Implement RFC7617-compliant multi-domain basic authentication

The https://datatracker.ietf.org/doc/html/rfc7617#section-2.2
defines multi-domain authentication behaviour and authentication
scopes for basic authentication. This change improves the
implementation of the multi-domain matching to be RC7617 compliant

* path matching (including longest match)
* scheme validation matching

Closes: pypa#10902

docs/html/topics/authentication.md

@@ -16,6 +16,33 @@ token as the "username" and do not provide a password:
 https://0123456789abcdef@pypi.company.com/simple
 ```
 
+When you specify several indexes, each of them can come with its own
+authentication information. When the domains and schemes of multiple
+indexes partially overlap, you can specify different authentication for each of them
+For example you can have:
+
+```
+PIP_INDEX_URL=https://build:password1@pkgs.dev.azure.com/feed1
+PIP_EXTRA_INDEX_URL=https://build:password2@pkgs.dev.azure.com/feed2
+```
+
+If you specify multiple identical index URLs with different authentication information,
+authentication from the first index will be used.
+
+In compliance with [RFC7617](https://datatracker.ietf.org/doc/html/rfc7617#section-2.2) if the indexes
+overlap, the authentication information from the prefix-match will be reused for the longer index if
+the longer index does not contain the authentication information. In case multiple indexes are
+prefix-matching, the authentication of the first of the longest matching prefixes is used.
+
+For example in this case, build:password authentication will be used when authenticating with the extra
+index URL.
+
+```
+PIP_INDEX_URL=https://build:password@pkgs.dev.azure.com/
+PIP_EXTRA_INDEX_URL=https://pkgs.dev.azure.com/feed1
+```
+
+
 ### Percent-encoding special characters
 
 ```{versionadded} 10.0

news/10904.feature.rst

@@ -0,0 +1,3 @@
+When you specify multiple indexes with common domains and schemes, basic authentication for each of
+the indexes can be different - compliant with RFC 7617. Previous versions of pip reused basic
+authentication credentials for all urls within the same domain.

src/pip/_internal/network/auth.py

@@ -5,7 +5,7 @@
 """
 
 import urllib.parse
-from typing import Any, Dict, List, Optional, Tuple
+from typing import Any, List, Optional, Tuple
 
 from pip._vendor.requests.auth import AuthBase, HTTPBasicAuth
 from pip._vendor.requests.models import Request, Response
@@ -76,7 +76,7 @@ def __init__(
     ) -> None:
         self.prompting = prompting
         self.index_urls = index_urls
-        self.passwords: Dict[str, AuthInfo] = {}
+        self.passwords: List[Tuple[str, AuthInfo]] = []
         # When the user is prompted to enter credentials and keyring is
         # available, we will offer to save them. If the user accepts,
         # this value is set to the credentials they entered. After the
@@ -163,6 +163,49 @@ def _get_new_credentials(
 
         return username, password
 
+    def _get_matching_credentials(self, original_url: str) -> Optional[AuthInfo]:
+        """
+        Find matching credentials based on the longest matching prefix found.
+
+        According to https://datatracker.ietf.org/doc/html/rfc7617#section-2.2
+        the authentication scope is defined by removing the last part after
+        the last "/" of the resource - but in case of `pip` the end of path is always
+        "/project", so we can treat the `original_url` as full `authentication scope'
+        for the request. The RFC specifies that prefix matching of the scope is
+        also within the protection scope specified by the realm value, so we can
+        safely assume that if we find any match, the longest matching prefix is
+        the right authentication information we should use.
+
+        The specification does not decide which of the matching scopes should be
+        used if there are more of them. Our decision is to choose the longest
+        matching one. In case exactly the same prefix will be added several times,
+        the authentication information from the first one is used.
+
+        According to the RFC, the authentication scope includes both scheme and netloc,
+        Both have to match in order to share the authentication scope.
+
+        :param original_url: original URL of the request
+        :return: Stored Authentication info matching the authentication scope or
+                 None if not found
+        """
+        max_matched_prefix_length = 0
+        matched_auth_info = None
+        no_auth_url = remove_auth_from_url(original_url)
+        parsed_original = urllib.parse.urlparse(no_auth_url)
+        for prefix, auth_info in self.passwords:
+            parsed_stored = urllib.parse.urlparse(prefix)
+            if (
+                parsed_stored.netloc != parsed_original.netloc
+                or parsed_stored.scheme != parsed_original.scheme
+            ):
+                # only consider match when both scheme and netloc match
+                continue
+            if no_auth_url.startswith(prefix):
+                if len(prefix) > max_matched_prefix_length:
+                    matched_auth_info = auth_info
+                    max_matched_prefix_length = len(prefix)
+        return matched_auth_info
+
     def _get_url_and_credentials(
         self, original_url: str
     ) -> Tuple[str, Optional[str], Optional[str]]:
@@ -184,12 +227,14 @@ def _get_url_and_credentials(
         # Do this if either the username or the password is missing.
         # This accounts for the situation in which the user has specified
         # the username in the index url, but the password comes from keyring.
-        if (username is None or password is None) and netloc in self.passwords:
-            un, pw = self.passwords[netloc]
-            # It is possible that the cached credentials are for a different username,
-            # in which case the cache should be ignored.
-            if username is None or username == un:
-                username, password = un, pw
+        if username is None or password is None:
+            matched_credentials = self._get_matching_credentials(original_url)
+            if matched_credentials:
+                un, pw = matched_credentials
+                # It is possible that the cached credentials are for a
+                # different username, in which case the cache should be ignored.
+                if username is None or username == un:
+                    username, password = un, pw
 
         if username is not None or password is not None:
             # Convert the username and password if they're None, so that
@@ -200,7 +245,7 @@ def _get_url_and_credentials(
             password = password or ""
 
             # Store any acquired credentials.
-            self.passwords[netloc] = (username, password)
+            self.passwords.append((remove_auth_from_url(url), (username, password)))
 
         assert (
             # Credentials were found
@@ -273,7 +318,9 @@ def handle_401(self, resp: Response, **kwargs: Any) -> Response:
         # Store the new username and password to use for future requests
         self._credentials_to_save = None
         if username is not None and password is not None:
-            self.passwords[parsed.netloc] = (username, password)
+            self.passwords.append(
+                (remove_auth_from_url(resp.url), (username, password))
+            )
 
             # Prompt to save the password to keyring
             if save and self._should_save_password_to_keyring():

tests/unit/test_network_auth.py

@@ -1,10 +1,12 @@
 import functools
 from typing import Any, List, Optional, Tuple
+from unittest.mock import Mock, PropertyMock
 
 import pytest
 
 import pip._internal.network.auth
 from pip._internal.network.auth import MultiDomainBasicAuth
+from src.pip._vendor.requests import PreparedRequest
 from tests.lib.requests_mocks import MockConnection, MockRequest, MockResponse
 
 
@@ -50,40 +52,143 @@ def test_get_credentials_parses_correctly(
         (username is None and password is None)
         or
         # Credentials were found and "cached" appropriately
-        auth.passwords["example.com"] == (username, password)
+        (url, (username, password)) in auth.passwords
     )
 
 
+def test_handle_prompt_for_password_successful() -> None:
+    auth = MultiDomainBasicAuth()
+    resp = Mock()
+    resp.status_code = 401
+    resp.url = "http://example.com"
+    resp.raw = Mock()
+    resp.content = PropertyMock()
+    resp.content = ""
+    resp.request = PreparedRequest()
+    resp.request.headers = {}
+    auth._prompt_for_password = Mock()  # type: ignore[assignment]
+    auth._prompt_for_password.return_value = ("user", "password", True)
+    auth.handle_401(resp)
+    auth._prompt_for_password.assert_called_with("example.com")
+    expected = ("http://example.com", ("user", "password"))
+    assert len(auth.passwords) == 1
+    assert auth.passwords[0] == expected
+
+
+def test_handle_prompt_for_password_unsuccessful() -> None:
+    auth = MultiDomainBasicAuth()
+    resp = Mock()
+    resp.status_code = 401
+    resp.url = "http://example.com"
+    resp.raw = Mock()
+    resp.content = PropertyMock()
+    resp.content = ""
+    resp.request = PreparedRequest()
+    resp.request.headers = {}
+    auth._prompt_for_password = Mock()  # type: ignore[assignment]
+    auth._prompt_for_password.return_value = (None, None, False)
+    auth.handle_401(resp)
+    auth._prompt_for_password.assert_called_with("example.com")
+    assert len(auth.passwords) == 0
+
+
 def test_get_credentials_not_to_uses_cached_credentials() -> None:
     auth = MultiDomainBasicAuth()
-    auth.passwords["example.com"] = ("user", "pass")
+    auth.passwords.append(("http://example.com", ("user", "pass")))
 
     got = auth._get_url_and_credentials("http://foo:bar@example.com/path")
     expected = ("http://example.com/path", "foo", "bar")
     assert got == expected
 
 
-def test_get_credentials_not_to_uses_cached_credentials_only_username() -> None:
+def test_get_credentials_not_to_use_cached_credentials_only_username() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("https://example.com", ("user", "pass")))
+
+    got = auth._get_url_and_credentials("https://foo@example.com/path")
+    expected = ("https://example.com/path", "foo", "")
+    assert got == expected
+
+
+def test_multi_domain_credentials_match() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+    auth.passwords.append(("http://example.com/path", ("user", "pass2")))
+
+    got = auth._get_url_and_credentials("http://user@example.com/path")
+    expected = ("http://example.com/path", "user", "pass2")
+    assert got == expected
+
+
+def test_multi_domain_credentials_longest_match() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+    auth.passwords.append(("http://example.com/path", ("user", "pass2")))
+    auth.passwords.append(("http://example.com/path/subpath", ("user", "pass3")))
+
+    got = auth._get_url_and_credentials("http://user@example.com/path")
+    expected = ("http://example.com/path", "user", "pass2")
+    assert got == expected
+
+
+def test_multi_domain_credentials_partial_match_only() -> None:
     auth = MultiDomainBasicAuth()
-    auth.passwords["example.com"] = ("user", "pass")
+    auth.passwords.append(("http://example.com/path1", ("user", "pass")))
 
-    got = auth._get_url_and_credentials("http://foo@example.com/path")
-    expected = ("http://example.com/path", "foo", "")
+    got = auth._get_url_and_credentials("http://example.com/path2")
+    expected = ("http://example.com/path2", None, None)
     assert got == expected
 
 
 def test_get_credentials_uses_cached_credentials() -> None:
     auth = MultiDomainBasicAuth()
-    auth.passwords["example.com"] = ("user", "pass")
+    auth.passwords.append(("https://example.com", ("user", "pass")))
+
+    got = auth._get_url_and_credentials("https://example.com/path")
+    expected = ("https://example.com/path", "user", "pass")
+    assert got == expected
+
+
+def test_get_credentials_not_uses_cached_credentials_different_scheme_http() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+
+    got = auth._get_url_and_credentials("https://example.com/path")
+    expected = ("https://example.com/path", None, None)
+    assert got == expected
+
+
+def test_get_credentials_not_uses_cached_credentials_different_scheme_https() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("https://example.com", ("user", "pass")))
 
     got = auth._get_url_and_credentials("http://example.com/path")
-    expected = ("http://example.com/path", "user", "pass")
+    expected = ("http://example.com/path", None, None)
     assert got == expected
 
 
 def test_get_credentials_uses_cached_credentials_only_username() -> None:
     auth = MultiDomainBasicAuth()
-    auth.passwords["example.com"] = ("user", "pass")
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+
+    got = auth._get_url_and_credentials("http://user@example.com/path")
+    expected = ("http://example.com/path", "user", "pass")
+    assert got == expected
+
+
+def test_get_credentials_uses_cached_credentials_wrong_username() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+
+    got = auth._get_url_and_credentials("http://user2@example.com/path")
+    expected = ("http://example.com/path", "user2", "")
+    assert got == expected
+
+
+def test_get_credentials_added_multiple_times() -> None:
+    auth = MultiDomainBasicAuth()
+    auth.passwords.append(("http://example.com", ("user", "pass")))
+    auth.passwords.append(("http://example.com", ("user", "pass2")))
 
     got = auth._get_url_and_credentials("http://user@example.com/path")
     expected = ("http://example.com/path", "user", "pass")