Fix #1101 - Properly handle a <base> without a href

pypa · Jun 13, 2014 · e05a759 · e05a759
1 parent a2c008f
commit e05a759
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 3 deletions.
diff --git a/CHANGES.txt b/CHANGES.txt
@@ -51,6 +51,9 @@
 * Fixed :issue:`1319`. Allow use of Zip64 extension in Wheels and other zip
   files. (:pull:`1868`)
 
+* Fixed :issue:`1101`. Properly handle an index or --find-links target which
+  has a <base> without a href attribute. (:pull:`1869`)
+
 
 **1.5.7**
 

diff --git a/pip/index.py b/pip/index.py
@@ -820,9 +820,12 @@ def api_version(self):
     @property
     def base_url(self):
         if not hasattr(self, "_base_url"):
-            base = self.parsed.find(".//base")
-            if base is not None and base.get("href"):
-                self._base_url = base.get("href")
+            bases = [
+                x for x in self.parsed.findall(".//base")
+                if x.get("href") is not None
+            ]
+            if bases and bases[0].get("href"):
+                self._base_url = bases[0].get("href")
             else:
                 self._base_url = self.url
         return self._base_url

diff --git a/tests/unit/test_index.py b/tests/unit/test_index.py
@@ -1,3 +1,5 @@
+import pytest
+
 from pip.download import PipSession
 from pip.index import package_to_requirement, HTMLPage
 from pip.index import PackageFinder, Link, INSTALLED_VERSION
@@ -80,3 +82,27 @@ def test_ext_fragment(self):
 
     def test_ext_query(self):
         assert '.whl' == Link('http://yo/wheel.whl?a=b').ext
+
+
+@pytest.mark.parametrize(
+    ("html", "url", "expected"),
+    [
+        ("<html></html>", "https://example.com/", "https://example.com/"),
+        (
+            "<html><head>"
+            "<base href=\"https://foo.example.com/\">"
+            "</head></html>",
+            "https://example.com/",
+            "https://foo.example.com/",
+        ),
+        (
+            "<html><head>"
+            "<base><base href=\"https://foo.example.com/\">"
+            "</head></html>",
+            "https://example.com/",
+            "https://foo.example.com/",
+        ),
+    ],
+)
+def test_base_url(html, url, expected):
+    assert HTMLPage(html, url).base_url == expected