summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
...
* BUG: Replace deprecated PyCObject by PyCapsule for Python >= 3.1.Charles Harris2010-02-2321-1122/+1329
|
* Add very basic sanity test for getting include directories (related to ↵mdroe2010-02-221-0/+10
| | | | ticket #1405)
* Add regression test for ticket #1405mdroe2010-02-221-0/+6
|
* Fix bad usage of namespace alias changechanley2010-02-221-5/+5
|
* BUG: Ignore "invalid value" from abs in testing/utils.pyCharles Harris2010-02-211-2/+7
|
* BUG: Workarounds for isfinite/isinf invalid values.Charles Harris2010-02-211-6/+6
|
* BUG: Fix unhandled divide-by-zero errors and invalid values in tests. Add someCharles Harris2010-02-211-50/+116
| | | | tests to check that divide-by-zero errors are raised when desired.
* BUG: Fix possibly unbalanced seterr calls. Fixes change of divide state afterCharles Harris2010-02-213-59/+78
| | | | running np.test().
* BUG: Fix possibly unbalanced seterr pairs. Fixes change in state of invalidCharles Harris2010-02-211-36/+124
| | | | after running test().
* BUG: More workarounds for np.isinf warning in tests.Charles Harris2010-02-212-9/+21
|
* BUG: Work around warning raised by np.isinf(np.inf). The isinf warning ↵Charles Harris2010-02-211-7/+15
| | | | | | | should be fixed at a lower level and tested separately. In the meantime, this fix avoids a lot of irrelevant warning messages in the tests.
* WHT: Remove trailing whitespace.Charles Harris2010-02-211-5/+4
|
* ENH: core: export format string as UString in numpymemoryview (Py2.6)Pauli Virtanen2010-02-211-1/+1
|
* BUG: core: fix PEP 3118 tests to account for variation in native long sizesPauli Virtanen2010-02-211-19/+26
|
* ENH: core: better support for native vs. standard sizes and alignments in ↵Pauli Virtanen2010-02-213-29/+196
| | | | | | | the PEP 3118 interface Also, try to produce minimal buffer format strings without unnecessary alignment characters.
* BUG: core: use C int size specifiers in the PEP 3118 tests, to make things ↵Pauli Virtanen2010-02-211-27/+35
| | | | work on 32 and 64 bits platforms
* 3K: BUG: alias divide -> true_divide correctly; just replacing the name in ↵Pauli Virtanen2010-02-212-1/+8
| | | | generate_umath.py is not the correct thing to do
* ENH: ma: use domains also in _DomainedBinaryOperation.__call__ (previously ↵Pauli Virtanen2010-02-211-0/+4
| | | | they were used only in __array_wrap__, which wouldn't be called from __div__ et al. which call the domained operations directly)
* BUG: fix ppc detection macro on AIX.David Cournapeau2010-02-211-1/+2
|
* ENH: testing: always enable --detailed-errors assert introspection in nosePauli Virtanen2010-02-211-0/+3
|
* ENH: core: use assert_equal(x, y) instead of assert x==y -- test failures ↵Pauli Virtanen2010-02-211-36/+36
| | | | are easier to trace that way
* BUG: Fix the invalid value warnings in test_umath.pyCharles Harris2010-02-211-19/+39
|
* DEP: Fix one more deprecated unittest function.Charles Harris2010-02-211-1/+1
|
* STY: Use import numpy as np. This seems to fix an import error introduced byCharles Harris2010-02-211-5/+13
| | | | | 2to3, but that may have been an artifact from a previous build. In anycase, no harm done.
* STY: Use explicit itertools namespace for all itertools functions.Charles Harris2010-02-211-6/+4
|
* BUG: fix numarray._capi compilationPauli Virtanen2010-02-211-1/+1
|
* STY: Remove unneeded import.Charles Harris2010-02-211-5/+0
|
* PY3: The izip function is no longer available in itertools. Be explicit aboutCharles Harris2010-02-211-16/+17
| | | | using itertools.izip to 2to3 can make the fix.
* WHT: Whitespace removal.Charles Harris2010-02-211-2/+1
|
* 3K: ENH: ma: fix str vs bytes and int issues in ma testsPauli Virtanen2010-02-213-8/+18
|
* BUG: ma: fix inoperative error state set/restorePauli Virtanen2010-02-211-5/+0
|
* BUG: ma: _check_fill_value shouldn't rely on implicit array() string castingPauli Virtanen2010-02-212-13/+14
| | | | | | | Since array([12345678.9, 'a']) == array(['12345678', 'a'], dtype='|S8') ie., automatic string conversion uses only the size of the minimal data type, not the size of the string representation, code should not rely on array() casting items implicitly to object arrays.
* 3K: ENH: core: support BOOL and VOID better in PEP 3118 buffersPauli Virtanen2010-02-213-12/+30
|
* 3K: ENH: core: void.item() -> byte array; simple memoryview -> BytesPauli Virtanen2010-02-211-16/+23
|
* 3K: ma: implement __rtruediv__, __rfloordiv__, __rpow__ for masked arraysPauli Virtanen2010-02-211-0/+12
|
* 3K: lib: implement __rtruediv__ for poly1dPauli Virtanen2010-02-211-0/+2
|
* 3K: core: mark int('0xff',16) test as known failure on Python3 -- we don't ↵Pauli Virtanen2010-02-211-0/+3
| | | | anyway support that call signature for all integer types even on Py2
* 3K: ma: solve some bytes vs. str issues in testsPauli Virtanen2010-02-212-18/+28
|
* 3K: ENH: ma: implement data-preserving __ifloordiv__ and __itruediv__Pauli Virtanen2010-02-211-0/+30
|
* 3K: ENH: make numpy.numarray to importPauli Virtanen2010-02-211-0/+3
|
* 3K: ENH: move numarray includes under numarray/include so that the 'numpy' ↵Pauli Virtanen2010-02-219-3/+3
| | | | directory does not confuse 2to3's relative import conversion
* 3K: ENH: make oldnumeric to import -- but don't implement ↵Pauli Virtanen2010-02-212-6/+21
| | | | oldnumeric.Unpickler as it's not straightforward to do
* 3K: ENH: core: follow Python in formatting negative zeros in (-0+x*j)Pauli Virtanen2010-02-211-0/+4
|
* 3K: lib: fix bytes vs str issues in testsPauli Virtanen2010-02-212-8/+11
|
* 3K: lib: zip is izip on Py3Pauli Virtanen2010-02-211-0/+4
|
* 3K: lib: missing_values in genfromtxt is never a basestring on Py3Pauli Virtanen2010-02-212-2/+2
|
* 3K: lib: fix sort(key=) issuePauli Virtanen2010-02-211-5/+3
|
* 3K: linalg: fix integer division issue and testsPauli Virtanen2010-02-212-2/+4
|
* 3K: polynomial: make tests to importPauli Virtanen2010-02-213-3/+0
|
* 3K: lib: poly1d __div__ -> __truediv__, and fix its doctestsPauli Virtanen2010-02-212-7/+9
|
View Lorry Controller Status