Script 'mail_helper' called by obssrc
Hello community,

here is the log from the commit of package ollama for openSUSE:Factory checked 
in at 2026-04-20 16:11:29
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/ollama (Old)
 and      /work/SRC/openSUSE:Factory/.ollama.new.11940 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Package is "ollama"

Mon Apr 20 16:11:29 2026 rev:60 rq:1348062 version:0.21.0

Changes:
--------
--- /work/SRC/openSUSE:Factory/ollama/ollama.changes    2026-04-11 
22:31:21.076604284 +0200
+++ /work/SRC/openSUSE:Factory/.ollama.new.11940/ollama.changes 2026-04-20 
16:11:45.414288324 +0200
@@ -1,0 +2,32 @@
+Fri Apr 17 07:40:40 UTC 2026 - Adrian Schröter <[email protected]>
+
+- Update to version 0.21.0
+  * launch: skip unchanged integration rewrite configration
+  * launch/openclaw: fix --yes flag behaviour to skip channels configuration
+  * launch: OpenCode inline config
+  * launch: add hermes by
+  * launch: always list cloud recommendations first
+  * cmd/launch: add Copilot CLI integration
+- fixing vulkan dependencies, should also fix runtime
+  (so far ollama claimed to use vulkan, but actually didn't)
+- check for missing symbols in cuda and vulkan backend if available
+
+-------------------------------------------------------------------
+Thu Apr 16 18:28:01 UTC 2026 - Egbert Eich <[email protected]>
+
+- Add HIP compiler flag `--offload-compress` for builds with ROCm
+  support. This will allow the package to build against ROCm 7.2
+  from science:GPU:ROCm.
+
+-------------------------------------------------------------------
+Wed Apr 15 16:14:31 UTC 2026 - Adrian Schröter <[email protected]>
+
+- Update to version 0.20.7
+  * Fix quality of gemma:e2b and gemma:e4b when thinking is disabled
+  * ROCm: Update to ROCm 7.2.1 on Linux by @saman-amd
+  * Gemma 4 tool calling ability is improved and updated to use Google's 
latest post-launch fixes
+  * Parallel tool calling improved for streaming responses
+  * Hermes agent Ollama integration guide is now available
+  * Ollama app is updated to fix image attachment errors
+
+-------------------------------------------------------------------

Old:
----
  ollama-0.20.5.tar.gz

New:
----
  ollama-0.21.0.tar.gz

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Other differences:
------------------
++++++ ollama.spec ++++++
--- /var/tmp/diff_new_pack.YHz6Ju/_old  2026-04-20 16:11:46.758343663 +0200
+++ /var/tmp/diff_new_pack.YHz6Ju/_new  2026-04-20 16:11:46.762343828 +0200
@@ -35,7 +35,7 @@
 %define cuda_version %{cuda_version_major}-%{cuda_version_minor}
 
 Name:           ollama
-Version:        0.20.5
+Version:        0.21.0
 Release:        0
 Summary:        Tool for running AI models on-premise
 License:        MIT
@@ -46,6 +46,7 @@
 Source3:        %{name}-user.conf
 Source4:        sysconfig.%{name}
 Patch0:         fix-mlxrunner-tests.diff
+BuildRequires:  ccache
 BuildRequires:  cmake >= 3.24
 BuildRequires:  git-core
 BuildRequires:  ninja
@@ -57,6 +58,11 @@
 BuildRequires:  group(render)
 BuildRequires:  group(video)
 BuildRequires:  pkgconfig(vulkan)
+# For Vulkan, or we have missing symbols
+BuildRequires:  glslang-devel
+BuildRequires:  libggml-base0
+BuildRequires:  vulkan-tools
+
 Requires:       group(render)
 Requires:       group(video)
 Recommends:     ( %{name}-vulkan or %{name}-cuda or %{name}-rocm )
@@ -164,7 +170,8 @@
     -DGGML_BACKEND_DIR=%{_libdir}/ollama \
     
%{?with_cuda:-DCMAKE_CUDA_COMPILER=/usr/local/cuda-%{cuda_version_major}.%{cuda_version_minor}/bin/nvcc}
 \
     %{?with_rocm:-DCMAKE_HIP_COMPILER=%rocmllvm_bindir/clang++ \
-       -DAMDGPU_TARGETS=%{rocm_gpu_list_default}} \
+       -DAMDGPU_TARGETS=%{rocm_gpu_list_default} \
+       -DCMAKE_HIP_FLAGS=--offload-compress} \
     %{nil}
 %cmake_build
 
@@ -197,6 +204,16 @@
 %endif
 go test -v ./...
 
+# verify for missing symbols to avoid shipping a broken version, which would
+# silently not load
+export LD_LIBRARY_PATH=%buildroot%_libdir/ollama
+for backend in cuda vulkan; do
+  file=%buildroot%_libdir/ollama/libggml-${backend}.so
+  test -e "$file" || continue
+  ldd -r "$file" | grep "undefined symbol:" && exit 1
+done
+exit 0
+
 %pre -f %{name}.pre
 %service_add_pre %{name}.service
 

++++++ ollama-0.20.5.tar.gz -> ollama-0.21.0.tar.gz ++++++
/work/SRC/openSUSE:Factory/ollama/ollama-0.20.5.tar.gz 
/work/SRC/openSUSE:Factory/.ollama.new.11940/ollama-0.21.0.tar.gz differ: char 
16, line 1

++++++ vendor.tar.zstd ++++++
Binary files /var/tmp/diff_new_pack.YHz6Ju/_old and 
/var/tmp/diff_new_pack.YHz6Ju/_new differ

Reply via email to