[C++] utf8_slice_codeunits doesn't support stop/step array type

### Describe the enhancement requested

Currently arrow compute `utf8_slice_codeunits` doesn't stop/step parameters with array type. When running following code, it'll raise `TypeError`
```python
import pyarrow as pa
import pyarrow.compute as pc

animals = pa.array(["A11A", "Horse", "Brittle stars", "Centipede"])
pc.utf8_slice_codeunits(animals, start=0, stop=pa.array([1, 2, 3, 4]))
```

Exception:
```
Traceback (most recent call last):
File "/Users/chaokunyang/anaconda3/envs/py3.8/lib/python3.8/site-packages/IPython/core/interactiveshell.py", line 3460, in run_code
File "/var/folders/zz/mnr4zfjn45j2d_vl88bz8hm80000gp/T/ipykernel_13874/31766106.py", line 1, in
pc.utf8_slice_codeunits(animals, start=0, stop=pa.array([1, 2, 3, 4]))
File "/Users/chaokunyang/anaconda3/envs/py3.8/lib/python3.8/site-packages/pyarrow/compute.py", line 252, in wrapper
File "/Users/chaokunyang/anaconda3/envs/py3.8/lib/python3.8/site-packages/pyarrow/compute.py", line 215, in _handle_options
def _wrap_function(name, func):
File "pyarrow/_compute.pyx", line 1168, in pyarrow._compute.SliceOptions.__init__
File "pyarrow/_compute.pyx", line 1147, in pyarrow._compute._SliceOptions._set_options
TypeError: an integer is required
```

It would be nice to add such support into arrow compute.

### Component(s)

C++

Start with the pyarrow.compute.utf8_slice_codeunits entry point and the SliceOptions handling named in the traceback. Reproduce the example with an array-valued stop, then investigate the corresponding Arrow compute path and add regression coverage for array-valued stop and step parameters. Done means the example no longer raises TypeError and produces the expected sliced array.

Written by the indexing model from the issue text.