fix(handler): fix UBI PEB size calculation.

Computing the most frequent interval can only work if we return each and
every interval value observed in a list. Returning it in a set hides the
repeating interval patterns.

The 'get_intervals' function now _actually_ acts like numpy.diff.

Author: qkaiser

tests/test_iter_utils.py

@@ -6,11 +6,11 @@
 @pytest.mark.parametrize(
     "values, expected",
     [
-        ([], set()),
-        ([0, 0], {0}),
-        ([0, 0, 0], {0}),
-        ([1, 2, 3], {1}),
-        ([1, 5, 8, 8, 10, 15], {4, 3, 0, 2, 5}),
+        ([], []),
+        ([0, 0], [0]),
+        ([0, 0, 0], [0, 0]),
+        ([1, 2, 3], [1, 1]),
+        ([1, 5, 8, 8, 10, 15], [4, 3, 0, 2, 5]),
     ],
 )
 def test_get_internals(values, expected):

unblob/iter_utils.py

@@ -1,5 +1,5 @@
 import itertools
-from typing import List, Set
+from typing import List
 
 
 def pairwise(iterable):
@@ -10,7 +10,7 @@ def pairwise(iterable):
     return zip(a, b)
 
 
-def get_intervals(values: List[int]) -> Set[int]:
+def get_intervals(values: List[int]) -> List[int]:
     """Get all the intervals between numbers.
 
     It's similar to numpy.diff function.
@@ -20,7 +20,7 @@ def get_intervals(values: List[int]) -> Set[int]:
     >>> get_intervals([1, 4, 5, 6, 10])
     [3, 1, 1, 4]
     """
-    all_diffs = set()
+    all_diffs = []
     for value, next_value in pairwise(values):
-        all_diffs.add(next_value - value)
+        all_diffs.append(next_value - value)
     return all_diffs