GH-49563: [C++][CMake] Remove clang/infer tools detection

[alinaliBQ]

Rationale for this change

GH-49563

What changes are included in this PR?

This issue occurred without any code changes, so I think it is due to an macOS update.

After a GitHub macOS update, the CLANG_TOOLS_PATH path is resolved as /usr/local/bin/clang-format in Intel CI. Before the GitHub macOS update, CLANG_TOOLS_PATH was resolved as /usr/local/opt/llvm@18/bin and we did not have the issue of libunwind linking dynamically to ODBC dylib.

We don't know why clang tools change is related to libunwind linking but we don't need to detect clang/infer tools in our CMake now. (We migrated to pre-commit from Archery for linting.) So we can fix this issue by removing clang/infer tools detection from our CMake configuration.

Are these changes tested?

• Tested in CI

Are there any user-facing changes?

N/A

• GitHub Issue: [C++][FlightRPC][ODBC] Unexpected dependency libunwind in macOS Intel CI #49563

[github-actions]

⚠️ GitHub issue #49563 has been automatically assigned in GitHub to PR creator.

[alinaliBQ]

@kou @lidavidm This PR is ready for review, please have a look

[kou]

Hmm. I don't know why this fixes the problem... The build doesn't use LLVM because Gandiva is off. Why do clang based tools relate to linked libraries...?

Anyway, we don't need them. Can we solve the problem by removing clang-tools and infer-tools related codes something like the following?

diff --git a/cpp/CMakeLists.txt b/cpp/CMakeLists.txt
index ea15bb7066..a77ed39eac 100644
--- a/cpp/CMakeLists.txt
+++ b/cpp/CMakeLists.txt
@@ -211,17 +211,6 @@ else()
   set(MSVC_TOOLCHAIN FALSE)
 endif()
 
-find_package(ClangTools)
-find_package(InferTools)
-if("$ENV{CMAKE_EXPORT_COMPILE_COMMANDS}" STREQUAL "1"
-   OR CLANG_TIDY_FOUND
-   OR INFER_FOUND)
-  # Generate a Clang compile_commands.json "compilation database" file for use
-  # with various development tools, such as Vim's YouCompleteMe plugin.
-  # See http://clang.llvm.org/docs/JSONCompilationDatabase.html
-  set(CMAKE_EXPORT_COMPILE_COMMANDS 1)
-endif()
-
 # Needed for Gandiva.
 # Use the first Python installation on PATH, not the newest one
 set(Python3_FIND_STRATEGY "LOCATION")
@@ -649,22 +638,6 @@ if(UNIX)
                     VERBATIM)
 endif(UNIX)
 
-#
-# "make infer" target
-#
-
-if(${INFER_FOUND})
-  # runs infer capture
-  add_custom_target(infer ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                          ${CMAKE_BINARY_DIR}/compile_commands.json 1)
-  # runs infer analyze
-  add_custom_target(infer-analyze ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                                  ${CMAKE_BINARY_DIR}/compile_commands.json 2)
-  # runs infer report
-  add_custom_target(infer-report ${BUILD_SUPPORT_DIR}/run-infer.sh ${INFER_BIN}
-                                 ${CMAKE_BINARY_DIR}/compile_commands.json 3)
-endif()
-
 #
 # Link targets
 #

git rm cpp/cmake_modules/FindClangTools.cmake
git rm cpp/cmake_modules/FindInferTools.cmake

[alinaliBQ]

Sure, let's try that. I have pushed the changes

[alinaliBQ]

@kou The new approach worked, I have updated the PR description, thanks!

[kou]

+1

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit bbfb242.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 10 possible false positives for unstable benchmarks that are known to sometimes produce them.