[ARROW-10193] [Python] Segfault when converting to fixed size binary array

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.0.0
• Component/s: Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/26198
• Assignee: Krisztian Szucs
• Reporter: Krisztian Szucs
• Created: 06/Oct/20 12:56
• Updated: 11/Jan/23 08:11
• Resolved: 06/Oct/20 17:42

Description

Reproducer:

data = [b'\x19h\r\x9e\x00\x00\x00\x00\x01\x9b\x9fA']
assert len(data[0]) == 12
ty = pa.binary(12)
arr = pa.array(data, type=ty)

Trace:

pyarrow/tests/test_convert_builtin.py::test_fixed_size_binary_length_check ../src/arrow/array/builder_binary.cc:53:  Check failed: (size) == (byte_width_) Appending wrong size to FixedSizeBinaryBuilder
0   libarrow.200.0.0.dylib              0x000000010e7f9704 _ZN5arrow4util7CerrLog14PrintBackTraceEv + 52
1   libarrow.200.0.0.dylib              0x000000010e7f9622 _ZN5arrow4util7CerrLogD2Ev + 98
2   libarrow.200.0.0.dylib              0x000000010e7f9585 _ZN5arrow4util7CerrLogD1Ev + 21
3   libarrow.200.0.0.dylib              0x000000010e7f95ac _ZN5arrow4util7CerrLogD0Ev + 28
4   libarrow.200.0.0.dylib              0x000000010e7f9492 _ZN5arrow4util8ArrowLogD2Ev + 82
5   libarrow.200.0.0.dylib              0x000000010e7f94c5 _ZN5arrow4util8ArrowLogD1Ev + 21
6   libarrow.200.0.0.dylib              0x000000010e303ec1 _ZN5arrow22FixedSizeBinaryBuilder14CheckValueSizeEx + 209
7   libarrow.200.0.0.dylib              0x000000010e30c361 _ZN5arrow22FixedSizeBinaryBuilder12UnsafeAppendEN6nonstd7sv_lite17basic_string_viewIcNSt3__111char_traitsIcEEEE + 49
8   libarrow_python.200.0.0.dylib       0x000000010b4efa7d _ZN5arrow2py20PyPrimitiveConverterINS_19FixedSizeBinaryTypeEvE6AppendEP7_object + 813

The input const char* value gets implicitly casted to string_view which makes the length check fail in debug builds.

Issue Links

• links to: GitHub Pull Request #8360