commits
(cherry picked from commit ba1bad39d82f3ed2ba09faf725c787d927dcef00)
(cherry picked from commit 213d00a8fa88f6a7bb07e6da71e741de6d1850f7)
(cherry picked from commit 8870079d3b2e7064efb98629b80686a15bdbad56)
(cherry picked from commit 7c87d328a88b1d3dfe7e94cf7b08be1aca987e04)
(cherry picked from commit 0e6a84a5e5429a498ea15acfced1af46032340e3)
(cherry picked from commit 44859d9bc791970803805ae8de3895e052c36217)
Changelog: https://github.com/quantumlib/Cirq/releases/tag/v1.5.0
(cherry picked from commit 00040b39679cb1c1631934ff903e9d781438c486)
(cherry picked from commit cc6a8582f6c8693c3c43aeac1d414216b3f19a69)
(cherry picked from commit cce58bb9599b08a9175a4da44056187fc12cc6f1)
(cherry picked from commit fd7165462c810f1e07366ca930e83d2c5042c229)
(cherry picked from commit 9be97d3de7eeeff251aa1470ce4935d5d702eb3f)
This seems to be a left-over from before the performance comparison was
changed to a difference-per-chunk analysis.
(cherry picked from commit edaf51cb83e3bef31ba33ce7327dfc9ac7f48843)
`OF_OVERLAY` breaks kernel compilation for version 5.10
(beginning with 5.10.236), but only on x86_64.
Luckily, `OF_OVERLAY` is of not real interest on x86_64,
so this commit unbreaks the build by just disabling that option
if the target is x86_64 *and* the kernel version is below 5.15.
v5.15 contains a patch that turns the compilation error into a
warning, so this measure is not needed from that version on:
https://git.kernel.org/linus/5aad03685185b5133a28e1ee1d4e98d3fd3642a3
The reasons for the build failure are not
fully clear at the time of this writing;
apparently, a change in `pahole` is the core cause.
For details see the issue and in particular this comment:
https://github.com/NixOS/nixpkgs/issues/403985#issuecomment-2863338895
If that is correct and `pahole` gets fixed,
we can probably revert the commit at hand.
(cherry picked from commit 1e6c4457678c650bf0577a019d825effe637641e)
Currently, is it possible for stdenv.hostPlatform.isEfi to be true but
the kernel to have no EFI support. This mainly occurs on armv6l and
causes systemd in initrd to fail to build because it attempts to include
the non-existent efivarfs module in the initrd.
To fix this, automatically enable CONFIG_EFI on platforms where isEfi is
true. An alternative would be to make isEfi false on armv6l, but EFI is
in fact usable on that platform.
To prove this, I successfully booted NixOS on a Raspberry Pi Zero using
systemd-boot and U-Boot's EFI support. I confirmed that efivarfs works
as well.
(cherry picked from commit 3aff7b471c61b26a0be6f1bdc35ee8871d0168b7)
(cherry picked from commit 858ce19170ef26018a28edc1c14b3a31f666a5df)
(cherry picked from commit 3b7f48048a82c328e61f6d57da116dcc1057b647)
(cherry picked from commit 4fb15d7946122a9af71df839aa36c3f3ee526b30)
(cherry picked from commit c02eaae23d3dc262cca8e0f674abfab94c2eb1ef)
(cherry picked from commit 39f51ddad7a5890d6eaecc26ce5cf595ffa7f34f)
(cherry picked from commit de6905ee3749082bb256ddbc30010b79c489f5c6)
Security fixes.
Release announcement: https://blog.floorp.app/en/release/11.26.2.html
Git changelog: https://github.com/Floorp-Projects/Floorp/compare/v11.26.1...v11.26.2
Signed-off-by: Christoph Heiss <christoph@c8h4.io>
(cherry picked from commit d1bc54ae6c7870053c5038710de8a9847b434700)
(cherry picked from commit aac2e41ce04c2dfab59497a741d34ab6994e9512)
(cherry picked from commit 9b8ced6e698f3cd5d92882bd7e10e7ddbbe7d19e)
Diff: https://gitlab.freedesktop.org/pipewire/wireplumber/-/compare/0.5.8...0.5.10
(cherry picked from commit 49842e98a730949807386d1698f5dc662439f37b)
(cherry picked from commit fa2bee6de36a6da85b23907a0b1d4f89cc90116b)
(cherry picked from commit 13c6d411e70ec7f7e17d4ab37c121cdb8187ef71)
(cherry picked from commit 4b9c17e546a18eeffb0182e95c6d1fbf80cf8720)
(cherry picked from commit 60cbd815539f8aa7b3d7f96ff41f8ee8d2f328e8)
(cherry picked from commit f9d66c36aaf338217f325797578a52b28dd464f8)
(cherry picked from commit edc5b6b0a59f74de17ee406c9e37b1e4bd17ba3a)
Fixes MMSA-2025-00491 (rated as a critical vuln).
(cherry picked from commit 45e21093c32f7da9f3083c97f2091b88f04f116a)
Fixes MMSA-2025-00491 (rated as a critical vuln).
(cherry picked from commit 0c65612e31db41d111f80f110d9f623433c49bcf)
https://git.lix.systems/lix-project/lix/issues/568
https://git.lix.systems/lix-project/lix/issues/832
(cherry picked from commit ccd5688c946fb85e7125a6384b38f863361215a8)
Also fix and enable strictDeps.
(cherry picked from commit d0027f93ac816422a13cf062e82c345e85538c5f)
https://about.gitlab.com/releases/2025/05/21/patch-release-gitlab-18-0-1-released/
(cherry picked from commit 3b258c1272ce786f33c78a7f4b68be54e17e8c1f)
(cherry picked from commit c1ef83481f55c21a750c406d1a7625928db40a50)
(cherry picked from commit f25948845aba15048f05d9516f94debcbc695d27)
Diff: https://github.com/markqvist/NomadNet/compare/refs/tags/0.6.2...refs/tags/0.7.0
Changelog: https://github.com/markqvist/NomadNet/releases/tag/0.7.0
(cherry picked from commit ae2ba4778a6c6ea8e28753b992fba7898b49c1a7)
Diff: https://github.com/markqvist/Reticulum/compare/refs/tags/0.9.5...refs/tags/0.9.6
Changelog: https://github.com/markqvist/Reticulum/blob/0.9.6/Changelog.md
(cherry picked from commit e2294866565604678635ccaced608c495edded56)
(cherry picked from commit ba1bad39d82f3ed2ba09faf725c787d927dcef00)
`OF_OVERLAY` breaks kernel compilation for version 5.10
(beginning with 5.10.236), but only on x86_64.
Luckily, `OF_OVERLAY` is of not real interest on x86_64,
so this commit unbreaks the build by just disabling that option
if the target is x86_64 *and* the kernel version is below 5.15.
v5.15 contains a patch that turns the compilation error into a
warning, so this measure is not needed from that version on:
https://git.kernel.org/linus/5aad03685185b5133a28e1ee1d4e98d3fd3642a3
The reasons for the build failure are not
fully clear at the time of this writing;
apparently, a change in `pahole` is the core cause.
For details see the issue and in particular this comment:
https://github.com/NixOS/nixpkgs/issues/403985#issuecomment-2863338895
If that is correct and `pahole` gets fixed,
we can probably revert the commit at hand.
(cherry picked from commit 1e6c4457678c650bf0577a019d825effe637641e)
Currently, is it possible for stdenv.hostPlatform.isEfi to be true but
the kernel to have no EFI support. This mainly occurs on armv6l and
causes systemd in initrd to fail to build because it attempts to include
the non-existent efivarfs module in the initrd.
To fix this, automatically enable CONFIG_EFI on platforms where isEfi is
true. An alternative would be to make isEfi false on armv6l, but EFI is
in fact usable on that platform.
To prove this, I successfully booted NixOS on a Raspberry Pi Zero using
systemd-boot and U-Boot's EFI support. I confirmed that efivarfs works
as well.
(cherry picked from commit 3aff7b471c61b26a0be6f1bdc35ee8871d0168b7)