camsoda and pornhub shorties

check.py

@@ -64,6 +64,23 @@ class Results:
 
 _BROWSER_UA = "Mozilla/5.0 (X11; Linux x86_64; rv:146.0) Gecko/20100101 Firefox/146.0"
 
+# Domains known to be Cloudflare-protected and return 403/connection-refused to direct
+# HTTP checks. URL reachability failures for these hosts are downgraded to warnings.
+_CF_PROTECTED_HOSTS = {
+    "www.camsoda.com",
+    "camsoda.com",
+}
+
+
+def _is_cf_protected(url: str) -> bool:
+    """Return True if the URL's host is known to be CF-protected."""
+    try:
+        from urllib.parse import urlparse
+        host = urlparse(url).hostname or ""
+        return host in _CF_PROTECTED_HOSTS
+    except Exception:
+        return False
+
 
 def http_ok(url: str, headers: dict | None = None) -> tuple[bool, int]:
     """Return (ok, http_status). Tries HEAD then ranged GET on 405."""
@@ -234,7 +251,10 @@ def check_video(video: dict, channel_id: str, results: Results, run_ytdlp: bool)
     else:
         ok, code = http_ok(vurl)
         if not ok:
-            results.err(channel_id, f"{label}: url unreachable HTTP={code}: {vurl}")
+            if _is_cf_protected(vurl):
+                results.warn(channel_id, f"{label}: url unreachable HTTP={code} (CF-protected host, expected): {vurl}")
+            else:
+                results.err(channel_id, f"{label}: url unreachable HTTP={code}: {vurl}")
         else:
             results.info(channel_id, f"{label}: url OK (HTTP {code})")
 
@@ -275,11 +295,18 @@ def check_video(video: dict, channel_id: str, results: Results, run_ytdlp: bool)
         results.info(channel_id, f"{label}: yt-dlp extract {ytdlp_url}")
         yt, stderr = ytdlp_extract(ytdlp_url)
         if yt is None:
-            results.err(
-                channel_id,
-                f"{label}: yt-dlp failed for {ytdlp_url}"
-                + (f": {stderr[:300]}" if stderr else ""),
-            )
+            if _is_cf_protected(ytdlp_url):
+                results.warn(
+                    channel_id,
+                    f"{label}: yt-dlp failed for {ytdlp_url} (CF-protected host, expected)"
+                    + (f": {stderr[:200]}" if stderr else ""),
+                )
+            else:
+                results.err(
+                    channel_id,
+                    f"{label}: yt-dlp failed for {ytdlp_url}"
+                    + (f": {stderr[:300]}" if stderr else ""),
+                )
         else:
             yt_title = (yt.get("title") or "").strip()
             api_title = (video.get("title") or "").strip()