tests: Ignore Textual snapshot mismatches on 3.7

The snapshots we've generated using current versions of Textual aren't
expected to match anymore on Python 3.7, as Textual dropped support for
Python 3.7 in the 0.44 release. However, we'd still like to run our
snapshot tests on Python 3.7, to confirm that no unexpected exceptions
occur and that the app doesn't crash. So, allow `snap_compare()` to
drive the application, but always return `True` on Python 3.7 as long as
no exception was raised.

Signed-off-by: Matt Wozniski <[email protected]>

Author: godlygeek

tests/conftest.py

@@ -1,5 +1,4 @@
 import socket
-import sys
 
 import pytest
 
@@ -11,12 +10,3 @@ def free_port():
     port_number = s.getsockname()[1]
     s.close()
     return port_number
-
-
-def pytest_configure(config):
-    # Several of the tree reporter tests require Textual 0.48, which does not
-    # support Python 3.7, but skipping those tests causes the test suite to
-    # fail due to unused snapshots. Override the configuration for Python 3.7
-    # so that unused snapshots are a warning, not an error.
-    if sys.version_info < (3, 8):
-        config.option.warn_unused_snapshots = True

tests/unit/test_tree_reporter.py

@@ -1532,6 +1532,15 @@ def test_render_runs_the_app(self):
 
 @pytest.fixture
 def compare(monkeypatch, tmp_path, snap_compare):
+    # The snapshots we've generated using current versions of Textual aren't
+    # expected to match anymore on Python 3.7, as Textual dropped support for
+    # Python 3.7 in the 0.44 release. However, we'd still like to run our
+    # snapshot tests on Python 3.7, to confirm that no unexpected exceptions
+    # occur and that the app doesn't crash. So, allow `snap_compare()` to drive
+    # the application, but always return `True` on Python 3.7 as long as no
+    # exception was raised.
+    succeed_even_if_mismatched = sys.version_info < (3, 8)
+
     def compare_impl(
         allocations: Iterator[AllocationRecord],
         press: Iterable[str] = (),
@@ -1550,21 +1559,20 @@ def compare_impl(
         with monkeypatch.context() as app_patch:
             app_patch.setitem(globals(), app_global, app)
             tmp_main.write_text(f"from {__name__} import {app_global} as app")
-            return snap_compare(
-                str(tmp_main),
-                press=press,
-                terminal_size=terminal_size,
-                run_before=run_before,
+            return (
+                snap_compare(
+                    str(tmp_main),
+                    press=press,
+                    terminal_size=terminal_size,
+                    run_before=run_before,
+                )
+                or succeed_even_if_mismatched
             )
 
     yield compare_impl
 
 
 class TestTUILooks:
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_basic(self, compare):
         # GIVEN
         code = dedent(
@@ -1599,10 +1607,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.48 or higher, which doesn't support 3.7",
-    )
     def test_basic_node_selected_not_leaf(self, compare):
         # GIVEN
         code = dedent(
@@ -1637,10 +1641,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[*["down"] * 2])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_basic_node_selected_leaf(self, compare):
         # GIVEN
         code = dedent(
@@ -1675,10 +1675,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[*["down"] * 3])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_two_chains(self, compare):
         # GIVEN
         code = dedent(
@@ -1726,10 +1722,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.48 or higher, which doesn't support 3.7",
-    )
     def test_two_chains_after_expanding_second(self, compare):
         # GIVEN
         code = dedent(
@@ -1779,10 +1771,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[*["down"] * 4, "e"])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_hide_import_system(self, compare):
         # GIVEN
         code = dedent(
@@ -1833,10 +1821,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=["i"])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_show_uninteresting(self, compare):
         # GIVEN
         code = dedent(
@@ -1887,10 +1871,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=["u"])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_show_uninteresting_and_hide_import_system(self, compare):
         # GIVEN
         code = dedent(
@@ -1942,10 +1922,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=["u", "i"])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.48 or higher, which doesn't support 3.7",
-    )
     def test_select_screen(self, tmp_path, compare):
         # GIVEN
         code = dedent(
@@ -1979,10 +1955,6 @@ def generate_primes():
             getlines.return_value = code.splitlines()
             assert compare(peak_allocations, press=[*["down"] * 3])
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_allocations_of_different_sizes(self, compare):
         # GIVEN
         peak_allocations = [
@@ -2003,10 +1975,6 @@ def test_allocations_of_different_sizes(self, compare):
             getlines.return_value = []
             assert compare(peak_allocations, press=[], terminal_size=(350, 100))
 
-    @pytest.mark.skipif(
-        sys.version_info < (3, 8),
-        reason="This test requires Textual 0.49 or higher, which doesn't support 3.7",
-    )
     def test_biggest_allocations(self, compare):
         # GIVEN
         peak_allocations = [

tests/unit/test_tui_reporter.py

@@ -1,5 +1,6 @@
 import asyncio
 import datetime
+import sys
 from io import StringIO
 from typing import Awaitable
 from typing import Callable
@@ -102,6 +103,15 @@ def get_current_snapshot(
 def compare(monkeypatch, tmp_path, snap_compare):
     monkeypatch.setattr(memray.reporters.tui, "datetime", FakeDatetime)
 
+    # The snapshots we've generated using current versions of Textual aren't
+    # expected to match anymore on Python 3.7, as Textual dropped support for
+    # Python 3.7 in the 0.44 release. However, we'd still like to run our
+    # snapshot tests on Python 3.7, to confirm that no unexpected exceptions
+    # occur and that the app doesn't crash. So, allow `snap_compare()` to drive
+    # the application, but always return `True` on Python 3.7 as long as no
+    # exception was raised.
+    succeed_even_if_mismatched = sys.version_info < (3, 8)
+
     def compare_impl(
         cmdline_override: Optional[str] = None,
         press: Iterable[str] = (),
@@ -128,11 +138,14 @@ async def run_before_wrapper(pilot) -> None:
         with monkeypatch.context() as app_patch:
             app_patch.setitem(globals(), app_global, app)
             tmp_main.write_text(f"from {__name__} import {app_global} as app")
-            return snap_compare(
-                str(tmp_main),
-                press=press,
-                terminal_size=terminal_size,
-                run_before=run_before_wrapper,
+            return (
+                snap_compare(
+                    str(tmp_main),
+                    press=press,
+                    terminal_size=terminal_size,
+                    run_before=run_before_wrapper,
+                )
+                or succeed_even_if_mismatched
             )
 
     yield compare_impl