fix: validate canonical CORS origins

Authored-By: @architect <architect@fritzlab.net>

scripts/utils.py

@@ -1,5 +1,6 @@
 """Shared utilities for the site-publish action."""
 
+import ipaddress
 import os
 import re
 import shutil
@@ -105,22 +106,47 @@ def _strings(value, label):
 
 
 def _cors_origins(value, label):
-    origins = _strings(value or [], label)
-    if len(origins) != len(set(origins)):
-        raise ConfigError(f"{label} must not contain duplicates")
+    origins = _list(value, label)
+    if any(not isinstance(item, str) or not item for item in origins):
+        raise ConfigError(f"{label} must be a list of non-empty strings")
+    canonical = []
     for origin in origins:
         if origin == "*":
+            canonical.append(origin)
             continue
         parsed = urlparse(origin)
         try:
             port = parsed.port
         except ValueError:
-            port = None
+            raise ConfigError(f"{label} must contain '*' or canonical HTTPS origins") from None
         if parsed.scheme != "https" or not parsed.hostname or parsed.path or parsed.params or (
             parsed.query or parsed.fragment or parsed.username or parsed.password
-            or (":" in parsed.netloc and port is None and not parsed.netloc.endswith("]"))
         ):
-            raise ConfigError(f"{label} must contain '*' or HTTPS origins")
+            raise ConfigError(f"{label} must contain '*' or canonical HTTPS origins")
+        try:
+            address = ipaddress.ip_address(parsed.hostname)
+        except ValueError:
+            try:
+                hostname = parsed.hostname.encode("idna").decode("ascii")
+            except UnicodeError:
+                raise ConfigError(
+                    f"{label} must contain '*' or canonical HTTPS origins"
+                ) from None
+            try:
+                _hostname(hostname, f"{label} hostname")
+            except ConfigError:
+                raise ConfigError(
+                    f"{label} must contain '*' or canonical HTTPS origins"
+                ) from None
+        else:
+            hostname = f"[{address.compressed}]" if address.version == 6 else address.compressed
+        canonical.append(f"https://{hostname}{f':{port}' if port not in (None, 443) else ''}")
+    if len(canonical) != len(set(canonical)):
+        raise ConfigError(f"{label} must not contain duplicate canonical origins")
+    if "*" in canonical and len(canonical) != 1:
+        raise ConfigError(f"{label} wildcard must be the only origin")
+    if origins != canonical:
+        raise ConfigError(f"{label} must contain '*' or canonical HTTPS origins")
     return origins
 
 
@@ -328,7 +354,10 @@ def _artifact(item, index):
         "website_authority": authority,
         "credentials": normalized_credentials,
         "cache_rules": cache_rules,
-        "cors_origins": _cors_origins(item.get("cors_origins"), f"{label}.cors_origins"),
+        "cors_origins": (
+            _cors_origins(item["cors_origins"], f"{label}.cors_origins")
+            if "cors_origins" in item else []
+        ),
     }
 
 

tests/test_contract.py

@@ -123,14 +123,51 @@ class ConfigContractTests(unittest.TestCase):
             lambda raw: raw["artifacts"][1].__setitem__(
                 "cors_origins", ["http://consumer.example"]
             ),
-            "must contain '\\*' or HTTPS origins",
+            "must contain '\\*' or canonical HTTPS origins",
+        )
+        for origin in (
+            "https://consumer example",
+            "https://Consumer.example",
+            "https://consumer.example:443",
+            "https://consumer.example/",
+            "https://consumer.example:invalid",
+        ):
+            with self.subTest(origin=origin):
+                self.assert_invalid(
+                    lambda raw, origin=origin: raw["artifacts"][1].__setitem__(
+                        "cors_origins", [origin]
+                    ),
+                    "must contain '\\*' or canonical HTTPS origins",
+                )
+        self.assert_invalid(
+            lambda raw: raw["artifacts"][1].__setitem__(
+                "cors_origins", ["https://consumer.example", "https://consumer.example:443"]
+            ),
+            "must not contain duplicate canonical origins",
         )
         self.assert_invalid(
             lambda raw: raw["artifacts"][1].__setitem__(
-                "cors_origins", ["https://consumer.example", "https://consumer.example"]
+                "cors_origins", ["*", "https://consumer.example"]
             ),
-            "must not contain duplicates",
+            "wildcard must be the only origin",
         )
+        raw = copy.deepcopy(self.raw)
+        raw["artifacts"][1]["cors_origins"] = ["https://[2602:817:3000::1]:8443"]
+        cfg = normalize_site_config(raw, "baseline.fritzlab.net")
+        self.assertEqual(
+            ["https://[2602:817:3000::1]:8443"],
+            next(a for a in cfg["artifacts"] if a["name"] == "distributions")["cors_origins"],
+        )
+
+    def test_declared_cors_origins_must_be_a_list(self):
+        for value in (None, False, 0, "", {}):
+            with self.subTest(value=value):
+                self.assert_invalid(
+                    lambda raw, value=value: raw["artifacts"][1].__setitem__(
+                        "cors_origins", value
+                    ),
+                    "cors_origins must be a list",
+                )
 
     def test_protected_route_rejects_wildcard_cors(self):
         self.assert_invalid(