Script 'mail_helper' called by obssrc Hello community, here is the log from the commit of package python-awkward for openSUSE:Factory checked in at 2026-05-20 15:26:33 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Comparing /work/SRC/openSUSE:Factory/python-awkward (Old) and /work/SRC/openSUSE:Factory/.python-awkward.new.1966 (New) ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "python-awkward" Wed May 20 15:26:33 2026 rev:42 rq:1354203 version:2.9.0 Changes: -------- --- /work/SRC/openSUSE:Factory/python-awkward/python-awkward.changes 2025-10-16 17:39:35.258520664 +0200 +++ /work/SRC/openSUSE:Factory/.python-awkward.new.1966/python-awkward.changes 2026-05-20 15:27:20.425108902 +0200 @@ -1,0 +2,152 @@ +Tue May 19 21:01:38 UTC 2026 - Atri Bhattacharya <[email protected]> + +- Use `--ignore-glob` to disable tests from test-cuda* dirs. + +------------------------------------------------------------------- +Sun Apr 19 08:21:20 UTC 2026 - Dirk Müller <[email protected]> + +- be compatible with quilt(1) again + +------------------------------------------------------------------- +Sat Apr 11 17:35:48 UTC 2026 - Dirk Müller <[email protected]> + +- update to 2.9.0: + * feat: drop Python 3.9 + * fix: pass the right input type to + `awkward_NumpyArray_subrange_equal*` kernels + * fix: define behavior class within the test function in + `test_1511_set_attribute.py` to make it thread-local + * fix: properly respect requested byteorder in + `ak.from_buffers` for ndarray buffers + * fix: use `set_mask` instead of `set_base_mask` in `to_cudf` + * chore: use "M8" dtype instead of " + +------------------------------------------------------------------- +Sat Mar 14 21:46:19 UTC 2026 - Dirk Müller <[email protected]> + +- update to 2.8.12: + * ## New features + * feat: improve merging of `NumpyArray`s and simplification of + `UnionArray`s + fix `ak.almost_equal` for `UnionArray`s + * feat: Add `ak.sort()` for CUDA backend + * feat: Make `ak.combinations` faster on GPU by using + `cp.searchsorted` to compute output list indexes + * feat: add argmax implementation using CCCL + * feat: make `ak.combinations` faster on GPU by using + `cp.searchsorted` to compute output list indexes + * feat: add argmin implementation using CCCL + * ## Bug-fixes and performance + * fix: incorrect type passed down for the `mergeastable` + argument + * fix: fix `ak.validity_error` for typetracer-backed layouts + with unknown length + * fix: make `ak.enforce_type` work with typetracers + * fix: `ArrayBuilder.show()` formatter forwarding + * fix: fix and test `Form.length_zero/one_array()` + * fix: make named axis thread-safe and start testing using + multiple threads in CI + * ## Other + * ci: use codecov token + * chore(ci): use macOS to test ML dependencies + * chore(deps): bump the actions group across 1 directory with 5 + updates + * ci: xfail failing cudf tests until fixed + * chore: update pre-commit hooks + * chore: request for tensorflow in requirements only if python + < 3.14 + * chore: fix noxfile on Intel macOS + * docs: add X0708a as a contributor for code + * docs: add a comparison of different argmax implementations + (using unary_transform, segmented_reduce and cuda-kernels) + * docs: Document GPU backend and cuDF compatibility limitations +- update to 2.8.11: + * ## New features + * feat: add behavior cls validation hook + * feat: add a playground for cccl studies + * feat: add cuda.compute implementation to CCCL study + * feat: ability to optionally disable the caching of virtual + arrays on themselves + * ## Bug-fixes and performance + * fix: string array numpy conversion fails with int32 offsets + from parquet + * fix: only check length against empty form in + `ak.from_buffers` if it is known + * fix: typestr respects now limit_cols for .show and + _repr_mimebundle_ + * fix: `ArrayBuilder.__bool__ ` raises `TypeError` when builder + length is 1 + * fix: prevent teardown from crashing GPU test suite + * fix: hide awkward.ak_* and awkward.str.akstr_* from 'awkward' + and 'awkward.str' module + * fix: align multi-level indices before merge in + `to_dataframe()` + * fix: numba test failures on Python 3.14 + * fix: incorrect passing of `Index` inside nplike function in + `to_IndexedOptionArray64` + * fix: weights should not be raised to the power n in the n-th + moment calculation + * fix: `ak.Record` creation from iterable with nested dicts + * fix: keep track of buffer key instead of field path for + VirtualNDArrays & PlaceholderArrays + * fix: typetracer nplike's `all` function leads to infinite + recursion for `axis=None` + * ## Other + * ci: fix compilation on self-hosted CI and test with multiple + versions of CUDA + * chore: update acknowledgements to include NumFOCUS + affiliation + * chore: update pre-commit hooks + * chore: pytest log_level is better than log_cli_level + * chore: update requirements files for latest possible python + versions + * chore(deps): bump the actions group with 2 updates + * chore: update pre-commit hooks + * chore: Remove CUDA 11 support + * chore(ci): use macos-15-intel for x86 wheel building and + testing + * docs: add new versions to version selector + * docs: 3.14 is supported + * docs: add T90REAL as a contributor for code + * docs: Add dependency and version change guidelines + * docs: add shwina as a contributor for code + * docs: Add an example summation reducer using cccl +- update to 2.8.10: + * ## New features + * feat: preserve field order when replacing existing fields in + with_field + * feat: add to_/from_safetensors + * ## Bug-fixes and performance + * fix: update image cell in thinking-in-arrays.ipynb + * fix: prefix with an underscore unused variables + * fix(ci): fix issues with dependencies for GPU tests + * fix: convert integer-like objects to python integers in + shapes + * fix: avoid recursion limit when concatenating `EmptyArray`s + * fix: only check `VirtualNDArray`'s `shape` versus `count` in + `from_buffers` if both are known + * fix: preserve the type when indexed by an empty array + * fix: ensure masked string/bytestring dtypes can hold 'nan' + fill value… + * fix: add missing units to datetime/timedelta in a test + * fix: pass attrs parameter to wrap_layout in from_* + * fix: use jax's `__setitem__` syntax in + `ak.merge_union_of_records` to make it work with the jax + backend + * ## Other + * chore(deps): bump pypa/cibuildwheel from 3.1 to 3.2 in the + actions group + * chore(deps): bump the actions group with 2 updates + * chore: update pre-commit hooks + * chore: mac os 13 hosted runner image is closing down + * chore: improve pyproject.toml from sp-repo-review checks + * chore(deps): bump astral-sh/setup-uv from 6 to 7 in the + actions group + * chore: small inconsistency/typo with the non-jax case below + * chore(deps): bump the actions group with 2 updates + * docs: add TaiSakuma as a contributor for code + * docs: add DylanModesitt as a contributor for code + * docs: add JamesBrofos as a contributor for code + * test: add cuda tests for virtual arrays + * test: add a test on flattering empty arrays + +------------------------------------------------------------------- Old: ---- awkward-2.8.9.tar.gz New: ---- awkward-2.9.0.tar.gz ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Other differences: ------------------ ++++++ python-awkward.spec ++++++ --- /var/tmp/diff_new_pack.W2gAkI/_old 2026-05-20 15:27:20.977131641 +0200 +++ /var/tmp/diff_new_pack.W2gAkI/_new 2026-05-20 15:27:20.981131806 +0200 @@ -1,7 +1,7 @@ # # spec file for package python-awkward # -# Copyright (c) 2025 SUSE LLC and contributors +# Copyright (c) 2026 SUSE LLC and contributors # # All modifications and additions to the file contributed by third parties # remain the property of their copyright owners, unless otherwise agreed @@ -17,21 +17,20 @@ %global flavor @BUILD_FLAVOR@%{nil} -%if "%{flavor}" == "" -%define psuffix %{nil} -%bcond_with test -BuildArch: noarch -%else +%if "%{flavor}" == "test" +%define psuffix -%{flavor} %bcond_without test # Test suite fails on numerous tests when trying to convert 64-bit types ExcludeArch: %{ix86} %{arm32} -%define psuffix -%{flavor} +%else +%define psuffix %{nil} +%bcond_with test %endif -%define awkward_cpp_version 50 +%define awkward_cpp_version 52 %{?sle15_python_module_pythons} Name: python-awkward%{psuffix} -Version: 2.8.9 +Version: 2.9.0 Release: 0 Summary: Manipulate arrays of complex data structures as easily as Numpy License: BSD-3-Clause @@ -104,7 +103,7 @@ donttest="$donttest or (test_1294_to_and_from_parquet and test_recordarray)" donttest="$donttest or (test_1440_start_v2_to_parquet and test_recordarray)" # no cupy / cuda on obs -%pytest -n auto --ignore tests-cuda/ --ignore tests-cuda-kernels/ -k "not ($donttest)" +%pytest -k "not ($donttest)" --ignore-glob=tests-cuda* %endif %if !%{with test} ++++++ awkward-2.8.9.tar.gz -> awkward-2.9.0.tar.gz ++++++ ++++ 21867 lines of diff (skipped)
