Address infinite loop when zipfile begins with more than one leading …

…slash.

Alternate and more surgical fix for #119. Ref python/cpython#123270

pyproject.toml

@@ -37,7 +37,6 @@ test = [
 	"pytest-ignore-flaky",
 	"jaraco.test",
 	"importlib_resources; python_version < '3.9'",
-	"pytest-timeout",
 ]
 
 doc = [

tests/test_path.py

@@ -586,7 +586,6 @@ def test_getinfo_missing(self, alpharep):
             alpharep.getinfo('does-not-exist')
 
     @pytest.mark.xfail(reason="python/cpython#123270")
-    @pytest.mark.timeout(1)
     def test_malformed_paths(self):
         """
         Path should handle malformed paths.

zipp/__init__.py

@@ -46,7 +46,7 @@ def _parents(path):
 def _ancestry(path):
     """
     Given a path with elements separated by
-    posixpath.sep, generate all elements of that path
+    posixpath.sep, generate all elements of that path.
 
     >>> list(_ancestry('b/d'))
     ['b/d', 'b']
@@ -58,9 +58,14 @@ def _ancestry(path):
     ['b']
     >>> list(_ancestry(''))
     []
+
+    Multiple separators are treated like a single.
+
+    >>> list(_ancestry('//b//d///f//'))
+    ['//b//d///f', '//b//d', '//b']
     """
     path = path.rstrip(posixpath.sep)
-    while path and path != posixpath.sep:
+    while path and not path.endswith(posixpath.sep):
         yield path
         path, tail = posixpath.split(path)