summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
* BUG: Fix check for singleton dimensions in einsum.Ryan Soklaski2018-06-011-1/+1
| | | | | Was checking array elements rather than array shapes for singleton dimensions.
* TST: Test einsum optimize broadcasting errorRyan Soklaski2018-06-011-3/+11
|
* Merge pull request #11211 from charris/backport-11193Charles Harris2018-05-312-2/+5
|\ | | | | BUG: Fix reference count/memory leak exposed by better testing
| * BUG: reference count exposed by better testingMarten van Kerkwijk2018-05-312-2/+5
| | | | | | | | | | | | Also for __array_ufunc__ overrides, sig and signature cannot be passed in both. This was incorrectly coded and slipped through as it was not tested.
* | BUG: delimiter/comments in genfromtxt should be encodedAllan Haldane2018-05-302-0/+9
|/ | | | Fixes #11028
* BUG: reduce using SSE only warns if inside SSE loopmattip2018-05-303-3/+10
|
* Merge pull request #11198 from charris/backport-11036Charles Harris2018-05-3010-147/+103
|\ | | | | BUG: optimizing compilers can reorder call to npy_get_floatstatus
| * BUG: optimizing compilers can reorder call to npy_get_floatstatusCharles Harris2018-05-3010-147/+103
| | | | | | | | | | | | | | | | We should find a more generic and explicit way to prevent optimizing compilers from reordering the call to npy_get_floatstatus. To reproduce the problem, clang or gcc-8.1 are required. Confirmed that this fixes the problem using clang-6.0
* | BUG: Python2 doubles don't print correctly in interactive shellAllan Haldane2018-05-292-10/+44
|/ | | | Fixes #11031
* Merge pull request #11104 from ahaldane/fixup_hack_double_double_1.14.4Charles Harris2018-05-291-8/+43
|\ | | | | BUG: str of DOUBLE_DOUBLE format wrong on ppc64
| * BUG: str of DOUBLE_DOUBLE format is not implementedAllan Haldane2018-05-151-8/+43
| | | | | | | | | | | | This commit implements backup code that at least prints the correct numerical value, but not always formatted correctly. Fixes #11103
* | BUG: void dtype setup checked offset not actual pointer for alignmentSebastian Berg2018-05-281-17/+18
| | | | | | | | This commit also adds an assert and closes gh-11088
* | Merge pull request #11174 from charris/backport-11169Charles Harris2018-05-271-0/+24
|\ \ | | | | | | MAINT: add sanity-checks to be run at import time
| * | MAINT: add sanity-checks to be run at import timePauli Virtanen2018-05-271-0/+24
| | | | | | | | | | | | This checks for potential BLAS issues, which are useful to catch early.
* | | Merge pull request #11170 from charris/backport-11168Charles Harris2018-05-271-1/+36
|\ \ \ | |_|/ |/| | TST: linalg: add regression test for gh-8577
| * | TST: linalg: add regression test for gh-8577Pauli Virtanen2018-05-271-1/+36
| |/ | | | | | | | | Add regression test that checks for certain bugs where results from sdot change if certain libraries are imported first.
* | MAINT: `multiarray_tests` was renamed in master.backport_10824Charles Harris2018-04-251-1/+1
| | | | | | Use the old name in the backport.
* | STY: Fix long line.Charles Harris2018-04-251-1/+2
| |
* | fix from reviewmattip2018-04-252-16/+18
| |
* | BUG: test, fix PyArray_DiscardWritebackIfCopy refcount issue and documentmattip2018-04-253-2/+45
|/
* Merge pull request #10947 from ahaldane/fix_style_arg_1.14Charles Harris2018-04-242-0/+5
|\ | | | | BUG: 'style' arg to array2string broken in legacy mode (1.14 Backport)
| * BUG: 'style' arg to array2string broken in legacy modeAllan Haldane2018-04-212-0/+5
| | | | | | | | Fixes #10934
* | Merge pull request #10962 from ahaldane/backport_10860Allan Haldane2018-04-232-2/+39
|\ \ | | | | | | (Backport 1.14.3) BUG: core: fix NPY_TITLE_KEY macro on pypy
| * | BUG: core: fix NPY_TITLE_KEY macro on pypyPauli Virtanen2018-04-232-2/+39
| |/ | | | | | | | | | | | | | | | | | | | | | | On Pypy, dictionary keys do not necessarily preserve object identity. This however was assumed by the NPY_TITLE_KEY macro, which relies on descriptor.c:568 using the same 'title' object both as a dictionary key as an entry in the tuple inserted. Since the items in the field dict are unique, value identity is however sufficient for the NPY_TITLE_KEY macro. On PyPy, fix the macro by comparing values instead.
* | Merge pull request #10961 from ahaldane/backport_10869Allan Haldane2018-04-231-0/+2
|\ \ | | | | | | (Backport 1.14.3) BUG: Fix encoding regression in ma/bench.py (Issue #10868)
| * | BUG: Fix encoding regression in ma/bench.py (Issue #10868)Jonathan March2018-04-231-0/+2
| |/
* | Merge pull request #10960 from ahaldane/backport_10916Allan Haldane2018-04-231-1/+1
|\ \ | | | | | | (Backport 1.14.3) Add missing underscore to prototype in check_embedded_lapack
| * | Add missing underscore to prototype in check_embedded_lapackMalcolm Smith2018-04-231-1/+1
| |/
* | BUG: test, fix for missing flags['WRITEBACKIFCOPY'] keymattip2018-04-232-1/+3
|/
* BUG: for 1.14 back-compat, accept list-of-lists in fromrecordsAllan Haldane2018-04-132-24/+40
| | | | Fixes #10870
* BUG: floating types should override tp_printAllan Haldane2018-04-082-0/+53
| | | | Fixes #10753
* Merge pull request #10727 from charris/backport-10650Charles Harris2018-03-112-2/+2
|\ | | | | BUG: Fix missing NPY_VISIBILITY_HIDDEN on npy_longdouble_to_PyLong
| * BUG: Fix missing NPY_VISIBILITY_HIDDEN on npy_longdouble_to_PyLongEric Wieser2018-03-112-2/+2
| | | | | | | | | | | | Fixes #10648 NPY_VISIBILITY_HIDDEN is used by memoverlap.{c,h}, so we should be using the same thing here
* | Merge pull request #10726 from charris/backport-10657Charles Harris2018-03-111-5/+25
|\ \ | | | | | | BUG: Fix f2py generated code to work on PyPy
| * | BUG: f2py: fix f2py generated code to work on PypyPauli Virtanen2018-03-111-5/+25
| |/ | | | | | | | | | | | | | | F2py generates code that uses PyTuple_SetItem on a tuple that has been "used", which is invalid on Pypy. Add #ifdefs that make the generated code convert the object to a list, use PyList_SetItem, and then convert it back to a tuple.
* | BUG: dragon4 fractional output mode adds too many trailing zerosAllan Haldane2018-03-112-27/+38
|/ | | | Fixes #10713
* BUG: Further back-compat fix for subclassed array reprAllan Haldane2018-02-272-13/+47
| | | | Fixes #10663
* BUG: break cyclic refs in recursive closuresAllan Haldane2018-02-184-9/+37
| | | | Fixes #10620
* Revert "Merge pull request #10622 from ↵Allan Haldane2018-02-182-103/+88
| | | | | | | ahaldane/fix_arrayprint_recursive_closure_backport" This reverts commit d0ba54f958efe07d3a1fc94bf61d3723ef89c267, reversing changes made to eaac472cbe4ecb1d5b825c6b77b97bf0a11865a9.
* BUG: Correctly identify comma seperated dtype stringsSimon Gibbons2018-02-182-1/+6
| | | | | | | | | | | | | When parsing dtype strings, we should only consider them to be comma seperated if there are commas not present in a pair of square brackets. Whilst we had a check for this already in the code there was an off by 1 bug where we failed to consider the first character of the string. This would lead to an infinite recursion when trying to parse strings of the form `[i8,f8]`. Fixes: #10440
* BUG: deallocate recursive closure in arrayprint.pyAllan Haldane2018-02-172-88/+103
| | | | Fixes #10620
* Merge pull request #10612 from charris/backport-10381Charles Harris2018-02-163-15/+48
|\ | | | | BUG/ENH: Improve output for structured non-void types
| * ENH: Always show dtype fields in the array repr, even for non-voidEric Wieser2018-02-162-7/+13
| |
| * BUG: Fix crash on non-void structured array reprEric Wieser2018-02-162-7/+29
| | | | | | | | Fixes gh-9821
| * BUG: Show the base of a compound dtype even when it doesn't subclass voidEric Wieser2018-02-162-2/+7
| |
* | Merge pull request #10610 from charris/backport-10477Charles Harris2018-02-1611-819/+256
|\ \ | | | | | | BUG: Align type definition with generated lapack
| * | BUG: Align type definition with generated lapack (#10477)xoviat2018-02-1611-819/+256
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * Align type definitions * Regenerate sources * Replace BytesIO * Consolidate executables * Create directories on PY2 * Revise step name * Consolidate directory creation * Don't catch makedirs errors * Revise step name * Add header source
* | Merge pull request #10608 from charris/backport-10588Charles Harris2018-02-162-15/+30
|\ \ | | | | | | BUG: Revert sort optimization in np.unique.
| * | BUG: Revert sort optimization in np.unique.Charles Harris2018-02-162-15/+30
| |/ | | | | | | | | | | | | | | | | | | Backport of #10588. The optimization was to sort integer subarrays by treating them as strings of unsigned bytes. That worked fine for finding the unique subarrays, but the sort order of the results could be unexpected. Closes #10495.
* | BUG: infinite recursion in str of 0d subclassesAllan Haldane2018-02-162-6/+63
|/ | | | Fixes #10360
View Lorry Controller Status