refactor(http): adopt native HTTPX2 client interface

Remove the obsolete Requests-shaped compatibility layers from both the outbound client and its HTTP test harness. Native HTTPX2 request and response semantics reduce duplicate adapters, improve typing, and preserve redirect streaming and SSRF address-pinning coverage.

Author: nijel

weblate/accounts/tests/test_avatars.py

@@ -16,7 +16,7 @@
 from weblate.accounts.apps import check_avatars
 from weblate.auth.models import User
 from weblate.trans.tests.test_views import FixtureTestCase
-from weblate.utils.tests import http_mock as responses
+from weblate.utils.tests import http_mock
 
 TEST_URL = (
     "https://www.gravatar.com/avatar/55502f40dc8b7c769880b10874abc9d0?d=identicon&s=32"
@@ -44,13 +44,13 @@ def test_fallback_avatar_uses_stable_url(self) -> None:
             "https://cdn.example.com/static/api-32.png",
         )
 
-    @responses.activate
+    @http_mock.activate
     def test_avatar(self) -> None:
         image = Image.new("RGB", (32, 32))
         storage = BytesIO()
         image.save(storage, "PNG")
         imagedata = storage.getvalue()
-        responses.add(responses.GET, TEST_URL, body=imagedata)
+        http_mock.register("GET", TEST_URL, content=imagedata)
         # Real user
         response = self.client.get(
             reverse("user_avatar", kwargs={"user": self.user.username, "size": 32})
@@ -64,9 +64,9 @@ def test_avatar(self) -> None:
         self.assert_png(response)
         self.assertEqual(response.content, imagedata)
 
-    @responses.activate
+    @http_mock.activate
     def test_avatar_error(self) -> None:
-        responses.add(responses.GET, TEST_URL, status=503)
+        http_mock.register("GET", TEST_URL, status_code=503)
         # Choose different username to avoid using cache
         self.user.username = "test2"
         self.user.save()
@@ -86,13 +86,13 @@ def test_avatar_check_handles_http_error(self) -> None:
         self.assertEqual(len(errors), 1)
         self.assertEqual(errors[0].id, "weblate.E018")
 
-    @responses.activate
+    @http_mock.activate
     def test_avatar_rejects_unsupported_size_before_fetch(self) -> None:
         response = self.client.get(
             reverse("user_avatar", kwargs={"user": self.user.username, "size": 999})
         )
         self.assertEqual(response.status_code, 404)
-        self.assertEqual(len(responses.calls), 0)
+        self.assertEqual(len(http_mock.calls), 0)
 
     def test_anonymous_avatar(self) -> None:
         anonymous = User.objects.get(username="anonymous")