apacheGH-34638: [C++][Docs] Add documentation for minimal build flags (…

…apache#34693) ### Rationale for this change There's an ongoing effort to support smaller builds of Arrow C++ for simple format integration so explicitly documenting the necessary flags would be useful. Also, some recent changes to the build system may require disclosure. ### What changes are included in this PR? Adds documentation to `cpp/building.rst` regarding: - Minimal cmake flags for basic format support - Recent changes to `ARROW_COMPUTE` and how compute kernels are included in builds. ### Are these changes tested? N/A ### Are there any user-facing changes? N/A * Closes: apache#34638 Authored-by: benibus <[email protected]> Signed-off-by: Weston Pace <[email protected]>
JackDrogon · Mar 24, 2023 · e7d6c13 · e7d6c13
1 parent 071ea1b
commit e7d6c13
Showing 1 changed file with 14 additions and 1 deletion.
diff --git a/docs/source/developers/cpp/building.rst b/docs/source/developers/cpp/building.rst
@@ -317,7 +317,7 @@ several optional system components which you can opt into building by passing
 boolean flags to ``cmake``.
 
 * ``-DARROW_BUILD_UTILITIES=ON`` : Build Arrow commandline utilities
-* ``-DARROW_COMPUTE=ON``: Computational kernel functions and other support
+* ``-DARROW_COMPUTE=ON``: Build all computational kernel functions
 * ``-DARROW_CSV=ON``: CSV reader module
 * ``-DARROW_CUDA=ON``: CUDA integration for GPU development. Depends on NVIDIA
   CUDA toolkit. The CUDA toolchain used to build the library can be customized
@@ -369,6 +369,19 @@ build times if they are not required for your application:
 
 * ``-DARROW_IPC=ON``: build the IPC extensions
 
+.. note::
+   If your use-case is limited to reading/writing Arrow data then the default
+   options should be sufficient. However, if you wish to build any tests/benchmarks
+   then ``ARROW_JSON`` is also required (it will be enabled automatically).
+   If extended format support is desired then adding ``ARROW_PARQUET``, ``ARROW_CSV``,
+    ``ARROW_JSON``, or ``ARROW_ORC`` shouldn't enable any additional components.
+
+.. note::
+   In general, it's a good idea to enable ``ARROW_COMPUTE`` if you anticipate using
+   any compute kernels beyond ``cast``. While there are (as of 12.0.0) a handful of
+   additional kernels built in by default, this list may change in the future as it's
+   partly based on kernel usage in the current format implementations.
+
 .. warning::
    Plasma is deprecated as of Arrow 10.0.0, and will be removed in 12.0.0 or so.