fix(activitypub): Drop cached URLs without scheme

- Validate cached file/reply URLs start with http(s) and regenerate if stale
- Add tests for scheme-less cache entries
- Update changelog for 1.3.5 fix

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## 1.3.5
+
+### Fixed
+
+- **Stale cached URLs without scheme**: `file_to_url` and
+  `reply_file_to_url` now validate that cached URLs in `file_urls.json`
+  have a proper `https://` (or `http://`) scheme. Entries cached before
+  the URL normalization fix (v1.3.3) are discarded and regenerated,
+  ensuring ActivityPub object IDs are always fully qualified URLs.
+
 ## 1.3.4
 
 ### Fixed

madblog/activitypub/_integration.py

@@ -224,7 +224,7 @@ class ActivityPubIntegration(ActivityPubRepliesMixin, StartupSyncMixin):
 
     def file_to_url(self, filepath: str) -> str:
         stored = self._get_file_url(filepath)
-        if stored:
+        if stored and re.match(r"^https?://", stored):
             return stored
 
         # Generate the base URL

madblog/activitypub/_replies.py

@@ -1,5 +1,6 @@
 import logging
 import os
+import re
 
 from datetime import datetime, timezone
 from typing import Callable
@@ -79,6 +80,9 @@ class ActivityPubRepliesMixin(ActivityPubPublishMixin):
             if "/None/" in stored or stored.endswith("/None"):
                 logger.info("Clearing stale URL mapping with None: %s", stored)
                 self._remove_reply_file_url(filepath)
+            elif not re.match(r"^https?://", stored):
+                logger.info("Clearing stale URL without scheme: %s", stored)
+                self._remove_reply_file_url(filepath)
             else:
                 return stored
 

tests/test_activitypub.py

@@ -1560,6 +1560,51 @@ class ActivityPubPublishTest(unittest.TestCase):
             "https://persist.example/users/frank",
         )
 
+    # ------------------------------------------------------------------
+    # URL cache validation
+    # ------------------------------------------------------------------
+
+    @skip_if_no_pubby
+    def test_file_to_url_discards_cached_url_without_scheme(self):
+        """A cached URL missing https:// is discarded and regenerated."""
+        integration, _, test_file, _ = self._make_integration()
+
+        # Simulate a stale cache entry from before _normalize_url existed
+        integration._set_file_url(str(test_file), "example.com/article/test-post")
+
+        url = integration.file_to_url(str(test_file))
+        self.assertTrue(
+            url.startswith("https://"),
+            f"Expected URL with scheme, got: {url}",
+        )
+
+    @skip_if_no_pubby
+    def test_reply_file_to_url_discards_cached_url_without_scheme(self):
+        """A cached reply URL missing https:// is discarded and regenerated."""
+        integration, _, _, pages_dir = self._make_integration()
+
+        replies_dir = pages_dir.parent / "replies"
+        replies_dir.mkdir()
+        integration.replies_dir = str(replies_dir)
+
+        art_dir = replies_dir / "some-article"
+        art_dir.mkdir()
+        reply_file = art_dir / "my-reply.md"
+        reply_file.write_text(
+            "[//]: # (reply-to: https://remote.example/post/1)\n\nGreat post!"
+        )
+
+        # Simulate a stale cache entry without scheme
+        integration._set_reply_file_url(
+            str(reply_file), "example.com/reply/some-article/my-reply"
+        )
+
+        url = integration.reply_file_to_url(str(reply_file))
+        self.assertTrue(
+            url.startswith("https://"),
+            f"Expected URL with scheme, got: {url}",
+        )
+
     # ------------------------------------------------------------------
     # Non-blocking / concurrency
     # ------------------------------------------------------------------