Add CDATA section handlers

diff --git a/expat/xmlparse/xmlparse.c b/expat/xmlparse/xmlparse.c
index 37f0aa794db9d678ec3bcc9e35bda0fb24f5b4b4..e5e77309056caa90cade2a4364187f4fe5f79bc4 100755 (executable)
--- a/expat/xmlparse/xmlparse.c
+++ b/expat/xmlparse/xmlparse.c
@@ -268,6 +268,8 @@ typedef struct {
   XML_CharacterDataHandler characterDataHandler;
   XML_ProcessingInstructionHandler processingInstructionHandler;
   XML_CommentHandler commentHandler;
+  XML_StartCdataSectionHandler startCdataSectionHandler;
+  XML_EndCdataSectionHandler endCdataSectionHandler;
   XML_DefaultHandler defaultHandler;
   XML_UnparsedEntityDeclHandler unparsedEntityDeclHandler;
   XML_NotationDeclHandler notationDeclHandler;
@@ -319,6 +321,8 @@ typedef struct {
 #define characterDataHandler (((Parser *)parser)->characterDataHandler)
 #define processingInstructionHandler (((Parser *)parser)->processingInstructionHandler)
 #define commentHandler (((Parser *)parser)->commentHandler)
+#define startCdataSectionHandler (((Parser *)parser)->startCdataSectionHandler)
+#define endCdataSectionHandler (((Parser *)parser)->endCdataSectionHandler)
 #define defaultHandler (((Parser *)parser)->defaultHandler)
 #define unparsedEntityDeclHandler (((Parser *)parser)->unparsedEntityDeclHandler)
 #define notationDeclHandler (((Parser *)parser)->notationDeclHandler)
@@ -385,6 +389,8 @@ XML_Parser XML_ParserCreate(const XML_Char *encodingName)
   characterDataHandler = 0;
   processingInstructionHandler = 0;
   commentHandler = 0;
+  startCdataSectionHandler = 0;
+  endCdataSectionHandler = 0;
   defaultHandler = 0;
   unparsedEntityDeclHandler = 0;
   notationDeclHandler = 0;
@@ -477,6 +483,8 @@ XML_Parser XML_ExternalEntityParserCreate(XML_Parser oldParser,
   XML_CharacterDataHandler oldCharacterDataHandler = characterDataHandler;
   XML_ProcessingInstructionHandler oldProcessingInstructionHandler = processingInstructionHandler;
   XML_CommentHandler oldCommentHandler = commentHandler;
+  XML_StartCdataSectionHandler oldStartCdataSectionHandler = startCdataSectionHandler;
+  XML_EndCdataSectionHandler oldEndCdataSectionHandler = endCdataSectionHandler;
   XML_DefaultHandler oldDefaultHandler = defaultHandler;
   XML_ExternalEntityRefHandler oldExternalEntityRefHandler = externalEntityRefHandler;
   XML_UnknownEncodingHandler oldUnknownEncodingHandler = unknownEncodingHandler;
@@ -494,6 +502,8 @@ XML_Parser XML_ExternalEntityParserCreate(XML_Parser oldParser,
   characterDataHandler = oldCharacterDataHandler;
   processingInstructionHandler = oldProcessingInstructionHandler;
   commentHandler = oldCommentHandler;
+  startCdataSectionHandler = oldStartCdataSectionHandler;
+  endCdataSectionHandler = oldEndCdataSectionHandler;
   defaultHandler = oldDefaultHandler;
   externalEntityRefHandler = oldExternalEntityRefHandler;
   unknownEncodingHandler = oldUnknownEncodingHandler;
@@ -612,6 +622,14 @@ void XML_SetCommentHandler(XML_Parser parser,
   commentHandler = handler;
 }
 
+void XML_SetCdataSectionHandler(XML_Parser parser,
+                               XML_StartCdataSectionHandler start,
+                               XML_EndCdataSectionHandler end)
+{
+  startCdataSectionHandler = start;
+  endCdataSectionHandler = end;
+}
+
 void XML_SetDefaultHandler(XML_Parser parser,
                           XML_DefaultHandler handler)
 {
@@ -1255,8 +1273,23 @@ doContent(XML_Parser parser,
     case XML_TOK_CDATA_SECT_OPEN:
       {
        enum XML_Error result;
-       if (characterDataHandler)
+       if (startCdataSectionHandler)
+         startCdataSectionHandler(handlerArg);
+#if 0
+       /* Suppose you doing a transformation on a document that involves
+          changing only the character data.  You set up a defaultHandler
+          and a characterDataHandler.  The defaultHandler simply copies
+          characters through.  The characterDataHandler does the transformation
+          and writes the characters out escaping them as necessary.  This case
+          will fail to work if we leave out the following two lines (because &
+          and < inside CDATA sections will be incorrectly escaped).
+
+          However, now we have a start/endCdataSectionHandler, so it seems
+          easier to let the user deal with this. */
+
+       else if (characterDataHandler)
          characterDataHandler(handlerArg, dataBuf, 0);
+#endif
        else if (defaultHandler)
          reportDefault(parser, enc, s, next);
        result = doCdataSection(parser, enc, &next, end, nextPtr);
@@ -1614,8 +1647,13 @@ enum XML_Error doCdataSection(XML_Parser parser,
     *eventEndPP = next;
     switch (tok) {
     case XML_TOK_CDATA_SECT_CLOSE:
-      if (characterDataHandler)
+      if (endCdataSectionHandler)
+       endCdataSectionHandler(handlerArg);
+#if 0
+      /* see comment under XML_TOK_CDATA_SECT_OPEN */
+      else if (characterDataHandler)
        characterDataHandler(handlerArg, dataBuf, 0);
+#endif
       else if (defaultHandler)
        reportDefault(parser, enc, s, next);
       *startPtr = next;

diff --git a/expat/xmlparse/xmlparse.h b/expat/xmlparse/xmlparse.h
index 57549e52aeea554b8a17ad729fb201b4b5a8e47f..473f2afea5cfd501bb1d026232da186afe875db1 100755 (executable)
--- a/expat/xmlparse/xmlparse.h
+++ b/expat/xmlparse/xmlparse.h
@@ -107,6 +107,9 @@ typedef void (*XML_ProcessingInstructionHandler)(void *userData,
 /* data is 0 terminated */
 typedef void (*XML_CommentHandler)(void *userData, const XML_Char *data);
 
+typedef void (*XML_StartCdataSectionHandler)(void *userData);
+typedef void (*XML_EndCdataSectionHandler)(void *userData);
+
 /* This is called for any characters in the XML document for
 which there is no applicable handler.  This includes both
 characters that are part of markup which is of a kind that is
@@ -255,6 +258,11 @@ void XMLPARSEAPI
 XML_SetCommentHandler(XML_Parser parser,
                       XML_CommentHandler handler);
 
+void XMLPARSEAPI
+XML_SetCdataSectionHandler(XML_Parser parser,
+                          XML_StartCdataSectionHandler start,
+                          XML_EndCdataSectionHandler end);
+
 /* This sets the default handler and also inhibits expansion of internal entities.
 The entity reference will be passed to the default handler. */
 

diff --git a/expat/xmlwf/xmlwf.c b/expat/xmlwf/xmlwf.c
index d3c89de1fe63fd5167099f38e8536bc4345c3fbe..5525d7baa5ca3035708a4b0324cc5719c593607b 100755 (executable)
--- a/expat/xmlwf/xmlwf.c
+++ b/expat/xmlwf/xmlwf.c
@@ -334,6 +334,24 @@ void metaComment(XML_Parser parser, const XML_Char *data)
   fputts(T("/>\n"), fp);
 }
 
+static
+void metaStartCdataSection(XML_Parser parser)
+{
+  FILE *fp = XML_GetUserData(parser);
+  fputts(T("<startcdata"), fp);
+  metaLocation(parser);
+  fputts(T("/>\n"), fp);
+}
+
+static
+void metaEndCdataSection(XML_Parser parser)
+{
+  FILE *fp = XML_GetUserData(parser);
+  fputts(T("<endcdata"), fp);
+  metaLocation(parser);
+  fputts(T("/>\n"), fp);
+}
+
 static
 void metaCharacterData(XML_Parser parser, const XML_Char *s, int len)
 {
@@ -695,6 +713,7 @@ int tmain(int argc, XML_Char **argv)
        XML_SetElementHandler(parser, metaStartElement, metaEndElement);
        XML_SetProcessingInstructionHandler(parser, metaProcessingInstruction);
        XML_SetCommentHandler(parser, metaComment);
+       XML_SetCdataSectionHandler(parser, metaStartCdataSection, metaEndCdataSection);
        XML_SetCharacterDataHandler(parser, metaCharacterData);
        XML_SetUnparsedEntityDeclHandler(parser, metaUnparsedEntityDecl);
        XML_SetNotationDeclHandler(parser, metaNotationDecl);