add support for processing instructions (that are under the root node)

Author: keith-hall

lxml_parser.py

@@ -41,9 +41,13 @@ class LocationAwareComment(etree.CommentBase):
     tag_pos = None
 
 
+class LocationAwareProcessingInstruction(etree.PIBase):
+    tag_pos = None
+
+
 # http://stackoverflow.com/questions/36246014/lxml-use-default-class-element-lookup-and-treebuilder-parser-target-at-the-sam
 class LocationAwareXMLParser:
-    RE_SPLIT_XML = re.compile('<!\[CDATA\[|\]\]>|[<>]')
+    RE_SPLIT_XML = re.compile(r'<!\[CDATA\[|\]\]>|[<>]')
 
     def __init__(self, position_offset = 0, **parser_options):
         def getLocation(index=None):
@@ -163,6 +167,13 @@ def element_end(self, tag, location=None):
     def text_data(self, data, location=None):
         self._text.append(data)
 
+    def pi(self, target, data, location=None):
+        if self._most_recent is not None:
+            self._flush()
+            self._appendNode(self.create_pi(target, data))
+            self._most_recent.tag_pos = location
+            self._in_tail = True
+    
     def comment(self, text, location=None):
         if self._most_recent is not None:
             self._flush()
@@ -173,6 +184,9 @@ def comment(self, text, location=None):
     def create_comment(self, text):
         return LocationAwareComment(text)
 
+    def create_pi(self, target, data):
+        return LocationAwareProcessingInstruction(target, data)
+    
     def _appendNode(self, node):
         if self._element_stack: # if we have anything on the stack
             self._element_stack[-1].append(node) # append the node as a child to the last/top element on the stack
@@ -206,7 +220,7 @@ def lxml_etree_parse_xml_string_with_location(xml_chunks, position_offset = 0, s
 def getNodeTagRange(node, position_type):
     """Given a node and position type (open or close), return the node's position."""
     pos = None
-    if isinstance(node, LocationAwareComment):
+    if isinstance(node, LocationAwareComment) or isinstance(node, LocationAwareProcessingInstruction):
         pos = node.tag_pos
     else:
         pos = getattr(node, position_type + '_tag_pos')

sublime_lxml.py

@@ -105,6 +105,8 @@ def get_nodes_from_document(nodes):
             element = node.getparent() # get the parent
             if element is None: # some nodes are not actually part of the original document we parsed, for example when using the substring function. so there is no way to find the original node, and therefore the location
                 continue
+        elif isinstance(node, etree.PIBase):
+            element = node
         elif isinstance(node, etree.CommentBase):
             element = node
         elif isinstance(node, etree.ElementBase):
@@ -154,7 +156,7 @@ def ensureTagNameEndPosIsSet(node, open_pos):
             if next_node is not None:
                 text_end_pos = getNodeTagRegion(view, next_node, 'open').begin()
             yield sublime.Region(text_begin_pos, text_end_pos)
-        elif isinstance(node, etree.CommentBase):
+        elif isinstance(node, etree.CommentBase) or isinstance(node, etree.PIBase):
             yield open_pos
         elif attr_name is None or attribute_position_type is None or attribute_position_type in ('element', 'parent'):
             # position type 'open' <|name| attr1="test"></name> "Goto name in open tag"

xpath.py

@@ -958,7 +958,7 @@ def completions_axis_specifiers():
         return [(completion + '\taxis', completion + '::') for completion in completions]
 
     def completions_node_types():
-        completions = ['text', 'node', 'comment'] # 'processing-instruction' is also a valid XPath node type, but not parsed into an ElementTree, so useless to show it in suggestions/completions
+        completions = ['text', 'node', 'comment', 'processing-instruction']
         return [(completion + '\tnode type', completion + '()') for completion in completions]
 
     def completions_functions():