@@ -6,9 +6,13 @@ NumPy 1.26.1 Release Notes
6
6
7
7
NumPy 1.26.1 is a maintenance release that fixes bugs and regressions
8
8
discovered after the 1.26.0 release. In addition, it adds new functionality for
9
- detecting BLAS and LAPACK when building from source. The 1.26.release series is
10
- the last planned minor release series before NumPy 2.0. The Python versions
11
- supported by this release are 3.9-3.12.
9
+ detecting BLAS and LAPACK when building from source. Highlights are:
10
+
11
+ - Improved detection of BLAS and LAPACK libraries for meson builds
12
+ - Pickle compatibility with the upcoming NumPy 2.0.
13
+
14
+ The 1.26.release series is the last planned minor release series before NumPy
15
+ 2.0. The Python versions supported by this release are 3.9-3.12.
12
16
13
17
14
18
Build system changes
@@ -46,9 +50,9 @@ to control BLAS/LAPACK selection and behavior:
46
50
releases.
47
51
- ``-Dallow-noblas ``: if set to ``true ``, allow NumPy to build with its
48
52
internal (very slow) fallback routines instead of linking against an external
49
- BLAS/LAPACK library. *The default for this flag may be changed to ``false ``
53
+ BLAS/LAPACK library. *The default for this flag may be changed to ``true ``
50
54
in a future 1.26.x release, however for 1.26.1 we'd prefer to keep it as
51
- ``true `` because if failures to detect an installed library are happening,
55
+ ``false `` because if failures to detect an installed library are happening,
52
56
we'd like a bug report for that, so we can quickly assess whether the new
53
57
auto-detection machinery needs further improvements. *
54
58
- ``-Dmkl-threading ``: to select the threading layer for MKL. There are four
@@ -58,3 +62,62 @@ to control BLAS/LAPACK selection and behavior:
58
62
- ``-Dblas-symbol-suffix ``: manually select the symbol suffix to use for the
59
63
library - should only be needed for linking against libraries built in a
60
64
non-standard way.
65
+
66
+
67
+ New features
68
+ ============
69
+
70
+ ``numpy._core `` submodule stubs
71
+ -------------------------------
72
+
73
+ ``numpy._core `` submodule stubs were added to provide compatibility with
74
+ pickled arrays created using NumPy 2.0 when running Numpy 1.26.
75
+
76
+
77
+ Contributors
78
+ ============
79
+
80
+ A total of 13 people contributed to this release. People with a "+" by their
81
+ names contributed a patch for the first time.
82
+
83
+ * Andrew Nelson
84
+ * Anton Prosekin +
85
+ * Charles Harris
86
+ * Chongyun Lee +
87
+ * Ivan A. Melnikov +
88
+ * Jake Lishman +
89
+ * Mahder Gebremedhin +
90
+ * Mateusz Sokół
91
+ * Matti Picus
92
+ * Munira Alduraibi +
93
+ * Ralf Gommers
94
+ * Rohit Goswami
95
+ * Sayed Adel
96
+
97
+
98
+ Pull requests merged
99
+ ====================
100
+
101
+ A total of 20 pull requests were merged for this release.
102
+
103
+ * `#24742 <https://github.com/numpy/numpy/pull/24742 >`__: MAINT: Update cibuildwheel version
104
+ * `#24748 <https://github.com/numpy/numpy/pull/24748 >`__: MAINT: fix version string in wheels built with setup.py
105
+ * `#24771 <https://github.com/numpy/numpy/pull/24771 >`__: BLD, BUG: Fix build failure for host flags e.g. ``-march=native ``...
106
+ * `#24773 <https://github.com/numpy/numpy/pull/24773 >`__: DOC: Updated the f2py docs to remove a note on -fimplicit-none
107
+ * `#24776 <https://github.com/numpy/numpy/pull/24776 >`__: BUG: Fix SIMD f32 trunc test on s390x when baseline is none
108
+ * `#24785 <https://github.com/numpy/numpy/pull/24785 >`__: BLD: add libquadmath to licences and other tweaks (#24753)
109
+ * `#24786 <https://github.com/numpy/numpy/pull/24786 >`__: MAINT: Activate ``use-compute-credits `` for Cirrus.
110
+ * `#24803 <https://github.com/numpy/numpy/pull/24803 >`__: BLD: updated vendored-meson/meson for mips64 fix
111
+ * `#24804 <https://github.com/numpy/numpy/pull/24804 >`__: MAINT: fix licence path win
112
+ * `#24813 <https://github.com/numpy/numpy/pull/24813 >`__: BUG: Fix order of Windows OS detection macros.
113
+ * `#24831 <https://github.com/numpy/numpy/pull/24831 >`__: BUG, SIMD: use scalar cmul on bad Apple clang x86_64 (#24828)
114
+ * `#24840 <https://github.com/numpy/numpy/pull/24840 >`__: BUG: Fix DATA statements for f2py
115
+ * `#24870 <https://github.com/numpy/numpy/pull/24870 >`__: API: Add ``NumpyUnpickler `` for backporting
116
+ * `#24872 <https://github.com/numpy/numpy/pull/24872 >`__: MAINT: Xfail test failing on PyPy.
117
+ * `#24879 <https://github.com/numpy/numpy/pull/24879 >`__: BLD: fix math func feature checks, fix FreeBSD build, add CI...
118
+ * `#24899 <https://github.com/numpy/numpy/pull/24899 >`__: ENH: meson: implement BLAS/LAPACK auto-detection and many CI...
119
+ * `#24902 <https://github.com/numpy/numpy/pull/24902 >`__: DOC: add a 1.26.1 release notes section for BLAS/LAPACK build...
120
+ * `#24906 <https://github.com/numpy/numpy/pull/24906 >`__: MAINT: Backport ``numpy._core `` stubs. Remove ``NumpyUnpickler ``
121
+ * `#24911 <https://github.com/numpy/numpy/pull/24911 >`__: MAINT: Bump pypa/cibuildwheel from 2.16.1 to 2.16.2
122
+ * `#24912 <https://github.com/numpy/numpy/pull/24912 >`__: BUG: loongarch doesn't use REAL(10)
123
+
0 commit comments