summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* REL: Numpy 1.11.1 release.v1.11.1Charles Harris2016-06-253-5/+8
|
* Merge pull request #7737 from charris/backport-7735Charles Harris2016-06-122-7/+14
|\ | | | | Backport 7735, BUG: fix issue on OS X with Python 3.x, npymath.ini not installed.
| * BUG: fix issue on OS X with Python 3.x where npymath.ini was not installed.Ralf Gommers2016-06-122-7/+14
|/ | | | | Closes gh-7707. Happened for `pip install .` or another location (including a remote git repo).
* Merge pull request #7731 from charris/backport-7723Charles Harris2016-06-111-1/+1
|\ | | | | Backport 7723, Change mkl_info.dir_env_var from MKL to MKLROOT
| * Change mkl_info.dir_env_var from MKL to MKLROOTRay Donnelly2016-06-111-1/+1
|/ | | | Reference: https://software.intel.com/en-us/node/528500
* Merge pull request #7724 from charris/backport-7719Charles Harris2016-06-102-1/+34
|\ | | | | Backport 7719, BUG: Fix segfaults in np.random.shuffle
| * BUG: Fix segfaults in np.random.shuffleSimon Gibbons2016-06-102-1/+34
|/ | | | | | | | | | | | | | | | np.random.shuffle will allocate a buffer based on the size of the first element of an array of strings. If the first element is smaller than another in the array this buffer will overflow, causing a segfault when garbage is collected. Additionally if the array contains objects then one would be left in the buffer and have it's refcount erroniously decrimented on function exit, causing that object to be deallocated too early. To fix this we change the buffer to be an array of int8 of the the size of the array's dtype, which sidesteps both issues. Fixes #7710
* Merge pull request #7682 from charris/backport-7681Charles Harris2016-05-261-20/+20
|\ | | | | Backport 7681, DOC: Update 1.11.1 release notes.
| * DOC: Update 1.11.1 release notes.Charles Harris2016-05-261-20/+20
|/ | | | [ci skip]
* Merge pull request #7680 from charris/backport-7678Charles Harris2016-05-262-0/+13
|\ | | | | Backport 7678, BUG: Fix np.clip bug NaN handling for Visual Studio 2015
| * MAINT: Fix up C++ style comment in arraytypes.c.src.Charles Harris2016-05-261-6/+7
| |
| * Fix np.clip bug NaN handling for Visual Studio 2015Loïc Estève2016-05-262-0/+12
|/ | | | by disabling loop vectorization and add test.
* Merge pull request #7677 from charris/backport-7673Charles Harris2016-05-252-0/+29
|\ | | | | Backport 7673, DOC: Create Numpy 1.11.1 release notes.
| * DOC: Create Numpy 1.11.1 release notes.Charles Harris2016-05-252-0/+29
| | | | | | | | [ci skip]
* | Merge pull request #7676 from charris/backport-7675Charles Harris2016-05-252-3/+8
|\ \ | |/ |/| Backport 7675, BUG: fix handling of right edge of final bin.
| * BUG: fix handling of right edge of final bin.Robert Kern2016-05-252-3/+8
|/
* Merge pull request #7671 from charris/backport-7669Charles Harris2016-05-242-2/+20
|\ | | | | Backport 7669, BUG: boolean assignment no GIL release when transfer needs API
| * BUG: boolean assignment no GIL release when transfer needs APISebastian Berg2016-05-242-2/+20
| | | | | | | | | | | | | | This caused bugs when the iteration does not need the API, but the data copy does. closes gh-7666
* | Merge pull request #7670 from charris/backport-7667Charles Harris2016-05-242-6/+28
|\ \ | |/ |/| Backport 7667, BUG: correct initial index estimate in histogram.
| * ENH: correct initial index estimate in histogram.Robert Kern2016-05-242-7/+18
| |
| * TST: Failing test for histogram.Robert Kern2016-05-241-1/+12
| |
* | Merge pull request #7665 from charris/backport-7658Charles Harris2016-05-232-3/+17
|\ \ | |/ |/| Backport 7658, BUG: fix incorrect printing of 1D masked arrays
| * BUG: fix incorrect printing of 1D masked arraysSimon Conseil2016-05-232-3/+17
|/ | | | | | | | | | Ref #7621. #6748 added `np.ma.MaskedArray._print_width` which is used to cut a masked array before printing it (to save memory and cpu time during the conversion to the object dtype). But this doesn't work correctly for 1D arrays, for which up to 1000 values can be printed before cutting the array. So this commit adds a new class variable `_print_width_1d` to handle the 1D case separately.
* Merge pull request #7660 from ahaldane/fix_7493_1.11Charles Harris2016-05-232-1/+12
|\ | | | | Backport 7659, BUG: Temporary fix for str(mvoid) for object field types
| * BUG: Temporary fix for str(mvoid) for object field typesAllan Haldane2016-05-222-1/+12
|/ | | | | | | | | Fixes #7493, using a temporary hack, to be properly fixed later (eg with #6053) Printing a Masked-Void instance broke if the instance has a field of Object dtype because assignment involving structured dtypes with objects doesn't work. Fix is to use dtype-transfer code which avoid the bug.
* Merge pull request #7656 from charris/backport-7655Charles Harris2016-05-222-4/+4
|\ | | | | Backport 7655, BLD: Remove Intel compiler flag -xSSE4.2
| * BLD: Remove Intel compiler flag -xSSE4.2Charles Harris2016-05-222-4/+4
| | | | | | | | | | | | | | The consensus seems to be that hardcoding SSE4.2 results in poor code for architectures lacking the feature. Closes #7287.
* | Merge pull request #7654 from charris/backport-7635Charles Harris2016-05-212-3/+20
|\ \ | |/ |/| Backport 7635, BUG: ma.median of 1d array should return a scalar
| * STY: ma.extras.median: avoid indexing with listAmit Aronovitch2016-05-211-2/+2
| | | | | | | | Recommended type for nd-indexing is a tuple. See #4434
| * BUG: ma.median of 1d array should return a scalarAmit Aronovitch2016-05-212-1/+18
|/ | | | | | Fixes #5969. Performance fix #4760 had caused wrong shaped results in the 1D case. This fix restores the original 1D behavior.
* Merge pull request #7638 from charris/backport-7634Charles Harris2016-05-151-3/+5
|\ | | | | Backport 7634, BLD: correct C compiler customization in system_info.py
| * BLD: correct C compiler customization in system_info.py Closes gh-7606.Ralf Gommers2016-05-141-3/+5
|/
* Merge pull request #7608 from charris/backport-7587Charles Harris2016-05-062-2/+48
|\ | | | | Backport 7587, BUG: linalg.norm(): Don't convert object arrays to float
| * TST: More tests for linalg.norm() with object arraysMatthias Geier2016-05-061-1/+41
| |
| * BUG: linalg.norm(): Don't convert object arrays to floatMatthias Geier2016-05-062-2/+8
|/ | | | This fixes #7575 (Regression in linalg.norm() using dtype=object).
* Merge pull request #7597 from charris/backport-7586Charles Harris2016-05-022-3/+9
|\ | | | | Backport 7586, Make np.ma.take works on scalars
| * BUG: Ensure mask is preserved on scalarsEric Wieser2016-05-021-3/+5
| | | | | | | | | | | | By promoting and demoting between scalars and arrays where appropriate See #7585
| * TST: Verify np.ma.take works on scalarsEric Wieser2016-05-021-0/+4
|/ | | | See #7585
* Merge pull request #7590 from charris/backport-7584Charles Harris2016-04-293-9/+12
|\ | | | | Backport 7584, BLD: fix configparser.InterpolationSyntaxError
| * BLD: fix configparser.InterpolationSyntaxErrorSorin Sbarnea2016-04-293-9/+12
|/ | | | Closes #7572 inability to install in virtualenvs with percent in their path.
* Merge pull request #7578 from charris/backport-7568Charles Harris2016-04-271-7/+14
|\ | | | | Backport 7568, Fix OverflowError in Python 3.x. in swig interface.
| * BUG: Fix OverflowError in Python 3.x in swig interface.Mathieu Lamarre2016-04-271-7/+14
|/ | | | | | | | | | | | | | | The error occurs When values above 0x7FFFFFF are passed to a function accepting "unsigned int". This a port of a fix in pyprimtype.swg from which several code snippets where copy pasted into swig/pyfragments.swg. Please see SWIG changes log (2015-12-23) for more details: http://www.swig.org/Release/CHANGES.current 2015-12-23: ahnolds [Python] Fixes for conversion of signed and unsigned integer types ...
* Merge pull request #7558 from charris/backport-7358Charles Harris2016-04-182-1/+14
|\ | | | | Backport 7358, BUG: constant padding expected wrong type in constant_values
| * BUG: constant padding expected wrong type in constant_valueschiffa2016-04-181-1/+1
| | | | | | | | | | Constant padding on 4 sides of a 2d array expected a numpy ndarray, and not a ndarray like (tuple, list, ...) Detailed description is in the issue #7353
| * TST: added a test for constant padding on 4 sides of a 2d arraychiffa2016-04-181-0/+13
|/ | | | This test exposes padding bug described in the issue #7353
* Merge pull request #7551 from charris/backport-7549Charles Harris2016-04-151-6/+30
|\ | | | | Backport 7549, BUG: allow graceful recovery for no compiler
| * BUG: allow graceful recovery for no compilerMatthew Brett2016-04-151-6/+30
|/ | | | | | | | | | If there is no compiler on Linux, the error we were getting was CompileError rather than the OSError the test was expecting. This had the nasty side-effect of leaving us in a deleted temporary directory, causing later test failures. Try a check to see if we have a compiler and skip otherwise.
* Merge pull request #7535 from charris/backport-7518Charles Harris2016-04-101-2/+2
|\ | | | | Backport 7518, BUG: Extend glibc complex trig functions blacklist to glibc < 2.18.
| * BUG: Extend glibc complex trig functions blacklist to glibc < 2.18.Nikola Forró2016-04-091-2/+2
|/ | | | | | | The library complex trig functions are inaccurate also in glibc versions 2.16 and 2.17, so extend the blacklist. Closes #7517.
* Merge pull request #7530 from charris/backport-7529Charles Harris2016-04-082-5/+11
|\ | | | | Backport 7529, BUG: Floating exception with invalid axis in np.lexsort
View Lorry Controller Status