summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* add -DNDEBUG for MSVCmesonMatti Picus2023-01-161-0/+4
|
* Merge branch 'main' into mesonmattip2022-12-0175-520/+768
|\
| * Merge pull request #22693 from mattip/backport-mesonRalf Gommers2022-11-302-2/+2
| |\ | | | | | | | | | | | | MAINT: unify NPY_NO_SIGNAL macros [ci skip]
| | * MAINT: unify NPY_NO_SIGNAL macrosmattip2022-11-302-2/+2
| | |
| * | Merge pull request #22644 from seberg/exceptionsMatti Picus2022-11-3017-190/+250
| |\ \ | | | | | | | | API: Add new `np.exceptions` namespace for errors and warnings
| | * | Update numpy/exceptions.pySebastian Berg2022-11-301-2/+0
| | | | | | | | | | | | Co-authored-by: Matti Picus <matti.picus@gmail.com>
| | * | MAINT: (meson) Add exceptions and expctions.pyi to pure python sourcesSebastian Berg2022-11-301-0/+2
| | | |
| | * | MAINT: Fixup new ufunc AxisError use to use `numpy.exceptions`Sebastian Berg2022-11-301-3/+2
| | | |
| | * | DOC: Update messages to NumPy 1.25 although I hope its 2.0 :)Sebastian Berg2022-11-301-3/+3
| | | |
| | * | DOC: Add release notes for `np.exceptions` namespaceSebastian Berg2022-11-301-0/+7
| | | |
| | * | DOC: Document exceptions and warnings in the refguideSebastian Berg2022-11-302-8/+41
| | | | | | | | | | | | | | | | | | | | AxisError did exist, but e.g. ComplexWarning wasn't even properly included.
| | * | TYP: Modify typing stubs for new `np.exceptions` moduleSebastian Berg2022-11-303-0/+10
| | | |
| | * | MAINT: Move ModuleDeprecationWarning and ModuleDeprecationWarningSebastian Berg2022-11-303-33/+43
| | | | | | | | | | | | | | | | Both should now live in the "exceptions" module.
| | * | API: Add new exceptions module and move exception exposed via numericSebastian Berg2022-11-3010-150/+151
| | |/ | | | | | | | | | This means moving ComplexWarning, TooHardError, and AxisError.
| * | Merge pull request #22687 from rgommers/fix-newer-setuptoolsMatti Picus2022-11-302-49/+66
| |\ \ | | | | | | | | BLD: revert adding PEP 621 metadata, it confuses setuptools
| | * | STY: satisfy linterStefan van der Walt2022-11-291-1/+2
| | | |
| | * | BLD: add workaround in setup.py for newer setuptoolsRalf Gommers2022-11-291-0/+2
| | | |
| | * | BLD: revert adding PEP 621 metadata, it confuses setuptoolsRalf Gommers2022-11-291-49/+63
| | | | | | | | | | | | | | | | See comments on gh-22663
| * | | Merge pull request #22694 from seberg/asv-quick-testCharles Harris2022-11-301-2/+2
| |\ \ \ | | |_|/ | |/| | CI: Make benchmark asv run quick to only check that benchmarks work
| | * | CI: Make benchmark asv run quick to only check that benchmarks workSebastian Berg2022-11-301-2/+2
| |/ / | | | | | | | | | | | | | | | | | | This means that benchmark results are effectively useless but I do not think we use them anyway right now. IT could be useful to have an asv running as a chron job with an in depth test. (or maybe also on request for a given PR?)
| * | Merge pull request #22689 from seberg/meson-utilsSebastian Berg2022-11-301-0/+1
| |\ \ | | | | | | | | MAINT: Add `np._utils` to meson
| | * | MAINT: Add `np._utils` to mesonSebastian Berg2022-11-301-0/+1
| |/ / | | | | | | | | | | | | | | | Odd PR timing/based on an older branch or seems to have hid, this necessary addition when adding `_util`. (Or we just missed it next to a lint failure)
| * | Merge pull request #22619 from seberg/move-set_moduleMatti Picus2022-11-2928-61/+72
| |\ \ | | | | | | | | MAINT: Move set_module from numpy.core to numpy._utils
| | * | STY: Fix indentation of _utilsSebastian Berg2022-11-251-12/+12
| | | |
| | * | MAINT: Move _inspect and _pep440 from compat to _utilsSebastian Berg2022-11-259-10/+11
| | | | | | | | | | | | | | | | | | | | Note that unfortunately, compat does expose _inspect as well, so the import remains (just the definition place moves).
| | * | MAINT: make `_utils` a full moduleSebastian Berg2022-11-253-24/+28
| | | |
| | * | MAINT: Move set_module to numpy.core to use without C importSebastian Berg2022-11-2420-52/+58
| | | |
| * | | Merge pull request #22533 from ngoldbaum/ufunc-and-function-listingSebastian Berg2022-11-296-0/+136
| |\ \ \ | | |_|/ | |/| | API: Add numpy.testing.overrides to aid testing of custom array containers
| | * | API: Add numpy.testing.overrides to aid testing of custom array containersNathan Goldbaum2022-11-166-0/+136
| | | | | | | | | | | | | | | | Closes #15544
| * | | Merge pull request #22637 from seberg/cython_long_tRalf Gommers2022-11-283-4/+15
| |\ \ \ | | | | | | | | | | API: (cython) remove `long_t` and `ulong_t`
| | * | | API: (cython) remove `long_t` and `ulong_t`Sebastian Berg2022-11-213-4/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | They are both very confusing aliases. Unfortunately, I did not find a way to give a more informative cython compilation error. Thus, I pasted the expected error message in the hope it will be easy to google. The release notes are long and maybe confusing. I really want to prepare us for switching the default integer to (probably) `intp` and this could be rather disprutpive for Cython modules if, so I thought it might be good to hint towards that, but maybe that is too much?
| * | | | Merge pull request #22685 from mattip/force-cirrusMatti Picus2022-11-283-48/+14
| |\ \ \ \ | | | | | | | | | | | | BUILD: move wheel uploads to cirrus for aarch64
| | * | | | BUILD: more cirrus CI env tweaksmattip2022-11-281-1/+3
| | | | | |
| | * | | | BUILD: tweak configuration, run set_travis_varsmattip2022-11-281-4/+6
| | | | | |
| | * | | | BUILD: disable aarch64 on travis, parse CIRRUS env variables for wheel uploadmattip2022-11-282-44/+6
| | | | | |
| * | | | | ENH: Slightly improve error when gufunc axes has wrong size (#22675)Sebastian Berg2022-11-283-17/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * DOC: Slightly improve error when gufunc axes has wrong size My hope was to tweak it into something useful that: a @= b can raise when `b` should have two dimensions and has two axes specified but actually only has one. I didn't succeed, but I still think it a slight improvement to give the ufunc name and the actual core dimensions. * ENH: Use AxisError when gufunc axes appear wrong due to the number of entries This allows catching the error relatively targeted for in-place matmul `a @= b` which may use this path. * MAINT: Restore most TypeErrors (a bit more compexl than nice, but...) * DOC: add a release note Co-authored-by: mattip <matti.picus@gmail.com>
| * | | | | MAINT: npymath cleanups for isnan, isinf, isinfinite, signbit, nextafter ↵Matti Picus2022-11-289-201/+16
| |/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (#22684) * make isnan, isinf, isfinite, signbit, nextafter aliases * fixes from review Co-authored-by: Sebastian Berg <sebastianb@nvidia.com>
| * | | | Merge pull request #22650 from andyfaff/cirrusMatti Picus2022-11-282-0/+126
| |\ \ \ \ | | | | | | | | | | | | CI: Add cirrus-ci to test linux_aarch64
| | * | | | remove TODOMatti Picus2022-11-271-1/+0
| | | | | |
| | * | | | Add encrypted tokensMatti Picus2022-11-271-3/+3
| | | | | |
| | * | | | CI: split into separate runsAndrew Nelson2022-11-261-1/+3
| | | | | |
| | * | | | CI: change to numpy/numpyAndrew Nelson2022-11-231-1/+1
| | | | | |
| | * | | | CI: enable uploadAndrew Nelson2022-11-221-9/+8
| | | | | |
| | * | | | CI: initial cirrus configAndrew Nelson2022-11-222-0/+126
| | | | | |
| * | | | | Merge pull request #22670 from MatteoRaso/masterMatti Picus2022-11-272-1/+12
| |\ \ \ \ \ | | | | | | | | | | | | | | BUG: Polynomials now copy properly (#22669)
| | * | | | | Added pickle test for polynomialsMatteo Raso2022-11-261-0/+6
| | | | | | |
| | * | | | | BUG: Polynomials now copy properly (#22669)Matteo Raso2022-11-252-1/+6
| | | |_|/ / | | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On line 502, self.symbol.copy() was called, which causes an AttributeError, since self.symbol is a string, so it doesn't have a copy() method. To fix it, I simply removed the copy() and directly assigned the string.
| * | | | | Merge pull request #22676 from seberg/issue-22672Matti Picus2022-11-272-1/+11
| |\ \ \ \ \ | | | | | | | | | | | | | | BUG: Ensure string aliases `"int0"`, etc. remain valid for now
| | * | | | | Update numpy/core/tests/test_dtype.pySebastian Berg2022-11-271-1/+1
| | | | | | | | | | | | | | | | | | | | | Co-authored-by: Matti Picus <matti.picus@gmail.com>
| | * | | | | BUG: Ensure string aliases `"int0"`, etc. remain valid for nowSebastian Berg2022-11-252-1/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | int0 and uint0 were accidentally dropped, the others just added as a test. We did successfully remove many Numeric types before, so these could probably be deprecated (it is a bit more annoying to do). These could probably just be removed, scikit-learn only noticed it in a single test (scipy probably not at all). But maybe not in 1.24
View Lorry Controller Status