summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
...
* BUG: numpy.ma set fpu exception error state to ignore all at import time.David Cournapeau2009-12-081-4/+0
|
* * _iotools.LineSplitter : prevent the first and/or last empty tab-separated ↵pierregm2009-12-082-54/+65
| | | | columns to be dropped
* BUG: fix test_buffer_hashlib for 64-bit systemsPauli Virtanen2009-12-071-1/+1
|
* BUG: Fix bugs in complex pow (fixes #1313)Pauli Virtanen2009-12-073-9/+34
| | | | Thanks to Francesc Alted.
* BUG: make assert_equal and assert_almost_equal always display err_msgPauli Virtanen2009-12-071-8/+5
|
* Cleanup r7978 (add err_msg support to assert_mask_equalmdroe2009-12-071-2/+2
|
* * Bugfix #1173 and #1174 (Michael D.)pierregm2009-12-072-1/+2
|
* Remove reliance on integer division by zero returning zero.Charles Harris2009-12-071-1/+3
|
* BUG: fix undefined symbol on Py < 2.6Pauli Virtanen2009-12-061-0/+7
|
* BUG: add a work-around for #1312 -- don't define bf_releasebuffer at allPauli Virtanen2009-12-066-34/+73
| | | | | | | | | | | | | | | The problem is that PyArg_ParseTuple("s#", ...) does not accept objects with the new buffer interface, if they have bf_releasebuffer defined. We now work around this by not defining bf_releasebuffer, and releasing any info associated with the buffer interface when array or its descriptor is deallocated. The shape+strides information for arrays stays static due to the refcounting lock. The format information for dtypes may change if field names are changed on-the-fly. XXX: dtype field name changes are not currently handled.
* 3K: more careful error checking in flagsobject.c:arrayflags_getitem for ↵Pauli Virtanen2009-12-061-1/+7
| | | | Unicode input
* Mark a non-implemented test as a knownfailurePauli Virtanen2009-12-061-0/+1
|
* ENH: core: add tests for Unicode (or Bytes on Py3) in field names and titlesPauli Virtanen2009-12-061-0/+34
|
* 3K: core: some PyString conversion in arrayobject.cPauli Virtanen2009-12-061-3/+3
|
* core/buffer: cache format string etc. in the ndarray objectPauli Virtanen2009-12-063-21/+78
|
* 3K: fix ma/test_mrecords.py so that it does not fail in test setup on Py3KPauli Virtanen2009-12-061-3/+11
|
* 3K: add missing includes for NPY_SEPARATE_COMPILATIONPauli Virtanen2009-12-0610-0/+18
|
* 3K: core: make comparison between Unicode and Bytes NotImplemented, as it is ↵Pauli Virtanen2009-12-061-0/+10
| | | | in Py3K
* 3K: core: PyString conversion in convert.cPauli Virtanen2009-12-062-8/+18
|
* 3K: core: fix IO locale tests -- apparently some changes in Py3 NosePauli Virtanen2009-12-061-12/+8
|
* 3K: core: try to retain backward compatibility with Python 2 picklesPauli Virtanen2009-12-064-16/+68
| | | | | | | | | | This entails using Unicode for the endian character, and accepting Unicode data as input. We assume the user always uses encoding='latin1' for load, if backwards compatibility is desired! Tests are fixed according to the above.
* BUG: fix comparison between tuple and int in test_numerictypes.pyPauli Virtanen2009-12-061-1/+1
|
* 3K: core: make defchararray to importPauli Virtanen2009-12-061-1/+4
|
* 3K: core: replace StringType in records.py by strPauli Virtanen2009-12-061-1/+1
|
* 3K: core: stub out UCS size check on Py3 in test_unicodePauli Virtanen2009-12-061-2/+7
| | | | | | Py3 str does not implement the buffer interface, so the previous way to do the check does not work any more. The check needs to be fixed later on -- this commit is just to make nosetests able to import the test file.
* 3K: core: make type codes of typeinfo dict Unicode in arraytypes.c.srcPauli Virtanen2009-12-061-0/+32
|
* 3K: core: PyString conversion in ufunc_object.cPauli Virtanen2009-12-061-4/+4
|
* 3K: core: PyString conversion in multiarraymodule.cPauli Virtanen2009-12-061-11/+23
|
* 3K: core: do not swallow exceptions needlessly in test_multiarrayPauli Virtanen2009-12-061-2/+3
|
* 3K: compat: add getexception to compat.py3kPauli Virtanen2009-12-061-1/+6
|
* 3K: core: Remove 'getbuffer' and 'newbuffer' from multiarray in Py3Pauli Virtanen2009-12-062-24/+12
| | | | | Py3 has the new memoryview object, which should obsolete these functions.
* 3K: core: remove unnecessary #warningsPauli Virtanen2009-12-062-7/+0
|
* 3K: test_errstate should check floor divide, to avoid spurious fp exceptionsPauli Virtanen2009-12-061-2/+2
|
* 3K: core: PyString conversion in ufunc_object.cPauli Virtanen2009-12-061-16/+16
|
* 3K: core: bytes vs. str fixes in memmap.pyPauli Virtanen2009-12-061-2/+4
|
* 3K: core: write only bytes to files in test_multiarray.pyPauli Virtanen2009-12-061-8/+9
|
* 3K: core: everything should be Bytes in _internal.pyPauli Virtanen2009-12-061-43/+43
|
* 3K: core: PyString conversion in hashdescr.cPauli Virtanen2009-12-061-1/+1
|
* 3K: core: cPickle is no more, use pickle in methods.cPauli Virtanen2009-12-061-0/+8
|
* 3K: core: PyString conversion in conversion_utils.cPauli Virtanen2009-12-061-3/+30
|
* 3K: core: PyString conversion in flagsobject.cPauli Virtanen2009-12-061-8/+34
|
* 3K: fix test_complex_inf_nanPauli Virtanen2009-12-061-1/+1
|
* 3K: core: fix a few Name and SyntaxErrors in testsPauli Virtanen2009-12-062-2/+9
|
* 3K: core: PyString conversion in scalartypes.c.srcPauli Virtanen2009-12-061-7/+15
|
* Add more Py3K notesPauli Virtanen2009-12-061-307/+0
|
* 3K: rename compat.isfile to isfileobj to avoid confusion with os.pathPauli Virtanen2009-12-063-8/+11
|
* 3K: lib: some fixes to lib.format on strings vs bytes; and file objectsPauli Virtanen2009-12-061-6/+6
|
* 3K: add some compatibility tools to numpy.compatPauli Virtanen2009-12-062-0/+25
|
* py3k_notes: some notes on PyOSPauli Virtanen2009-12-061-0/+10
|
* 3K: do not use PyOS_ascii_strtod -- it segfaults and is deprecatedPauli Virtanen2009-12-061-0/+8
|
View Lorry Controller Status