[Python-checkins] bpo-30485: support a default prefix mapping in ElementPath by passing None as prefix (#1823)

[Stefan Behnel]

https://github.com/python/cpython/commit/e9927e1820caea01e576141d9a623ea394d43dad
commit: e9927e1820caea01e576141d9a623ea394d43dad
branch: master
author: Stefan Behnel <stefan_ml at behnel.de>
committer: GitHub <noreply at github.com>
date: 2019-04-14T10:09:09+02:00
summary:

bpo-30485: support a default prefix mapping in ElementPath by passing None as prefix (#1823)

files:
A Misc/NEWS.d/next/Library/2019-04-13-23-42-33.bpo-30485.JHhjJS.rst
M Doc/library/xml.etree.elementtree.rst
M Lib/test/test_xml_etree.py
M Lib/xml/etree/ElementPath.py

diff --git a/Doc/library/xml.etree.elementtree.rst b/Doc/library/xml.etree.elementtree.rst
index 9bee0eadc289..c83e719e959a 100644
--- a/Doc/library/xml.etree.elementtree.rst
+++ b/Doc/library/xml.etree.elementtree.rst
@@ -764,7 +764,8 @@ Element Objects
       Finds the first subelement matching *match*.  *match* may be a tag name
       or a :ref:`path <elementtree-xpath>`.  Returns an element instance
       or ``None``.  *namespaces* is an optional mapping from namespace prefix
-      to full name.
+      to full name.  Pass ``None`` as prefix to move all unprefixed tag names
+      in the expression into the given namespace.
 
 
    .. method:: findall(match, namespaces=None)
@@ -772,7 +773,8 @@ Element Objects
       Finds all matching subelements, by tag name or
       :ref:`path <elementtree-xpath>`.  Returns a list containing all matching
       elements in document order.  *namespaces* is an optional mapping from
-      namespace prefix to full name.
+      namespace prefix to full name.  Pass ``None`` as prefix to move all
+      unprefixed tag names in the expression into the given namespace.
 
 
    .. method:: findtext(match, default=None, namespaces=None)
@@ -782,7 +784,8 @@ Element Objects
       of the first matching element, or *default* if no element was found.
       Note that if the matching element has no text content an empty string
       is returned. *namespaces* is an optional mapping from namespace prefix
-      to full name.
+      to full name.  Pass ``None`` as prefix to move all unprefixed tag names
+      in the expression into the given namespace.
 
 
    .. method:: getchildren()
diff --git a/Lib/test/test_xml_etree.py b/Lib/test/test_xml_etree.py
index bdcd4e0d19a7..2f7a3b60b22d 100644
--- a/Lib/test/test_xml_etree.py
+++ b/Lib/test/test_xml_etree.py
@@ -2463,6 +2463,12 @@ def test_findall_different_nsmaps(self):
         nsmap = {'xx': 'Y'}
         self.assertEqual(len(root.findall(".//xx:b", namespaces=nsmap)), 1)
         self.assertEqual(len(root.findall(".//b", namespaces=nsmap)), 2)
+        nsmap = {'xx': 'X', None: 'Y'}
+        self.assertEqual(len(root.findall(".//xx:b", namespaces=nsmap)), 2)
+        self.assertEqual(len(root.findall(".//b", namespaces=nsmap)), 1)
+        nsmap = {'xx': 'X', '': 'Y'}
+        with self.assertRaisesRegex(ValueError, 'namespace prefix'):
+            root.findall(".//xx:b", namespaces=nsmap)
 
     def test_bad_find(self):
         e = ET.XML(SAMPLE_XML)
diff --git a/Lib/xml/etree/ElementPath.py b/Lib/xml/etree/ElementPath.py
index ef32917b14d4..0e3854f9db22 100644
--- a/Lib/xml/etree/ElementPath.py
+++ b/Lib/xml/etree/ElementPath.py
@@ -71,16 +71,22 @@
     )
 
 def xpath_tokenizer(pattern, namespaces=None):
+    default_namespace = namespaces.get(None) if namespaces else None
     for token in xpath_tokenizer_re.findall(pattern):
         tag = token[1]
-        if tag and tag[0] != "{" and ":" in tag:
-            try:
+        if tag and tag[0] != "{":
+            if ":" in tag:
                 prefix, uri = tag.split(":", 1)
-                if not namespaces:
-                    raise KeyError
-                yield token[0], "{%s}%s" % (namespaces[prefix], uri)
-            except KeyError:
-                raise SyntaxError("prefix %r not found in prefix map" % prefix) from None
+                try:
+                    if not namespaces:
+                        raise KeyError
+                    yield token[0], "{%s}%s" % (namespaces[prefix], uri)
+                except KeyError:
+                    raise SyntaxError("prefix %r not found in prefix map" % prefix) from None
+            elif default_namespace:
+                yield token[0], "{%s}%s" % (default_namespace, tag)
+            else:
+                yield token
         else:
             yield token
 
@@ -264,10 +270,19 @@ def __init__(self, root):
 
 def iterfind(elem, path, namespaces=None):
     # compile selector pattern
-    cache_key = (path, None if namespaces is None
-                            else tuple(sorted(namespaces.items())))
     if path[-1:] == "/":
         path = path + "*" # implicit all (FIXME: keep this?)
+
+    cache_key = (path,)
+    if namespaces:
+        if '' in namespaces:
+            raise ValueError("empty namespace prefix must be passed as None, not the empty string")
+        if None in namespaces:
+            cache_key += (namespaces[None],) + tuple(sorted(
+                item for item in namespaces.items() if item[0] is not None))
+        else:
+            cache_key += tuple(sorted(namespaces.items()))
+
     try:
         selector = _cache[cache_key]
     except KeyError:
diff --git a/Misc/NEWS.d/next/Library/2019-04-13-23-42-33.bpo-30485.JHhjJS.rst b/Misc/NEWS.d/next/Library/2019-04-13-23-42-33.bpo-30485.JHhjJS.rst
new file mode 100644
index 000000000000..6c82efd3e009
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2019-04-13-23-42-33.bpo-30485.JHhjJS.rst
@@ -0,0 +1,3 @@
+Path expressions in xml.etree.ElementTree can now avoid explicit namespace
+prefixes for tags (or the "{namespace}tag" notation) by passing a default
+namespace with a 'None' prefix.