[Python][C++] Segfault reading bz2 compressed csv file with CSVStreamingReader

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

I am getting reproducible segfaults when trying to close a CSVWriter in the simplest code possible.

```python
import pyarrow as pa
from pyarrow import csv
import argparse

parser = argparse.ArgumentParser(description="Converts a potentially bz2/gz compressed csv file to a tsv/smi file with no header and just 'smiles' and 'id' columns.")
parser.add_argument("input_file", help="Path to the input bz2 file")
parser.add_argument("output_file", help="Path to the output csv file")
parser.add_argument("--chunksize", type=int, default=2**30, help="Size of block to process at a time")
parser.add_argument("--num_chunks", type=int, default=0, help="Number of chunks to process")

args = parser.parse_args()

input_file = args.input_file
output_file = args.output_file
chunksize = args.chunksize

readopt = csv.ReadOptions(block_size=chunksize)
parseopt = csv.ParseOptions(delimiter="\t")
convertopt = csv.ConvertOptions(column_types={"smiles": pa.string(), "id": pa.string()}, include_columns=["smiles", "id"])

reader = csv.open_csv(
input_file,
read_options=readopt,
parse_options=parseopt,
convert_options=convertopt
)

writeopt = csv.WriteOptions(include_header=False, delimiter="\t", quoting_style="none")
# schema is just smiles and id column, both strings
schema = pa.schema([
('smiles', pa.string()),
('id', pa.string())
])

with csv.CSVWriter(sink=output_file, schema=schema, write_options=writeopt) as writer:

cnt = 0
while cnt < args.num_chunks or args.num_chunks == 0:
try:
writer.write_batch(reader.read_next_batch())
cnt += 1
print("Finished chunk ", cnt)
except StopIteration:
break

print("Done. Closing file.")
```

Output for num_chunks = 2 e.g.:
It hangs for a long time after "Closing file."
```
Finished chunk 1
Finished chunk 2
Done. Closing file.
Segmentation fault (core dumped)
```

linux aarch64, pyarrow 17 (tried both pip and conda-forge), also tried many combinations of filesystem, pa.OSFile, batch_sizes, closing manually vs contexts, etc.
My input file is b2zipped.

The culprit seems to be the limiting of the number of chunks that I added for testing. Writing ALL chunks seems to work.

### Component(s)

Python