[C++] Better support optional start/stop in "utf8_slice_codeunits" kernel

[jorisvandenbossche]

Describe the bug, including details regarding any error messages, version, and platform.

There have been various (slightly different) bugs reported about using "utf8_slice_codeunits" with optional start or stop. The stop argument is already optional and translated into the largest int to indicate to always slice until the end, but that internal "workaround" also produces some bugs in the current implementation due to integer overflows.

Potentially, we could use a different mechanism to signal a default start/stop, such as using std::optional<int64_t> instead of std::numeric_limits<int64_t>::max()

Listing the related issues:

• https://github.com/apache/arrow/issues/28940
• https://github.com/apache/arrow/issues/14991
• https://github.com/apache/arrow/issues/34917
• https://github.com/apache/arrow/issues/34928

The option class is also used for "binary_slice" kernel.

Component(s)

C++

[pitrou]

cc @benibus . This would slightly break the C++ API so we have to make sure this would make things better.

[benibus]

If we were to modify SliceOptions directly then that would also affect the ascii kernel as well. Not opposed to doing that, just wanted to give a heads-up.

In any case, I'll probably try this with a distinct options class first to see how it goes with one of the kernels. Then we can determine if the breaking change is justified.

[pitrou]

If we were to modify SliceOptions directly then that would also affect the ascii kernel as well. Not opposed to doing that, just wanted to give a heads-up.

I think it's fine to affect both kernels, and actually it's quite logical as well.

If std::optional had been available to us before, we would probably have used it from the start here.