Commits on Jul 29, 2016
  1. @charris

    Merge pull request #7871 from juliantaylor/introsort2

    BUG: handle introsort depth limit properly
    charris committed on GitHub Jul 29, 2016
  2. @juliantaylor

    BENCH: add more sort benchmarks

    benchmark for random, all equal elements, many equal elements and median
    of 3 worst case
    juliantaylor committed Jul 27, 2016
  3. @juliantaylor

    BUG: handle introsort depth limit properly

    the larger partition needs its own depth limit else it grows faster than
    logarithmic.
    Also increase the stack size to hold up to 64 instead of 50 pointer
    pairs else it might overflow on arrays larger than than 2^50 elements.
    juliantaylor committed Jul 26, 2016
Commits on Jul 28, 2016
  1. @charris

    Merge pull request #7879 from dunneff/patch-1

    DOC: fix typo in documentation of loadtxt (closes #7878)
    charris committed on GitHub Jul 28, 2016
  2. @charris

    Merge pull request #7699 from nolta/preserve-lib-order

    BLD: preserve library order
    charris committed on GitHub Jul 28, 2016
  3. @dunneff

    DOC: fix typo in documentation of loadtxt (closes #7878)

    dunneff committed on GitHub Jul 28, 2016
Commits on Jul 25, 2016
  1. @charris

    Merge pull request #7866 from ssanderson/document-runtests-extra-argv

    Document runtests extra argv
    charris committed on GitHub Jul 25, 2016
  2. @charris

    Merge pull request #7853 from charris/fix-novalue-import

        BUG: Make sure numpy globals keep identity after reload.
    charris committed on GitHub Jul 25, 2016
  3. @ssanderson

    STY: Use consistent variable convention.

    ssanderson committed on GitHub Jul 25, 2016
  4. @ssanderson

    DOC: Use a shorter example.

    ssanderson committed on GitHub Jul 25, 2016
  5. @charris

    Merge pull request #7863 from juliantaylor/introsort

    ENH: turn quicksort into introsort
    charris committed on GitHub Jul 25, 2016
Commits on Jul 23, 2016
  1. @ssanderson

    DOC: Add example to runtests usage doc.

    ssanderson committed Jul 23, 2016
  2. @ssanderson

    DOC: Document extra args forwarding to nose.

    This is useful for things like passing ``--pdb`` to make nose drop into
    a pdb post-mortem on exception.
    ssanderson committed Jul 23, 2016
Commits on Jul 21, 2016
  1. @juliantaylor

    ENH: turn quicksort into introsort

    Introsort is regular quicksort but changing to a heapsort when not
    enough progress is made. This retains the good quicksort performance
    while changing the worst case runtime from O(N^2) to O(N*log(N))
    juliantaylor committed Jul 21, 2016
  2. @juliantaylor
Commits on Jul 20, 2016
  1. @charris

    BUG: Raise RuntimeError when reloading numpy is attempted.

    There seems to be little use in reloading numpy as any changed modules
    that are imported into numpy would need to be reloaded first in order to
    see any changes. Furthermore, reloading causes problems as global
    classes defined in numpy/__init__.py change their identity. Hence we
    raise a RuntimeError when an attempt to reload numpy is made.
    
    Closes #7844.
    charris committed Jul 20, 2016
Commits on Jul 19, 2016
  1. @charris

    Merge pull request #7840 from llasram/atlas-version-check

    Fix ATLAS version detection
    charris committed on GitHub Jul 19, 2016
  2. @njsmith

    Merge pull request #7848 from charris/imp-module-deprecation

    MAINT: Fix remaining uses of deprecated Python imp module.
    njsmith committed on GitHub Jul 18, 2016
  3. @njsmith

    Merge pull request #7839 from mattip/pypy-fixes

    Pypy fixes
    njsmith committed on GitHub Jul 18, 2016
Commits on Jul 18, 2016
  1. @charris

    MAINT: Fix remaining uses of deprecated Python imp module.

    charris committed Jul 18, 2016
  2. @charris

    Merge pull request #7791 from bertrand-l/feature/f2py-issues-7683

    f2py.compile issues (#7683)
    charris committed on GitHub Jul 18, 2016
  3. @mattip

    ENH: Support for python runtime with no Makefile, like PyPy

    mattip committed Jul 15, 2016
  4. @mattip

    ENH: set reasonable defaults for slots before PyType_Ready is called,…

    … enables overriding them later
    mattip committed Jul 15, 2016
  5. @bertrand-l

    MAINT,DOC: add to compat.py3k a function to load modules. Fix some do…

    …c for f2py.compile (issue #7683)
    bertrand-l committed Jun 13, 2016
Commits on Jul 16, 2016
  1. @njsmith

    Merge pull request #7842 from stefanv/versionadded_typos

    Fix versionadded tags
    njsmith committed on GitHub Jul 15, 2016
Commits on Jul 15, 2016
  1. @stefanv

    Fix versionadded tags

    stefanv committed Jul 15, 2016
  2. @llasram

    Restore get_output

    The ATLAS version check still depends on this function.
    llasram committed Jul 15, 2016
Commits on Jul 14, 2016
  1. @jaimefrio

    Merge pull request #7834 from madphysicist/patch-1

    DOC: Added an example
    jaimefrio committed on GitHub Jul 14, 2016
  2. @madphysicist

    DOC: Added an example

    It is not 100% clear that this will work properly for scalars, so I added an example that I thought was useful.
    madphysicist committed on GitHub Jul 14, 2016
Commits on Jul 11, 2016
  1. @charris

    Merge pull request #7820 from charris/fix-empty-array-allocation-size

    MAINT: Allocate fewer bytes for empty arrays.
    charris committed on GitHub Jul 11, 2016
  2. @rgommers

    Merge pull request #7818 from emmenlau/master

    MAINT: fix for gh-7809 (distutils): check that _tmpdirs is not None
    rgommers committed on GitHub Jul 11, 2016
Commits on Jul 10, 2016
  1. @juliantaylor

    Merge pull request #7817 from charris/fix-mul-with-overflow

    BUG: Make sure npy_mul_with_overflow_<type> detects overflow.
    juliantaylor committed on GitHub Jul 11, 2016
  2. @charris

    MAINT: Allocate fewer bytes for empty arrays.

    Currently the number of bytes allocated is the product of the non-zero
    dimensions x element_size, which can create huge memory allocations for
    empty arrays. Change that to just allocate enough for one element. Some
    allocation is needed for the array.data attribute to work correctly.
    
    Note: Could probably allocate a minimal number of bytes, say 16 as
    dictated by alignment, but the old intended number is used here as a
    more conservative fix.
    
    Closes #7813.
    charris committed Jul 10, 2016
  3. @charris

    BUG: Ignore a DeprecationWarning.

    Changing the data attribute is deprecated and that warning is raised in
    the test_multiarray_writable_attributes_deletion test in
    numpy/core/tests/test_multiarray.py
    charris committed Jul 10, 2016
  4. @charris

    BUG: Make sure npy_mul_with_overflow_<type> detects overflow.

    The previous check for overflow would fail with signed integers as it
    was designed to check for overflow of the larger corresponding unsigned
    type.
    charris committed Jul 10, 2016