[Gnash-commit] gnash ChangeLog server/asobj/xml.cpp server/aso...

[Sandro Santilli]

CVSROOT:        /sources/gnash
Module name:    gnash
Changes by:     Sandro Santilli <strk>  08/04/08 10:09:21

Modified files:
        .              : ChangeLog 
        server/asobj   : xml.cpp xml.h 
        testsuite/actionscript.all: XML.as 
        testsuite/swfdec: PASSING 

Log message:
                * server/asobj/xml.{cpp,h}: rework parsing so it doesn't
                  try to recover at first, but does in a more compatible way
                  on second try. Fixes parsing of STUDENT.xml in bug #22819.
                * testsuite/actionscript.all/XML.as: parsing of gnashrc.xml work
                  more compatibly now.
                * testsuite/swfdec/PASSING: string-object-tostring-{5,6,7,8}.

CVSWeb URLs:
http://cvs.savannah.gnu.org/viewcvs/gnash/ChangeLog?cvsroot=gnash&r1=1.6214&r2=1.6215
http://cvs.savannah.gnu.org/viewcvs/gnash/server/asobj/xml.cpp?cvsroot=gnash&r1=1.77&r2=1.78
http://cvs.savannah.gnu.org/viewcvs/gnash/server/asobj/xml.h?cvsroot=gnash&r1=1.32&r2=1.33
http://cvs.savannah.gnu.org/viewcvs/gnash/testsuite/actionscript.all/XML.as?cvsroot=gnash&r1=1.56&r2=1.57
http://cvs.savannah.gnu.org/viewcvs/gnash/testsuite/swfdec/PASSING?cvsroot=gnash&r1=1.124&r2=1.125

Patches:
Index: ChangeLog
===================================================================
RCS file: /sources/gnash/gnash/ChangeLog,v
retrieving revision 1.6214
retrieving revision 1.6215
diff -u -b -r1.6214 -r1.6215
--- ChangeLog   8 Apr 2008 02:56:28 -0000       1.6214
+++ ChangeLog   8 Apr 2008 10:09:20 -0000       1.6215
@@ -1,3 +1,12 @@
+2008-04-07 Sandro Santilli <address@hidden>
+
+       * server/asobj/xml.{cpp,h}: rework parsing so it doesn't
+         try to recover at first, but does in a more compatible way
+         on second try. Fixes parsing of STUDENT.xml in bug #22819.
+       * testsuite/actionscript.all/XML.as: parsing of gnashrc.xml work
+         more compatibly now.
+       * testsuite/swfdec/PASSING: string-object-tostring-{5,6,7,8}.
+
 2008-04-07 Russ Nelson <address@hidden>
 
        * GAH!  Checked in too much.

Index: server/asobj/xml.cpp
===================================================================
RCS file: /sources/gnash/gnash/server/asobj/xml.cpp,v
retrieving revision 1.77
retrieving revision 1.78
diff -u -b -r1.77 -r1.78
--- server/asobj/xml.cpp        28 Mar 2008 16:23:08 -0000      1.77
+++ server/asobj/xml.cpp        8 Apr 2008 10:09:21 -0000       1.78
@@ -60,6 +60,9 @@
 // Define this to enable verbosity of XML loads
 //#define DEBUG_XML_LOADS 1
 
+// Define this to enable verbosity of XML parsing
+//#define DEBUG_XML_PARSE 1
+
 static as_object* getXMLInterface();
 static void attachXMLInterface(as_object& o);
 static void attachXMLProperties(as_object& o);
@@ -83,8 +86,8 @@
 XML::XML() 
     :
     XMLNode(getXMLInterface()),
-    _doc(0),
-    _firstChild(0),
+    //_doc(0),
+    //_firstChild(0),
     _loaded(-1), 
     _status(sOK),
     _loadThreads(),
@@ -105,8 +108,8 @@
 XML::XML(const std::string& xml_in)
     :
     XMLNode(getXMLInterface()),
-    _doc(0),
-    _firstChild(0),
+    //_doc(0),
+    //_firstChild(0),
     _loaded(-1), 
     _status(sOK),
     _loadThreads(),
@@ -236,16 +239,18 @@
     xmlChar *ptr = NULL;
     boost::intrusive_ptr<XMLNode> child;
 
-//    log_debug(_("Created new element for %s at %p"), node->name, element);
-
-//    log_debug(_("%s: extracting node %s"), __FUNCTION__, node->name);
+#ifdef DEBUG_XML_PARSE
+    log_debug(_("%s: extracting node %s"), __FUNCTION__, node->name);
+#endif
 
     // See if we have any Attributes (properties)
     attr = node->properties;
     while (attr != NULL)
     {
-        //log_debug(_("extractNode %s has property %s, value is %s"),
-        //          node->name, attr->name, attr->children->content);
+#ifdef DEBUG_XML_PARSE
+        log_debug(_("extractNode %s has property %s, value is %s"),
+                  node->name, attr->name, attr->children->content);
+#endif
         
         std::ostringstream name, content;
 
@@ -254,8 +259,11 @@
         
         XMLAttr attrib(name.str(), content.str());
 
-        //log_debug(_("\tPushing attribute %s for element %s has value %s"),
-        //        attr->name, node->name, attr->children->content);
+#ifdef DEBUG_XML_PARSE
+        log_debug(_("\tPushing attribute %s for element %s has value %s, next 
attribute is %p"),
+                attr->name, node->name, attr->children->content, attr->next);
+#endif
+
         element._attributes.push_back(attrib);
         attr = attr->next;
     }
@@ -290,7 +298,9 @@
                {
                        if ( in.str().find_first_not_of(" \n\t\r") == 
std::string::npos )
                        {
+#ifdef DEBUG_XML_PARSE
                                log_debug("Text node value consists in blanks 
only, discarding");
+#endif
                                xmlFree(ptr);
                                return false;
                        }
@@ -307,7 +317,10 @@
     {
         child = new XMLNode();
         child->setParent(&element);
-        if ( extractNode(*child, childnode, mem) ) 
element._children.push_back(child);
+        if ( extractNode(*child, childnode, mem) )
+       {
+               element._children.push_back(child);
+       }
         childnode = childnode->next;
     }
 
@@ -316,27 +329,22 @@
 
 /*private*/
 bool
-XML::parseDoc(xmlDocPtr document, bool mem)
+XML::parseDoc(xmlNodePtr cur, bool mem)
 {
-    //GNASH_REPORT_FUNCTION;  
-
-    xmlNodePtr cur;
-
-    if (document == 0) {
-        log_error(_("Can't load XML file"));
-        return false;
-    }
-
-    cur = xmlDocGetRootElement(document);
+    GNASH_REPORT_FUNCTION;  
   
-    if (cur != NULL)
+    while (cur)
     {
         boost::intrusive_ptr<XMLNode> child = new XMLNode();
         child->setParent(this);
+#ifdef DEBUG_XML_PARSE
+        log_debug("\tParsing top-level node %s", cur->name);
+#endif
         if ( extractNode(*child, cur, mem) ) 
        {
                _children.push_back(child);
        }
+       cur = cur->next;
     }  
 
     return true;
@@ -367,16 +375,36 @@
     
     initParser();
 
-    //_doc = xmlParseMemory(xml_in.c_str(), xml_in.size());
-    _doc = xmlReadMemory(xml_in.c_str(), xml_in.size(), NULL, NULL, 
getXMLOptions());
-    if (_doc == 0) {
-        log_error(_("Can't parse XML data"));
+    xmlNodePtr firstNode; 
+
+    xmlDocPtr doc = xmlReadMemory(xml_in.c_str(), xml_in.size(), NULL, NULL, 
getXMLOptions()); // do NOT recover here !
+    if ( doc )
+    {
+        firstNode = doc->children; // xmlDocGetRootElement(doc);
+    }
+    else
+    {
+        log_debug(_("malformed XML, trying to recover"));
+        int ret = xmlParseBalancedChunkMemoryRecover(NULL, NULL, NULL, 0, 
(const xmlChar*)xml_in.c_str(), &firstNode, 1);
+        log_debug("xmlParseBalancedChunkMemoryRecover returned %d", ret);
+        if ( ! firstNode )
+        {
+            log_error(_("unrecoverable malformed XML 
(xmlParseBalancedChunkMemoryRecover returned %d)."), ret);
         return false;
     }
+        else
+        {
+            log_error(_("recovered malformed XML."));
+        }
+    }
+
+
+
+    bool ret = parseDoc(firstNode, true);
 
-    bool ret = parseDoc(_doc, true);
     xmlCleanupParser();
-    xmlFreeDoc(_doc);
+    if ( doc ) xmlFreeDoc(doc); // TOCHECK: can it be freed before ?
+    else if ( firstNode ) xmlFreeNodeList(firstNode);
     xmlMemoryDump();
 
 #ifndef USE_DMALLOC
@@ -1022,12 +1050,8 @@
 bool
 XML::ignoreWhite() const
 {
-    // TODO: initialize this thing once...
-    std::string propname;
-    if ( VM::get().getSWFVersion() < 7 ) propname = "ignorewhite";
-    else propname = "ignoreWhite";
 
-       string_table::key propnamekey = 
VM::get().getStringTable().find(propname);
+       string_table::key propnamekey = 
VM::get().getStringTable().find("ignoreWhite");
     as_value val;
     if (!const_cast<XML*>(this)->get_member(propnamekey, &val) ) return false;
     return val.to_bool();
@@ -1037,8 +1061,11 @@
 int
 XML::getXMLOptions() const
 {
-    int options = XML_PARSE_RECOVER | XML_PARSE_NOWARNING
-               | XML_PARSE_NOERROR | XML_PARSE_NOCDATA;
+    int options = XML_PARSE_NOENT
+               //| XML_PARSE_RECOVER -- don't recover now, we'll call 
xmlParseBalancedChunkRecover later
+               //| XML_PARSE_NOWARNING
+               //| XML_PARSE_NOERROR
+               | XML_PARSE_NOCDATA;
     // Using libxml2 to convert CDATA nodes to text seems to be what is
     // required.
     

Index: server/asobj/xml.h
===================================================================
RCS file: /sources/gnash/gnash/server/asobj/xml.h,v
retrieving revision 1.32
retrieving revision 1.33
diff -u -b -r1.32 -r1.33
--- server/asobj/xml.h  28 Mar 2008 16:23:08 -0000      1.32
+++ server/asobj/xml.h  8 Apr 2008 10:09:21 -0000       1.33
@@ -175,10 +175,7 @@
     //
     /// This is the base method used by both parseXML() and load().
     ///
-    bool parseDoc(xmlDocPtr document, bool mem);
-
-    xmlDocPtr _doc;
-    xmlNodePtr _firstChild;
+    bool parseDoc(xmlNodePtr startNode, bool mem);
     
     // -1 if never asked to load anything
     //  0 if asked to load but not yet loaded (or failure)

Index: testsuite/actionscript.all/XML.as
===================================================================
RCS file: /sources/gnash/gnash/testsuite/actionscript.all/XML.as,v
retrieving revision 1.56
retrieving revision 1.57
diff -u -b -r1.56 -r1.57
--- testsuite/actionscript.all/XML.as   11 Mar 2008 19:31:48 -0000      1.56
+++ testsuite/actionscript.all/XML.as   8 Apr 2008 10:09:21 -0000       1.57
@@ -21,7 +21,7 @@
 // execute it like this gnash -1 -r 0 -v out.swf
 
 
-rcsid="$Id: XML.as,v 1.56 2008/03/11 19:31:48 strk Exp $";
+rcsid="$Id: XML.as,v 1.57 2008/04/08 10:09:21 strk Exp $";
 #include "check.as"
 //#include "dejagnu.as"
 #include "utils.as"
@@ -693,7 +693,7 @@
        myxml.status = status_backup;
 
 
-       //note("myxml.toString(): "+myxml.toString());
+       note("myxml.toString(): "+myxml.toString());
        check_equals(typeof(myxml.attributes), 'object');
        check(! myxml.attributes instanceof Object);
        check_equals(typeof(myxml.attributes.__proto__), 'undefined');
@@ -714,7 +714,7 @@
        {
                check_equals(myxml.childNodes.length, 
myxml.lastChildNodesCount);
        }
-       xcheck_equals(myxml.childNodes.length, 3); // gnash fails discarding 
the comment and the ending blanks
+       check_equals(myxml.childNodes.length, 3); // gnash fails discarding the 
comment and the ending blanks
 
        // We're done
        ++this.onLoadCalls;

Index: testsuite/swfdec/PASSING
===================================================================
RCS file: /sources/gnash/gnash/testsuite/swfdec/PASSING,v
retrieving revision 1.124
retrieving revision 1.125
diff -u -b -r1.124 -r1.125
--- testsuite/swfdec/PASSING    7 Apr 2008 22:01:55 -0000       1.124
+++ testsuite/swfdec/PASSING    8 Apr 2008 10:09:21 -0000       1.125
@@ -734,6 +734,10 @@
 stringlength-5.swf:bf014dcc64776fb859ff878286f56b16
 stringlength-6.swf:94b4de2a590c915696b4a783338a1f05
 stringlength-7.swf:140952977abb85b966f1e5c4447854ce
+string-object-tostring-5.swf:d45c1785719c3e2607f6ac10f48b2e80
+string-object-tostring-6.swf:f4970a48bfb3c5d5315699263c2a15a1
+string-object-tostring-7.swf:060a8f7ba61ae1501bdcb060cd321313
+string-object-tostring-8.swf:7741a0c0d26aef26bd46749457b0357a
 string-properties-5.swf:39cc1143e531659d37072534a3d3455e
 string-properties-6.swf:6e04c332f503311867082ab55f22b16b
 string-properties-7.swf:bf61a4b0a4cc67c8df63203407ac243c