From 0b2791b56f658b6ec6ca7377dbcd2b72e6a78193 Mon Sep 17 00:00:00 2001
From: wolfbeast <mcwerewolf@wolfbeast.com>
Date: Wed, 23 Oct 2019 10:32:13 +0200
Subject: Issue #1255 - Port upstream fix from libexpat

---
 parser/expat/lib/xmlparse.c | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

(limited to 'parser')

diff --git a/parser/expat/lib/xmlparse.c b/parser/expat/lib/xmlparse.c
index 93a817764..6ab140c89 100644
--- a/parser/expat/lib/xmlparse.c
+++ b/parser/expat/lib/xmlparse.c
@@ -335,7 +335,7 @@ initializeEncoding(XML_Parser parser);
 static enum XML_Error
 doProlog(XML_Parser parser, const ENCODING *enc, const char *s, 
          const char *end, int tok, const char *next, const char **nextPtr, 
-         XML_Bool haveMore);
+         XML_Bool haveMore, XML_Bool allowClosingDoctype);
 static enum XML_Error
 processInternalEntity(XML_Parser parser, ENTITY *entity, 
                       XML_Bool betweenDecl);
@@ -3760,7 +3760,7 @@ externalParEntProcessor(XML_Parser parser,
 
   processor = prologProcessor;
   return doProlog(parser, encoding, s, end, tok, next, 
-                  nextPtr, (XML_Bool)!ps_finalBuffer);
+                  nextPtr, (XML_Bool)!ps_finalBuffer, XML_TRUE);
 }
 
 static enum XML_Error PTRCALL
@@ -3810,7 +3810,7 @@ prologProcessor(XML_Parser parser,
   const char *next = s;
   int tok = XmlPrologTok(encoding, s, end, &next);
   return doProlog(parser, encoding, s, end, tok, next, 
-                  nextPtr, (XML_Bool)!ps_finalBuffer);
+                  nextPtr, (XML_Bool)!ps_finalBuffer, XML_TRUE);
 }
 
 static enum XML_Error
@@ -3821,7 +3821,8 @@ doProlog(XML_Parser parser,
          int tok,
          const char *next,
          const char **nextPtr,
-         XML_Bool haveMore)
+         XML_Bool haveMore,
+         XML_Bool allowClosingDoctype)
 {
 #ifdef XML_DTD
   static const XML_Char externalSubsetName[] = { '#' , '\0' };
@@ -3987,6 +3988,11 @@ doProlog(XML_Parser parser,
       }
       break;
     case XML_ROLE_DOCTYPE_CLOSE:
+      if (allowClosingDoctype != XML_TRUE) {
+        /* Must not close doctype from within expanded parameter entities */
+        return XML_ERROR_INVALID_TOKEN;
+      }
+
       if (doctypeName) {
         startDoctypeDeclHandler(handlerArg, doctypeName,
                                 doctypeSysid, doctypePubid, 0);
@@ -4892,7 +4898,7 @@ processInternalEntity(XML_Parser parser, ENTITY *entity,
   if (entity->is_param) {
     int tok = XmlPrologTok(internalEncoding, textStart, textEnd, &next);
     result = doProlog(parser, internalEncoding, textStart, textEnd, tok, 
-                      next, &next, XML_FALSE);
+                      next, &next, XML_FALSE, XML_FALSE);
   }
   else 
 #endif /* XML_DTD */
@@ -4959,7 +4965,7 @@ internalEntityProcessor(XML_Parser parser,
   if (entity->is_param) {
     int tok = XmlPrologTok(internalEncoding, textStart, textEnd, &next);
     result = doProlog(parser, internalEncoding, textStart, textEnd, tok, 
-                      next, &next, XML_FALSE);
+                      next, &next, XML_FALSE, XML_TRUE);
   }
   else
 #endif /* XML_DTD */
@@ -4986,7 +4992,7 @@ internalEntityProcessor(XML_Parser parser,
     processor = prologProcessor;
     tok = XmlPrologTok(encoding, s, end, &next);
     return doProlog(parser, encoding, s, end, tok, next, nextPtr, 
-                    (XML_Bool)!ps_finalBuffer);
+                    (XML_Bool)!ps_finalBuffer, XML_TRUE);
   }
   else
 #endif /* XML_DTD */
-- 
cgit v1.2.3


From 250875046a10ea5ca64d787a7870b1a3aba7aa3d Mon Sep 17 00:00:00 2001
From: Henri Sivonen <hsivonen@hsivonen.fi>
Date: Thu, 24 Oct 2019 11:43:25 +0200
Subject: Adjust tokenization of U+0000

---
 parser/html/nsHtml5Tokenizer.cpp | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

(limited to 'parser')

diff --git a/parser/html/nsHtml5Tokenizer.cpp b/parser/html/nsHtml5Tokenizer.cpp
index 4c815b0c0..5464d211d 100644
--- a/parser/html/nsHtml5Tokenizer.cpp
+++ b/parser/html/nsHtml5Tokenizer.cpp
@@ -1054,9 +1054,6 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
           }
           c = checkChar(buf, pos);
           switch(c) {
-            case '\0': {
-              NS_HTML5_BREAK(stateloop);
-            }
             case '-': {
               clearStrBufAfterOneHyphen();
               state = P::transition(mViewSource, NS_HTML5TOKENIZER_COMMENT_START, reconsume, pos);
@@ -1461,9 +1458,6 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
           NS_HTML5_BREAK(stateloop);
         }
         c = checkChar(buf, pos);
-        if (c == '\0') {
-          NS_HTML5_BREAK(stateloop);
-        }
         switch(c) {
           case ' ':
           case '\t':
@@ -1471,7 +1465,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
           case '\r':
           case '\f':
           case '<':
-          case '&': {
+          case '&':
+          case '\0': {
             emitOrAppendCharRefBuf(returnState);
             if (!(returnState & NS_HTML5TOKENIZER_DATA_AND_RCDATA_MASK)) {
               cstart = pos;
@@ -1519,9 +1514,6 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
             NS_HTML5_BREAK(stateloop);
           }
           c = checkChar(buf, pos);
-          if (c == '\0') {
-            NS_HTML5_BREAK(stateloop);
-          }
           int32_t hilo = 0;
           if (c <= 'z') {
             const int32_t* row = nsHtml5NamedCharactersAccel::HILO_ACCEL[c];
@@ -1556,9 +1548,6 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
             NS_HTML5_BREAK(stateloop);
           }
           c = checkChar(buf, pos);
-          if (c == '\0') {
-            NS_HTML5_BREAK(stateloop);
-          }
           entCol++;
           for (; ; ) {
             if (hi < lo) {
-- 
cgit v1.2.3