Merge pull request #95 from G-Research/fix-tests

Make the tests pass more consistently

Author: itamarst

.hypothesis/examples/098667f9ae029f55/b9fc09fff40c6832

@@ -1,6 +1,6 @@
 [package]
 name = "ahocorasick_rs"
-version = "0.20.0"
+version = "0.21.1"
 edition = "2021"
 authors = ["G-Research <[email protected]>", "Itamar Turner-Trauring <[email protected]>"]
 description = "Search a string for multiple substrings at once"

.hypothesis/examples/2568bb2bb8962727/b9fc09fff40c6832

@@ -7,7 +7,7 @@
 import pytest
 
 from hypothesis import strategies as st
-from hypothesis import given, assume
+from hypothesis import given
 
 from ahocorasick_rs import (
     AhoCorasick,
@@ -140,19 +140,18 @@ def test_unicode_extensive(
     Non-ASCII unicode patterns still give correct results for
     find_matches_as_indexes(), with property-testing.
     """
-    assume(pattern not in prefix)
-    assume(pattern not in suffix)
     haystack = prefix + pattern + suffix
     if store_patterns is None:
         ac = AhoCorasick([pattern])
     else:
         ac = AhoCorasick([pattern], store_patterns=store_patterns)
 
     index_matches = ac.find_matches_as_indexes(haystack)
-    expected = [pattern]
-    assert [i for (i, _, _) in index_matches] == [0]
-    assert [haystack[s:e] for (_, s, e) in index_matches] == expected
-    assert ac.find_matches_as_strings(haystack) == [pattern]
+    expected = {pattern}
+    assert {i for (i, _, _) in index_matches} == {0}
+    # Occasionally might get an overlap between haystack and preffix/suffix...
+    assert {haystack[s:e] for (_, s, e) in index_matches} == expected
+    assert set(ac.find_matches_as_strings(haystack)) == expected
 
 
 @pytest.mark.parametrize("bad_patterns", [[""], ["", "xx"], ["xx", ""]])

.hypothesis/examples/2be72fd93c3bfc88/b9fc09fff40c6832

@@ -7,7 +7,7 @@
 import pytest
 
 from hypothesis import strategies as st
-from hypothesis import given, assume
+from hypothesis import given
 
 from ahocorasick_rs import (
     BytesAhoCorasick,
@@ -123,15 +123,13 @@ def test_random_bytes_extensive(prefix: bytes, pattern: bytes, suffix: bytes) ->
     Random bytes patterns still give correct results for
     find_matches_as_indexes(), with property-testing.
     """
-    assume(pattern not in prefix)
-    assume(pattern not in suffix)
     haystack = prefix + pattern + suffix
     ac = BytesAhoCorasick([pattern])
 
     index_matches = ac.find_matches_as_indexes(haystack)
-    expected = [pattern]
-    assert [i for (i, _, _) in index_matches] == [0]
-    assert [haystack[s:e] for (_, s, e) in index_matches] == expected
+    assert {i for (i, _, _) in index_matches} == {0}
+    # Occasionally might get overlap between haystack and prefix/suffix...
+    assert {haystack[s:e] for (_, s, e) in index_matches} == {pattern}
 
 
 @pytest.mark.parametrize("bad_patterns", [[b""], [b"", b"xx"], [b"xx", b""]])