BUG: StringDType mis-handles 255-character ASCII strings during contiguous slice assignment #29261
Description
Describe the issue:
Slice-assigning values into a StringDType array silently corrupts the target elements whenever the source slice contains at least one ASCII string of exactly 255 characters.
-
The assignment raises no exception, but a subsequent equality check fails (AssertionError).
-
Repeating the same assignment twice makes the data compare equal, suggesting the first write leaves the internal storage in an inconsistent state.
Reproduce the code example:
import numpy as np
chunk_name = np.array(
[
'0' * 256,
'0' * 256,
'0' * 256,
'0' * 255,
'0' * 256,
'0' * 255,
'0' * 256,
'0' * 256,
'0' * 255,
],
dtype=np.dtypes.StringDType()
)
events_name = np.array(
[
'0' * 255,
'0' * 256,
'0' * 256,
'0' * 255,
'0' * 256,
'0' * 255,
'',
'',
'',
'',
''
],
dtype=np.dtypes.StringDType()
)
N = 9
events_name[1:N+1] = chunk_name
assert np.all(np.strings.equal(events_name[1:N+1], chunk_name))Error message:
Traceback (most recent call last):
File "/shared/home/Fortitude42/work/x.py", line 39, in <module>
assert np.all(np.strings.equal(events_name[1:N+1], chunk_name))Python and NumPy Versions:
- Python version: 3.11.12
- NumPy version: 2.2.5
Runtime Environment:
No response
Context for the issue:
No response