test: add doctest that would infinite-loop with naive shift fix

Addresses reviewer feedback by adding a test case where the mismatched
character ('C') appears to the RIGHT of the mismatch offset in the
pattern 'ABCB'. A naive shift that searches the entire pattern would
compute a negative shift and loop forever; the correct implementation
only searches to the left of the mismatch offset, guaranteeing shift>=1.

If the shift logic regresses, this doctest will hang instead of passing.

Author: almogyalin

strings/boyer_moore_search.py

@@ -116,6 +116,16 @@ def bad_character_heuristic(self) -> list[int]:
         >>> bms = BoyerMooreSearch(text="ABAABA", pattern="AB")
         >>> bms.bad_character_heuristic()
         [0, 3]
+
+        The text below contains "CCB" which causes the mismatched character 'C'
+        to appear to the RIGHT of the mismatch offset inside the pattern "ABCB".
+        A naive shift (searching the entire pattern) would compute a zero or
+        negative shift and loop forever; the correct shift only looks to the left
+        of the mismatch offset, guaranteeing forward progress.
+
+        >>> bms = BoyerMooreSearch(text="ABCBACCBABCB", pattern="ABCB")
+        >>> bms.bad_character_heuristic()
+        [0, 8]
         """
 
         positions = []