parser: Add more parser context accessors

Fixes #763.

HTMLparser.c

@@ -5927,13 +5927,7 @@ htmlCtxtParseDocument(htmlParserCtxtPtr ctxt, xmlParserInputPtr input)
     ctxt->html = INSERT_INITIAL;
     htmlParseDocument(ctxt);
 
-    if (ctxt->errNo != XML_ERR_NO_MEMORY) {
+    ret = xmlCtxtGetDocument(ctxt);
-        ret = ctxt->myDoc;
-    } else {
-        ret = NULL;
-        xmlFreeDoc(ctxt->myDoc);
-    }
-    ctxt->myDoc = NULL;
 
     /* assert(ctxt->inputNr == 1); */
     while (ctxt->inputNr > 0)

doc/libxml2-api.xml

@@ -690,11 +690,16 @@
      <exports symbol='xmlCtxtGetCatalogs' type='function'/>
      <exports symbol='xmlCtxtGetDeclaredEncoding' type='function'/>
      <exports symbol='xmlCtxtGetDict' type='function'/>
+     <exports symbol='xmlCtxtGetDocument' type='function'/>
      <exports symbol='xmlCtxtGetOptions' type='function'/>
      <exports symbol='xmlCtxtGetPrivate' type='function'/>
+     <exports symbol='xmlCtxtGetSaxHandler' type='function'/>
      <exports symbol='xmlCtxtGetStandalone' type='function'/>
      <exports symbol='xmlCtxtGetStatus' type='function'/>
+     <exports symbol='xmlCtxtGetValidCtxt' type='function'/>
      <exports symbol='xmlCtxtGetVersion' type='function'/>
+     <exports symbol='xmlCtxtIsHtml' type='function'/>
+     <exports symbol='xmlCtxtIsStopped' type='function'/>
      <exports symbol='xmlCtxtParseContent' type='function'/>
      <exports symbol='xmlCtxtParseDocument' type='function'/>
      <exports symbol='xmlCtxtParseDtd' type='function'/>
@@ -713,6 +718,7 @@
      <exports symbol='xmlCtxtSetOptions' type='function'/>
      <exports symbol='xmlCtxtSetPrivate' type='function'/>
      <exports symbol='xmlCtxtSetResourceLoader' type='function'/>
+     <exports symbol='xmlCtxtSetSaxHandler' type='function'/>
      <exports symbol='xmlCtxtUseOptions' type='function'/>
      <exports symbol='xmlCtxtValidateDocument' type='function'/>
      <exports symbol='xmlCtxtValidateDtd' type='function'/>
@@ -8451,6 +8457,11 @@ crash if you try to modify the tree)'/>
       <return type='xmlDictPtr' info='the dictionary.'/>
       <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
     </function>
+    <function name='xmlCtxtGetDocument' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.</info>
+      <return type='xmlDocPtr' info='the parsed document or NULL if a fatal error occurred when parsing. The document must be freed by the caller. Resets the context&apos;s document to NULL.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+    </function>
     <function name='xmlCtxtGetLastError' file='xmlerror' module='parserInternals'>
       <info>Get the last parsing error registered.</info>
       <return type='const xmlError *' info='NULL if no error occurred or a pointer to the error'/>
@@ -8466,6 +8477,11 @@ crash if you try to modify the tree)'/>
       <return type='void *' info='the private application data.'/>
       <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
     </function>
+    <function name='xmlCtxtGetSaxHandler' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.</info>
+      <return type='xmlSAXHandler *' info='the SAX handler struct. This is not a copy and must not be freed. Handlers can be updated.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+    </function>
     <function name='xmlCtxtGetStandalone' file='parser' module='parser'>
       <info>Available since 2.14.0.</info>
       <return type='int' info='the value from the standalone document declaration.'/>
@@ -8476,11 +8492,26 @@ crash if you try to modify the tree)'/>
       <return type='int' info='a bitmask of XML_STATUS_* flags ORed together.'/>
       <arg name='ctxt' type='xmlParserCtxt *' info='an XML parser context'/>
     </function>
+    <function name='xmlCtxtGetValidCtxt' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.</info>
+      <return type='xmlValidCtxtPtr' info='the validation context.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+    </function>
     <function name='xmlCtxtGetVersion' file='parser' module='parser'>
       <info>Available since 2.14.0.</info>
       <return type='const xmlChar *' info='the version from the XML declaration.'/>
       <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
     </function>
+    <function name='xmlCtxtIsHtml' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.</info>
+      <return type='int' info='1 if this is a HTML parser context, 0 otherwise.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+    </function>
+    <function name='xmlCtxtIsStopped' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.</info>
+      <return type='int' info='1 if the parser is stopped, 0 otherwise.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+    </function>
     <function name='xmlCtxtParseContent' file='parser' module='parser'>
       <info>Parse a well-balanced chunk of XML matching the &apos;content&apos; production.  Namespaces in scope of @node and entities of @node&apos;s document are recognized. When validating, the DTD of @node&apos;s document is used.  Always consumes @input even in error case.  Available since 2.14.0.</info>
       <return type='xmlNodePtr' info='a node list or NULL in case of error.'/>
@@ -8632,6 +8663,12 @@ crash if you try to modify the tree)'/>
       <arg name='loader' type='xmlResourceLoader' info='callback'/>
       <arg name='vctxt' type='void *' info='user data'/>
     </function>
+    <function name='xmlCtxtSetSaxHandler' file='parser' module='parserInternals'>
+      <info>Available since 2.14.0.  Set the SAX handler struct to a copy of @sax.</info>
+      <return type='int' info='0 on success or -1 if arguments are invalid or a memory allocation failed.'/>
+      <arg name='ctxt' type='xmlParserCtxtPtr' info='parser context'/>
+      <arg name='sax' type='const xmlSAXHandler *' info='SAX handler'/>
+    </function>
     <function name='xmlCtxtUseOptions' file='parser' module='parser'>
       <info>DEPRECATED: Use xmlCtxtSetOptions.  Applies the options to the parser context. The following options are never cleared and can only be enabled:  XML_PARSE_NOERROR XML_PARSE_NOWARNING XML_PARSE_NONET XML_PARSE_NSCLEAN XML_PARSE_NOCDATA XML_PARSE_COMPACT XML_PARSE_OLD10 XML_PARSE_HUGE XML_PARSE_OLDSAX XML_PARSE_IGNORE_ENC XML_PARSE_BIG_LINES</info>
       <return type='int' info='0 in case of success, the set of unknown or unimplemented options in case of error.'/>

include/libxml/parser.h

@@ -1405,6 +1405,19 @@ XMLPUBFUN xmlDictPtr
 XMLPUBFUN void
 		xmlCtxtSetDict		(xmlParserCtxtPtr ctxt,
 					 xmlDictPtr);
+XMLPUBFUN xmlSAXHandler *
+		xmlCtxtGetSaxHandler	(xmlParserCtxtPtr ctxt);
+XMLPUBFUN int
+		xmlCtxtSetSaxHandler	(xmlParserCtxtPtr ctxt,
+					 const xmlSAXHandler *sax);
+XMLPUBFUN xmlDocPtr
+		xmlCtxtGetDocument	(xmlParserCtxtPtr ctxt);
+XMLPUBFUN int
+		xmlCtxtIsHtml		(xmlParserCtxtPtr ctxt);
+XMLPUBFUN int
+		xmlCtxtIsStopped	(xmlParserCtxtPtr ctxt);
+XMLPUBFUN xmlValidCtxtPtr
+		xmlCtxtGetValidCtxt	(xmlParserCtxtPtr ctxt);
 XMLPUBFUN const xmlChar *
 		xmlCtxtGetVersion	(xmlParserCtxtPtr ctxt);
 XMLPUBFUN const xmlChar *

parser.c

@@ -13816,17 +13816,9 @@ xmlCtxtParseDocument(xmlParserCtxtPtr ctxt, xmlParserInputPtr input)
 
     xmlParseDocument(ctxt);
 
-    if ((ctxt->wellFormed) ||
+    ret = xmlCtxtGetDocument(ctxt);
-        ((ctxt->recovery) && (!xmlCtxtIsCatastrophicError(ctxt)))) {
+    if ((ret == NULL) && (ctxt->errNo == XML_ERR_OK))
-        ret = ctxt->myDoc;
+        xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR, "unknown error\n");
-    } else {
-        if (ctxt->errNo == XML_ERR_OK)
-            xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR, "unknown error\n");
-
-        ret = NULL;
-	xmlFreeDoc(ctxt->myDoc);
-    }
-    ctxt->myDoc = NULL;
 
     /* assert(ctxt->inputNr == 1); */
     while (ctxt->inputNr > 0)

parserInternals.c

@@ -3162,6 +3162,130 @@ xmlCtxtSetDict(xmlParserCtxtPtr ctxt, xmlDictPtr dict) {
     ctxt->dict = dict;
 }
 
+/**
+ * xmlCtxtGetSaxHandler:
+ * @ctxt:  parser context
+ *
+ * Available since 2.14.0.
+ *
+ * Returns the SAX handler struct. This is not a copy and must not
+ * be freed. Handlers can be updated.
+ */
+xmlSAXHandler *
+xmlCtxtGetSaxHandler(xmlParserCtxtPtr ctxt) {
+    if (ctxt == NULL)
+        return(NULL);
+
+    return(ctxt->sax);
+}
+
+/**
+ * xmlCtxtSetSaxHandler:
+ * @ctxt:  parser context
+ * @sax:  SAX handler
+ *
+ * Available since 2.14.0.
+ *
+ * Set the SAX handler struct to a copy of @sax.
+ *
+ * Returns 0 on success or -1 if arguments are invalid or a memory
+ * allocation failed.
+ */
+int
+xmlCtxtSetSaxHandler(xmlParserCtxtPtr ctxt, const xmlSAXHandler *sax) {
+    xmlSAXHandler *copy;
+
+    if ((ctxt == NULL) || (sax == NULL))
+        return(-1);
+
+    copy = xmlMalloc(sizeof(*copy));
+    if (copy == NULL)
+        return(-1);
+
+    memcpy(copy, sax, sizeof(*copy));
+    ctxt->sax = copy;
+
+    return(0);
+}
+
+/**
+ * xmlCtxtGetDocument:
+ * @ctxt:  parser context
+ *
+ * Available since 2.14.0.
+ *
+ * Returns the parsed document or NULL if a fatal error occurred when
+ * parsing. The document must be freed by the caller. Resets the
+ * context's document to NULL.
+ */
+xmlDocPtr
+xmlCtxtGetDocument(xmlParserCtxtPtr ctxt) {
+    xmlDocPtr doc;
+
+    if (ctxt == NULL)
+        return(NULL);
+
+    if ((ctxt->wellFormed) ||
+        (((ctxt->recovery) || (ctxt->html)) &&
+         (!xmlCtxtIsCatastrophicError(ctxt)))) {
+        doc = ctxt->myDoc;
+    } else {
+        doc = NULL;
+        xmlFreeDoc(ctxt->myDoc);
+    }
+    ctxt->myDoc = NULL;
+
+    return(doc);
+}
+
+/**
+ * xmlCtxtIsHtml:
+ * @ctxt:  parser context
+ *
+ * Available since 2.14.0.
+ *
+ * Returns 1 if this is a HTML parser context, 0 otherwise.
+ */
+int
+xmlCtxtIsHtml(xmlParserCtxtPtr ctxt) {
+    if (ctxt == NULL)
+        return(0);
+
+    return(ctxt->html ? 1 : 0);
+}
+
+/**
+ * xmlCtxtIsStopped:
+ * @ctxt:  parser context
+ *
+ * Available since 2.14.0.
+ *
+ * Returns 1 if the parser is stopped, 0 otherwise.
+ */
+int
+xmlCtxtIsStopped(xmlParserCtxtPtr ctxt) {
+    if (ctxt == NULL)
+        return(0);
+
+    return(PARSER_STOPPED(ctxt));
+}
+
+/**
+ * xmlCtxtGetValidCtxt:
+ * @ctxt:  parser context
+ *
+ * Available since 2.14.0.
+ *
+ * Returns the validation context.
+ */
+xmlValidCtxtPtr
+xmlCtxtGetValidCtxt(xmlParserCtxtPtr ctxt) {
+    if (ctxt == NULL)
+        return(NULL);
+
+    return(&ctxt->vctxt);
+}
+
 /************************************************************************
  *									*
  *		Handling of node information				*

testapi.c

@@ -11937,6 +11937,38 @@ test_xmlCtxtGetDeclaredEncoding(void) {
 }
 
 
+static int
+test_xmlCtxtGetDocument(void) {
+    int test_ret = 0;
+
+    int mem_base;
+    xmlDocPtr ret_val;
+    xmlParserCtxtPtr ctxt; /* parser context */
+    int n_ctxt;
+
+    for (n_ctxt = 0;n_ctxt < gen_nb_xmlParserCtxtPtr;n_ctxt++) {
+        mem_base = xmlMemBlocks();
+        ctxt = gen_xmlParserCtxtPtr(n_ctxt, 0);
+
+        ret_val = xmlCtxtGetDocument(ctxt);
+        desret_xmlDocPtr(ret_val);
+        call_tests++;
+        des_xmlParserCtxtPtr(n_ctxt, ctxt, 0);
+        xmlResetLastError();
+        if (mem_base != xmlMemBlocks()) {
+            printf("Leak of %d blocks found in xmlCtxtGetDocument",
+	           xmlMemBlocks() - mem_base);
+	    test_ret++;
+            printf(" %d", n_ctxt);
+            printf("\n");
+        }
+    }
+    function_tests++;
+
+    return(test_ret);
+}
+
+
 static int
 test_xmlCtxtGetOptions(void) {
     int test_ret = 0;
@@ -12001,6 +12033,16 @@ test_xmlCtxtGetPrivate(void) {
 }
 
 
+static int
+test_xmlCtxtGetSaxHandler(void) {
+    int test_ret = 0;
+
+
+    /* missing type support */
+    return(test_ret);
+}
+
+
 static int
 test_xmlCtxtGetStandalone(void) {
     int test_ret = 0;
@@ -12069,6 +12111,16 @@ test_xmlCtxtGetStatus(void) {
 }
 
 
+static int
+test_xmlCtxtGetValidCtxt(void) {
+    int test_ret = 0;
+
+
+    /* missing type support */
+    return(test_ret);
+}
+
+
 static int
 test_xmlCtxtGetVersion(void) {
     int test_ret = 0;
@@ -12101,6 +12153,70 @@ test_xmlCtxtGetVersion(void) {
 }
 
 
+static int
+test_xmlCtxtIsHtml(void) {
+    int test_ret = 0;
+
+    int mem_base;
+    int ret_val;
+    xmlParserCtxtPtr ctxt; /* parser context */
+    int n_ctxt;
+
+    for (n_ctxt = 0;n_ctxt < gen_nb_xmlParserCtxtPtr;n_ctxt++) {
+        mem_base = xmlMemBlocks();
+        ctxt = gen_xmlParserCtxtPtr(n_ctxt, 0);
+
+        ret_val = xmlCtxtIsHtml(ctxt);
+        desret_int(ret_val);
+        call_tests++;
+        des_xmlParserCtxtPtr(n_ctxt, ctxt, 0);
+        xmlResetLastError();
+        if (mem_base != xmlMemBlocks()) {
+            printf("Leak of %d blocks found in xmlCtxtIsHtml",
+	           xmlMemBlocks() - mem_base);
+	    test_ret++;
+            printf(" %d", n_ctxt);
+            printf("\n");
+        }
+    }
+    function_tests++;
+
+    return(test_ret);
+}
+
+
+static int
+test_xmlCtxtIsStopped(void) {
+    int test_ret = 0;
+
+    int mem_base;
+    int ret_val;
+    xmlParserCtxtPtr ctxt; /* parser context */
+    int n_ctxt;
+
+    for (n_ctxt = 0;n_ctxt < gen_nb_xmlParserCtxtPtr;n_ctxt++) {
+        mem_base = xmlMemBlocks();
+        ctxt = gen_xmlParserCtxtPtr(n_ctxt, 0);
+
+        ret_val = xmlCtxtIsStopped(ctxt);
+        desret_int(ret_val);
+        call_tests++;
+        des_xmlParserCtxtPtr(n_ctxt, ctxt, 0);
+        xmlResetLastError();
+        if (mem_base != xmlMemBlocks()) {
+            printf("Leak of %d blocks found in xmlCtxtIsStopped",
+	           xmlMemBlocks() - mem_base);
+	    test_ret++;
+            printf(" %d", n_ctxt);
+            printf("\n");
+        }
+    }
+    function_tests++;
+
+    return(test_ret);
+}
+
+
 static int
 test_xmlCtxtParseContent(void) {
     int test_ret = 0;
@@ -12714,6 +12830,49 @@ test_xmlCtxtSetResourceLoader(void) {
 }
 
 
+#define gen_nb_const_xmlSAXHandler_ptr 1
+#define gen_const_xmlSAXHandler_ptr(no, nr) NULL
+#define des_const_xmlSAXHandler_ptr(no, val, nr)
+
+static int
+test_xmlCtxtSetSaxHandler(void) {
+    int test_ret = 0;
+
+    int mem_base;
+    int ret_val;
+    xmlParserCtxtPtr ctxt; /* parser context */
+    int n_ctxt;
+    const xmlSAXHandler * sax; /* SAX handler */
+    int n_sax;
+
+    for (n_ctxt = 0;n_ctxt < gen_nb_xmlParserCtxtPtr;n_ctxt++) {
+    for (n_sax = 0;n_sax < gen_nb_const_xmlSAXHandler_ptr;n_sax++) {
+        mem_base = xmlMemBlocks();
+        ctxt = gen_xmlParserCtxtPtr(n_ctxt, 0);
+        sax = gen_const_xmlSAXHandler_ptr(n_sax, 1);
+
+        ret_val = xmlCtxtSetSaxHandler(ctxt, sax);
+        desret_int(ret_val);
+        call_tests++;
+        des_xmlParserCtxtPtr(n_ctxt, ctxt, 0);
+        des_const_xmlSAXHandler_ptr(n_sax, sax, 1);
+        xmlResetLastError();
+        if (mem_base != xmlMemBlocks()) {
+            printf("Leak of %d blocks found in xmlCtxtSetSaxHandler",
+	           xmlMemBlocks() - mem_base);
+	    test_ret++;
+            printf(" %d", n_ctxt);
+            printf(" %d", n_sax);
+            printf("\n");
+        }
+    }
+    }
+    function_tests++;
+
+    return(test_ret);
+}
+
+
 static int
 test_xmlCtxtUseOptions(void) {
     int test_ret = 0;
@@ -13415,10 +13574,6 @@ test_xmlNewParserCtxt(void) {
 }
 
 
-#define gen_nb_const_xmlSAXHandler_ptr 1
-#define gen_const_xmlSAXHandler_ptr(no, nr) NULL
-#define des_const_xmlSAXHandler_ptr(no, val, nr)
-
 static int
 test_xmlNewSAXParserCtxt(void) {
     int test_ret = 0;
@@ -15489,7 +15644,7 @@ static int
 test_parser(void) {
     int test_ret = 0;
 
-    if (quiet == 0) printf("Testing parser : 88 of 104 functions ...\n");
+    if (quiet == 0) printf("Testing parser : 92 of 110 functions ...\n");
     test_ret += test_xmlByteConsumed();
     test_ret += test_xmlCleanupGlobals();
     test_ret += test_xmlClearNodeInfoSeq();
@@ -15498,11 +15653,16 @@ test_parser(void) {
     test_ret += test_xmlCreatePushParserCtxt();
     test_ret += test_xmlCtxtGetCatalogs();
     test_ret += test_xmlCtxtGetDeclaredEncoding();
+    test_ret += test_xmlCtxtGetDocument();
     test_ret += test_xmlCtxtGetOptions();
     test_ret += test_xmlCtxtGetPrivate();
+    test_ret += test_xmlCtxtGetSaxHandler();
     test_ret += test_xmlCtxtGetStandalone();
     test_ret += test_xmlCtxtGetStatus();
+    test_ret += test_xmlCtxtGetValidCtxt();
     test_ret += test_xmlCtxtGetVersion();
+    test_ret += test_xmlCtxtIsHtml();
+    test_ret += test_xmlCtxtIsStopped();
     test_ret += test_xmlCtxtParseContent();
     test_ret += test_xmlCtxtParseDocument();
     test_ret += test_xmlCtxtParseDtd();
@@ -15519,6 +15679,7 @@ test_parser(void) {
     test_ret += test_xmlCtxtSetOptions();
     test_ret += test_xmlCtxtSetPrivate();
     test_ret += test_xmlCtxtSetResourceLoader();
+    test_ret += test_xmlCtxtSetSaxHandler();
     test_ret += test_xmlCtxtUseOptions();
     test_ret += test_xmlCtxtValidateDocument();
     test_ret += test_xmlCtxtValidateDtd();