try some more stuff

hypothesis-python/RELEASE.rst

@@ -0,0 +1,3 @@
+RELEASE_TYPE: patch
+
+This patch makes some internal changes for compatibility with Crosshair.

hypothesis-python/src/hypothesis/internal/conjecture/engine.py

@@ -515,7 +515,9 @@ def test_function(self, data: ConjectureData) -> None:
                 # drive the ir tree through the test function to convert it
                 # to a buffer
                 initial_origin = data.interesting_origin
-                initial_traceback = getattr(data.extra_information, "_expected_traceback", None)
+                initial_traceback = getattr(
+                    data.extra_information, "_expected_traceback", None
+                )
                 data = ConjectureData.for_ir_tree(data.examples.ir_tree_nodes)
                 self.__stoppable_test_function(data)
                 data.freeze()
@@ -527,7 +529,11 @@ def test_function(self, data: ConjectureData) -> None:
                         data.status.INVALID: "failed filters",
                         data.status.OVERRUN: "overran",
                     }[data.status]
-                    wrapped_tb = "" if initial_traceback is None else textwrap.indent(initial_traceback, "  | ")
+                    wrapped_tb = (
+                        ""
+                        if initial_traceback is None
+                        else textwrap.indent(initial_traceback, "  | ")
+                    )
                     raise FlakyReplay(
                         f"Inconsistent results from replaying a failing test case!\n"
                         f"{wrapped_tb}on backend={self.settings.backend!r} but "

hypothesis-python/tests/cover/test_randoms.py

@@ -333,6 +333,7 @@ def test_randbytes_have_right_length(rnd, n):
     assert len(rnd.randbytes(n)) == n
 
 
+@pytest.mark.skipif(settings._current_profile == "crosshair", reason="takes hours")
 @given(any_random)
 def test_can_manage_very_long_ranges_with_step(rnd):
     i = rnd.randrange(0, 2**256, 3)

hypothesis-python/tests/cover/test_stateful.py

@@ -1325,4 +1325,6 @@ def rule1(self, data):
         assert data
 
 
-TestLotsOfEntropyPerStepMachine = LotsOfEntropyPerStepMachine.TestCase
+@pytest.mark.skipif(Settings._current_profile == "crosshair", reason="takes hours")
+def test_lots_of_entropy():
+    run_state_machine_as_test(LotsOfEntropyPerStepMachine)

hypothesis-python/tests/nocover/test_characters.py

@@ -11,10 +11,10 @@
 import string
 from encodings.aliases import aliases
 
-from hypothesis import given, strategies as st
-from hypothesis import settings
 import pytest
 
+from hypothesis import given, settings, strategies as st
+
 from tests.common.utils import Why, xfail_on_crosshair
 
 IDENTIFIER_CHARS = string.ascii_letters + string.digits + "_"
@@ -52,7 +52,6 @@ def _enc(cdc):
 assert len(lots_of_encodings) > 100  # sanity-check
 
 
-
 @pytest.mark.skipif(settings._current_profile == "crosshair", reason="takes 2000s")
 @given(data=st.data(), codec=st.sampled_from(lots_of_encodings))
 def test_can_constrain_characters_to_codec(data, codec):

hypothesis-python/tests/nocover/test_collective_minimization.py

@@ -16,8 +16,10 @@
 
 from tests.common import standard_types
 from tests.common.debug import minimal
+from tests.common.utils import Why, xfail_on_crosshair
 
 
+@xfail_on_crosshair(Why.floats, strict=False)
 @pytest.mark.parametrize("spec", standard_types, ids=repr)
 def test_can_collectively_minimize(spec):
     n = 10

hypothesis-python/tests/nocover/test_database_usage.py

@@ -112,15 +112,15 @@ def condition(x):
     assert len(all_values(database)) < original
 
 
-@xfail_on_crosshair(Why.symbolic_outside_context)
 def test_respects_max_examples_in_database_usage():
     key = b"a database key"
     database = InMemoryExampleDatabase()
     do_we_care = True
-    counter = [0]
+    counter = 0
 
     def check(x):
-        counter[0] += 1
+        nonlocal counter
+        counter += 1
         return do_we_care and has_a_non_zero_byte(x)
 
     def stuff():
@@ -138,9 +138,9 @@ def stuff():
         stuff()
     assert len(all_values(database)) > 10
     do_we_care = False
-    counter[0] = 0
+    counter = 0
     stuff()
-    assert counter == [10]
+    assert counter == 10
 
 
 def test_does_not_use_database_when_seed_is_forced(monkeypatch):

hypothesis-python/tests/nocover/test_sampled_from.py

@@ -131,6 +131,7 @@ def test_flags_minimizes_bit_count():
     )
 
 
+@xfail_on_crosshair(Why.undiscovered)
 def test_flags_finds_all_bits_set():
     assert find_any(st.sampled_from(LargeFlag), lambda f: f == ~LargeFlag(0))
 

hypothesis-python/tests/nocover/test_simple_numbers.py

@@ -17,6 +17,7 @@
 from hypothesis.strategies import floats, integers, lists
 
 from tests.common.debug import minimal
+from tests.common.utils import Why, xfail_on_crosshair
 
 
 def test_minimize_negative_int():
@@ -116,6 +117,7 @@ def test_can_minimal_infinite_negative_float():
     assert minimal(floats(), lambda x: x < -sys.float_info.max)
 
 
+@xfail_on_crosshair(Why.floats)
 def test_can_minimal_float_on_boundary_of_representable():
     minimal(floats(), lambda x: x + 1 == x and not math.isinf(x))
 
@@ -153,6 +155,7 @@ def test_minimal_fractional_float():
     assert minimal(floats(), lambda x: x >= 1.5) == 2
 
 
+@xfail_on_crosshair(Why.undiscovered)
 def test_minimizes_lists_of_negative_ints_up_to_boundary():
     result = minimal(
         lists(integers(), min_size=10),

hypothesis-python/tests/numpy/test_from_dtype.py

@@ -69,7 +69,7 @@ def test_is_t(x):
     test_is_t()
 
 
-@settings(max_examples=100)
+@settings(max_examples=100, deadline=None)
 @given(nps.nested_dtypes(max_itemsize=400), st.data())
 def test_infer_strategy_from_dtype(dtype, data):
     # Given a dtype