summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
...
* | | | | | Fix segfaultJay Bourque2013-05-141-1/+1
|/ / / / / | | | | | | | | | | | | | | | outer_steps array should be size 4 since op_count can be either 2 or 4
* | | | | BLD: Fixes for building on win32 with MSVC 2008Mark Wiebe2013-05-134-150/+186
| | | | |
* | | | | Merge pull request #3322 from ContinuumIO/array_equal_fixCharles Harris2013-05-132-2/+8
|\ \ \ \ \ | |_|/ / / |/| | | | Fix array_equal and array_equiv issue
| * | | | Make sure comparison result is arrayJay Bourque2013-05-131-1/+1
| | | | |
| * | | | Fix array_equal and array_equiv issueJay Bourque2013-05-102-2/+8
| | | | | | | | | | | | | | | | | | | | array_equal() and array_equiv() don't work for string arrays and record arrays. This is because those methods use numpy.equals ufunc for comparison, and there are no equal ufuncs registered for strings and record arrays. Replace numpy.equals with '==' array operation, which handles strings and record arrays.
* | | | | TST: linalg: make sure that running the xerbla test is safePauli Virtanen2013-05-121-1/+43
| | | | |
* | | | | Merge pull request #2701 from seberg/indexingseberg2013-05-116-52/+392
|\ \ \ \ \ | | | | | | | | | | | | Fancy Indexing enhancements and bug fixes
| * | | | | MAINT: Remove unnecessary and wrong check for mapiterswapSebastian Berg2013-05-071-5/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | With the slight modification of the MapIter API to allow newaxis in fancy indexes, the MapIterObject consec field is already the axis to transpose to, and with a leading newaxis, even if iteraxes[0] is 0, consec may not be 0.
| * | | | | TST: Add fancy indexing test for tuple indexingSebastian Berg2013-04-121-9/+322
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is a first version of fancy indexing tests by mimicing the behavoir in python. Note that error messages are not tested.
| * | | | | MNT: Tiny cleanup for mapping.c and commentsSebastian Berg2013-04-122-9/+12
| | | | | |
| * | | | | BUG: Do not ignore subspace unless it is true 0-dSebastian Berg2013-04-121-15/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is more important when allowing newaxis, but already before lead to successful indexing when it should not be suspected. Also some maintenance.
| * | | | | BUG: Check for maximum dimension for fancy indexing.Sebastian Berg2013-04-121-1/+10
| | | | | |
| * | | | | ENH: Allow newaxis in combination with fancy indexing.Sebastian Berg2013-04-122-21/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit allows the use of None when also using fancy indexing. To achieve this mit->consec is used to note the denote the transposition axis (which means that if its consecutive but at the start, mit->consec==0, and no transposition is necessary).
| * | | | | BUG: Fix indexing 0-sized arrays using empty fancy-indexesSebastian Berg2013-04-123-12/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Due to subspace creation for fancy indexing selecting the 0s element on every fancy dimension, this failed if the dimension was empty. Propagating check_index to disable the unnecessary check in this case.
* | | | | | Merge pull request #3316 from seberg/issue-3314njsmith2013-05-103-24/+23
|\ \ \ \ \ \ | | | | | | | | | | | | | | BUG: Fix 0-d array special case from reductions.
| * | | | | | BUG: Fix 0-d array special case from reductions.Sebastian Berg2013-05-083-24/+23
| | |_|/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | This channels scalars through the usual reduction machinery and modifies it slightly to correctly support scalar reductions of identity-less ufuncs.
* | | | | | Merge pull request #3325 from seberg/remove-gh-356-special-caseCharles Harris2013-05-102-12/+6
|\ \ \ \ \ \ | | | | | | | | | | | | | | MAINT: Remove unnecessary 0-d special case.
| * | | | | | MAINT: Remove unnecessary 0-d special case.Sebastian Berg2013-05-102-12/+6
| | |_|_|_|/ | |/| | | | | | | | | | | | | | | | This special case was made unnecessary by changes to the nditer.
* | | | | | Merge pull request #3323 from ContinuumIO/broadcast_arrays_fixCharles Harris2013-05-102-1/+12
|\ \ \ \ \ \ | |/ / / / / |/| | | | | Broadcast arrays fix
| * | | | | Add testJay Bourque2013-05-101-0/+8
| | | | | |
| * | | | | Fix issue with broadcast_arrays() and user defined dtypesJay Bourque2013-05-101-1/+4
| | |_|/ / | |/| | | | | | | | | | | | | broadcast_arrays() does not handle struct and custom dtypes correctly. Dtype of returned broadcasted arrays is always '|V8'. Fix broadcast_arrays() so that dtype of returned arrays is correct dtype for user defined dtypes.
* | | | | Merge pull request #359 from jayvius/operand-flagsCharles Harris2013-05-106-23/+222
|\ \ \ \ \ | |/ / / / |/| | | | Allow ufunc operand flags to be set
| * | | | Remove unneeded blank lineJay Bourque2013-05-101-1/+0
| | | | |
| * | | | Add null check after allocating op_flagsJay Bourque2013-05-092-0/+6
| | | | |
| * | | | Whitespace changesJay Bourque2013-05-091-5/+5
| | | | |
| * | | | Fix formatting issuesJay Bourque2013-05-063-23/+20
| | | | |
| * | | | Fix python 2/3 issues in operand_flag_tests moduleJay Bourque2013-05-061-6/+25
| | | | |
| * | | | Add test for setting operand flags through c apiJay Bourque2013-05-063-0/+102
| | | | |
| * | | | Remove guard against read+write output operandsJay Bourque2013-05-061-18/+0
| | | | |
| * | | | correctly initialize op_flags and iter_flags in frompyfunc()Jay Bourque2013-05-061-1/+3
| | | | |
| * | | | initialize ufunc op_flags pointer in frompyfunc()Jay Bourque2013-05-061-0/+1
| | | | |
| * | | | minor fixes for previous ufunc operand flag changejayvius2013-05-061-13/+18
| | | | |
| * | | | This change allows ufunc operand flags to be set in addition to thejayvius2013-05-062-16/+102
| |/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | default operand flags set by the nditer object. For example, after calling PyUFunc_FromFuncAndData() to create a ufunc object, the first operand can be set to be read/write like so: ufunc->op_flags[0] = NPY_ITER_READWRITE Setting flags for output operands is not supported at this time and will raise an exception. This change also allows the global nditer flags to be set, so after setting the read/write flag, the reduce ok flag could be set like so: ufunc->iter_flags = NPY_ITER_REDUCE_OK
* | | | BUG: testing: always enable --exePauli Virtanen2013-05-091-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | Setuptools tends to set +x to the installed test scripts, which makes numpy.test() to not run any tests. Having --exe always enabled is not problematic because only files matching 'test_*.py' are looked into.
* | | | TST: add a test for issue #3312Edward Catmur2013-05-071-0/+5
| | | |
* | | | BUG: fix potentially infinite recursion in VOID_nonzero. Fixes #3312.Edward Catmur2013-05-071-1/+1
| |_|/ |/| |
* | | STY: pep8 and pep257 audit.Josh Warner (Mac)2013-05-071-89/+101
|/ / | | | | | | | | | | Audit numpy/lib/arraypad.py for pep8 and pep257 compliance. Also fix a few minor docstring corrections converting ] into ) or vice versa.
* | Merge pull request #3182 from juliantaylor/debug-extRalf Gommers2013-05-052-6/+40
|\ \ | | | | | | get_shared_lib_extension(): strip debug extension from so ext
| * | BUG: hardcode some known shared library extensionsJulian Taylor2013-04-282-6/+40
| | | | | | | | | | | | | | | | | | | | | | | | the configuration variables are not a reliable mean to get the shared library extension. darwin, windows and debug linux are wrong in these variables. SHLIB_SUFFIX is also wrong as of python 3.3.1 closes #3057
* | | Merge pull request #3264 from endolith/patch-1Charles Harris2013-05-041-8/+5
|\ \ \ | | | | | | | | DOC: Change example to demonstrate function
| * | | DOC: Change example to demonstrate functionendolith2013-04-191-8/+5
| | | | | | | | | | | | "a * 0.5" example might as well be new_func(a) directly, it doesn't demonstrate the purpose of apply_along_axis().
* | | | BLD: fix Bento build after umath_linalg move from core to linalg module.Ralf Gommers2013-05-043-29/+6
| | | |
* | | | Merge pull request #3303 from ContinuumIO/rational_fixCharles Harris2013-05-031-22/+37
|\ \ \ \ | | | | | | | | | | Fix clang build issue
| * | | | Fix potential reference leakJay Bourque2013-05-031-0/+6
| | | | |
| * | | | Refactor previous commitJay Bourque2013-05-031-12/+9
| | | | |
| * | | | Do not return anything for python 2Jay Bourque2013-05-031-20/+32
| | | | |
* | | | | MAINT: Remove unneeded version check from multiarray/common.c.Arink Verma2013-05-041-4/+0
|/ / / / | | | | | | | | | | | | | | | | A version check for Python >= 2.6 is no longer needed as versions < 2.6 are no longer supported.
* | | | BUG: gh-3300, Fix missing return value in test_rational.c.src.Charles Harris2013-05-031-7/+7
| | | | | | | | | | | | | | | | | | | | | | | | Some error returns were returning nothing rather than NULL. Closes #3300.
* | | | Merge pull request #3261 from keflavich/f77_ext_bugfixCharles Harris2013-05-021-2/+2
|\ \ \ \ | | | | | | | | | | bugfix in build_ext: check that ext has an attribute before trying to us it
| * | | | reduces excess cleverness from 1/2 to 1/3Adam Ginsburg2013-04-211-2/+2
| | | | |
View Lorry Controller Status