Imported Upstream version 2.9.5_rc2

[platform/upstream/libxml2.git] / parserInternals.c

diff --git a/parserInternals.c b/parserInternals.c
index 81b0e0f..43a0f5a 100644 (file)
--- a/parserInternals.c
+++ b/parserInternals.c
@@ -701,6 +701,16 @@ xmlCurrentChar(xmlParserCtxtPtr ctxt, int *len) {
     return((int) *ctxt->input->cur);
 encoding_error:
     /*
+     * An encoding problem may arise from a truncated input buffer
+     * splitting a character in the middle. In that case do not raise
+     * an error but return 0 to endicate an end of stream problem
+     */
+    if (ctxt->input->end - ctxt->input->cur < 4) {
+       *len = 0;
+       return(0);
+    }
+
+    /*
      * If we detect an UTF8 error that probably mean that the
      * input encoding didn't get properly advertised in the
      * declaration header. Report the error and switch the encoding
@@ -710,21 +720,9 @@ encoding_error:
     {
         char buffer[150];
 
-        if (ctxt->input->cur[1] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X EOF\n",
-                     ctxt->input->cur[0]);
-        } else if (ctxt->input->cur[2] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X EOF\n",
-                     ctxt->input->cur[0], ctxt->input->cur[1]);
-        } else if (ctxt->input->cur[3] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X 0x%02X EOF\n",
-                     ctxt->input->cur[0], ctxt->input->cur[1],
-                     ctxt->input->cur[2]);
-        } else {
-           snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
-                    ctxt->input->cur[0], ctxt->input->cur[1],
-                    ctxt->input->cur[2], ctxt->input->cur[3]);
-        }
+       snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
+                       ctxt->input->cur[0], ctxt->input->cur[1],
+                       ctxt->input->cur[2], ctxt->input->cur[3]);
        __xmlErrEncoding(ctxt, XML_ERR_INVALID_CHAR,
                     "Input is not proper UTF-8, indicate encoding !\n%s",
                     BAD_CAST buffer, NULL);
@@ -814,6 +812,17 @@ xmlStringCurrentChar(xmlParserCtxtPtr ctxt, const xmlChar * cur, int *len)
     *len = 1;
     return ((int) *cur);
 encoding_error:
+
+    /*
+     * An encoding problem may arise from a truncated input buffer
+     * splitting a character in the middle. In that case do not raise
+     * an error but return 0 to endicate an end of stream problem
+     */
+    if ((ctxt == NULL) || (ctxt->input == NULL) ||
+        (ctxt->input->end - ctxt->input->cur < 4)) {
+       *len = 0;
+       return(0);
+    }
     /*
      * If we detect an UTF8 error that probably mean that the
      * input encoding didn't get properly advertised in the
@@ -824,19 +833,9 @@ encoding_error:
     {
         char buffer[150];
 
-        if (cur[1] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X EOF\n",
-                     cur[0]);
-        } else if (cur[2] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X EOF\n",
-                     cur[0], cur[1]);
-        } else if (cur[3] == 0) {
-            snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X 0x%02X EOF\n",
-                     cur[0], cur[1], cur[2]);
-        } else {
-           snprintf(&buffer[0], 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
-                    cur[0], cur[1], cur[2], cur[3]);
-        }
+       snprintf(buffer, 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
+                       ctxt->input->cur[0], ctxt->input->cur[1],
+                       ctxt->input->cur[2], ctxt->input->cur[3]);
        __xmlErrEncoding(ctxt, XML_ERR_INVALID_CHAR,
                     "Input is not proper UTF-8, indicate encoding !\n%s",
                     BAD_CAST buffer, NULL);