From 250faf3c832d998baa559ca1a1c61935235aba20 Mon Sep 17 00:00:00 2001
From: Nick Wellnhofer <wellnhofer@aevum.de>
Date: Thu, 20 Apr 2023 12:35:21 +0200
Subject: [PATCH] parser: Fix regression in xmlParserNodeInfo accounting

Commit 62150ed2 broke begin_pos and begin_line when extra node info was
recorded.

Fixes #523.
---
 SAX2.c   | 20 ++------------------
 parser.c | 53 +++++++++++++++++++++++++----------------------------
 2 files changed, 27 insertions(+), 46 deletions(-)

diff --git a/SAX2.c b/SAX2.c
index 4b09fa6e..b6be1a6c 100644
--- a/SAX2.c
+++ b/SAX2.c
@@ -1830,13 +1830,6 @@ xmlSAX2EndElement(void *ctx, const xmlChar *name ATTRIBUTE_UNUSED)
 	xmlGenericError(xmlGenericErrorContext, "SAX.xmlSAX2EndElement(%s)\n", name);
 #endif
 
-    /* Capture end position and add node */
-    if (cur != NULL && ctxt->record_info) {
-      ctxt->nodeInfo->end_pos = ctxt->input->cur - ctxt->input->base;
-      ctxt->nodeInfo->end_line = ctxt->input->line;
-      ctxt->nodeInfo->node = cur;
-      xmlParserAddNodeInfo(ctxt, ctxt->nodeInfo);
-    }
     ctxt->nodemem = -1;
 
 #ifdef LIBXML_VALID_ENABLED
@@ -2479,24 +2472,15 @@ xmlSAX2EndElementNs(void *ctx,
 		    const xmlChar * URI ATTRIBUTE_UNUSED)
 {
     xmlParserCtxtPtr ctxt = (xmlParserCtxtPtr) ctx;
-    xmlParserNodeInfo node_info;
-    xmlNodePtr cur;
 
     if (ctx == NULL) return;
-    cur = ctxt->node;
-    /* Capture end position and add node */
-    if ((ctxt->record_info) && (cur != NULL)) {
-        node_info.end_pos = ctxt->input->cur - ctxt->input->base;
-        node_info.end_line = ctxt->input->line;
-        node_info.node = cur;
-        xmlParserAddNodeInfo(ctxt, &node_info);
-    }
     ctxt->nodemem = -1;
 
 #ifdef LIBXML_VALID_ENABLED
     if (ctxt->validate && ctxt->wellFormed &&
         ctxt->myDoc && ctxt->myDoc->intSubset)
-        ctxt->valid &= xmlValidateOneElement(&ctxt->vctxt, ctxt->myDoc, cur);
+        ctxt->valid &= xmlValidateOneElement(&ctxt->vctxt, ctxt->myDoc,
+                                             ctxt->node);
 #endif /* LIBXML_VALID_ENABLED */
 
     /*
diff --git a/parser.c b/parser.c
index 78c384f2..9ed64c9e 100644
--- a/parser.c
+++ b/parser.c
@@ -9945,7 +9945,7 @@ xmlParseElementStart(xmlParserCtxtPtr ctxt) {
     const xmlChar *URI = NULL;
     xmlParserNodeInfo node_info;
     int line, tlen = 0;
-    xmlNodePtr ret;
+    xmlNodePtr cur;
     int nsNr = ctxt->nsNr;
 
     if (((unsigned int) ctxt->nameNr > xmlParserMaxDepth) &&
@@ -9987,7 +9987,7 @@ xmlParseElementStart(xmlParserCtxtPtr ctxt) {
         return(-1);
     }
     nameNsPush(ctxt, name, prefix, URI, line, ctxt->nsNr - nsNr);
-    ret = ctxt->node;
+    cur = ctxt->node;
 
 #ifdef LIBXML_VALID_ENABLED
     /*
@@ -10020,17 +10020,23 @@ xmlParseElementStart(xmlParserCtxtPtr ctxt) {
 	spacePop(ctxt);
 	if (nsNr != ctxt->nsNr)
 	    nsPop(ctxt, ctxt->nsNr - nsNr);
-	if ( ret != NULL && ctxt->record_info ) {
-	   node_info.end_pos = ctxt->input->consumed +
-			      (CUR_PTR - ctxt->input->base);
-	   node_info.end_line = ctxt->input->line;
-	   node_info.node = ret;
-	   xmlParserAddNodeInfo(ctxt, &node_info);
+	if (cur != NULL && ctxt->record_info) {
+            node_info.node = cur;
+            node_info.end_pos = ctxt->input->consumed +
+                                (CUR_PTR - ctxt->input->base);
+            node_info.end_line = ctxt->input->line;
+            xmlParserAddNodeInfo(ctxt, &node_info);
 	}
 	return(1);
     }
     if (RAW == '>') {
         NEXT1;
+        if (cur != NULL && ctxt->record_info) {
+            node_info.node = cur;
+            node_info.end_pos = 0;
+            node_info.end_line = 0;
+            xmlParserAddNodeInfo(ctxt, &node_info);
+        }
     } else {
         xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_GT_REQUIRED,
 		     "Couldn't find end of Start Tag %s line %d\n",
@@ -10044,17 +10050,6 @@ xmlParseElementStart(xmlParserCtxtPtr ctxt) {
 	spacePop(ctxt);
 	if (nsNr != ctxt->nsNr)
 	    nsPop(ctxt, ctxt->nsNr - nsNr);
-
-	/*
-	 * Capture end position and add node
-	 */
-	if ( ret != NULL && ctxt->record_info ) {
-	   node_info.end_pos = ctxt->input->consumed +
-			      (CUR_PTR - ctxt->input->base);
-	   node_info.end_line = ctxt->input->line;
-	   node_info.node = ret;
-	   xmlParserAddNodeInfo(ctxt, &node_info);
-	}
 	return(-1);
     }
 
@@ -10069,8 +10064,7 @@ xmlParseElementStart(xmlParserCtxtPtr ctxt) {
  */
 static void
 xmlParseElementEnd(xmlParserCtxtPtr ctxt) {
-    xmlParserNodeInfo node_info;
-    xmlNodePtr ret = ctxt->node;
+    xmlNodePtr cur = ctxt->node;
 
     if (ctxt->nameNr <= 0) {
         if ((RAW == '<') && (NXT(1) == '/'))
@@ -10091,14 +10085,17 @@ xmlParseElementEnd(xmlParserCtxtPtr ctxt) {
 #endif /* LIBXML_SAX1_ENABLED */
 
     /*
-     * Capture end position and add node
+     * Capture end position
      */
-    if ( ret != NULL && ctxt->record_info ) {
-       node_info.end_pos = ctxt->input->consumed +
-                          (CUR_PTR - ctxt->input->base);
-       node_info.end_line = ctxt->input->line;
-       node_info.node = ret;
-       xmlParserAddNodeInfo(ctxt, &node_info);
+    if (cur != NULL && ctxt->record_info) {
+        xmlParserNodeInfoPtr node_info;
+
+        node_info = (xmlParserNodeInfoPtr) xmlParserFindNodeInfo(ctxt, cur);
+        if (node_info != NULL) {
+            node_info->end_pos = ctxt->input->consumed +
+                                 (CUR_PTR - ctxt->input->base);
+            node_info->end_line = ctxt->input->line;
+        }
     }
 }