fix: RSS feeds show stale data for rebuilt packages

CHANGES/1374.bugfix

@@ -0,0 +1 @@
+RSS feeds now reflect the most recently added file for each release, produce unique guids when content is updated, and support up to 500 entries in `updates.xml`.

pulp_python/app/pypi/feeds.py

@@ -2,7 +2,7 @@
 from email.utils import getaddresses
 from urllib.parse import urljoin
 
-from django.db.models import F, FilteredRelation, Min, Q
+from django.db.models import F, FilteredRelation, Max, Min, Q
 from django.http.response import HttpResponse, HttpResponseNotFound
 from django.utils.decorators import method_decorator
 from django.utils.feedgenerator import Rss201rev2Feed
@@ -15,7 +15,7 @@
 from pulp_python.app.cache import PythonApiCache, find_base_path_cached
 from pulp_python.app.pypi.views import PyPIMixin, _etag_func
 
-UPDATES_LIMIT = 100
+UPDATES_LIMIT = 500
 PACKAGES_LIMIT = 40
 PROJECT_RELEASES_LIMIT = 40
 RSS_CONTENT_TYPE = "application/rss+xml; charset=utf-8"
@@ -66,7 +66,7 @@ def iter_releases(content, repo_ver, name_normalized=None, limit=UPDATES_LIMIT):
         qs.order_by()
         .values("name_normalized", "version")
         .annotate(
-            added_at=Min("file_added_at"),
+            added_at=Max("file_added_at"),
             name=Min("name"),
             summary=Min("summary"),
             author_email=Min("author_email"),
@@ -82,7 +82,7 @@ def iter_projects(content, repo_ver, limit=PACKAGES_LIMIT):
         qs.order_by()
         .values("name_normalized")
         .annotate(
-            added_at=Min("file_added_at"),
+            added_at=Max("file_added_at"),
             name=Min("name"),
             summary=Min("summary"),
             author_email=Min("author_email"),
@@ -98,7 +98,7 @@ def _item_dict(title, link, description, author_email, pubdate):
         "description": sanitize_xml_text(description),
         "author_email": format_author(author_email),
         "pubdate": pubdate,
-        "unique_id": link,
+        "unique_id": f"{link}#{pubdate.isoformat()}",
     }
 
 
@@ -118,7 +118,7 @@ def render_rss(title, link, description, items):
             author_email=item["author_email"],
             pubdate=item["pubdate"],
             unique_id=item["unique_id"],
-            unique_id_is_permalink=True,
+            unique_id_is_permalink=False,
         )
     return feed.writeString("utf-8")
 

pulp_python/tests/functional/api/test_pypi_feeds.py

@@ -125,8 +125,40 @@ def test_pinned_version_feeds(
 
     item = _parse_items(_get_feed(distro, "rss/updates.xml", bindings_cfg))[0]
     assert item.findtext("link").endswith("pypi/shelf-reader/0.1/json")
-    assert item.findtext("guid").endswith("pypi/shelf-reader/0.1/json")
+    assert "pypi/shelf-reader/0.1/json#" in item.findtext("guid")
 
     python_content_factory(TWINE_WHEEL_FILENAME, url=TWINE_WHEEL_URL, repository=repo)
     update_titles = _titles(_parse_items(_get_feed(distro, "rss/updates.xml", bindings_cfg)))
     assert update_titles == ["shelf-reader 0.1"]
+
+
+@pytest.mark.parallel
+def test_new_file_for_existing_version_updates_guid(
+    bindings_cfg, python_content_factory, python_empty_repo_distro
+):
+    """Adding a new file for an existing (name, version) produces a new guid and updated date."""
+    repo, distro = python_empty_repo_distro()
+
+    python_content_factory(PYTHON_EGG_FILENAME, url=PYTHON_EGG_URL, repository=repo)
+
+    items = _parse_items(_get_feed(distro, "rss/updates.xml", bindings_cfg))
+    assert len(items) == 1
+    first_guid = items[0].findtext("guid")
+    first_date = items[0].findtext("pubDate")
+    assert "pypi/shelf-reader/0.1/json" in first_guid
+
+    python_content_factory(PYTHON_WHEEL_FILENAME, url=PYTHON_WHEEL_URL, repository=repo)
+
+    items = _parse_items(_get_feed(distro, "rss/updates.xml", bindings_cfg))
+    assert len(items) == 1
+    second_guid = items[0].findtext("guid")
+    second_date = items[0].findtext("pubDate")
+
+    assert second_guid != first_guid
+    assert second_date >= first_date
+
+    release_items = _parse_items(
+        _get_feed(distro, "rss/project/shelf-reader/releases.xml", bindings_cfg)
+    )
+    assert len(release_items) == 1
+    assert release_items[0].findtext("guid") == second_guid