Fix CVE-2016-0718: Expat XML Parser Crashes on Malformed Input

Includes all changes made on the cve-2016-0718-fix-2-2-1 branch
that was merged into the master branch in commit
be4b1c06daba1849b8ff5e00bae5caf47f6c39fd
in repository http://git.code.sf.net/p/expat/code_git

(cherry picked from commit 7b64b72da916c9c9c285939c11f81de938268fa9)

Bug: 28698301
Change-Id: I9fe93040b3a9afb04be63c5fd453a2b99c01fb37
(cherry picked from commit 4bf8f12b1d505770ce1fd3b5aa33bd7abe46ee74)
diff --git a/lib/xmlparse.c b/lib/xmlparse.c
index e308c79..18bfb7e 100644
--- a/lib/xmlparse.c
+++ b/lib/xmlparse.c
@@ -2426,11 +2426,11 @@
           for (;;) {
             int bufSize;
             int convLen;
-            XmlConvert(enc,
+            const enum XML_Convert_Result convert_res = XmlConvert(enc,
                        &fromPtr, rawNameEnd,
                        (ICHAR **)&toPtr, (ICHAR *)tag->bufEnd - 1);
             convLen = (int)(toPtr - (XML_Char *)tag->buf);
-            if (fromPtr == rawNameEnd) {
+            if ((convert_res == XML_CONVERT_COMPLETED) || (convert_res == XML_CONVERT_INPUT_INCOMPLETE)) {
               tag->name.strLen = convLen;
               break;
             }
@@ -2651,11 +2651,11 @@
           if (MUST_CONVERT(enc, s)) {
             for (;;) {
               ICHAR *dataPtr = (ICHAR *)dataBuf;
-              XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd);
+              const enum XML_Convert_Result convert_res = XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd);
               *eventEndPP = s;
               charDataHandler(handlerArg, dataBuf,
                               (int)(dataPtr - (ICHAR *)dataBuf));
-              if (s == next)
+              if ((convert_res == XML_CONVERT_COMPLETED) || (convert_res == XML_CONVERT_INPUT_INCOMPLETE))
                 break;
               *eventPP = s;
             }
@@ -3261,11 +3261,11 @@
           if (MUST_CONVERT(enc, s)) {
             for (;;) {
               ICHAR *dataPtr = (ICHAR *)dataBuf;
-              XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd);
+              const enum XML_Convert_Result convert_res = XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd);
               *eventEndPP = next;
               charDataHandler(handlerArg, dataBuf,
                               (int)(dataPtr - (ICHAR *)dataBuf));
-              if (s == next)
+              if ((convert_res == XML_CONVERT_COMPLETED) || (convert_res == XML_CONVERT_INPUT_INCOMPLETE))
                 break;
               *eventPP = s;
             }
@@ -5342,6 +5342,7 @@
               const char *s, const char *end)
 {
   if (MUST_CONVERT(enc, s)) {
+    enum XML_Convert_Result convert_res;
     const char **eventPP;
     const char **eventEndPP;
     if (enc == encoding) {
@@ -5354,11 +5355,11 @@
     }
     do {
       ICHAR *dataPtr = (ICHAR *)dataBuf;
-      XmlConvert(enc, &s, end, &dataPtr, (ICHAR *)dataBufEnd);
+      convert_res = XmlConvert(enc, &s, end, &dataPtr, (ICHAR *)dataBufEnd);
       *eventEndPP = s;
       defaultHandler(handlerArg, dataBuf, (int)(dataPtr - (ICHAR *)dataBuf));
       *eventPP = s;
-    } while (s != end);
+    } while ((convert_res != XML_CONVERT_COMPLETED) && (convert_res != XML_CONVERT_INPUT_INCOMPLETE));
   }
   else
     defaultHandler(handlerArg, (XML_Char *)s, (int)((XML_Char *)end - (XML_Char *)s));
@@ -6163,8 +6164,8 @@
   if (!pool->ptr && !poolGrow(pool))
     return NULL;
   for (;;) {
-    XmlConvert(enc, &ptr, end, (ICHAR **)&(pool->ptr), (ICHAR *)pool->end);
-    if (ptr == end)
+    const enum XML_Convert_Result convert_res = XmlConvert(enc, &ptr, end, (ICHAR **)&(pool->ptr), (ICHAR *)pool->end);
+    if ((convert_res == XML_CONVERT_COMPLETED) || (convert_res == XML_CONVERT_INPUT_INCOMPLETE))
       break;
     if (!poolGrow(pool))
       return NULL;
@@ -6248,8 +6249,13 @@
     }
   }
   if (pool->blocks && pool->start == pool->blocks->s) {
+    BLOCK *temp;
     int blockSize = (int)(pool->end - pool->start)*2;
-    BLOCK *temp = (BLOCK *)
+
+    if (blockSize < 0)
+      return XML_FALSE;
+
+    temp = (BLOCK *)
       pool->mem->realloc_fcn(pool->blocks,
                              (offsetof(BLOCK, s)
                               + blockSize * sizeof(XML_Char)));
@@ -6264,6 +6270,10 @@
   else {
     BLOCK *tem;
     int blockSize = (int)(pool->end - pool->start);
+
+    if (blockSize < 0)
+      return XML_FALSE;
+
     if (blockSize < INIT_BLOCK_SIZE)
       blockSize = INIT_BLOCK_SIZE;
     else
diff --git a/lib/xmltok.c b/lib/xmltok.c
index bf09dfc..cb98ce1 100644
--- a/lib/xmltok.c
+++ b/lib/xmltok.c
@@ -318,39 +318,55 @@
   UTF8_cval4 = 0xf0
 };
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 utf8_toUtf8(const ENCODING *enc,
             const char **fromP, const char *fromLim,
             char **toP, const char *toLim)
 {
+  enum XML_Convert_Result res = XML_CONVERT_COMPLETED;
   char *to;
   const char *from;
   if (fromLim - *fromP > toLim - *toP) {
     /* Avoid copying partial characters. */
+    res = XML_CONVERT_OUTPUT_EXHAUSTED;
     for (fromLim = *fromP + (toLim - *toP); fromLim > *fromP; fromLim--)
       if (((unsigned char)fromLim[-1] & 0xc0) != 0x80)
         break;
   }
-  for (to = *toP, from = *fromP; from != fromLim; from++, to++)
+  for (to = *toP, from = *fromP; (from < fromLim) && (to < toLim); from++, to++)
     *to = *from;
   *fromP = from;
   *toP = to;
+
+  if ((to == toLim) && (from < fromLim))
+    return XML_CONVERT_OUTPUT_EXHAUSTED;
+  else
+    return res;
 }
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 utf8_toUtf16(const ENCODING *enc,
              const char **fromP, const char *fromLim,
              unsigned short **toP, const unsigned short *toLim)
 {
+  enum XML_Convert_Result res = XML_CONVERT_COMPLETED;
   unsigned short *to = *toP;
   const char *from = *fromP;
-  while (from != fromLim && to != toLim) {
+  while (from < fromLim && to < toLim) {
     switch (((struct normal_encoding *)enc)->type[(unsigned char)*from]) {
     case BT_LEAD2:
+      if (fromLim - from < 2) {
+        res = XML_CONVERT_INPUT_INCOMPLETE;
+        break;
+      }
       *to++ = (unsigned short)(((from[0] & 0x1f) << 6) | (from[1] & 0x3f));
       from += 2;
       break;
     case BT_LEAD3:
+      if (fromLim - from < 3) {
+        res = XML_CONVERT_INPUT_INCOMPLETE;
+        break;
+      }
       *to++ = (unsigned short)(((from[0] & 0xf) << 12)
                                | ((from[1] & 0x3f) << 6) | (from[2] & 0x3f));
       from += 3;
@@ -358,8 +374,14 @@
     case BT_LEAD4:
       {
         unsigned long n;
-        if (to + 1 == toLim)
+        if (toLim - to < 2) {
+          res = XML_CONVERT_OUTPUT_EXHAUSTED;
           goto after;
+        }
+        if (fromLim - from < 4) {
+          res = XML_CONVERT_INPUT_INCOMPLETE;
+          goto after;
+        }
         n = ((from[0] & 0x7) << 18) | ((from[1] & 0x3f) << 12)
             | ((from[2] & 0x3f) << 6) | (from[3] & 0x3f);
         n -= 0x10000;
@@ -377,6 +399,7 @@
 after:
   *fromP = from;
   *toP = to;
+  return res;
 }
 
 #ifdef XML_NS
@@ -425,7 +448,7 @@
   STANDARD_VTABLE(sb_) NORMAL_VTABLE(utf8_)
 };
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 latin1_toUtf8(const ENCODING *enc,
               const char **fromP, const char *fromLim,
               char **toP, const char *toLim)
@@ -433,30 +456,35 @@
   for (;;) {
     unsigned char c;
     if (*fromP == fromLim)
-      break;
+      return XML_CONVERT_COMPLETED;
     c = (unsigned char)**fromP;
     if (c & 0x80) {
       if (toLim - *toP < 2)
-        break;
+        return XML_CONVERT_OUTPUT_EXHAUSTED;
       *(*toP)++ = (char)((c >> 6) | UTF8_cval2);
       *(*toP)++ = (char)((c & 0x3f) | 0x80);
       (*fromP)++;
     }
     else {
       if (*toP == toLim)
-        break;
+        return XML_CONVERT_OUTPUT_EXHAUSTED;
       *(*toP)++ = *(*fromP)++;
     }
   }
 }
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 latin1_toUtf16(const ENCODING *enc,
                const char **fromP, const char *fromLim,
                unsigned short **toP, const unsigned short *toLim)
 {
-  while (*fromP != fromLim && *toP != toLim)
+  while (*fromP < fromLim && *toP < toLim)
     *(*toP)++ = (unsigned char)*(*fromP)++;
+
+  if ((*toP == toLim) && (*fromP < fromLim))
+    return XML_CONVERT_OUTPUT_EXHAUSTED;
+  else
+    return XML_CONVERT_COMPLETED;
 }
 
 #ifdef XML_NS
@@ -483,13 +511,18 @@
   STANDARD_VTABLE(sb_)
 };
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 ascii_toUtf8(const ENCODING *enc,
              const char **fromP, const char *fromLim,
              char **toP, const char *toLim)
 {
-  while (*fromP != fromLim && *toP != toLim)
+  while (*fromP < fromLim && *toP < toLim)
     *(*toP)++ = *(*fromP)++;
+
+  if ((*toP == toLim) && (*fromP < fromLim))
+    return XML_CONVERT_OUTPUT_EXHAUSTED;
+  else
+    return XML_CONVERT_COMPLETED;
 }
 
 #ifdef XML_NS
@@ -536,13 +569,14 @@
 }
 
 #define DEFINE_UTF16_TO_UTF8(E) \
-static void  PTRCALL \
+static enum XML_Convert_Result  PTRCALL \
 E ## toUtf8(const ENCODING *enc, \
             const char **fromP, const char *fromLim, \
             char **toP, const char *toLim) \
 { \
-  const char *from; \
-  for (from = *fromP; from != fromLim; from += 2) { \
+  const char *from = *fromP; \
+  fromLim = from + (((fromLim - from) >> 1) << 1);  /* shrink to even */ \
+  for (; from < fromLim; from += 2) { \
     int plane; \
     unsigned char lo2; \
     unsigned char lo = GET_LO(from); \
@@ -552,7 +586,7 @@
       if (lo < 0x80) { \
         if (*toP == toLim) { \
           *fromP = from; \
-          return; \
+          return XML_CONVERT_OUTPUT_EXHAUSTED; \
         } \
         *(*toP)++ = lo; \
         break; \
@@ -562,7 +596,7 @@
     case 0x4: case 0x5: case 0x6: case 0x7: \
       if (toLim -  *toP < 2) { \
         *fromP = from; \
-        return; \
+        return XML_CONVERT_OUTPUT_EXHAUSTED; \
       } \
       *(*toP)++ = ((lo >> 6) | (hi << 2) |  UTF8_cval2); \
       *(*toP)++ = ((lo & 0x3f) | 0x80); \
@@ -570,7 +604,7 @@
     default: \
       if (toLim -  *toP < 3)  { \
         *fromP = from; \
-        return; \
+        return XML_CONVERT_OUTPUT_EXHAUSTED; \
       } \
       /* 16 bits divided 4, 6, 6 amongst 3 bytes */ \
       *(*toP)++ = ((hi >> 4) | UTF8_cval3); \
@@ -580,7 +614,11 @@
     case 0xD8: case 0xD9: case 0xDA: case 0xDB: \
       if (toLim -  *toP < 4) { \
         *fromP = from; \
-        return; \
+        return XML_CONVERT_OUTPUT_EXHAUSTED; \
+      } \
+      if (fromLim - from < 4) { \
+        *fromP = from; \
+        return XML_CONVERT_INPUT_INCOMPLETE; \
       } \
       plane = (((hi & 0x3) << 2) | ((lo >> 6) & 0x3)) + 1; \
       *(*toP)++ = ((plane >> 2) | UTF8_cval4); \
@@ -596,20 +634,32 @@
     } \
   } \
   *fromP = from; \
+  if (from < fromLim) \
+    return XML_CONVERT_INPUT_INCOMPLETE; \
+  else \
+    return XML_CONVERT_COMPLETED; \
 }
 
 #define DEFINE_UTF16_TO_UTF16(E) \
-static void  PTRCALL \
+static enum XML_Convert_Result  PTRCALL \
 E ## toUtf16(const ENCODING *enc, \
              const char **fromP, const char *fromLim, \
              unsigned short **toP, const unsigned short *toLim) \
 { \
+  enum XML_Convert_Result res = XML_CONVERT_COMPLETED; \
+  fromLim = *fromP + (((fromLim - *fromP) >> 1) << 1);  /* shrink to even */ \
   /* Avoid copying first half only of surrogate */ \
   if (fromLim - *fromP > ((toLim - *toP) << 1) \
-      && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) \
+      && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) { \
     fromLim -= 2; \
-  for (; *fromP != fromLim && *toP != toLim; *fromP += 2) \
+    res = XML_CONVERT_INPUT_INCOMPLETE; \
+  } \
+  for (; *fromP < fromLim && *toP < toLim; *fromP += 2) \
     *(*toP)++ = (GET_HI(*fromP) << 8) | GET_LO(*fromP); \
+  if ((*toP == toLim) && (*fromP < fromLim)) \
+    return XML_CONVERT_OUTPUT_EXHAUSTED; \
+  else \
+    return res; \
 }
 
 #define SET2(ptr, ch) \
@@ -1288,7 +1338,7 @@
   return (c & ~0xFFFF) || checkCharRefNumber(c) < 0;
 }
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 unknown_toUtf8(const ENCODING *enc,
                const char **fromP, const char *fromLim,
                char **toP, const char *toLim)
@@ -1299,21 +1349,21 @@
     const char *utf8;
     int n;
     if (*fromP == fromLim)
-      break;
+      return XML_CONVERT_COMPLETED;
     utf8 = uenc->utf8[(unsigned char)**fromP];
     n = *utf8++;
     if (n == 0) {
       int c = uenc->convert(uenc->userData, *fromP);
       n = XmlUtf8Encode(c, buf);
       if (n > toLim - *toP)
-        break;
+        return XML_CONVERT_OUTPUT_EXHAUSTED;
       utf8 = buf;
       *fromP += (AS_NORMAL_ENCODING(enc)->type[(unsigned char)**fromP]
                  - (BT_LEAD2 - 2));
     }
     else {
       if (n > toLim - *toP)
-        break;
+        return XML_CONVERT_OUTPUT_EXHAUSTED;
       (*fromP)++;
     }
     do {
@@ -1322,13 +1372,13 @@
   }
 }
 
-static void PTRCALL
+static enum XML_Convert_Result PTRCALL
 unknown_toUtf16(const ENCODING *enc,
                 const char **fromP, const char *fromLim,
                 unsigned short **toP, const unsigned short *toLim)
 {
   const struct unknown_encoding *uenc = AS_UNKNOWN_ENCODING(enc);
-  while (*fromP != fromLim && *toP != toLim) {
+  while (*fromP < fromLim && *toP < toLim) {
     unsigned short c = uenc->utf16[(unsigned char)**fromP];
     if (c == 0) {
       c = (unsigned short)
@@ -1340,6 +1390,11 @@
       (*fromP)++;
     *(*toP)++ = c;
   }
+
+  if ((*toP == toLim) && (*fromP < fromLim))
+    return XML_CONVERT_OUTPUT_EXHAUSTED;
+  else
+    return XML_CONVERT_COMPLETED;
 }
 
 ENCODING *
@@ -1503,7 +1558,7 @@
 {
   const ENCODING **encPtr;
 
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   encPtr = enc->encPtr;
   if (ptr + 1 == end) {
diff --git a/lib/xmltok.h b/lib/xmltok.h
index ca867aa..752007e 100644
--- a/lib/xmltok.h
+++ b/lib/xmltok.h
@@ -130,6 +130,12 @@
                                const char *,
                                const char **);
 
+enum XML_Convert_Result {
+  XML_CONVERT_COMPLETED = 0,
+  XML_CONVERT_INPUT_INCOMPLETE = 1,
+  XML_CONVERT_OUTPUT_EXHAUSTED = 2  /* and therefore potentially input remaining as well */
+};
+
 struct encoding {
   SCANNER scanners[XML_N_STATES];
   SCANNER literalScanners[XML_N_LITERAL_TYPES];
@@ -158,12 +164,12 @@
                             const char *ptr,
                             const char *end,
                             const char **badPtr);
-  void (PTRCALL *utf8Convert)(const ENCODING *enc,
+  enum XML_Convert_Result (PTRCALL *utf8Convert)(const ENCODING *enc,
                               const char **fromP,
                               const char *fromLim,
                               char **toP,
                               const char *toLim);
-  void (PTRCALL *utf16Convert)(const ENCODING *enc,
+  enum XML_Convert_Result (PTRCALL *utf16Convert)(const ENCODING *enc,
                                const char **fromP,
                                const char *fromLim,
                                unsigned short **toP,
diff --git a/lib/xmltok_impl.c b/lib/xmltok_impl.c
index 9c2895b..6c5a3ba 100644
--- a/lib/xmltok_impl.c
+++ b/lib/xmltok_impl.c
@@ -93,13 +93,13 @@
 PREFIX(scanComment)(const ENCODING *enc, const char *ptr,
                     const char *end, const char **nextTokPtr)
 {
-  if (ptr != end) {
+  if (ptr < end) {
     if (!CHAR_MATCHES(enc, ptr, ASCII_MINUS)) {
       *nextTokPtr = ptr;
       return XML_TOK_INVALID;
     }
     ptr += MINBPC(enc);
-    while (ptr != end) {
+    while (ptr < end) {
       switch (BYTE_TYPE(enc, ptr)) {
       INVALID_CASES(ptr, nextTokPtr)
       case BT_MINUS:
@@ -147,7 +147,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     case BT_PERCNT:
       if (ptr + MINBPC(enc) == end)
@@ -233,7 +233,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_S: case BT_CR: case BT_LF:
@@ -242,7 +242,7 @@
         return XML_TOK_INVALID;
       }
       ptr += MINBPC(enc);
-      while (ptr != end) {
+      while (ptr < end) {
         switch (BYTE_TYPE(enc, ptr)) {
         INVALID_CASES(ptr, nextTokPtr)
         case BT_QUEST:
@@ -305,7 +305,7 @@
 PREFIX(cdataSectionTok)(const ENCODING *enc, const char *ptr,
                         const char *end, const char **nextTokPtr)
 {
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   if (MINBPC(enc) > 1) {
     size_t n = end - ptr;
@@ -348,7 +348,7 @@
     ptr += MINBPC(enc);
     break;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
 #define LEAD_CASE(n) \
     case BT_LEAD ## n: \
@@ -391,11 +391,11 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_S: case BT_CR: case BT_LF:
-      for (ptr += MINBPC(enc); ptr != end; ptr += MINBPC(enc)) {
+      for (ptr += MINBPC(enc); ptr < end; ptr += MINBPC(enc)) {
         switch (BYTE_TYPE(enc, ptr)) {
         case BT_S: case BT_CR: case BT_LF:
           break;
@@ -432,7 +432,7 @@
 PREFIX(scanHexCharRef)(const ENCODING *enc, const char *ptr,
                        const char *end, const char **nextTokPtr)
 {
-  if (ptr != end) {
+  if (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     case BT_DIGIT:
     case BT_HEX:
@@ -441,7 +441,7 @@
       *nextTokPtr = ptr;
       return XML_TOK_INVALID;
     }
-    for (ptr += MINBPC(enc); ptr != end; ptr += MINBPC(enc)) {
+    for (ptr += MINBPC(enc); ptr < end; ptr += MINBPC(enc)) {
       switch (BYTE_TYPE(enc, ptr)) {
       case BT_DIGIT:
       case BT_HEX:
@@ -464,7 +464,7 @@
 PREFIX(scanCharRef)(const ENCODING *enc, const char *ptr,
                     const char *end, const char **nextTokPtr)
 {
-  if (ptr != end) {
+  if (ptr < end) {
     if (CHAR_MATCHES(enc, ptr, ASCII_x))
       return PREFIX(scanHexCharRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
     switch (BYTE_TYPE(enc, ptr)) {
@@ -474,7 +474,7 @@
       *nextTokPtr = ptr;
       return XML_TOK_INVALID;
     }
-    for (ptr += MINBPC(enc); ptr != end; ptr += MINBPC(enc)) {
+    for (ptr += MINBPC(enc); ptr < end; ptr += MINBPC(enc)) {
       switch (BYTE_TYPE(enc, ptr)) {
       case BT_DIGIT:
         break;
@@ -506,7 +506,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_SEMI:
@@ -529,7 +529,7 @@
 #ifdef XML_NS
   int hadColon = 0;
 #endif
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
 #ifdef XML_NS
@@ -716,7 +716,7 @@
   hadColon = 0;
 #endif
   /* we have a start-tag */
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
 #ifdef XML_NS
@@ -740,7 +740,7 @@
     case BT_S: case BT_CR: case BT_LF:
       {
         ptr += MINBPC(enc);
-        while (ptr != end) {
+        while (ptr < end) {
           switch (BYTE_TYPE(enc, ptr)) {
           CHECK_NMSTRT_CASES(enc, ptr, end, nextTokPtr)
           case BT_GT:
@@ -785,7 +785,7 @@
 PREFIX(contentTok)(const ENCODING *enc, const char *ptr, const char *end,
                    const char **nextTokPtr)
 {
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   if (MINBPC(enc) > 1) {
     size_t n = end - ptr;
@@ -832,7 +832,7 @@
     ptr += MINBPC(enc);
     break;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
 #define LEAD_CASE(n) \
     case BT_LEAD ## n: \
@@ -895,7 +895,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_SEMI:
@@ -921,7 +921,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_CR: case BT_LF: case BT_S:
@@ -941,7 +941,7 @@
                 const char *ptr, const char *end,
                 const char **nextTokPtr)
 {
-  while (ptr != end) {
+  while (ptr < end) {
     int t = BYTE_TYPE(enc, ptr);
     switch (t) {
     INVALID_CASES(ptr, nextTokPtr)
@@ -973,7 +973,7 @@
                   const char **nextTokPtr)
 {
   int tok;
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   if (MINBPC(enc) > 1) {
     size_t n = end - ptr;
@@ -1141,7 +1141,7 @@
     *nextTokPtr = ptr;
     return XML_TOK_INVALID;
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     CHECK_NAME_CASES(enc, ptr, end, nextTokPtr)
     case BT_GT: case BT_RPAR: case BT_COMMA:
@@ -1204,10 +1204,10 @@
                           const char *end, const char **nextTokPtr)
 {
   const char *start;
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   start = ptr;
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
 #define LEAD_CASE(n) \
     case BT_LEAD ## n: ptr += n; break;
@@ -1262,10 +1262,10 @@
                        const char *end, const char **nextTokPtr)
 {
   const char *start;
-  if (ptr == end)
+  if (ptr >= end)
     return XML_TOK_NONE;
   start = ptr;
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
 #define LEAD_CASE(n) \
     case BT_LEAD ## n: ptr += n; break;
@@ -1326,7 +1326,7 @@
       end = ptr + n;
     }
   }
-  while (ptr != end) {
+  while (ptr < end) {
     switch (BYTE_TYPE(enc, ptr)) {
     INVALID_CASES(ptr, nextTokPtr)
     case BT_LT:
@@ -1373,7 +1373,7 @@
 {
   ptr += MINBPC(enc);
   end -= MINBPC(enc);
-  for (; ptr != end; ptr += MINBPC(enc)) {
+  for (; ptr < end; ptr += MINBPC(enc)) {
     switch (BYTE_TYPE(enc, ptr)) {
     case BT_DIGIT:
     case BT_HEX:
@@ -1760,7 +1760,7 @@
     case BT_CR:
       pos->lineNumber++;
       ptr += MINBPC(enc);
-      if (ptr != end && BYTE_TYPE(enc, ptr) == BT_LF)
+      if (ptr < end && BYTE_TYPE(enc, ptr) == BT_LF)
         ptr += MINBPC(enc);
       pos->columnNumber = (XML_Size)-1;
       break;