commit 1d7d90ea955628adf7444c2b9b537a3d3fcb17be
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Thu Sep 10 05:07:22 2026 +0200

dev-python/fsspec: trim 2025.9.0 comment prose

commit 21623c783feea285d9c45da0041ac0797c1c644f
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Sep 5 10:08:38 2026 +0200

dev-python/fsspec: Keyword 2025.9.0 for ~arm64

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

commit a56cc515b869a1ee7d5ad17a12577e999c7a0a13
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat Aug 29 14:33:15 2026 +0200

dev-python/fsspec: document why this is older than ::gentoo

No version change. nvchecker reports fsspec 2025.9.0 -> 2026.7.0 on every
run, and ::gentoo already ships 2026.6.0 and 2026.7.0, so the obvious reading
of this package is "stale overlay copy, bump or drop it". Both would break
the tree, and nothing in the ebuild said so.

The package exists only to satisfy sci-ml/datasets-4.3.0's
`<=dev-python/fsspec-2025.9.0-r0` ceiling, which neither ::gentoo version can
meet. No other consumer in the overlay has a CEILING: dask, lightning,
pytorch and bigcode-eval declare a floor, and huggingface_hub and
sitka-spruce are unversioned. Either way they all resolve against ::gentoo's
newer fsspec.

datasets-5.0.1 raised that ceiling to <=2026.6.0, so the exit condition is
concrete and worth recording: when datasets-4.3.0 is eventually dropped (it
is kept as the last of the 4.x series), this package can be dropped with it.

commit 114aaca5ed2bf2ff309f65f3cada17f5ed64bfee
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sun Aug 23 09:37:49 2026 +0200

dev-python/fsspec: add 2025.9.0