- Require libffi when we build openmp for offloading.
- Update llvm17.keyring from upstream.
- Correct target name for libomptarget in file list (backport from llvm18)
- Enable lldb on s390x and ppc64le (bsc#1232906).

• Files Changed
• Browse Source

- Add clang-Fix-build-with-GCC-14-on-ARM.patch to fix build with
  GCC 14 on aarch64. (boo#1229868, gh#llvm/llvm-project#78691)

• Files Changed
• Browse Source

- Add minor version to Python shebangs. (boo#1212476)
- Remove shebang and executable bit where not needed.

• Files Changed
• Browse Source

- Use old conflicts/provides for python3-clang on Leap.

• Files Changed
• Browse Source

- Explicitly define PYTHON3_EXECUTABLE, since python 3.11 is being
  pulled in by cmake in SLE-15-SP6.
- Fix PowerPC triples: use powerpc[64[le]] instead of ppc[64[le]].
- Require 8 GB memory for 64-bit architectures to prevent OOM.
- Use %patch -P N instead of deprecated %patchN.

• Files Changed
• Browse Source

- Use %_target_* for determining the compiler host triple (boo#1218701)

• Files Changed
• Browse Source

- Update to version 17.0.6.
  * This release contains bug-fixes for the LLVM 17.0.0 release.
    This release is API and ABI compatible with 17.0.0.
- Rebase patches:
  * llvm-do-not-install-static-libraries.patch
  * llvm-normally-versioned-libllvm.patch
- Use major version only instead of full (patch-level) version for
  versioned executables: we don't support parallel installation of
  multiple versions of the same LLVM major version. (boo#1217450)
  Also drop the clang-X.Y symlink that was introduced for
  boo#1012260, because Y is always 0.
- Manage clang-cpp with update-alternatives like other binaries.

• Files Changed
• Browse Source

- Update to version 17.0.5.
  * This release contains bug-fixes for the LLVM 17.0.0 release.
    This release is API and ABI compatible with 17.0.0.
- Rebase llvm-do-not-install-static-libraries.patch.
- Also test clang-tools-extra (at least most parts) and lld.
- Adapt test in lld-default-sha1.patch.
- Don't disable testing if qemu_user_space_build has been set to 0.

• Files Changed
• Browse Source

- Update to version 17.0.4.
  * This release contains bug-fixes for the LLVM 17.0.0 release.
    This release is API and ABI compatible with 17.0.0.
- Rebase llvm-do-not-install-static-libraries.patch.
- Include libomptarget.devicertl.a in libomp*-devel, which is
  required for (GPU) offloading.

• Files Changed
• Browse Source

- Update to version 17.0.3.
  * This release contains bug-fixes for the LLVM 17.0.0 release.
    This release is API and ABI compatible with 17.0.0.
- Rebase llvm-do-not-install-static-libraries.patch.
- Disable zstd support for now regardless of whether libzstd-devel
  is installed. (boo#1216121)
- Limit number of LTO threads on 32-bit architectures to prevent
  out-of-memory during linking.

• Files Changed
• Browse Source

- Update to version 17.0.2.
  * This release contains bug-fixes for the LLVM 17.0.0 release.
    This release is API and ABI compatible with 17.0.0.
- Rebase llvm-do-not-install-static-libraries.patch.

• Files Changed
• Browse Source

- Initial release based on version 17.0.1.
  * For details, see the release notes:
    - https://releases.llvm.org/17.0.1/docs/ReleaseNotes.html
    - https://releases.llvm.org/17.0.1/tools/clang/docs/ReleaseNotes.html
    - https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/ReleaseNotes.html
    - https://releases.llvm.org/17.0.1/projects/libcxx/docs/ReleaseNotes.html
    - https://releases.llvm.org/17.0.1/tools/lld/docs/ReleaseNotes.html
- Rebase patches:
  * libcxx-test-library-path.patch
  * llvm_build_tablegen_component_as_shared_library.patch
  * llvm-do-not-install-static-libraries.patch
  * llvm-normally-versioned-libllvm.patch
  * llvm-remove-clang-only-flags.patch
  * opt-viewer-Find-style-css-in-usr-share.patch
- Simplify check-no-llvm-exegesis.patch by removing test directory.
- Drop patches that have landed upstream:
  * openmp-drop-rpath.patch
- Add openmp-dont-run-gpu-arch.patch: the binaries are not
  available in our build, and neither are the toolchains they test
  for (NV CUDA and AMD HSA).

• Browse Source