commit 827ac3c720b884f20df52f97e3b1c28b6cba46f6
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Sep 5 10:34:20 2026 +0200

sci-ml/modelscope-hub: Keyword 0.4.0 for ~arm64

Pure-Python PEP 517 package with architecture-independent output.

commit e4f02962344750d6e4a95aaa76eb4321d8bda3f7
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Sep 5 10:34:20 2026 +0200

sci-ml/modelscope-hub: Keyword 0.3.1 for ~arm64

Pure-Python PEP 517 package with architecture-independent output.

commit f97a5ba8870203895da334c30fdfa8652a31cdbc
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Wed Sep 2 09:19:18 2026 +0200

sci-ml/modelscope-hub: drop 0.3.0

Retention: 0.3.1 and 0.4.0 are the newest two, and 0.3.0 differed from
0.3.1 in nothing this overlay carries. sci-ml/modelscope floor-pins
>=0.0.7 and >=0.2.0 across its two ebuilds, both satisfied by what
remains.

commit 9fb4d3977dc1b8690ea5f98d7bbe20d0fa458f77
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Wed Sep 2 09:19:18 2026 +0200

sci-ml/modelscope-hub: add 0.4.0

Adds the cryptography dependency 0.4.0 introduces, and deselects a test
file the sdist cannot run.

cryptography>=41 is a new base Requires-Dist and is genuinely
unconditional rather than an optional integration: the new agent_idp
module imports cryptography.hazmat at module scope for agent-token
signing, and __init__.py pulls agent_idp in at package root, so a bare
`import modelscope_hub` needs it.

test_openapi_coverage.py is new in 0.4.0 and cannot work from a PyPI
sdist. All six of its cases read tests/data/openapi.json, a vendored
copy of the live OpenAPI document, and the sdist ships no tests/data/
directory at all -- there is no openapi.json anywhere in it. They guard
upstream's own repo against spec drift rather than testing the built
package, so deselecting costs no coverage of what we ship. Left the
existing `-m "not remote"` filter alone.

Verified with FEATURES=test: 296 passed / 6 errors on FileNotFoundError
before, 261 passed / 41 deselected and a clean phase after. Both
numbers account for the same 302 collected cases.

commit c2127c1fae66b09c8521e0f11692c2b3b8965dc7
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Sep 1 11:19:19 2026 +0200

sci-ml/modelscope-hub: drop 0.1.9 0.2.0

Retention: keep the last two (0.3.0, 0.3.1). Both modelscope consumers
use >= floors (1.39.0 wants >=0.0.7, 1.39.1 wants >=0.2.0) that the
retained pair satisfies, and no atom names either dropped version.

There is no previous major series to anchor -- upstream has never left
0.x, so bin/retention-audit.py treats all four as one series.

commit 1c6b421370be33b52bf9ab1708f261acb79b8538
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Sep 1 11:19:01 2026 +0200

sci-ml/modelscope-hub: add 0.3.1

[project.dependencies] and the dev extra are identical to 0.3.0, and the
[tool.pytest.ini_options] remote marker the ebuild deselects is still
declared upstream, so nothing in the ebuild needed adjusting.

Test suite run at build time: 177 passed.

commit 547e4b2999e953d5644e7c693a0373bc8db5532e
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 29 14:33:15 2026 +0200

sci-ml/modelscope-hub: add 0.3.0

Straight bump on the dependency side: 0.3.0's Requires-Dist is identical to
0.2.0's (requests, tqdm, filelock, urllib3, same floors).

Also runs the test suite, which had never been wired up. The sdist has
shipped a tests/ tree since 0.2.0 while the ebuild neither ran it nor
RESTRICTed it, so every build emitted the "package seems to contain tests but
they are not enabled" QA notice and 133 real tests went unrun. Upstream marks
the cases that hit the live ModelScope API with a `remote` pytest marker and
documents them as needing .env credentials, so the suite is deselected with
`epytest -m "not remote"` rather than RESTRICTed wholesale. 133 pass, 0 fail.

Only dev-python/responses is declared as a test dep, floored at upstream's
>=0.20. Upstream's `dev` extra also lists pytest-mock, but the suite never
touches the `mocker` fixture -- every test mocks with stdlib unittest.mock --
so transcribing that extra would add a dependency nothing uses. Verified by
running the full suite without it: still 133 passed.

EPYTEST_PLUGINS=() because responses is used as a plain library
(@responses.activate), not as a pytest plugin; this also clears the
EPyTestPluginsSuggestion that enabling tests otherwise raises.

commit cea4b9b5a6ac396ec156df400c6d164a75a26800
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 1 23:08:39 2026 +0200

sci-ml/modelscope-hub: drop 0.1.8

Superseded; retention keeps 0.1.9 (last 0.1.x) and 0.2.0.

commit f108a2baa39d77c08c0ebca7cbdc2f41804e42ce
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 1 23:08:33 2026 +0200

sci-ml/modelscope-hub: add 0.2.0

Runtime dependencies are unchanged from 0.1.9 (upstream requires_dist
matches RDEPEND despite the minor bump), so the ebuild is copied verbatim.

commit 8b982cbc2a1b7cae1e552acae97c3fb9518fe00f
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 1 02:19:30 2026 +0200

sci-ml/modelscope-hub: drop 0.1.6, 0.1.7

Superseded; retention keeps the last two releases (0.1.8, 0.1.9). The
only consumer (sci-ml/modelscope) floors at >=0.0.7.

commit ec78ef846156176260dc4770fbea3d7d36101392
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 1 02:19:25 2026 +0200

sci-ml/modelscope-hub: add 0.1.9

Runtime dependencies are unchanged from 0.1.8 (upstream requires_dist
matches RDEPEND), so the ebuild is copied verbatim.

commit d26e488d095396a8343a1053601f18f59982a8df
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Mon Jul 27 10:04:12 2026 +0200

*/*: drop dead Python targets from PYTHON_COMPAT

139 ebuilds across 65 packages still listed python3_10, python3_11 or
python3_13t. None of the three is a supported implementation any more:
gentoo's python-utils-r1 sets _PYTHON_ALL_IMPLS to python3_ plus
python3_t, and everything older falls into the branch commented
"implementations deprecated prior to EAPI 9 are fatal".

Inert today, which is why nothing caught it. Targets outside
_PYTHON_ALL_IMPLS are silently dropped rather than rejected, so they
never became USE flags and never produced a dependency atom -- an ebuild
declaring python3_ was only ever offering 12, 13 and 14.

Two reasons to clean them up anyway. The eclass branch they land in
dies for any EAPI other than 7 or 8, so all 139 are waiting to break on
the first EAPI 9 bump. And pkgcheck's OldPythonCompat is a git-scoped
check that only inspects ebuilds touched in a commit range, so it never
reports these on a full-tree scan and would instead fire on whichever
unrelated commit next touches one of them.

No revbump, because nothing observable changes. Verified on a sample
spanning every transformation shape that the python USE flags portage
offers are byte-identical before and after, and that a pkgcheck run
over all 65 packages returns a set-identical list of findings.

The eight python2_7 ebuilds are deliberately untouched: that target is
also absent from _PYTHON_ALL_IMPLS, but they reach it through the
_PYTHON_ALLOW_PY27 hatch in the overlay's forked python-utils-r1_py2.

commit 089c9dea02b9d5aaed94dee0ad82c20bbf0cecdb
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Jul 21 11:45:35 2026 +0200

sci-ml/modelscope-hub: add 0.1.8

commit 0919c0066104add251ff4b5a3500a7103d44e8f5
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Tue Jul 7 19:53:22 2026 +0200

sci-ml/modelscope-hub: add 0.1.7

commit 2ffc86e9225d997d413971ab3235f9f750555d70
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Jul 4 14:13:27 2026 +0200

sci-ml/modelscope-hub: new package, add 0.1.6