Commit Graph

484 Commits

Author SHA1 Message Date
Sandro
599c7643ca oneMath: init at 0.9 (#514640) 2026-08-16 15:29:19 +00:00
Aliaksandr
9921d05da3 treewide: replace stdenv.is* with stdenv.hostPlatform.is*
Assisted-by: claude-code with claude-opus-4-8
2026-08-12 20:27:00 +02:00
whispers
c7404c9232 libappindicator-gtk3: make alias for libappindicator 2026-08-10 23:03:31 -04:00
Ben Siraphob
5506449838 pkgs/by-name: fix typos
Assisted-by: Claude Code (claude-opus-5)
2026-08-05 09:56:33 -07:00
R. Ryantm
ed50c12bce onionshare: 2.6.4 -> 2.6.5 2026-08-04 19:21:30 +00:00
2kybe3
9dc557378d treewide: follow some GitHub redirects for fetchFromGitHub
generated using https://git.kybe.xyz/2kybe3/nixpkgs-github-redirect

continuation of: #538981
2026-08-01 19:26:08 +02:00
nixpkgs-ci[bot]
7616a5eb42 Merge master into staging-next 2026-07-31 18:31:04 +00:00
nixpkgs-ci[bot]
6a56b427eb oniux: 0.10.0 -> 0.12.0 (#525531) 2026-07-31 17:41:24 +00:00
nixpkgs-ci[bot]
ef1c4fdfdc Merge master into staging-next 2026-07-31 00:33:19 +00:00
R. Ryantm
c32fa56fd8 oniux: 0.10.0 -> 0.12.0 2026-07-30 19:09:14 +00:00
dotlambda
7a7e106c31 onednn: 3.12.2 -> 3.13 (#544875) 2026-07-30 18:59:45 +00:00
nixpkgs-ci[bot]
772be08438 Merge master into staging-next 2026-07-30 00:29:51 +00:00
Gaetan Lepage
8dd979fe72 onnxruntime: fix cuda build 2026-07-29 20:54:13 +00:00
Sandro
da3aedc384 onetbb: never treat warnings as fatal (#537640) 2026-07-25 12:44:36 +00:00
R. Ryantm
79adfd14c4 onednn: 3.12.2 -> 3.13 2026-07-23 12:18:15 +00:00
Yohann Boniface
a96a7b9dc6 treewide: use finalAttrs.finalPackage.doCheck (#543265) 2026-07-23 11:35:08 +00:00
Martin Weinelt
e22d795c6a onednn: 3.11.2 -> 3.12.2 (#511076) 2026-07-21 14:14:11 +00:00
Christian Kögler
e5dd6b2d01 onnxruntime: 1.26.0 -> 1.27.1 (#541892) 2026-07-20 15:30:35 +00:00
Arnout Engelen
3160431a46 onlyoffice-documentserver: update nodejs
Fixes #477803
2026-07-19 00:15:31 +01:00
Peder Bergebakken Sundt
a0d022d0f4 treewide: use finalAttrs.finalPackage.doCheck 2026-07-18 18:23:21 +02:00
Eman Resu
ff094f60ad treewide: replace more singleton license lists 2026-07-15 14:04:39 -04:00
kilyanni
a27bda6f62 oneMath: init at 0.9 2026-07-15 14:17:58 +02:00
Attila Oláh
513cb1ebd7 onnxruntime: 1.26.0 -> 1.27.1
Changelog: https://github.com/microsoft/onnxruntime/releases/tag/v1.27.1
2026-07-14 19:46:30 +02:00
eymeric
5690096bbf onlyoffice-documentserver: revert back to standard icu
Reverts onlyoffice-documentserver to use the standard icu version again (which is icu76/icu77) as nodejs reverted to icu77 again.

This keeps the robust symlinking/development header fix from f9f60a4 to facilitate any future overrides.
2026-07-13 11:08:16 +07:00
R. Ryantm
ef89318b1c onedrivegui: 1.3.1 -> 1.3.2 2026-07-05 17:50:28 +00:00
R. Ryantm
fbaf84e753 onednn: 3.11.2 -> 3.12.2 2026-07-05 00:19:08 +00:00
Arne Keller
9a0289a82a onedrive: 2.5.10 -> 2.5.11 (#536968) 2026-07-04 14:19:38 +00:00
Arnout Engelen
9dab67c16b onlyoffice-documentserver: use icu78 (#537160) 2026-07-03 07:07:50 +00:00
whispers
14700de3af onetbb: never treat warnings as fatal
onetbb often triggers compiler warnings upstream, and works around them
in an ad-hoc, per version manner:
88482f5f1a
bdbec20606
we encountered this failing while preparing for a gcc 16 upgrade in
Nixpkgs, and the same issue came up for gcc 15 and similar (#446139). it
is likely to come up again, as -Werror is extremely susceptible to
compiler and library changes (though onetbb admittedly has few
dependencies). additionally, while it may be useful for upstream onetbb,
it seems to provide little value to us downstream; we just end up
working around it. thus, we never treat warnings as errors by disabling
upstream's cmake flag for this purpose.
2026-07-01 22:20:47 -04:00
Gaétan Lepage
59f622ab22 onnxruntime: fix CUDAExecutionProvider loading error (#536479) 2026-07-01 07:23:52 +00:00
Arnout Engelen
f9f60a4608 onlyoffice-documentserver: use icu78
This is needed so that doctrenderer uses the same version of icu
as the nodejs.libv8 . This is a little scary because it's a different
version of icu than boost is built with, but seems to work out in
practice. After #520553 this should ideally be in sync again.

Fixes #536824
2026-06-30 21:40:23 +02:00
Daniel Fullmer
c78f5f413f onnxruntime: fix CUDAExecutionProvider loading error
onnxruntime fails to load the CUDAExecutionProvider with an libonnxruntime_providers_cuda.so: undefined symbol: nvrtcGetProgramLogSize error.

cudnn-frontend >=1.19.0 now includes symbols provided by nvrtc but does
not ensure it is linked against it. Recent commits to onnxruntime add
the CUDA::nvrtc link target to resolve the issue. This adds the same
change here.
2026-06-30 11:21:18 -07:00
R. Ryantm
204b7f504e onedrive: 2.5.10 -> 2.5.11 2026-06-30 09:15:24 +00:00
winston
34b22c7d81 onnxruntime: always enable RTTI with OpenVINO
needed for `{ openvinoSupport = true; pythonSupport = false; }`
2026-06-29 20:00:49 +02:00
winston
15b29960b8 onnxruntime: add openvino support
Co-authored-by: Martin Weinelt <hexa@darmstadt.ccc.de>
2026-06-29 20:00:48 +02:00
Gaetan Lepage
3fecd68e9d cudaPackages_13_3.cccl: init at 13.3.3.3.1 2026-06-21 16:56:43 +00:00
TomaSajt
0747e884ad online-judge-template-g-generator: migrate to by-name 2026-06-15 16:46:41 +02:00
Philipp Bartsch
51c432fc5e oniux: make use of compund licenses 2026-06-12 23:58:16 +02:00
dotlambda
b390caf7d6 onionshare: 2.6.3 -> 2.6.4 (#530043) 2026-06-10 01:24:33 +00:00
Adam C. Stephens
213c2036d8 onedriver: bump to fuse3 (#529241) 2026-06-09 21:00:47 +00:00
Robert Schütz
aeaabce790 onionshare: 2.6.3 -> 2.6.4
Diff: https://github.com/onionshare/onionshare/compare/v2.6.3...v2.6.4

Changelog: https://github.com/onionshare/onionshare/releases/tag/v2.6.4
2026-06-09 12:00:36 -07:00
Christian Kögler
cc10077236 onnxruntime: use latest abseil and simplify Nix expression (#527839) 2026-06-09 17:41:08 +00:00
Luca Ruperto
4b9186cb1c onedriver: bump to fuse3 2026-06-07 21:59:53 +02:00
Christian Kögler
dd509002f8 onnxruntime: simplify Nix expression
ncclSupport is already guarded with cudaSupport
2026-06-04 05:32:18 +02:00
Christian Kögler
49666df296 onnxruntime: use latest abseil-cpp 2026-06-04 05:31:45 +02:00
Nick Cao
7f865aee5b onioncircuits: add missing dependency (#516672) 2026-06-01 21:33:52 +00:00
toyboot4e
025dd7a538 online-judge-verify-helper: remove from python3Packages 2026-05-30 17:18:44 +09:00
Christian Kögler
eab2efadd8 onetbb, tiledb: fix dylib install name on Darwin (#525497) 2026-05-30 07:30:25 +00:00
Christian Kögler
56ecb9c425 onetbb: set CMAKE_INSTALL_NAME_DIR on darwin 2026-05-30 08:29:25 +02:00
whispers
f571a6c560 onnxruntime: never treat warnings as fatal
onnxruntime is currently failing on staging with:

```
/build/source/onnxruntime/test/ir/graph_test.cc:253:26: error: 'void google::protobuf::RepeatedField< <template-parameter-1-1> >::Resize(int, const Element&) [with Element = long int]' is deprecated [-Werror=deprecated-declarations]
  253 |   m_indicies_data->Resize(static_cast<int>(indices.size()), 0);
      |   ~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- snip --
/build/source/onnxruntime/test/ir/graph_test.cc:257:16: error: 'void google::protobuf::RepeatedField< <template-parameter-1-1> >::Resize(int, const Element&) [with Element = long int]' is deprecated [-Werror=deprecated-declarations]
      |   ~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```

This is one of several cases of onnxruntime having broken on library
updates. This also occured in
d54a792625,
where we introduced a messy workaround by patching the library. We
believe this was a mistake.

In general, -Werror is highly susceptible to changes in libraries,
compilers, and other components. In Nixpkgs, we often move faster than
upstream, and that is extremely wont to cause breakage. Packages often
use warnings like deprecations to inform packages of changes in a
non-breaking way, but we are not able to meaningfully respond in
Nixpkgs. We can either just ignore the warnings, or patch away the
issues. The latter seems like a great deal of work for a
constantly-moving target (like the commit linked above), and provides
questionable utility.

Thus, we simply promote `--compile-no-warning-as-error` from being
CUDA-only to applying in all cases. This fixes the build failure on
staging as a side-effect of cleaning up this issue.

Alternatively, we could just use `-Wno-error=deprecated-declarations` or
patch, but we believe this is a more correct solution.
2026-05-29 10:08:39 -04:00