GH-50859: [C++][Parquet] Move JsonWriter to simdjson utilities

[rok]

Rationale for this change

Parquet uses JsonWriter when ARROW_JSON=OFF, but its implementation was only built with Arrow JSON, causing link failures.

What changes are included in this PR?

Move JsonWriter to the simdjson utilities and update its callers and CMake/Meson builds.

Are these changes tested?

Yes. CMake shared/static and Meson Parquet builds pass with JSON disabled. Unit tests and pre-commit checks also pass.

Are there any user-facing changes?

No. This only fixes the affected build configuration.

AI disclosure - this was AI generated to test alternative approach to #50900.

• GitHub Issue: [C++] parquet tools: linking error when ARROW_JSON=OFF with undefined reference to `arrow::json::JsonWriter::GetString() const' #50859

[pitrou]

The CI failures need fixing.

[rok]

@pitrou done.

[Reranko05]

@rok Since simdjson_internal.cc now exists, shouldn't all the non-template functions currently in simdjson_internal.h be moved into simdjson_internal.cc, leaving only declarations in the header?

[rok]

@rok Since simdjson_internal.cc now exists, shouldn't all the non-template functions currently in simdjson_internal.h be moved into simdjson_internal.cc, leaving only declarations in the header?

Done

[Reranko05]

LGTM!

[rok]

Again, this PR is purely AI generated. I will mark as ready for review once I review myself.

[pitrou]

@rok Are you willing to prioritize this?

[rok]

@pitrou I'll review this in about an hour and ping again

[tadeja]

@github-actions crossbow submit example-cpp-tutorial

[github-actions]

Revision: 973b4d2

Submitted crossbow builds: ursacomputing/crossbow @ actions-5765e8701c

Task	Status
example-cpp-tutorial

[rok]

Should we use arrow::simdjson here instead? Both should work, but arrow alieas would hide the vendored/system simdjson.

[wgtmac]

if(SIMDJSON_VENDORED)
  add_library(arrow::simdjson ALIAS simdjson)
else()
  add_library(arrow::simdjson ALIAS simdjson::simdjson)
endif()

It seems better to use arrow::simdjson based on the above statements? cc @kou

[rok]

Agreed, changed to arrow::simdjson.

[rok]

@github-actions crossbow submit example-cpp-tutorial

[github-actions]

Revision: 86e85bd

Submitted crossbow builds: ursacomputing/crossbow @ actions-e4578321bd

Task	Status
example-cpp-tutorial

[Copilot]

Pull request overview

Copilot reviewed 34 out of 34 changed files in this pull request and generated no new comments.

Suppressed comments (1)

cpp/src/parquet/encryption/key_material.h:79

• This removes an exported overload from the installed parquet/encryption/key_material.h, so existing source and binaries that call KeyMaterial::Parse(const arrow::json::internal::ObjectParser*) will no longer compile/link. That contradicts the stated absence of user-facing changes. Please preserve the existing parser type/overload while making its implementation available when simdjson is enabled, or explicitly treat and document this as an API/ABI break.

  static KeyMaterial Parse(const std::string& key_material_string);

[Copilot]

Pull request overview

Copilot reviewed 34 out of 34 changed files in this pull request and generated no new comments.

[Copilot]

Pull request overview

Copilot reviewed 34 out of 34 changed files in this pull request and generated no new comments.

[rok]

@github-actions crossbow submit -g cpp

[github-actions]

Revision: d15007a

Submitted crossbow builds: ursacomputing/crossbow @ actions-350780c611

Task	Status
example-cpp-minimal-build-static
example-cpp-minimal-build-static-system-dependency
example-cpp-tutorial
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-debian-13-cpp-amd64
test-debian-13-cpp-i386
test-debian-experimental-cpp-gcc-15
test-fedora-42-cpp
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-bundled
test-ubuntu-22.04-cpp-emscripten
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-bundled-offline
test-ubuntu-24.04-cpp-gcc-13-bundled
test-ubuntu-24.04-cpp-gcc-14
test-ubuntu-24.04-cpp-minimal-with-formats
test-ubuntu-24.04-cpp-thread-sanitizer

[pitrou]

Thanks a lot @rok !

[pitrou]

CI failures are unrelated to this PR.