
  5. BUG: core/umath: make complex number comparisons False when *either* …

    …element is nan
    This also fixes NaN propagation in np.maximum/minimum, for example in
    the case for
        maximum(1, complex(0, nan)) -> (0, nan)
    which previously yielded (1, 0).
    (cherry picked from commit r8508)
    pv committed Jul 24, 2010
  6. BUG: core: make set_string_function(None) to restore the functions to…

    … what they are on import (#1130)
    (cherry picked from commit r8507)
    pv committed Jul 24, 2010