test: add test coverage for tests/helpers.py

[yosofbadr]

Summary

Closes #9161

Adds comprehensive test coverage for tests/helpers.py as requested by @Secrus in #9149 (comment).

The existing tests/test_helpers.py had tests for only 3 of the 15+ helper functions/classes. This PR adds tests covering all remaining helpers:

• get_package - package creation with version and yanked status
• get_dependency - dependency creation with constraints, groups, optional, prereleases
• copy_path - file and directory copying with overwrite behavior
• MockDulwichRepo - path storage and mock git revision
• mock_clone - fixture-based git clone simulation
• TestLocker - lock state management, mock data, and write persistence
• TestRepository - package lookup (success and not-found) and link generation
• make_entry_point_from_plugin - entry point creation with and without dist/group
• mock_metadata_entry_points - metadata patching and group filtering
• with_working_directory - working directory context manager with optional copy
• set_keyring_backend - keyring backend switching with cache clearing
• pbs_installer_supported_arch - architecture support detection (supported, unsupported, case-insensitive)

Also adds additional edge-case tests for the previously covered helpers (flatten_dict, isolated_environment, switch_working_directory).

Total: 66 tests, all passing.

Test plan

• All 66 tests pass locally with pytest tests/test_helpers.py -v
• CI passes on all supported platforms

[sourcery-ai]

Hey - I've found 3 issues

Prompt for AI Agents

Please address the comments from this code review:

## Individual Comments

### Comment 1
<location path="tests/test_helpers.py" line_range="478-487" />
<code_context>
+class TestSetKeyringBackend:
</code_context>
<issue_to_address>
**issue (testing):** Test does not actually verify the cache-clearing behavior and leaves global keyring state mutated after the test.

This only checks that the global backend is set, not that `PoetryKeyring`’s availability cache is actually cleared and recomputed. Please also assert that `PoetryKeyring.is_available()` is recalculated (e.g., set a known value before `set_keyring_backend` and verify it changes afterward). In addition, the test mutates the process-wide keyring backend without restoring it; capture `keyring.get_keyring()` beforehand and restore it in a `try/finally` or fixture teardown to avoid cross-test interference.
</issue_to_address>

### Comment 2
<location path="tests/test_helpers.py" line_range="260-269" />
<code_context>
+class TestTestRepository:
</code_context>
<issue_to_address>
**suggestion (testing):** Repository tests don't cover version/constraint resolution behavior when multiple packages are present.

Since `TestRepository` is used to simulate real resolution behavior, please add a test with multiple versions of the same package and a constrained dependency. For example, add `foo-1.0.0` and `foo-2.0.0` and assert that `find_packages(get_dependency("foo", "<2.0"))` returns only the expected version(s).
</issue_to_address>

### Comment 3
<location path="tests/test_helpers.py" line_range="65-74" />
<code_context>
+# --- get_dependency ---
+
+
+class TestGetDependency:
+    def test_returns_dependency_with_wildcard_constraint(self) -> None:
+        dep = get_dependency("foo")
+        assert dep.name == "foo"
+        assert str(dep.constraint) == "*"
+
+    def test_returns_dependency_with_version_constraint(self) -> None:
+        dep = get_dependency("foo", ">=1.0")
+        assert dep.name == "foo"
+
+    def test_returns_dependency_with_dict_constraint(self) -> None:
+        dep = get_dependency("foo", {"version": ">=1.0", "python": "^3.10"})
+        assert dep.name == "foo"
+
+    def test_returns_optional_dependency(self) -> None:
+        dep = get_dependency("foo", ">=1.0", optional=True)
+        assert dep.is_optional()
+
+    def test_returns_dependency_with_groups(self) -> None:
+        dep = get_dependency("foo", ">=1.0", groups=["dev"])
+        assert dep.in_extras == []
+
+    def test_returns_dependency_allowing_prereleases(self) -> None:
</code_context>
<issue_to_address>
**issue (testing):** The `test_returns_dependency_with_groups` case doesn't assert anything about the `groups` argument behavior.

Here the test passes `groups=["dev"]` but only asserts `dep.in_extras == []`, which doesn’t exercise or document the `groups` behavior. Please add an assertion on whatever attribute actually represents `groups` (e.g. `dep.groups` or the equivalent) so this test will fail if group handling regresses, rather than giving a false sense of coverage.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue (testing): Test does not actually verify the cache-clearing behavior and leaves global keyring state mutated after the test.

This only checks that the global backend is set, not that PoetryKeyring’s availability cache is actually cleared and recomputed. Please also assert that PoetryKeyring.is_available() is recalculated (e.g., set a known value before set_keyring_backend and verify it changes afterward). In addition, the test mutates the process-wide keyring backend without restoring it; capture keyring.get_keyring() beforehand and restore it in a try/finally or fixture teardown to avoid cross-test interference.

[sourcery-ai]

suggestion (testing): Repository tests don't cover version/constraint resolution behavior when multiple packages are present.

Since TestRepository is used to simulate real resolution behavior, please add a test with multiple versions of the same package and a constrained dependency. For example, add foo-1.0.0 and foo-2.0.0 and assert that find_packages(get_dependency("foo", "<2.0")) returns only the expected version(s).

[sourcery-ai]

issue (testing): The test_returns_dependency_with_groups case doesn't assert anything about the groups argument behavior.

Here the test passes groups=["dev"] but only asserts dep.in_extras == [], which doesn’t exercise or document the groups behavior. Please add an assertion on whatever attribute actually represents groups (e.g. dep.groups or the equivalent) so this test will fail if group handling regresses, rather than giving a false sense of coverage.