- Large resync between W3C and Gnome tree

- configure.in: 2.1.0 prerelease
- example/Makefile.am example/gjobread.c tree.h: work on
  libxml1 libxml2 convergence.
- nanoftp, nanohttp.c: fixed stalled connections probs
- HTMLtree.c SAX.c : support for attribute without values in
  HTML for andersca
- valid.c: Fixed most validation + namespace problems
- HTMLparser.c: start document callback for andersca
- debugXML.c xpath.c: lots of XPath fixups from Picdar Technology
- parser.h, SAX.c: serious speed improvement for large
  CDATA blocks
- encoding.[ch] xmlIO.[ch]: Improved seriously saving to
  different encoding
- config.h.in parser.c xmllint.c: added xmlCheckVersion()
  and the LIBXML_TEST_VERSION macro
Daniel

entities.c

@@ -945,6 +945,86 @@ xmlEncodeEntitiesReentrant(xmlDocPtr doc, const xmlChar *input) {
     return(buffer);
 }
 
+/**
+ * xmlEncodeSpecialChars:
+ * @doc:  the document containing the string
+ * @input:  A string to convert to XML.
+ *
+ * Do a global encoding of a string, replacing the predefined entities
+ * this routine is reentrant, and result must be deallocated.
+ *
+ * Returns A newly allocated string with the substitution done.
+ */
+xmlChar *
+xmlEncodeSpecialChars(xmlDocPtr doc, const xmlChar *input) {
+    const xmlChar *cur = input;
+    xmlChar *buffer = NULL;
+    xmlChar *out = NULL;
+    int buffer_size = 0;
+    int html = 0;
+
+    if (input == NULL) return(NULL);
+    if (doc != NULL)
+        html = (doc->type == XML_HTML_DOCUMENT_NODE);
+
+    /*
+     * allocate an translation buffer.
+     */
+    buffer_size = 1000;
+    buffer = (xmlChar *) xmlMalloc(buffer_size * sizeof(xmlChar));
+    if (buffer == NULL) {
+	perror("malloc failed");
+	return(NULL);
+    }
+    out = buffer;
+
+    while (*cur != '\0') {
+        if (out - buffer > buffer_size - 10) {
+	    int index = out - buffer;
+
+	    growBufferReentrant();
+	    out = &buffer[index];
+	}
+
+	/*
+	 * By default one have to encode at least '<', '>', '"' and '&' !
+	 */
+	if (*cur == '<') {
+	    *out++ = '&';
+	    *out++ = 'l';
+	    *out++ = 't';
+	    *out++ = ';';
+	} else if (*cur == '>') {
+	    *out++ = '&';
+	    *out++ = 'g';
+	    *out++ = 't';
+	    *out++ = ';';
+	} else if (*cur == '&') {
+	    *out++ = '&';
+	    *out++ = 'a';
+	    *out++ = 'm';
+	    *out++ = 'p';
+	    *out++ = ';';
+	} else if (*cur == '"') {
+	    *out++ = '&';
+	    *out++ = 'q';
+	    *out++ = 'u';
+	    *out++ = 'o';
+	    *out++ = 't';
+	    *out++ = ';';
+	} else {
+	    /*
+	     * Works because on UTF-8, all extended sequences cannot
+	     * result in bytes in the ASCII range.
+	     */
+	    *out++ = *cur;
+	}
+	cur++;
+    }
+    *out++ = 0;
+    return(buffer);
+}
+
 /**
  * xmlCreateEntitiesTable:
  *