Accepting request 872790 from home:bnavigator:branches:devel:languages:python

- Update to 1.0.1
  * dask: avoid redundant scattering of large arguments to make a
    more efficient use of the network resources and avoid crashing
    dask with "OSError: [Errno 55] No buffer space available" or
    "ConnectionResetError: [Errno 104] connection reset by
    peer".
- Changees in 1.0.0
  * Make joblib.hash and joblib.Memory caching system compatible
    with numpy >= 1.20.0. Also make it explicit in the
    documentation that users should now expect to have their joblib.
    Memory cache invalidated when either joblib or a third party
    library involved in the cached values definition is upgraded.
    In particular, users updating joblib to a release that includes
    this fix will see their previous cache invalidated if they
    contained reference to numpy objects.
  * Remove deprecated check_pickle argument in delayed.
- Changes in 0.17.0
  * Fix a spurious invalidation of Memory.cache'd functions called
    with Parallel under Jupyter or IPython.
  * Bump vendored loky to 2.9.0 and cloudpickle to 1.6.0. In
    particular this fixes a problem to add compat for Python 3.9.
- Don't require optional NumPy for python36 tests in TW, because
  NumPy 1.20 dropped support for Python 3.6 (NEP 29)
- Drop joblib-disable-unrelialble-tests.patch, they are already
  used in pytest deselection parameter.
  * Do the same for disable_test_on_big_endian.patch.

OBS-URL: https://build.opensuse.org/request/show/872790
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python/python-joblib?expand=0&rev=47

disable_test_on_big_endian.patch

@@ -1,39 +0,0 @@
-From: Michel Normand <normand@linux.vnet.ibm.com>
-Subject: disable test on big endian
-Date: Fri, 21 Aug 2020 10:35:13 +0200
-
-disable test on big endian
-
-because still open
-https://github.com/joblib/joblib/issues/279
-
-Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
----
- joblib/test/test_numpy_pickle.py |    7 +++++++
- 1 file changed, 7 insertions(+)
-
-Index: joblib-0.16.0/joblib/test/test_numpy_pickle.py
-===================================================================
---- joblib-0.16.0.orig/joblib/test/test_numpy_pickle.py
-+++ joblib-0.16.0/joblib/test/test_numpy_pickle.py
-@@ -2,6 +2,7 @@
- 
- import copy
- import os
-+import sys
- import random
- import re
- import io
-@@ -426,6 +427,12 @@ def _check_pickle(filename, expected_lis
- 
- @with_numpy
- def test_joblib_pickle_across_python_versions():
-+    # temporarily disable this test on non little-endian machines
-+    # because next endianness lines not sufficient.
-+    # https://github.com/joblib/joblib/issues/279 still open.
-+    if sys.byteorder != 'little':
-+        raise SkipTest('Skipping this test on non little-endian machines')
-+
-     # We need to be specific about dtypes in particular endianness
-     # because the pickles can be generated on one architecture and
-     # the tests run on another one. See

joblib-0.16.0.tar.gz

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8f52bf24c64b608bf0b2563e0e47d6fcf516abc8cfafe10cfd98ad66d94f92d6
-size 349661

joblib-1.0.1.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c17567692206d2f3fb9ecf5e991084254fe631665c450b443761c4186a613f7
+size 316131

joblib-disable-unrelialble-tests.patch

@@ -1,28 +0,0 @@
---- joblib-0.16.0/joblib/test/test_memmapping.py.orig	2020-09-08 08:01:19.285304370 +0000
-+++ joblib-0.16.0/joblib/test/test_memmapping.py	2020-09-08 08:05:25.318014706 +0000
-@@ -512,6 +512,7 @@ def test_memmapping_temp_folder_thread_s
- @with_numpy
- @with_multiprocessing
- def test_multithreaded_parallel_termination_resource_tracker_silent():
-+    raise SkipTest('Skipping this test as it fails often in OBS')
-     # test that concurrent termination attempts of a same executor does not
-     # emit any spurious error from the resource_tracker. We test various
-     # situations making 0, 1 or both parallel call sending a task that will
-@@ -681,6 +682,8 @@ def test_memmap_returned_as_regular_arra
- @with_multiprocessing
- @parametrize("backend", ["multiprocessing", param("loky", marks=xfail)])
- def test_resource_tracker_silent_when_reference_cycles(backend):
-+    if backend == 'multiprocessing':
-+        raise SkipTest('Skipping this test as it fails often in OBS')
-     # There is a variety of reasons that can make joblib with loky backend
-     # output noisy warnings when a reference cycle is preventing a memmap from
-     # being garbage collected. Especially, joblib's main process finalizer
-@@ -778,6 +781,8 @@ def test_memmapping_pool_for_large_array
- @with_multiprocessing
- @parametrize("backend", ["multiprocessing", "loky"])
- def test_child_raises_parent_exits_cleanly(backend):
-+    if backend == 'multiprocessing':
-+        raise SkipTest('Skipping this test as it fails often in OBS')
-     # When a task executed by a child process raises an error, the parent
-     # process's backend is notified, and calls abort_everything.
-     # In loky, abort_everything itself calls shutdown(kill_workers=True) which

python-joblib.changes

@@ -1,3 +1,33 @@
+-------------------------------------------------------------------
+Mon Feb 15 23:26:21 UTC 2021 - Ben Greiner <code@bnavigator.de>
+
+- Update to 1.0.1
+  * dask: avoid redundant scattering of large arguments to make a
+    more efficient use of the network resources and avoid crashing
+    dask with "OSError: [Errno 55] No buffer space available" or
+    "ConnectionResetError: [Errno 104] connection reset by
+    peer".
+- Changees in 1.0.0
+  * Make joblib.hash and joblib.Memory caching system compatible
+    with numpy >= 1.20.0. Also make it explicit in the
+    documentation that users should now expect to have their joblib.
+    Memory cache invalidated when either joblib or a third party
+    library involved in the cached values definition is upgraded.
+    In particular, users updating joblib to a release that includes
+    this fix will see their previous cache invalidated if they
+    contained reference to numpy objects.
+  * Remove deprecated check_pickle argument in delayed.
+- Changes in 0.17.0
+  * Fix a spurious invalidation of Memory.cache'd functions called
+    with Parallel under Jupyter or IPython.
+  * Bump vendored loky to 2.9.0 and cloudpickle to 1.6.0. In
+    particular this fixes a problem to add compat for Python 3.9.
+- Don't require optional NumPy for python36 tests in TW, because
+  NumPy 1.20 dropped support for Python 3.6 (NEP 29)
+- Drop joblib-disable-unrelialble-tests.patch, they are already
+  used in pytest deselection parameter.
+  * Do the same for disable_test_on_big_endian.patch.
+
 -------------------------------------------------------------------
 Sun Dec  6 21:37:54 UTC 2020 - Benjamin Greiner <code@bnavigator.de>
 

python-joblib.spec

@@ -1,7 +1,7 @@
 #
 # spec file for package python-joblib
 #
-# Copyright (c) 2020 SUSE LLC
+# Copyright (c) 2021 SUSE LLC
 #
 # All modifications and additions to the file contributed by third parties
 # remain the property of their copyright owners, unless otherwise agreed
@@ -19,27 +19,25 @@
 %{?!python_module:%define python_module() python-%{**} python3-%{**}}
 %global skip_python2 1
 Name:           python-joblib
-Version:        0.16.0
+Version:        1.0.1
 Release:        0
 Summary:        Module for using Python functions as pipeline jobs
 License:        BSD-3-Clause
 Group:          Development/Languages/Python
 URL:            https://github.com/joblib/joblib
 Source:         https://files.pythonhosted.org/packages/source/j/joblib/joblib-%{version}.tar.gz
-Patch1:         disable_test_on_big_endian.patch
-# PATCH-FIX-OPENSUSE - Disable tests failing often in OBS
-Patch2:         joblib-disable-unrelialble-tests.patch
 BuildRequires:  %{python_module lz4}
-BuildRequires:  %{python_module numpy}
 BuildRequires:  %{python_module psutil}
 BuildRequires:  %{python_module pytest}
 BuildRequires:  %{python_module setuptools}
 BuildRequires:  %{python_module threadpoolctl}
 BuildRequires:  fdupes
 BuildRequires:  python-rpm-macros
-Requires:       python-lz4
+BuildRequires:  %{python_module numpy if (%python-base without python36-base)}
+Recommends:     python-lz4
 Recommends:     python-numpy
 Recommends:     python-psutil
+Suggests:       python-dask-distributed
 BuildArch:      noarch
 %python_subpackages
 
@@ -58,10 +56,6 @@ Joblib can handle large data and has specific optimizations for `numpy` arrays.
 
 %prep
 %setup -q -n joblib-%{version}
-%patch1 -p1
-%ifarch aarch64 %arm ppc64 %{ix86}
-%patch2 -p1
-%endif
 
 %build
 %python_build
@@ -93,6 +87,9 @@ export LANG=en_US.UTF-8
 #  test_hashes_are_different_between_c_and_fortran_contiguous_arrays
 #  test_hashes_stay_the_same_with_numpy_objects
 #  test_non_contiguous_array_pickling
+#
+# always fails:
+# test_parallel_call_cached_function_defined_in_jupyter
 DISABLED_TESTS="test_hash_numpy_noncontiguous or \
                 test_hashes_are_different_between_c_and_fortran_contiguous_arrays or \
                 test_hashes_stay_the_same_with_numpy_objects or \
@@ -104,7 +101,11 @@ DISABLED_TESTS="test_hash_numpy_noncontiguous or \
                 test_hash_numpy_noncontiguous or \
                 test_hashes_are_different_between_c_and_fortran_contiguous_arrays or \
                 test_hashes_stay_the_same_with_numpy_objects or \
-                test_non_contiguous_array_pickling"
+                test_non_contiguous_array_pickling or \
+                test_parallel_call_cached_function_defined_in_jupyter"
+if [ $(python3 -c 'import sys; print(sys.byteorder)') != "little" ]; then
+  DISABLED_TESTS+=" or test_joblib_pickle_across_python_versions"
+fi                
 %pytest -k "not ($DISABLED_TESTS)"
 
 %files %{python_files}