Use packaging.version.Version to sort filenames by the version of the package they represent. Alternate implementation of that proposed in #829. Also ref #629.

2 files changed, 36 insertions, 5 deletions

diff --git a/CHANGES.rst b/CHANGES.rst
index 3ef933e5..8caa2138 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -2,6 +2,14 @@
 CHANGES
 =======
 
+v28.8.0
+-------
+
+* #629: Per the discussion, refine the sorting to use version
+  value order for more accurate detection of the latest
+  available version when scanning for packages. See also
+  #829.
+
 v28.7.1
 -------
 
diff --git a/pkg_resources/__init__.py b/pkg_resources/__init__.py
index 12226d4b..a323857c 100644
--- a/pkg_resources/__init__.py
+++ b/pkg_resources/__init__.py
@@ -36,6 +36,7 @@ import plistlib
 import email.parser
 import tempfile
 import textwrap
+import itertools
 from pkgutil import get_importer
 
 try:
@@ -1966,6 +1967,32 @@ def find_nothing(importer, path_item, only=False):
 register_finder(object, find_nothing)
 
 
+def _by_version_descending(names):
+    """
+    Given a list of filenames, return them in descending order
+    by version number.
+
+    >>> names = 'bar', 'foo', 'Python-2.7.10.egg', 'Python-2.7.2.egg'
+    >>> _by_version_descending(names)
+    ['Python-2.7.10.egg', 'Python-2.7.2.egg', 'foo', 'bar']
+    >>> names = 'Setuptools-1.2.3b1.egg', 'Setuptools-1.2.3.egg'
+    >>> _by_version_descending(names)
+    ['Setuptools-1.2.3.egg', 'Setuptools-1.2.3b1.egg']
+    >>> names = 'Setuptools-1.2.3b1.egg', 'Setuptools-1.2.3.post1.egg'
+    >>> _by_version_descending(names)
+    ['Setuptools-1.2.3.post1.egg', 'Setuptools-1.2.3b1.egg']
+    """
+    def _by_version(name):
+        """
+        Parse each component of the filename
+        """
+        name, ext = os.path.splitext(name)
+        parts = itertools.chain(name.split('-'), [ext])
+        return [packaging.version.parse(part) for part in parts]
+
+    return sorted(names, key=_by_version, reverse=True)
+
+
 def find_on_path(importer, path_item, only=False):
     """Yield distributions accessible on a sys.path directory"""
     path_item = _normalize_cached(path_item)
@@ -1979,11 +2006,7 @@ def find_on_path(importer, path_item, only=False):
             )
         else:
             # scan for .egg and .egg-info in directory
-
-            path_item_entries = os.listdir(path_item)
-            # Reverse so we find the newest version of a distribution,
-            path_item_entries.sort()
-            path_item_entries.reverse()
+            path_item_entries = _by_version_descending(os.listdir(path_item))
             for entry in path_item_entries:
                 lower = entry.lower()
                 if lower.endswith('.egg-info') or lower.endswith('.dist-info'):