OverflowError: Python integer 256 out of bounds for uint8

Start in wfdb/io/annotation.py at proc_core_fields, using the traceback and the reported demo4.py invocation to reproduce the OverflowError. Compare the separate script's rdann result with the notebook behavior; done means demo4.py runs successfully on the sample data without the uint8 overflow.

I am breaking the demos in the notebook in individual scripts to get acquainted with the package to learn how it works and adapt it to my needs.

Running demo4 as a separate script with python I get this error traceback:

D:\Users\xxx\Work\__WFDB\wfdb-python>python demo4.py
Traceback (most recent call last):
  File "D:\Users\xxx\Work\__WFDB\wfdb-python\demo4.py", line 12, in <module>
    annotation = wfdb.rdann('sample-data/100', 'atr', sampfrom=100000, sampto=110000)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Users\xxx\Work\__WFDB\wfdb-python\wfdb\io\annotation.py", line 1953, in rdann
    (sample, label_store, subtype, chan, num, aux_note) = proc_ann_bytes(
                                                          ^^^^^^^^^^^^^^^
  File "D:\Users\xxx\Work\__WFDB\wfdb-python\wfdb\io\annotation.py", line 2154, in proc_ann_bytes
    sample_diff, current_label_store, bpi = proc_core_fields(filebytes, bpi)
                                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Users\xxx\Work\__WFDB\wfdb-python\wfdb\io\annotation.py", line 2240, in proc_core_fields
    sample_diff += int(filebytes[bpi, 0] + 256 * (filebytes[bpi, 1] & 3))
                                           ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
OverflowError: Python integer 256 out of bounds for uint8

The same happens with Linux (Ubuntu 22.04).

After changing the highlighted line it works like the notebook.

$ git diff
diff --git a/wfdb/io/annotation.py b/wfdb/io/annotation.py
index b398fa0..4d451b1 100644
--- a/wfdb/io/annotation.py
+++ b/wfdb/io/annotation.py
@@ -2237,7 +2237,7 @@ def proc_core_fields(filebytes, bpi):

     # Not a skip - it is the actual sample number + annotation type store value
     label_store = filebytes[bpi, 1] >> 2
-    sample_diff += int(filebytes[bpi, 0] + 256 * (filebytes[bpi, 1] & 3))
+    sample_diff += int(filebytes[bpi, 0]) + 256 * int(filebytes[bpi, 1] & 3)
     bpi = bpi + 1

     return sample_diff, label_store, bpi

EDIT:
Python versions
Windows: 3.12.4
Linux: 3.10.12