ARROW-16778: [C++] Fix build/unit test issues in msvc/win32

[avertleyb]

No description provided.

[github-actions]

Thanks for opening a pull request!

If this is not a minor PR. Could you open an issue for this pull request on JIRA? https://issues.apache.org/jira/browse/ARROW

Opening JIRAs ahead of time contributes to the Openness of the Apache Arrow project.

Then could you also rename pull request title in the following format?

ARROW-${JIRA_ID}: [${COMPONENT}] ${SUMMARY}

or

MINOR: [${COMPONENT}] ${SUMMARY}

See also:

• Other pull requests
• Contribution Guidelines - How to contribute patches

[github-actions]

https://issues.apache.org/jira/browse/ARROW-16778

[github-actions]

⚠️ Ticket has not been started in JIRA, please click 'Start Progress'.

[kou]

Could you also add a CI job for building 32bit binaries with Visual C++?

The following patch will work:

diff --git a/.github/workflows/cpp.yml b/.github/workflows/cpp.yml
index 82e99160c3..6eb2213c93 100644
--- a/.github/workflows/cpp.yml
+++ b/.github/workflows/cpp.yml
@@ -192,12 +192,15 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os:
-          - windows-2019
         include:
           - os: windows-2019
-            name: Windows 2019
+            name: Windows 2019 x64
             generator: Visual Studio 16 2019
+            platform: x64
+          - os: windows-2019
+            name: Windows 2019 Win32
+            generator: Visual Studio 16 2019
+            platform: Win32
     env:
       ARROW_BOOST_USE_SHARED: OFF
       ARROW_BUILD_BENCHMARKS: ON
@@ -222,7 +225,7 @@ jobs:
       ARROW_WITH_ZLIB: ON
       ARROW_WITH_ZSTD: ON
       BOOST_SOURCE: BUNDLED
-      CMAKE_ARGS: '-A x64 -DOPENSSL_ROOT_DIR=C:\Program Files\OpenSSL-Win64'
+      CMAKE_ARGS: '-A ${{ matrix.platform }} -DOPENSSL_ROOT_DIR=C:\Program Files\OpenSSL-Win64'
       CMAKE_GENERATOR: ${{ matrix.generator }}
       CMAKE_INSTALL_LIBDIR: bin
       CMAKE_INSTALL_PREFIX: /usr

[kou]

I'm not sure which is better for performance but it seems that we can use bit operations to split uint64_t value to uint32_t values:

  return ARROW_POPCOUNT32((bitmap >> 32) & UINT32_MAX) +
         ARROW_POPCOUNT32(bitmap & UINT32_MAX);

[avertleyb]

If performance is critical, I would assume this version would just add a couple of bit operation. But if you feel it looks cleaner I can change this. Please let me know.

[avertleyb]

@kou updated PR

[avertleyb]

I just found out that I need to follow certain process while submitting fixes to open source projects.

I will re-submit these changes soon.

[jmckenna]

I see that the win32 workflow added here, is now removed from main (this would explain why users cannot compile arrow per #32111 ). Can we add this workflow back into the main branch?

[kou]

This wasn't merged.
Do you want to work on this (#32111)?