[C++][Compute] "error: unused variable ‘is_utf8’" can occur when building the Arrow C++ main branch

[komainu8]

Problem

The following build error can occur when building the Arrow C++ main branch.

/arrow/cpp/src/arrow/compute/kernels/scalar_string_ascii.cc:1753:16: error: unused variable ‘is_utf8’ [-Werror=unused-variable]
 1753 |     const bool is_utf8 = is_string_or_string_view(batch[0].type()->id());
      |                ^~~~~~~

Reproduce

This error occurs under the following conditions:

• Use Debug build(Use -Werror)
• Specify ARROW_COMPUTE=ON
• Specify ARROW_WITH_RE2=OFF

Steps to reproduce:

mkdir arrow-cpp-build
git clone https://github.com/apache/arrow.git
cd arrow-cpp-build
cmake ../arrow/cpp --preset ninja-debug-minimal -DARROW_COMPUTE=ON -DARROW_WITH_RE2=OFF
cmake --build .

Cause

The is_utf8 variable introduced by commit 374db36 is unused when ARROW_WITH_RE2=OFF is specified.

  static Status Exec(KernelContext* ctx, const ExecSpan& batch, ExecResult* out) {
    const MatchSubstringOptions& options = MatchSubstringState::Get(ctx);
+   const bool is_utf8 = is_string_or_string_view(batch[0].type()->id());
    if (options.ignore_case) {
#ifdef ARROW_WITH_RE2
      ARROW_ASSIGN_OR_RAISE(auto matcher,
-                           FindSubstringRegex::Make(options, InputType::is_utf8, true));
-     applicator::ScalarUnaryNotNullStateful<OffsetType, InputType, FindSubstringRegex>
+                           FindSubstringRegex::Make(options, is_utf8, true));
+     applicator::ScalarUnaryNotNullStateful<OffsetType, InputPhysicalType,
+                                            FindSubstringRegex>
          kernel{std::move(matcher)};
      return kernel.Exec(ctx, batch, out);
#else
      return Status::NotImplemented("ignore_case requires RE2");
#endif
    }
-   applicator::ScalarUnaryNotNullStateful<OffsetType, InputType, FindSubstring> kernel{
-       FindSubstring(PlainSubstringMatcher(options))};
+   applicator::ScalarUnaryNotNullStateful<OffsetType, InputPhysicalType, FindSubstring>
+       kernel{FindSubstring(PlainSubstringMatcher(options))};
    return kernel.Exec(ctx, batch, out);
  }
};

Suggested fix

We can prevent this error by moving the declaration of is_utf8 inside the #ifdef ARROW_WITH_RE2 block as follows:

diff --git a/cpp/src/arrow/compute/kernels/scalar_string_ascii.cc b/cpp/src/arrow/compute/kernels/scalar_string_ascii.cc
index 06ec8c999c..8e7b626837 100644
--- a/cpp/src/arrow/compute/kernels/scalar_string_ascii.cc
+++ b/cpp/src/arrow/compute/kernels/scalar_string_ascii.cc
@@ -1750,9 +1750,9 @@ struct FindSubstringExec {

   static Status Exec(KernelContext* ctx, const ExecSpan& batch, ExecResult* out) {
     const MatchSubstringOptions& options = MatchSubstringState::Get(ctx);
-    const bool is_utf8 = is_string_or_string_view(batch[0].type()->id());
     if (options.ignore_case) {
 #ifdef ARROW_WITH_RE2
+      const bool is_utf8 = is_string_or_string_view(batch[0].type()->id());
       ARROW_ASSIGN_OR_RAISE(auto matcher,
                             FindSubstringRegex::Make(options, is_utf8, true));
       applicator::ScalarUnaryNotNullStateful<OffsetType, InputPhysicalType,

If this approach looks good to you, I'd be happy to open a PR.

Component(s)

C++

[raulcd]

Thanks @komainu8 ! The solution you are proposing sounds reasonable to me. Feel free to open a PR, I'll assign the issue to you as there's lots of bots looking for quick PRs lately :)

[komainu8]

@raulcd Thank you for your comment.
I've opened a PR at #50754.

[raulcd]

Issue resolved by pull request 50754
#50754