fixed a memory bug added test from Graham Bennett and regenerated the

* xmlschemas.c: fixed a memory bug
* doc/examples/reader4.c doc/examples/*: added test from Graham Bennett
  and regenerated the directory
Daniel

doc/examples/reader4.c

@@ -0,0 +1,122 @@
+/**
+ * section: xmlReader
+ * synopsis: Parse multiple XML files reusing an xmlReader
+ * purpose: Demonstrate the use of xmlReaderForFile() and
+ * xmlReaderNewFile to parse XML files while reusing the reader object
+ * and parser context.  (Note that the XMLReader functions require
+ * libxml2 version later than 2.6.)
+ * usage: reader4 <filename> [ filename ... ]
+ * test: reader4 test1.xml test2.xml test3.xml > reader4.tmp ; diff reader4.tmp reader4.res ; rm reader4.tmp
+ * author: Graham Bennett
+ * copy: see Copyright for the status of this software.
+ */
+
+#include <stdio.h>
+#include <libxml/xmlreader.h>
+
+#ifdef LIBXML_READER_ENABLED
+
+static void processDoc(xmlTextReaderPtr readerPtr) {
+    int ret;
+    xmlDocPtr docPtr;
+    const xmlChar *URL;
+
+    ret = xmlTextReaderRead(readerPtr);
+    while (ret == 1) {
+      ret = xmlTextReaderRead(readerPtr);
+    }
+
+    /*
+     * One can obtain the document pointer to get insteresting
+     * information about the document like the URL, but one must also
+     * be sure to clean it up at the end (see below).
+     */
+    docPtr = xmlTextReaderCurrentDoc(readerPtr);
+    if (NULL == docPtr) {
+      fprintf(stderr, "failed to obtain document\n");      
+      return;
+    }
+      
+    URL = docPtr->URL;
+    if (NULL == URL) {
+      fprintf(stderr, "Failed to obtain URL\n");      
+    }
+
+    if (ret != 0) {
+      fprintf(stderr, "%s: Failed to parse\n", URL);
+      return;
+    }
+
+    printf("%s: Processed ok\n", (const char *)URL);
+}
+
+int main(int argc, char **argv) {
+    xmlTextReaderPtr readerPtr;
+    int i;
+    xmlDocPtr docPtr;
+
+    if (argc < 2)
+        return(1);
+
+    /*
+     * this initialises the library and check potential ABI mismatches
+     * between the version it was compiled for and the actual shared
+     * library used.
+     */
+    LIBXML_TEST_VERSION
+
+    /*
+     * Create a new reader for the first file and process the
+     * document.
+     */
+    readerPtr = xmlReaderForFile(argv[1], NULL, 0);
+    if (NULL == readerPtr) {
+      fprintf(stderr, "%s: failed to create reader\n", argv[1]);      
+      return(1);
+    }
+    processDoc(readerPtr);
+
+    /*
+     * The reader can be reused for subsequent files.
+     */
+    for (i=2; i < argc; ++i) {
+      	xmlReaderNewFile(readerPtr, argv[i], NULL, 0);
+	if (NULL == readerPtr) {
+	  fprintf(stderr, "%s: failed to create reader\n", argv[i]);      
+	  return(1);
+	}
+        processDoc(readerPtr);
+    }
+
+    /*
+     * Since we've called xmlTextReaderCurrentDoc, we now have to
+     * clean up after ourselves.  We only have to do this the last
+     * time, because xmlReaderNewFile calls xmlCtxtReset which takes
+     * care of it.
+     */
+    docPtr = xmlTextReaderCurrentDoc(readerPtr);
+    if (docPtr != NULL)
+      xmlFreeDoc(docPtr);
+
+    /*
+     * Clean up the reader.
+     */
+    xmlFreeTextReader(readerPtr);
+
+    /*
+     * Cleanup function for the XML library.
+     */
+    xmlCleanupParser();
+    /*
+     * this is to debug memory for regression tests
+     */
+    xmlMemoryDump();
+    return(0);
+}
+
+#else
+int main(void) {
+    fprintf(stderr, "xmlReader support not compiled in\n");
+    exit(1);
+}
+#endif