GH-47848: [C++] Add TimeUnit::PICO

[tswast]

Rationale for this change

I'm working on a system that supports picosecond precision timestamps, and I'm interested in interoperability with the Arrow ecosystem. See #47848

What changes are included in this PR?

This PR only adds the new enum value TimeUnit::PICO. This enum value is unused for now, but I plan to introduce a timestamp128 type in the future which supports this unit type.

Are these changes tested?

I have updated and run the C++ test suite.

root@dc20d61d62a2:/github.com/apache/arrow/cpp/build# 
...
100% tests passed, 0 tests failed out of 91

Label Time Summary:
arrow-compute-tests    = 128.41 sec*proc (15 tests)
arrow-tests            =  65.74 sec*proc (38 tests)
arrow_acero            = 120.93 sec*proc (14 tests)
arrow_dataset          = 103.61 sec*proc (13 tests)
arrow_substrait        =   0.98 sec*proc (1 test)
filesystem             =   1.97 sec*proc (1 test)
parquet-tests          =  88.40 sec*proc (10 tests)
unittest               = 508.07 sec*proc (91 tests)

Total Test time (real) =  50.20 sec

Are there any user-facing changes?

This PR includes breaking changes to public APIs.

Adding a new enum variable did break some tests that assumed that all TimeUnit::values() are valid for use in the timestamp (64-bit) data type.

• GitHub Issue: timestamp128 data type with support for picosecond units #47848

[github-actions]

⚠️ GitHub issue #47848 has been automatically assigned in GitHub to PR creator.

[tswast]

I think this change isn't necessary for this particular PR, but it will be for timestamp128 support. Let me know if you want me to break this out.

[tswast]

Actually, maybe it is, because otherwise cpp/src/arrow/array/diff.cc won't compile. I needed to be able to handle all unit enum values, and I don't think it makes sense for picoseconds to be stored in anything smaller than int128.

[tswast]

Looks like I probably need to adjust some linker stuff on Windows? Lots of errors like this:

"C:\Program Files\Git\usr\bin\ccache.exe" C:\PROGRA~1\MICROS~2\2022\ENTERP~1\VC\Tools\MSVC\1444~1.352\bin\Hostx64\x64\cl.exe  /nologo /TP -DARROW_EXPORTING -DARROW_HAVE_RUNTIME_AVX2 -DARROW_HAVE_RUNTIME_AVX512 -DARROW_HAVE_RUNTIME_BMI2 -DARROW_HAVE_RUNTIME_SSE4_2 -DARROW_HAVE_SSE4_2 -DARROW_WITH_TIMING_TESTS -D_CRT_SECURE_NO_WARNINGS -D__SSE2__ -D__SSE4_1__ -D__SSE4_2__ -ID:\a\arrow\arrow\build\cpp\src -ID:\a\arrow\arrow\cpp\src -ID:\a\arrow\arrow\cpp\src\generated /D_SILENCE_TR1_NAMESPACE_DEPRECATION_WARNING /W3 /EHsc /wd5105 /bigobj /utf-8 /W3 /wd4365 /wd4267 /wd4838 /wd4800 /wd4996 /wd4065 /arch:SSE4.2  /Ob0 /Od /RTC1 /WX -std:c++17 -MDd -Zi /showIncludes /Fosrc\arrow\CMakeFiles\arrow_array_shared.dir\array\diff.cc.obj /Fdsrc\arrow\CMakeFiles\arrow_array_shared.dir\ /FS -c D:\a\arrow\arrow\cpp\src\arrow\array\diff.cc
D:\a\arrow\arrow\cpp\src\arrow/util/int128_internal.h(26): fatal error C1083: Cannot open include file: 'boost/multiprecision/cpp_int.hpp': No such file or directory

[wgtmac]

This seems to be a spec change. Perhaps it is better to discuss this on the [email protected] per the process https://arrow.apache.org/docs/format/Changing.html#discussion-and-voting-process