Script 'mail_helper' called by obssrc
Hello community,

here is the log from the commit of package python-numpy for openSUSE:Factory 
checked in at 2026-05-13 17:18:51
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/python-numpy (Old)
 and      /work/SRC/openSUSE:Factory/.python-numpy.new.1966 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Package is "python-numpy"

Wed May 13 17:18:51 2026 rev:146 rq:1352728 version:2.4.4

Changes:
--------
--- /work/SRC/openSUSE:Factory/python-numpy/python-numpy.changes        
2026-03-27 06:48:25.815310380 +0100
+++ /work/SRC/openSUSE:Factory/.python-numpy.new.1966/python-numpy.changes      
2026-05-13 17:19:23.467432200 +0200
@@ -1,0 +2,25 @@
+Mon May 11 16:06:40 UTC 2026 - John Paul Adrian Glaubitz 
<[email protected]>
+
+- Force "none" cpu-baseline on x86_64 and %{ix86} only
+- Ignore test_cpu_features test failure on s390x
+- Ignore test_einsum test failure on %{ix86}
+
+-------------------------------------------------------------------
+Wed Apr 15 11:38:04 UTC 2026 - Markéta Machová <[email protected]>
+
+- Update to 2.4.4
+  * np.linalg.norm returned different value after update to 2.4.2
+  * fix FNV-1a 64-bit selection by using NPY_SIZEOF_UINTP
+  * avoid warning on ufunc with where=True and no output
+  * document caveats of ndarray.resize on 3.14 and newer
+  * fix POWER VSX feature mapping
+  * numpy.i: Replace deprecated sprintf with snprintf
+
+-------------------------------------------------------------------
+Mon Mar 30 22:31:27 UTC 2026 - Matej Cepl <[email protected]>
+
+- Set `cpu_baseline` build option to `none` to be compatible with
+  the widest arrays of CPU (including `qemu` and older CPUs;
+  bsc#1261151).
+
+-------------------------------------------------------------------

Old:
----
  numpy-2.4.3.tar.gz

New:
----
  numpy-2.4.4.tar.gz

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Other differences:
------------------
++++++ python-numpy.spec ++++++
--- /var/tmp/diff_new_pack.34Q1wZ/_old  2026-05-13 17:19:25.907533434 +0200
+++ /var/tmp/diff_new_pack.34Q1wZ/_new  2026-05-13 17:19:25.911533600 +0200
@@ -30,7 +30,7 @@
 %{?sle15_python_module_pythons}
 
 Name:           python-numpy
-Version:        2.4.3
+Version:        2.4.4
 Release:        0
 Summary:        NumPy array processing for numbers, strings, records and 
objects
 License:        BSD-3-Clause
@@ -140,6 +140,12 @@
 export CXX=g++-12
 %endif
 
+# The "min" baseline is suitable for all non-x86 targets as it
+# defaults to the standard features on these architectures
+# see: https://numpy.org/doc/2.4/reference/simd/build-options.html#min
+%ifarch x86_64 %{ix86}
+export PIP_CONFIG_SETTINGS="setup-args=-Dcpu-baseline=none"
+%endif
 %pyproject_wheel
 
 %install
@@ -192,6 +198,7 @@
 # missing instruction set
 %ifarch s390x
 test_failok+=" or test_truncate_f32"
+test_failok+=" or test_cpu_features"
 %endif
 %ifarch %{ix86}
 # (arm 32-bit seems okay here)
@@ -199,6 +206,7 @@
 test_failok+=" or test_pareto"
 # gh#numpy/numpy#18388
 test_failok+=" or test_float_remainder_overflow"
+test_failok+=" or test_einsum"
 %endif
 %ifarch %{ix86} %{arm32}
 # too much memory for 32bit

++++++ numpy-2.4.3.tar.gz -> numpy-2.4.4.tar.gz ++++++
/work/SRC/openSUSE:Factory/python-numpy/numpy-2.4.3.tar.gz 
/work/SRC/openSUSE:Factory/.python-numpy.new.1966/numpy-2.4.4.tar.gz differ: 
char 21, line 1

Reply via email to