Unresolver: allow a full URL when un-resolving a domain (#11632)

This also checks that the URL has a valid protocol. This is needed for readthedocs/readthedocs-corporate#1887.

readthedocs/core/unresolver.py

@@ -208,7 +208,7 @@ def unresolve_url(self, url, append_indexhtml=True):
         parsed_url = urlparse(url)
         if parsed_url.scheme not in ["http", "https"]:
             raise InvalidSchemeError(parsed_url.scheme)
-        domain = self.get_domain_from_host(parsed_url.netloc)
+        domain = parsed_url.hostname
         unresolved_domain = self.unresolve_domain(domain)
         return self._unresolve(
             unresolved_domain=unresolved_domain,
@@ -551,8 +551,18 @@ def unresolve_domain(self, domain):
         Unresolve domain by extracting relevant information from it.
 
         :param str domain: Domain to extract the information from.
+         It can be a full URL, in that case, only the domain is used.
         :returns: A UnresolvedDomain object.
         """
+        parsed_domain = urlparse(domain)
+        if parsed_domain.scheme:
+            if parsed_domain.scheme not in ["http", "https"]:
+                raise InvalidSchemeError(parsed_domain.scheme)
+            domain = parsed_domain.hostname
+
+        if not domain:
+            raise InvalidSubdomainError(domain)
+
         public_domain = self.get_domain_from_host(settings.PUBLIC_DOMAIN)
         external_domain = self.get_domain_from_host(
             settings.RTD_EXTERNAL_VERSION_DOMAIN

readthedocs/rtd_tests/tests/test_unresolver.py

@@ -11,11 +11,13 @@
     InvalidExternalVersionError,
     InvalidPathForVersionedProjectError,
     InvalidSchemeError,
+    InvalidSubdomainError,
     SuspiciousHostnameError,
     TranslationNotFoundError,
     TranslationWithoutVersionError,
     VersionNotFoundError,
     unresolve,
+    unresolver,
 )
 from readthedocs.projects.constants import SINGLE_VERSION_WITHOUT_TRANSLATIONS
 from readthedocs.projects.models import Domain
@@ -372,8 +374,34 @@ def test_unresolve_invalid_scheme(self):
             "fttp://pip.readthedocs.io/en/latest/",
             "fttps://pip.readthedocs.io/en/latest/",
             "ssh://pip.readthedocs.io/en/latest/",
+            "javascript://pip.readthedocs.io/en/latest/",
             "://pip.readthedocs.io/en/latest/",
         ]
         for url in invalid_urls:
             with pytest.raises(InvalidSchemeError):
                 unresolve(url)
+
+        # A triple slash is interpreted as a URL without domain,
+        # we don't support that.
+        with pytest.raises(InvalidSubdomainError):
+            unresolve("https:///pip.readthedocs.io/en/latest/")
+
+    def test_unresolve_domain_with_full_url(self):
+        result = unresolver.unresolve_domain("https://pip.readthedocs.io/en/latest/")
+        self.assertIsNone(result.domain)
+        self.assertEqual(result.project, self.pip)
+        self.assertTrue(result.is_from_public_domain)
+        self.assertEqual(result.source_domain, "pip.readthedocs.io")
+
+    def test_unresolve_domain_with_full_url_invalid_protocol(self):
+        invalid_protocols = [
+            "fttp",
+            "fttps",
+            "ssh",
+            "javascript",
+        ]
+        for protocol in invalid_protocols:
+            with pytest.raises(InvalidSchemeError):
+                unresolver.unresolve_domain(
+                    f"{protocol}://pip.readthedocs.io/en/latest/"
+                )