Script 'mail_helper' called by obssrc Hello community, here is the log from the commit of package ollama for openSUSE:Factory checked in at 2026-10-17 21:24:09 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Comparing /work/SRC/openSUSE:Factory/ollama (Old) and /work/SRC/openSUSE:Factory/.ollama.new.3698736 (New) ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "ollama" Sat Oct 17 21:24:09 2026 rev:70 rq:1385170 version:0.40.5 Changes: -------- --- /work/SRC/openSUSE:Factory/ollama/ollama.changes 2026-09-04 12:41:27.896807177 +0200 +++ /work/SRC/openSUSE:Factory/.ollama.new.3698736/ollama.changes 2026-10-17 21:24:22.482307926 +0200 @@ -1,0 +2,129 @@ +Fri Oct 16 07:20:59 UTC 2026 - Adrian Schröter <[email protected]> + +- Updated to version 0.40.5 + * NVIDIA Nemotron 3.5 Super available (vision-language model for + coding, reasoning, tool use, and document understanding) + * Compact Nemotron 3.5 Bolt sibling released (74B parameters) + * Decision models now available on cloud via clef:cloud + * Continue OpenAI-compatible Responses conversations with custom + tool calls without errors + * Clearer error messages for cloud access failures + * Updated llama.cpp with Nemotron 3.5 support and MLX update +- Update generated using pbuild-ai + +------------------------------------------------------------------- +Wed Oct 14 15:05:32 UTC 2026 - Adrian Schröter <[email protected]> + +- Updated to version 0.40.4 + * Decision models can now be used directly from the terminal with + `ollama run` + * Kev decision models are now supported by the MLX runner on Apple + silicon + * ollama launch can now be used with usage credits without upgrading + to a paid plan + * ollama launch codex includes locally installed Ollama models in its + picker + * Improved inference performance by reusing verified manifests and + loaded models + * Added support for multimodal embeddings through llama-server + * Fixed loading multimodal embedding models with small context windows + * Improved Gemma 4 prefill performance on NVIDIA GPUs with the MLX + runner + * MLX runner now clips long Laya states instead of rejecting them + * Improved speculative decoding reliability in the MLX runner + * Updated llama.cpp with native decision model support + * Updated README and documentation with clearer privacy and + cloud-services wording +- Update generated using pbuild-ai +- added support for ccache builds (rocm goes down from 1.5h to 3 minutes) + +------------------------------------------------------------------- +Mon Oct 12 09:51:38 UTC 2026 - Adrian Schröter <[email protected]> + +- Updated to version 0.40.3 + * `ollama launch claude` no longer shows a warning about Claude + connectors + * `ollama launch codex` no longer warns about service tiers + * Paused background model upgrade for embedding and other request + performance +- Updated to version 0.40.2 + * Models downloaded with earlier versions are upgraded in the + background on first run + * `ollama list` no longer shows duplicate entries for upgraded + models + * `ollama launch claude` uses full context length +- Updated to version 0.40.1 + * Proxy cloud usage and balance APIs + * Fix clef head reads past 2GiB on Windows + * Remove account step from CLI onboarding + * Avoid symlinks on Windows + * Drop carried metal residency patch now that it is upstream +- Updated to version 0.40.0 + * Models run on MLX on Apple Silicon by default + * New embedding model: embeddinggemma-2 + * Decision models available on MLX (Nimble, Tev1, Clef) +- Updated to version 0.35.1 + * Web search up to ten searches per response + * Modelfiles support `CAPABILITY` declarations + * `ollama show` reports only `decision` capability for decision + models +- Updated to version 0.35.0 + * Support decision models through `/v1/systemone` (Nimble, Tev1) +- Updated to version 0.34.4 + * Structured outputs on thinking models apply in a single pass + * Fixed intermittent "model not found" errors with a large local + library + * Fixed macOS app becoming unresponsive when checking ChatGPT/Codex + status + * Qwen 3.8 prompt processing faster on Apple Silicon + * Gemma 4 on Apple Silicon picks best image resolution per image +- Update generated using pbuild-ai +- obsoletes fix-mlxrunner-tests.diff + +------------------------------------------------------------------- +Wed Oct 7 18:55:55 UTC 2026 - Christian Goll <[email protected]> + +- Enable ROCm builds for Tumbleweed + +------------------------------------------------------------------- +Tue Sep 29 10:00:00 UTC 2026 - Adrian Schröter <[email protected]> + +- Updated to version 0.34.3 + * 0.34.1: MLX safetensors ollama create no longer experimental. + GGUF model creation now requires using llama.cpp tooling + for safe + * 0.34.1: Improved MLX memory handling on Apple Silicon + * 0.34.2: Added first-run setup when running ollama, with options to sign + in or continue locally. Setup completion is shared with + * 0.34.2: Added ollama://apps to open the desktop app’s Apps page + directly on macOS and Windows. +- Update generated using pbuild-ai + +------------------------------------------------------------------- +Mon Sep 28 06:04:47 UTC 2026 - Egbert Eich <[email protected]> + +- Bump CUDA to 13.4 +- Fix fillups + +------------------------------------------------------------------- +Wed Sep 23 07:45:14 UTC 2026 - Adrian Schröter <[email protected]> + +- Updated to version 0.34.0 + * Ollama models can now be used directly in ChatGPT Desktop + on macOS + * Improved structured output performance on Apple Silicon + * OpenAI-compatible client tool search support + * OpenAI-compatible client response compaction support +- Updated to version 0.33.3 + * gemma4 now supports images and audio on MLX engine + * Report cached prompt tokens + * Honor GGUF model defined default parameters + * MLX, MLX-C, and llama.cpp updates +- Updated to version 0.33.2 + * Ollama's app now follows the system appearance again + * Fixed the macOS app to properly hand off to an already-running + instance instead of starting a second one + * The Claude Desktop proxy no longer interrupts in-flight + requests when the model catalog updates + +------------------------------------------------------------------- @@ -9 +138,2 @@ - * mlxrunner: avoid Metal GPU timeouts when loading models from slow storage + * mlxrunner: avoid Metal GPU timeouts when loading models from + slow storage @@ -124,6 +253,0 @@ - -------------------------------------------------------------------- -Tue Jul 7 15:20:44 UTC 2026 - Adrian Schröter <[email protected]> - -- Define a build counter sync tag, to ensure all flavors have - the same build number Old: ---- fix-mlxrunner-tests.diff ollama-0.33.1.tar.gz New: ---- AGENTS.md _agents.obscpio ollama-0.40.5.tar.gz ----------(Old B)---------- Old:- Update generated using pbuild-ai - obsoletes fix-mlxrunner-tests.diff ----------(Old E)---------- ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Other differences: ------------------ ++++++ ollama.spec ++++++ --- /var/tmp/diff_new_pack.eptTFG/_old 2026-10-17 21:24:23.728359907 +0200 +++ /var/tmp/diff_new_pack.eptTFG/_new 2026-10-17 21:24:23.730359991 +0200 @@ -15,19 +15,16 @@ # Please submit bugfixes or comments via https://bugs.opensuse.org/ # -#!BcntSyncTag: ollama -# We require too much memory atm. -# these jobs work on all supported %ifarch x86_64 %global jobs 3 %else %global jobs 4 %endif -%global llama_cpp_version b10630 -%global mlx_version c793734eb715dbcfdb1ced58e348ec53c2d7ed85 -%global mlx_c_version fba4470b89073180056c9ea46c443051375f7399 +%global llama_cpp_version b11669 +%global mlx_version bebd6422f8db88d56e6220520a35bcd146f823f8 +%global mlx_c_version a341b4925024b88b2c593468f16e12f5e17315da %{bcond_with rocm} %{bcond_with cuda} @@ -51,7 +48,9 @@ %if "%{flavor}" == "rocm" %global preset_flavor rocm_v7_2_linux -%if %{with rocm} +%if 0%{?suse_version} >= 1699 +%{bcond_without rocm} +# ROCm is only available for x86_64 ExclusiveArch: x86_64 %else ExclusiveArch: SKIP_IT @@ -74,11 +73,11 @@ %endif %define cuda_version_major 13 -%define cuda_version_minor 3 +%define cuda_version_minor 4 %define cuda_version %{cuda_version_major}-%{cuda_version_minor} Name: ollama -Version: 0.33.1 +Version: 0.40.5 Release: 0 Summary: Tool for running AI models on-premise License: MIT @@ -104,7 +103,6 @@ #!CreateArchive Source12: mlx-c-main.tar.xz -Patch0: fix-mlxrunner-tests.diff Patch1: disable-llama.cpp-ui.patch BuildRequires: ccache BuildRequires: cmake >= 3.24 @@ -164,6 +162,8 @@ BuildRequires: rocminfo %endif +Requires(pre): %fillup_prereq + %systemd_ordering # 32bit seems not to be supported anymore while riscv isn't yet. ExcludeArch: %{ix86} %{arm} riscv64 @@ -279,6 +279,7 @@ -DOLLAMA_BUILD_PARALLEL=%jobs \ -DFETCHCONTENT_SOURCE_DIR_LLAMA_CPP=$PWD/../git-modules/llama.cpp \ -DOLLAMA_MLX_BACKENDS="" \ + -DGGML_INTERNAL_SME=OFF \ %if "%{?flavor}" == "cuda" -DCMAKE_CUDA_COMPILER=/usr/local/cuda-%{cuda_version_major}.%{cuda_version_minor}/bin/nvcc \ -DOLLAMA_MLX_BACKENDS=cuda_v13 \ @@ -294,17 +295,6 @@ %{nil} # -# Main package -# -%if "%{?flavor}" == "" -%cmake_build -j %jobs - -cd .. -go build -ldflags="-X github.com/ollama/ollama/version.Version=%version" -trimpath -o %{name} . -exit 0 -%endif - -# # Cuda Flavor # %if "%{?flavor}" == "cuda" @@ -321,9 +311,21 @@ exit 0 %endif +%if "%{?flavor}" == "" +# +# Main package +# + +%cmake_build -j %jobs + +cd .. +go build -ldflags="-X github.com/ollama/ollama/version.Version=%version" -trimpath -o %{name} . + +%else # # Flavor package (vulkan/rocm) # + cd .. export OLLAMA_SKIP_CPU_GENERATE="1" cmake -S llama/server --preset %preset_flavor \ @@ -335,7 +337,7 @@ local_directory=build/llama-server-%preset_flavor cmake --build ${local_directory%_linux} -- -l %{jobs} - +%endif %install %if "%{?flavor}" == "test" @@ -432,6 +434,7 @@ %post %service_add_post %{name}.service +%fillup_only %preun %service_del_preun %{name}.service @@ -458,7 +461,7 @@ %{_libdir}/ollama/libggml-hip.so %endif -%if "%{flavor}" == "" && "%{flavor}" != "test" +%if "%{flavor}" == "" %files %doc README.md %license LICENSE ++++++ AGENTS.md ++++++ # ollama packaging rules for pbuild-ai ## Updating the package The only build recipe of this package is `ollama.spec` in the workspace root. After the `Version:` tag in `ollama.spec` has been changed, apply the OLLAMA_UPDATE skill (`_agents/skills/OLLAMA_VERSION_UPDATE.md`) by running the post-update script `_agents/skills/update_references.sh`. ++++++ _scmsync.obsinfo ++++++ --- /var/tmp/diff_new_pack.eptTFG/_old 2026-10-17 21:24:23.810363328 +0200 +++ /var/tmp/diff_new_pack.eptTFG/_new 2026-10-17 21:24:23.812363411 +0200 @@ -1,5 +1,5 @@ -mtime: 1788419298 -commit: c96108ab6c009e2e5c89a409a9fb99202328af538ecc4b079b1f331e2bb65cef +mtime: 1792144501 +commit: 7608850e4613850af1971c90698b620d001623a1f798e0dfac9b06a60cd7371f url: https://src.opensuse.org/AI/ollama.git revision: main ++++++ build.specials.obscpio ++++++ --- old/.gitignore 2026-09-03 09:08:18.000000000 +0200 +++ new/.gitignore 2026-10-16 11:55:01.000000000 +0200 @@ -8,3 +8,10 @@ mlx mlx-c +*.obsinfo +binaries +_service:* +*.orig +*.rej +*~ +osc-collab.* ++++++ build.specials.obscpio ++++++ diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn' '--exclude=.svnignore' old/.gitignore new/.gitignore --- old/.gitignore 1970-01-01 01:00:00.000000000 +0100 +++ new/.gitignore 2026-10-16 11:55:01.000000000 +0200 @@ -0,0 +1,17 @@ +*.obscpio +*.osc +_build.* +.pbuild +.assets +ollama +llama.cpp +mlx +mlx-c + +*.obsinfo +binaries +_service:* +*.orig +*.rej +*~ +osc-collab.* ++++++ llama.cpp-main.tar.xz ++++++ /work/SRC/openSUSE:Factory/ollama/llama.cpp-main.tar.xz /work/SRC/openSUSE:Factory/.ollama.new.3698736/llama.cpp-main.tar.xz differ: char 15, line 1 ++++++ mlx-c-main.tar.xz ++++++ /work/SRC/openSUSE:Factory/ollama/mlx-c-main.tar.xz /work/SRC/openSUSE:Factory/.ollama.new.3698736/mlx-c-main.tar.xz differ: char 15, line 1 ++++++ mlx-main.tar.xz ++++++ ++++ 64890 lines of diff (skipped) ++++++ ollama-0.33.1.tar.gz -> ollama-0.40.5.tar.gz ++++++ /work/SRC/openSUSE:Factory/ollama/ollama-0.33.1.tar.gz /work/SRC/openSUSE:Factory/.ollama.new.3698736/ollama-0.40.5.tar.gz differ: char 12, line 1 ++++++ vendor.tar.gz ++++++ /work/SRC/openSUSE:Factory/ollama/vendor.tar.gz /work/SRC/openSUSE:Factory/.ollama.new.3698736/vendor.tar.gz differ: char 13, line 1
