summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
...
| * BUG: core: make Numpy build again on Python 3 and 2.4Pauli Virtanen2011-02-026-57/+183
| |
| * BUG: Fix exception syntax to conform to python 2.4.Charles Harris2011-02-011-1/+1
| |
| * ENH: iter: Allow copies of read-only scalar arrays even if COPY isn't specifiedMark Wiebe2011-01-303-39/+79
| | | | | | | | | | | | | | When buffering, this reduces the work needed by a lot. For non-reduction operands, the iterator always provides contiguous data when an operand is buffered, but by making a copy the operand can be flagged as never needing buffering
| * BUG: iter: Fix checking of allocated output with op_axes specifiedMark Wiebe2011-01-302-58/+63
| |
| * BUG: core: Fix "a[...] += 1" with array scalar 'a'Mark Wiebe2011-01-303-4/+78
| | | | | | | | | | | | | | | | | | This was crashing, because it appears Python does this: 1. call the array_subscript to get a[...], then calls inplace add on it 2. call the array_ass_subscript with (a, Ellipsis, a), which for an object array was setting the single element of a to point to itself, and caused crashes. This case is now checked, and success is returned immediately when an array is assigned to itself this way.
| * WRN: core: Remove many warnings gcc produces with stricter flagsMark Wiebe2011-01-297-163/+183
| |
| * BUG: Fix crash in PyArray_Nonzero due to uninitialized variable usageMark Wiebe2011-01-294-19/+19
| |
| * STY: Rename NPY_ITER_DONT_REVERSE_AXES to NPY_ITER_DONT_NEGATE_STRIDES to be ↵Mark Wiebe2011-01-283-4/+4
| | | | | | | | more intuitive
| * Merge branch 'new_iterator' - new iterator, ufunc update, restore 1.5 ABIMark Wiebe2011-01-2862-4433/+24973
| |\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | New Iterator - Read doc/neps/new-iterator-ufunc.rst. UFunc Update - Change all ufunc functions to use the new iterator. This replaces the inline buffering with iterator buffering, except for the reductions and generalized ufunc which use updateifcopy at the moment. Also adds out= and order= parameters to all ufuncs. Restore 1.5 ABI - This was done by moving the new type numbers to the end of the type enumeration, and replacing all type promotion code with a table-based approach. The ArrFuncs was restored by putting the new type cast functions into the cast dictionary, originally designed just for custom types. Conflicts: numpy/core/src/multiarray/ctors.c numpy/core/tests/test_regression.py
| | * ENH: core: Add support for numpy.ravel(a, order='K')Mark Wiebe2011-01-288-72/+198
| | |
| | * ENH: core: Add dtype= and order= parameters to zeros_like, ones_like, and ↵Mark Wiebe2011-01-2810-109/+249
| | | | | | | | | | | | empty_like
| | * ENH: core: Add PyArray_NewLikeArray functionMark Wiebe2011-01-284-23/+138
| | | | | | | | | | | | | | | | | | This function implements the numpy.empty_like semantics, but supports the new NPY_KEEPORDER enumeration value as well as switching to a different data type.
| | * STY: core: Improve formating of ParseTuple argument listsMark Wiebe2011-01-281-119/+105
| | | | | | | | | | | | Now in methods.c each argument gets its own line, so its easier to validate.
| | * WRN: Fix some warnings and suspicious looking codeMark Wiebe2011-01-276-14/+13
| | |
| | * BUG: Single file build had wrong promotion table typeMark Wiebe2011-01-271-1/+1
| | |
| | * ENH: iter: Improve performance of buffering reduce operationsMark Wiebe2011-01-272-135/+840
| | |
| | * WRN: iter: Fix half-float warnings, other small tweaksMark Wiebe2011-01-275-27/+23
| | |
| | * ENH: iter: Add a per-operand REDUCE flagMark Wiebe2011-01-271-10/+20
| | |
| | * ENH: iter: Remove internal NPY_OP_ITFLAG_COPY flagMark Wiebe2011-01-271-20/+14
| | |
| | * ENH: iter: Change permutation to a fixed-size char arrayMark Wiebe2011-01-271-83/+57
| | |
| | * ENH: ufunc: Add new iterator version of generalized ufuncsMark Wiebe2011-01-275-280/+476
| | |
| | * ENH: core: Make NumPy trunk ABI-compatible with 1.5Mark Wiebe2011-01-277-279/+305
| | |
| | * ENH: ufunc: Move reduceat to its own function, have it use the iteratorMark Wiebe2011-01-271-218/+381
| | |
| | * BLD: Add einsum to the one file buildMark Wiebe2011-01-271-0/+1
| | |
| | * BUG: core: Fix things so scipy trunk passes all tests (but one)Mark Wiebe2011-01-278-64/+112
| | | | | | | | | | | | | | | | | | | | | With this patch, the latest scipy trunk (7087), built against NumPy 1.5.1, passes all tests when run against the numpy trunk. The single failing test, test_imresize, fails because it tests all float types, and the new 'half' type lacks the precision to pass that test.
| | * BUG: ufunc: Fix buffered reduceMark Wiebe2011-01-272-52/+65
| | |
| | * ENH: core: Make PyArray_ArrFuncs have the same structure as in 1.5 and earlierMark Wiebe2011-01-254-78/+103
| | |
| | * ENH: core: Shuffle type numbers so any new ones go on the end.Mark Wiebe2011-01-2510-114/+212
| | |
| | * ENH: ufunc: Remove dead code from ufunc_object.cMark Wiebe2011-01-252-2046/+20
| | |
| | * ENH: ufunc: Convert Reduceat to use the new iteratorMark Wiebe2011-01-252-247/+281
| | | | | | | | | | | | | | | | | | | | | Reduceat doesn't fit into the new iterator idea quite as well as the others, so it doesn't get the matching memory layout. Neither it nor accumulate get buffering, though accumulate could use buffering if it was implemented slightly differently.
| | * ENH: ufunc: Convert Accumulate to use the new iteratorMark Wiebe2011-01-253-293/+270
| | |
| | * ENH: iter: Add NpyIter_RemoveAxis function to the iterator APIMark Wiebe2011-01-255-31/+216
| | |
| | * ENH: ufunc: Release the GIL during ufunc computationsMark Wiebe2011-01-253-2/+104
| | |
| | * ENH: ufunc: Convert UFunc Reduce to use the new iteratorMark Wiebe2011-01-253-161/+534
| | |
| | * DOC: core: Document the new einsum functionMark Wiebe2011-01-243-2/+151
| | |
| | * ENH: core: Add SSE intrinsic support to some of the einsum core loopsMark Wiebe2011-01-244-105/+1408
| | |
| | * ENH: core: EinSum function worksMark Wiebe2011-01-243-16/+319
| | |
| | * ENH: core: Implement iterator set up for einsumMark Wiebe2011-01-245-165/+632
| | |
| | * ENH: core: Start einsum function, add copyright notices to filesMark Wiebe2011-01-2311-3/+928
| | |
| | * ENH: iter: Add ability to do buffered iteration of reductionsMark Wiebe2011-01-224-113/+396
| | | | | | | | | | | | | | | This also tightens the error checking of reduction operations. Now if you want to allow a reduction, you must specify NPY_ITER_REDUCE_OK.
| | * ENH: core: Make the debug printf tracing macros more systematicMark Wiebe2011-01-223-80/+104
| | |
| | * STY: iter: Change if statements in trivial iteration macrosMark Wiebe2011-01-221-5/+3
| | |
| | * ENH: iter: Improve broadcasting error messageMark Wiebe2011-01-222-25/+155
| | |
| | * ENH: ufunc: Add support for the sig= type_tuple parameter to new ufunc codeMark Wiebe2011-01-212-35/+357
| | |
| | * ENH: ufunc: Implement user loop searchingMark Wiebe2011-01-213-138/+344
| | |
| | * ENH: ufunc: Get all tests passing with new iterator ufunc except generalized ↵Mark Wiebe2011-01-216-23/+73
| | | | | | | | | | | | ufunc tests
| | * ENH: ufunc: Change object/reference handling code to treat NULL as Py_None ↵Mark Wiebe2011-01-213-49/+30
| | | | | | | | | | | | more consistently
| | * BUG: core: Fix issue converting nested list with a number to a string typeMark Wiebe2011-01-212-41/+44
| | |
| | * BUG: iter: Fix 0-size edge case with trivial iteration macrosMark Wiebe2011-01-214-4/+25
| | |
| | * ENH: ufunc: Make many more tests pass with the new ufunc codeMark Wiebe2011-01-2110-89/+258
| | |
View Lorry Controller Status