[xml] pre 2.7 sax behavior

[Rob Richards <rrichards ctindustries net>]

Apologies if this comes through twice. Originally sent from a different 
email address.

Attached is patch to provide pre 2.7 entity processing behavior when 
setting an XML_PARSE_OLDSAX option on the parser context.
The changes in entity processing made in 2.7 cause some serious breakage 
in the PHP SAX processor and no other alternative exists to fix this 
problem without libxml2 supporting the old behavior. In order not to 
degrade performance, this option is disabled by default.

Please let me know if anyone has or sees any issues with this patch?

Thanks,

Rob


Index: include/libxml/parser.h
===================================================================
--- include/libxml/parser.h     (revision 3804)
+++ include/libxml/parser.h     (working copy)
@@ -1096,7 +1096,8 @@
                                   crash if you try to modify the tree) */
     XML_PARSE_OLD10    = 1<<17,/* parse using XML-1.0 before update 5 */
     XML_PARSE_NOBASEFIX = 1<<18,/* do not fixup XINCLUDE xml:base uris */
-    XML_PARSE_HUGE      = 1<<19 /* relax any hardcoded limit from the parser */
+    XML_PARSE_HUGE      = 1<<19, /* relax any hardcoded limit from the parser */
+    XML_PARSE_OLDSAX    = 1<<20,/* parse using SAX2 interface from before 2.7.0 */
 } xmlParserOption;
 
 XMLPUBFUN void XMLCALL
Index: parser.c
===================================================================
--- parser.c    (revision 3804)
+++ parser.c    (working copy)
@@ -7047,9 +7047,11 @@
     /*
      * Predefined entites override any extra definition
      */
-    ent = xmlGetPredefinedEntity(name);
-    if (ent != NULL)
-        return(ent);
+    if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
+        ent = xmlGetPredefinedEntity(name);
+        if (ent != NULL)
+            return(ent);
+    }
 
     /*
      * Increate the number of entity references parsed
@@ -7063,6 +7065,9 @@
     if (ctxt->sax != NULL) {
        if (ctxt->sax->getEntity != NULL)
            ent = ctxt->sax->getEntity(ctxt->userData, name);
+       if ((ctxt->wellFormed == 1 ) && (ent == NULL) && 
+           (ctxt->options & XML_PARSE_OLDSAX))
+           ent = xmlGetPredefinedEntity(name);
        if ((ctxt->wellFormed == 1 ) && (ent == NULL) &&
            (ctxt->userData==ctxt)) {
            ent = xmlSAX2GetEntity(ctxt, name);
@@ -7135,6 +7140,7 @@
      */
     else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
             (ent != NULL) && (ent->content != NULL) &&
+            (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
             (xmlStrchr(ent->content, '<'))) {
        xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
     "'<' in entity '%s' is not allowed in attributes values\n", name);
@@ -7231,11 +7237,13 @@
     /*
      * Predefined entites override any extra definition
      */
-    ent = xmlGetPredefinedEntity(name);
-    if (ent != NULL) {
-        xmlFree(name);
-        *str = ptr;
-        return(ent);
+    if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
+        ent = xmlGetPredefinedEntity(name);
+        if (ent != NULL) {
+            xmlFree(name);
+            *str = ptr;
+            return(ent);
+        }
     }
 
     /*
@@ -7250,6 +7258,8 @@
     if (ctxt->sax != NULL) {
        if (ctxt->sax->getEntity != NULL)
            ent = ctxt->sax->getEntity(ctxt->userData, name);
+       if ((ent == NULL) && (ctxt->options & XML_PARSE_OLDSAX))
+           ent = xmlGetPredefinedEntity(name);
        if ((ent == NULL) && (ctxt->userData==ctxt)) {
            ent = xmlSAX2GetEntity(ctxt, name);
        }
@@ -7318,6 +7328,7 @@
      */
     else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
             (ent != NULL) && (ent->content != NULL) &&
+            (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
             (xmlStrchr(ent->content, '<'))) {
        xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
      "'<' in entity '%s' is not allowed in attributes values\n",
@@ -14211,6 +14222,10 @@
        ctxt->options |= XML_PARSE_HUGE;
         options -= XML_PARSE_HUGE;
     }
+    if (options & XML_PARSE_OLDSAX) {
+       ctxt->options |= XML_PARSE_OLDSAX;
+        options -= XML_PARSE_OLDSAX;
+    }
     ctxt->linenumbers = 1;
     return (options);
 }