ListArray.Append() was being indirectly invoked with Append(true, 0), which is a simple way to dynamically build a ListArray with unknown size. However ListViewArray needs the size included in the Append() API. In order for ListArray to be convertible to ListViewArray, it also needs the size assigned in Append(). That change is included here for ListArray and ListViewArray (MapArray and FixedSizeListArray don't support a parameterized Append() method). See this comment for more details:

Sounds good. Maybe add a brief comment about this? (why those AppendTo methods are defined here)

I suppose at the moment we get here, we know for sure that the object has a size? (eg a generator is iterable, but has no length. We do support generators as the main value passed to pa.array(..) (through ConvertToSequenceAndInferSize), but not nested in a list array, I think)

At the moment, AppendIterable() is only called for Sets and Dictionary values views. I confirmed generators are routed through AppendSequence() instead. I did give it a try as well:

>>> def gen():
...     yield [0]
...     yield [1]
...     yield [2]
...
>>> pa.array(gen(), type=pa.list_view(pa.int32()))
<pyarrow.lib.ListViewArray object at 0x12c96e6e0>
[
  [
    0
  ],
  [
    1
  ],
  [
    2
  ]
]
>>> pa.array([gen()], type=pa.list_view(pa.int32()))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "pyarrow/array.pxi", line 343, in pyarrow.lib.array
    result = _sequence_to_array(obj, mask, size, type, pool, c_from_pandas)
  File "pyarrow/array.pxi", line 42, in pyarrow.lib._sequence_to_array
    chunked = GetResultValue(
  File "pyarrow/error.pxi", line 154, in pyarrow.lib.pyarrow_internal_check_status
    return check_status(status)
  File "pyarrow/error.pxi", line 91, in pyarrow.lib.check_status
    raise convert_status(status)
pyarrow.lib.ArrowTypeError: Could not convert <generator object gen at 0x12bed75e0> with type generator: was not a sequence or recognized null for conversion to list type

Sounds good. Maybe add a brief comment about this? (why those AppendTo methods are defined here)

I would expect that this test needs some slight changes to work for the list view types? Because those use 3 instead of 2 buffers for the main array (apart from the child values)?

Maybe something like this instead? Because right now it doesn't really guarantee it was constructed correctly

Although this test might also not add really value compared to what is already tested in test_convert_builtin.py?

I will remove it, since I also think it is a redundant test. I duplicated this test from list/largelist unit tests unnecessarily.

Can you keep the case where type is None (inferred)? Eg add None to the parametrzation, and then after this line set ty to the default inferred one if it was None

Good catch! Will update