summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #12004 from charris/fix-failing-f2py-testCharles Harris2018-09-212-105/+112
|\ | | | | BUG: Fix f2py compile function testing.
| * BUG: Fix f2py test_compile_function.py tests.Charles Harris2018-09-211-74/+61
| | | | | | | | | | | | The tests were failing on platforms without a fortran compiler as that was not being checked for. Add that check among others and clean up the test functions a bit.
| * TST, MAINT: Separate f2py compile function testing into own fileCharles Harris2018-09-202-105/+125
| | | | | | | | | | | | | | Create new file `test_compile_function.py` for the test, transferring the content from `test_quoted_characters.py` where it was before. The tricky part here is maintaining the history in both files after the move.
* | Merge pull request #12008 from eric-wieser/cast-commentsMatti Picus2018-09-211-5/+3
|\ \ | | | | | | MAINT: Reassociate `np.cast` with the comment describing it
| * | MAINT: Reassociate `np.cast` with the comment describing itEric Wieser2018-09-201-5/+3
| | |
* | | Merge pull request #12009 from eric-wieser/actually-remove-_typestrMatti Picus2018-09-212-32/+0
|\ \ \ | |/ / |/| | MAINT: Eliminate the private `numerictypes._typestr`
| * | MAINT: Eliminate the private `numerictypes._typestr`Eric Wieser2018-09-202-33/+0
| | | | | | | | | | | | | | | This was used internally only by a private function that is also unused. The loop adding type aliases was shown by #12007 to have no effect.
| * | MAINT: Add a RuntimeError to prove `_typestr` is unusedEric Wieser2018-09-201-0/+1
|/ /
* | Merge pull request #12002 from numpy/revert-11934-revert-11693-accept-nep18Stephan Hoyer2018-09-201-1/+2
|\ \ | | | | | | Revert "NEP: Revert "NEP: Mark NEP 18 as accepted""
| * | Revert "NEP: Revert "NEP: Mark NEP 18 as accepted""Stephan Hoyer2018-09-201-1/+2
|/ /
* | Merge pull request #11997 from eric-wieser/string-helpersMarten van Kerkwijk2018-09-202-97/+105
|\ \ | | | | | | MAINT: Extract string helpers to a new private file
| * | MAINT: Extract string helpers to a new private fileEric Wieser2018-09-202-97/+105
| | |
* | | Merge pull request #11996 from eric-wieser/dtype-testsCharles Harris2018-09-201-35/+43
|\ \ \ | |/ / |/| | TST: Parametrize and break apart dtype tests
| * | TST: Parametrize and break apart dtype testsEric Wieser2018-09-201-35/+43
|/ / | | | | | | Now that we're using parametrize, may as well drop it in a few more places
* | Merge pull request #10998 from tylerjereddy/numeric_ops_goto_removalCharles Harris2018-09-191-1/+1
|\ \ | |/ |/| DOC: removed spurious FIXME comment in number.c
| * DOC: replaced spurious FIXME comment in number.cTyler Reddy2018-08-171-1/+1
| |
* | Merge pull request #11989 from mhvk/subclass-indexing-base-setting-fixCharles Harris2018-09-192-7/+25
|\ \ | | | | | | BUG: Ensure boolean indexing of subclasses sets base correctly.
| * | BUG: Ensure boolean indexing of subclasses sets base correctly.Marten van Kerkwijk2018-09-192-7/+25
|/ / | | | | | | | | | | | | | | | | Corrects a mistake in gh-11246, where the base for this case was set to the indexed subclass instance (like for a view). Since a copy is made in this case, the base should be the new ndarray that holds the indexed values. Thanks to @yymao for noticing and reporting the bug in astropy.
* | DOC: add docstrings for numeric types (#11858)Yannick Jadoul2018-09-184-48/+150
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * DOC: add docstring for np.float16 (see #10106) * DOC: add docstrings for uint8, uint16, uint32, uint64 (see #10106) * DOC: make numeric types' docstrings char codes formatting consistent * DOC: Adding documentation to C/Python compatible names of scalar types * DOC: Adding platform-dependent aliases to scalar type docstrings * DOC: Fix typos and minor issues in scalar type docstrings * DOC: Add hard-coded aliases to scalar type docstrings * DOC: Fix platform-dependent aliases in doc string of scalar types * DOC: Fix indentation of numeric types' docstring * TST: Add test checking for platform-dependent generation of numeric type docstrings * TST: Fix docstring test failures when PYTHONOPTIMIZE/Py_OptimizeFlag > 1 * TST: Moving check of PYTHONOPTIMIZE for test_numerictypes.py::TestDocStrings into pytest.mark.skip decorator * DOC: Cleanup code for adding docstrings to numeric types and mention canonical type name * DOC: Cleaning up access of numeric type objects and character codes in dynamic docstring generation * DOC: Prepend 'np.' to numeric type docstrings * MAINT: Remove dead code, rename function * DOC: Cleaning up numeric_type_aliases in core/_add_newdocs.py * DOC: Fixing dynamic aliases in docstrings for complex scalar types * DOC: Removing docstrings for complex scalar types in numpy/core/src/multiarray/scalartypes.c.src, as they are now set in numpy/core/_add_newdocs.py * DOC: Merging lists of different dynamic aliases into one, and prepending 'np.' to fixed aliases * DOC: Adding scalar type docstring improvements to release notes * DOC: Tweak release notes
* | Merge pull request #11961 from eric-wieser/pad-testsCharles Harris2018-09-181-1/+42
|\ \ | | | | | | TST: Add more tests for np.pad
| * | TST: Add a (failing) test for using np.pad with Fractions, which currently ↵Eric Wieser2018-09-171-0/+19
| | | | | | | | | | | | rounds to `float`
| * | TST: Add a regression test for gh-11216Lars G2018-09-171-1/+18
| | | | | | | | | | | | The test is marked xfail right now as it is not fixed in master
| * | TST: Add a test for pad-wrapping by length 0Lars G2018-09-151-0/+5
| | |
* | | Merge pull request #11973 from juliantaylor/thread-allocCharles Harris2018-09-182-11/+20
|\ \ \ | | | | | | | | BUG: fix cached allocations without the GIL
| * | | BUG: fix cached allocations without the GILJulian Taylor2018-09-182-11/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There currently only is a global cache for small allocations so the functions must be called while the GIL is held. Ensure this by checking the GIL state in debug mode (which is tested via a ci configuration). Closes gh-11942
* | | | Merge pull request #11976 from eric-wieser/tweak-block-error-checksCharles Harris2018-09-181-16/+35
|\ \ \ \ | | | | | | | | | | MAINT/DOC: Show the location of an empty list in np.block
| * | | | MAINT: Show the location of an empty list in np.blockEric Wieser2018-09-181-10/+22
| | | | |
| * | | | DOC: Use a standard-style docstring to explain `_block_check_depths_match`Eric Wieser2018-09-171-9/+16
| | | | |
* | | | | Merge pull request #11937 from tylerjereddy/rm_exec_command_7Charles Harris2018-09-182-11/+152
|\ \ \ \ \ | |/ / / / |/| | | | MAINT: remove exec_command() from f2py init
| * | | | MAINT: remove exec_command() from f2py initTyler Reddy2018-09-182-11/+152
|/ / / / | | | | | | | | | | | | | | | | * replaced the usage of exec_command() in the f2py init with standard library equivalent
* | | | Merge pull request #11935 from tylerjereddy/rm_exec_command_6Charles Harris2018-09-171-6/+27
|\ \ \ \ | | | | | | | | | | MAINT: remove usage of exec_command in config.py
| * | | | MAINT: remove usage of exec_command in config.pyTyler Reddy2018-09-121-6/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * replace usage of exec_command() with standard library equivalent in the distutils config module
* | | | | Merge pull request #11958 from hmaarrfk/test_pad_objectMatti Picus2018-09-171-0/+15
|\ \ \ \ \ | | | | | | | | | | | | TST: Add a test for np.pad where constant_values is an object
| * | | | | styleMark Harfouche2018-09-151-1/+1
| | | | | |
| * | | | | Pad start and end with different valuesMark Harfouche2018-09-151-3/+5
| | | | | |
| * | | | | Don't use None to valid the array as it is the "Default" for emptyMark Harfouche2018-09-151-5/+8
| | | | | |
| * | | | | MAINT: Add a test for np.pad where constant_values is an objectMark Harfouche2018-09-151-0/+10
| | |_|/ / | |/| | |
* | | | | MAINT: refactor design of recursive closures (#11910)Matti Picus2018-09-165-21/+55
| | | | |
* | | | | Merge pull request #11967 from mattip/PyArray_AdaptFlexibleDType-leakCharles Harris2018-09-161-6/+8
|\ \ \ \ \ | | | | | | | | | | | | BUG: fix refcount leak in PyArray_AdaptFlexibleDType
| * | | | | STY: fix from reviewmattip2018-09-171-1/+2
| | | | | |
| * | | | | BUG: fix refcount leak in PyArray_AdaptFlexibleDTypemattip2018-09-161-6/+7
|/ / / / /
* | | | | MAINT: remove exec_command usage in ibm.py (#11901)Tyler Reddy2018-09-151-4/+9
| | | | | | | | | | | | | | | | | | | | Replaced the usage of exec_command() in distutils ibm module with subprocess.check_output
* | | | | Merge pull request #11953 from tylerjereddy/issue_11947Charles Harris2018-09-151-0/+16
|\ \ \ \ \ | |/ / / / |/| | | | TST: codecov.yml improvements
| * | | | TST: codecov.yml improvementsTyler Reddy2018-09-141-0/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * we no longer require CircleCI jobs to pass for coverage reporting, and failures of Travis jobs other than the single coverage job should not interfere with coverage reporting now
* | | | | Merge pull request #11952 from jlmuir/npy-internal-build-define-fixCharles Harris2018-09-141-1/+1
|\ \ \ \ \ | |/ / / / |/| | | | MAINT: test NPY_INTERNAL_BUILD only if defined
| * | | | MAINT: test NPY_INTERNAL_BUILD only if definedJ. Lewis Muir2018-09-141-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When using GCC 4.4.7 on RHEL 6 to compile a C program that embeds Python and includes numpy/core/include/numpy/npy_common.h, and the compiler is configured to warn about all questionable constructions (-Wall) and warn about undefined identifiers (-Wundef), it will emit a warning like /tmp/python-2.7.15/lib/python2.7/site-packages/numpy/core/include/numpy/npy_common.h:17:5: warning: "NPY_INTERNAL_BUILD" is not defined The warning is because npy_common.h contains #if NPY_INTERNAL_BUILD #ifndef NPY_INLINE_MATH #define NPY_INLINE_MATH 1 #endif #endif Eliminate this warning by changing the #if expression to test NPY_INTERNAL_BUILD only if it is defined. (This preserves the original behavior because in the original, when NPY_INTERNAL_BUILD is undefined, the preprocessor treats it as 0 (false).)
* | | | | Merge pull request #11948 from shoyer/finish-nat-deprecationCharles Harris2018-09-143-83/+40
|\ \ \ \ \ | | | | | | | | | | | | DEP: finish making all comparisons to NaT false
| * | | | | DOC: release notes for expired NaT depreactionStephan Hoyer2018-09-141-1/+14
| | | | | |
| * | | | | DEP: finish making all comparisons to NaT falseStephan Hoyer2018-09-132-82/+26
| |/ / / / | | | | | | | | | | | | | | | Closes gh-11945
* | | | | Merge pull request #11943 from shoyer/nep-18-warn-about-changesStephan Hoyer2018-09-141-13/+41
|\ \ \ \ \ | | | | | | | | | | | | DOC: Add warning/clarification about backwards compat in NEP-18
View Lorry Controller Status