pygccxml-commit

[pygccxml-commit] SF.net SVN: pygccxml: [1003] pygccxml_dev/pygccxml/parser/etree_scanner.py

[<rom...@us...>]

Revision: 1003
          http://svn.sourceforge.net/pygccxml/?rev=1003&view=rev
Author:   roman_yakovenko
Date:     2007-04-19 02:44:45 -0700 (Thu, 19 Apr 2007)

Log Message:
-----------
small improvements to the code

Modified Paths:
--------------
    pygccxml_dev/pygccxml/parser/etree_scanner.py

Modified: pygccxml_dev/pygccxml/parser/etree_scanner.py
===================================================================
--- pygccxml_dev/pygccxml/parser/etree_scanner.py	2007-04-19 08:51:24 UTC (rev 1002)
+++ pygccxml_dev/pygccxml/parser/etree_scanner.py	2007-04-19 09:44:45 UTC (rev 1003)
@@ -7,36 +7,27 @@
 import xml.etree.cElementTree as ElementTree
         
 class etree_saxifier_t(object):
-    """Produces SAX events for an element and children."""
-    def __init__(self, element_or_tree, content_handler):
-        try:
-            element = element_or_tree.getroot()
-        except AttributeError:
-            element = element_or_tree
-        self._element = element
-        self._content_handler = content_handler        
+    def __init__(self, etree, handler):
+        self.__root_elem = etree.getroot()
+        self.__handler = handler        
         
     def saxify(self):
-        self._content_handler.startDocument()
-        self._recursive_saxify(self._element, {})
-        self._content_handler.endDocument()
+        self.__handler.startDocument()
+        self.__recursive_saxify( self.__root_elem )
+        self.__handler.endDocument()
 
-    def _recursive_saxify(self, element, prefixes):
-        attrs = {}
-        for attr in element.items():
-            attrs[ attr[0] ] = attr[1]
-
-        content_handler = self._content_handler
-        content_handler.startElement( element.tag, attrs )
+    def __recursive_saxify(self, element ):
+        self.__handler.startElement( element.tag, element.attrib )
         
         if element.text:
-            content_handler.characters(element.text)
-        for child in element:
-            self._recursive_saxify(child, prefixes)
-        content_handler.endElement( element.tag )
+            self.__handler.characters(element.text)
+            
+        map( self.__recursive_saxify, element )
+            
+        self.__handler.endElement( element.tag )
 
         if element.tail:
-            content_handler.characters(element.tail)
+            self.__handler.characters(element.tail)
 
 class etree_scanner_t( scanner.scanner_t ):
     def __init__(self, gccxml_file, decl_factory, *args ):


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

[pygccxml-commit] SF.net SVN: pygccxml: [1004] pygccxml_dev/pygccxml/parser/etree_scanner.py

[<rom...@us...>]

Revision: 1004
          http://svn.sourceforge.net/pygccxml/?rev=1004&view=rev
Author:   roman_yakovenko
Date:     2007-04-19 04:23:51 -0700 (Thu, 19 Apr 2007)

Log Message:
-----------
even faster implementation of xml parser

Modified Paths:
--------------
    pygccxml_dev/pygccxml/parser/etree_scanner.py

Modified: pygccxml_dev/pygccxml/parser/etree_scanner.py
===================================================================
--- pygccxml_dev/pygccxml/parser/etree_scanner.py	2007-04-19 09:44:45 UTC (rev 1003)
+++ pygccxml_dev/pygccxml/parser/etree_scanner.py	2007-04-19 11:23:51 UTC (rev 1004)
@@ -19,22 +19,37 @@
     def __recursive_saxify(self, element ):
         self.__handler.startElement( element.tag, element.attrib )
         
-        if element.text:
-            self.__handler.characters(element.text)
+        #~ if element.text:
+            #~ self.__handler.characters(element.text)
             
         map( self.__recursive_saxify, element )
             
         self.__handler.endElement( element.tag )
 
-        if element.tail:
-            self.__handler.characters(element.tail)
+        #~ if element.tail:
+            #~ self.__handler.characters(element.tail)
 
 class etree_scanner_t( scanner.scanner_t ):
     def __init__(self, gccxml_file, decl_factory, *args ):
         scanner.scanner_t.__init__( self, gccxml_file, decl_factory, *args )
     
-    def read( self ):
+    def read( self ):        
         tree = ElementTree.parse( self.gccxml_file )
         saxifier = etree_saxifier_t( tree, self )
-        saxifier.saxify()
+        saxifier.saxify()  
+
+class ietree_scanner_t( scanner.scanner_t ):
+    def __init__(self, gccxml_file, decl_factory, *args ):
+        scanner.scanner_t.__init__( self, gccxml_file, decl_factory, *args )
     
+    def read( self ):        
+        context = ElementTree.iterparse(self.gccxml_file, events=("start", "end"))
+        for event, elem in context:
+            if event == 'start':
+                self.startElement( elem.tag, elem.attrib )
+            else:
+                self.endElement( elem.tag )
+                elem.clear()
+        self.endDocument()
+    
+etree_scanner_t = ietree_scanner_t


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.