Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Start implementing padding modes for neighborhood iterator: 0, 1 and ↵ | David Cournapeau | 2009-07-21 | 2 | -2/+77 |
| | | | | constant padding implemented. | ||||
* | Rename zero member to constant. | David Cournapeau | 2009-07-21 | 3 | -6/+6 |
| | |||||
* | Add a deprecated decorator to avoid cluttering test run output while ↵ | David Cournapeau | 2009-07-21 | 2 | -0/+111 |
| | | | | checking for raised deprecation. | ||||
* | Add our own atan2 function, as a replacement for buggy implementations (like ↵ | David Cournapeau | 2009-07-21 | 1 | -0/+89 |
| | | | | the MS C runtime). | ||||
* | BUG: npy_math_private was using some endian-related macros without including ↵ | David Cournapeau | 2009-07-21 | 1 | -4/+6 |
| | | | | the necessary header + type fixes. | ||||
* | Add a few more sanity checks for atan2. | David Cournapeau | 2009-07-21 | 1 | -0/+5 |
| | |||||
* | Add atan2 tests for special values. | David Cournapeau | 2009-07-20 | 1 | -0/+64 |
| | |||||
* | Fix to setdiff1d (and masked version) + tests (#1133, by N.C.) | Robert Cimrman | 2009-07-20 | 4 | -4/+10 |
| | |||||
* | BUG: fix C99_FORMATS-related macros definition in both numscons/distutils ↵ | David Cournapeau | 2009-07-20 | 3 | -8/+7 |
| | | | | builds. | ||||
* | BUG: fix logic OR vs. bitwise OR. | David Cournapeau | 2009-07-20 | 1 | -1/+1 |
| | |||||
* | BUG: Fix numpyconfig subst dict for the case where isnan and co are not ↵ | David Cournapeau | 2009-07-20 | 1 | -3/+4 |
| | | | | available. | ||||
* | Handle nan entries for hypot. | David Cournapeau | 2009-07-20 | 1 | -1/+11 |
| | |||||
* | Add tests for hypot function. | David Cournapeau | 2009-07-20 | 1 | -0/+19 |
| | |||||
* | Only use npy_-prefixed functions in npy_math. | David Cournapeau | 2009-07-20 | 1 | -16/+16 |
| | |||||
* | Change how we decorate npy_math functions. | David Cournapeau | 2009-07-20 | 1 | -34/+39 |
| | | | | | | | | | | | | | | | | | | | | | | | | | For platforms without the necessary math functions, we used to first implement them by ourselves, and then decorate the math function with npy_ prefix, whether the undecorated functions are from the implementation or from ourselves. This is not always practical, because we may want to use our own implementation even if the platform has one implementation. Instead, when we define our own implementation, we directly define the decorated function, and only decorate functions when we use the platform implementation. Example: say the function hypot is available on the platform, but buggy. If we undefine HAVE_HYPOT, in the former scheme, we would have first: static double hypot(double x, double y); And then: double npy_hypot(double x, double y); But this would clash with the platform hypot. By defining directly hpy_hypot, and put our implementation there, we avoid this clash. | ||||
* | Add a few more math functions we forgot to test at configuration stage. | David Cournapeau | 2009-07-20 | 1 | -1/+1 |
| | |||||
* | Remove obsolete comments. | David Cournapeau | 2009-07-20 | 1 | -5/+0 |
| | |||||
* | Fix ws issues in npy_math sources. | David Cournapeau | 2009-07-20 | 1 | -4/+4 |
| | |||||
* | Add parisc and alpha to supported cpus. Patch from armin76, fixes ticket #1178. | Charles Harris | 2009-07-19 | 2 | -6/+8 |
| | |||||
* | Add a couple of digits to the defined constants in npy_math.h. Because it makes | Charles Harris | 2009-07-19 | 1 | -36/+36 |
| | | | | me feel a tad safer with quad precision. | ||||
* | core : fixed MaskedArray.__array_finalize__ when the mask is full of False ↵ | pierregm | 2009-07-18 | 4 | -10/+110 |
| | | | | | | (bugfix #1166) extras : introduced clump_masked and clump_unmasked | ||||
* | Update new correlation support. | David Cournapeau | 2009-07-18 | 4 | -59/+48 |
| | | | | | | | | | | | | Rename PyArray_Acorrelate to PyArray_Correlate2 to follow the current NumPy convention when adding new function and keeping the old one for backward compatibility. Add keyword to correlate to switch old/new behavior instead of adding acorrelate, which is not discoverable. Raise a DeprecationWarning when the old behavior is used - old behavior still the default, though. | ||||
* | Fix #728 again. This time don't use max on a partially-ordered set. | Travis Oliphant | 2009-07-16 | 2 | -13/+21 |
| | |||||
* | Fixed #1168: more robust NoseTester package name detection (patch by Gaël ↵ | Pauli Virtanen | 2009-07-12 | 1 | -6/+11 |
| | | | | Varoquaux) | ||||
* | Fixed #1162: make matrix_power accept lists etc. as input | Pauli Virtanen | 2009-07-12 | 2 | -1/+12 |
| | |||||
* | Address #1167: faster twodim_base.diag/eye implementation by Luca Citi + tests | Pauli Virtanen | 2009-07-12 | 2 | -23/+62 |
| | |||||
* | Fix argwhere for masked arrays. | Stefan van der Walt | 2009-07-08 | 2 | -1/+5 |
| | |||||
* | Improvement to fill_diagonal suggested by Anand Patil. | Stefan van der Walt | 2009-07-08 | 1 | -1/+1 |
| | |||||
* | Use npy_fabs instead of fabs in loops.c.src . | Charles Harris | 2009-07-08 | 1 | -2/+2 |
| | |||||
* | Enhancements to arraysetops (ticket #1133, by Neil Crighton) | Robert Cimrman | 2009-07-08 | 6 | -350/+362 |
| | |||||
* | MmmKay... Let's see if complex floor divide can be made a bit more robust ↵ | Charles Harris | 2009-07-07 | 1 | -6/+4 |
| | | | | | | against some of the buildbots. | ||||
* | Make complex floor_divide more robust for underflow/overflow. | Charles Harris | 2009-07-07 | 2 | -9/+30 |
| | | | | Add test of complex floor_divide. | ||||
* | Remove debug print committed by mistake. | David Cournapeau | 2009-07-07 | 1 | -1/+0 |
| | |||||
* | Fix setupscons.py for changed locations of generated header files. | David Cournapeau | 2009-07-07 | 1 | -4/+8 |
| | |||||
* | 'Install' generated headers for in-place build support. | David Cournapeau | 2009-07-07 | 1 | -2/+10 |
| | |||||
* | Fix installed location of generated headers. | David Cournapeau | 2009-07-07 | 1 | -3/+8 |
| | | | | | | Headers and .c files generated for the C API for multiarray and ufunc are now put in build_dir/numpy/core/include/numpy instead of build_dir/numpy/core. | ||||
* | Try another test fix for complex division. The current one works on my machine | Charles Harris | 2009-07-07 | 1 | -2/+5 |
| | | | | but fails on all the buildbots. | ||||
* | Fix in-place numscons build. | David Cournapeau | 2009-07-07 | 1 | -1/+4 |
| | | | | | The generated numpyconfig.h header is 'installed' in the build directory now, so that one can build C extensions against a numpy built in-place. | ||||
* | Replace another assert_equal in test_division_umath. | Charles Harris | 2009-07-07 | 1 | -1/+1 |
| | |||||
* | Use 1 instead of 1.0, hopefully compiler will convert to correct float type. | Charles Harris | 2009-07-07 | 1 | -2/+2 |
| | |||||
* | Use type extension on fabs. | Charles Harris | 2009-07-07 | 1 | -1/+1 |
| | |||||
* | Fix test of complex division to use assert_almost_equal. | Charles Harris | 2009-07-07 | 2 | -3/+3 |
| | |||||
* | Revert last change. The problem lies elsewhere. | Charles Harris | 2009-07-07 | 2 | -4/+12 |
| | | | | Add test for complex division. | ||||
* | Oops, branches reversed in fix for ticket #1159. Do it right. | Charles Harris | 2009-07-07 | 1 | -1/+1 |
| | |||||
* | Make complex division more robust against over/under flow. Fixes ticket #1159. | Charles Harris | 2009-07-07 | 1 | -3/+12 |
| | |||||
* | Remove unnecessary global.c file. | David Cournapeau | 2009-07-05 | 2 | -4/+1 |
| | |||||
* | Fix versioning and inclusion to ref. guide in docs for #1132 | Pauli Virtanen | 2009-07-04 | 2 | -8/+41 |
| | |||||
* | Add example to savez. | Stefan van der Walt | 2009-07-04 | 1 | -0/+6 |
| | |||||
* | Add tests for dictionary interface to NpzFile. | Stefan van der Walt | 2009-07-04 | 1 | -0/+26 |
| | |||||
* | More complete dictionary interface to NpzFile [patch by David Warde-Farley] | Stefan van der Walt | 2009-07-04 | 1 | -0/+21 |
| |