X-Git-Url: http://git.asbjorn.biz/?a=blobdiff_plain;f=lib%2Fas3%2Ftokenizer.lex;h=7e936b65cdb3834459d60686269523dbc741b285;hb=c63b2bf21dc1df9a736f0b4c08f6cba828cdab92;hp=805cf2e68ee516da0e3329719f3fc7d1df1356f6;hpb=62a83d162b254d91da418cee25f5b87b067a3f92;p=swftools.git

diff --git a/lib/as3/tokenizer.lex b/lib/as3/tokenizer.lex
index 805cf2e..7e936b6 100644
--- a/lib/as3/tokenizer.lex
+++ b/lib/as3/tokenizer.lex
@@ -42,7 +42,7 @@ static void dbg(const char*format, ...)
     if(as3_verbosity<3)
 	return;
     va_start(arglist, format);
-    vsprintf(buf, format, arglist);
+    vsnprintf(buf, sizeof(buf)-1, format, arglist);
     va_end(arglist);
     l = strlen(buf);
     while(l && buf[l-1]=='\n') {
@@ -76,6 +76,9 @@ void as3_buffer_input(void*buffer, int len)
     as3_in = 0;
 }
 
+//#undef BEGIN
+//#define BEGIN(x) {(yy_start) = 1 + 2 *x;dbg("entering state %d", x);}
+
 #define YY_INPUT(buf,result,max_size) { \
   if(!as3_buffer) { \
       errno=0; \
@@ -107,10 +110,10 @@ void handleInclude(char*text, int len, char quotes)
     } else {
         int i1=0,i2=len;
         // find start
-        while(!strchr(" \n\r\t", text[i1])) i1++;
+        while(!strchr(" \n\r\t\xa0", text[i1])) i1++;
         // strip
-        while(strchr(" \n\r\t", text[i1])) i1++;
-        while(strchr(" \n\r\t", text[i2-1])) i2--;
+        while(strchr(" \n\r\t\xa0", text[i1])) i1++;
+        while(strchr(" \n\r\t\xa0", text[i2-1])) i2--;
         if(i2!=len) text[i2]=0;
         filename = strdup(&text[i1]);
     }
@@ -123,7 +126,7 @@ void handleInclude(char*text, int len, char quotes)
     }
 
     yy_switch_to_buffer(yy_create_buffer( yyin, YY_BUF_SIZE ) );
-    //BEGIN(INITIAL); keep context
+    //BEGIN(DEFAULT); keep context
 }
 
 static int do_unescape(const char*s, const char*end, char*n) 
@@ -252,6 +255,12 @@ static void handleCData(char*s, int len)
     a3_lval.str.str = strdup_n(a3_lval.str.str, a3_lval.str.len);
 }
 
+static void handleRaw(char*s, int len)
+{
+    a3_lval.str.len = len;
+    a3_lval.str.str = strdup_n(s, a3_lval.str.len);
+}
+
 static void handleString(char*s, int len)
 {
     if(s[0]=='"') {
@@ -276,7 +285,6 @@ static inline int m(int type)
     return type;
 }
 
-
 static char numberbuf[64];
 static char*nrbuf()
 {
@@ -462,45 +470,47 @@ static inline void c() {
     current_column+=yyleng;
 }
 
-trie_t*active_namespaces = 0;
-/*void tokenizer_register_namespace(const char*id)
-{
-    trie_put(namespaces, id, 0);
-}
-void tokenizer_unregister_namespace(const char*id)
-{
-    trie_remove(namespaces, id);
-}*/
-static inline tokenizer_is_namespace(const char*id)
-{
-    return trie_contains(active_namespaces, id);
-}
-
 static inline int handleIdentifier()
 {
     char*s = malloc(yyleng+1);
     memcpy(s, yytext, yyleng);
     s[yyleng]=0;
     a3_lval.id = s;
-    if(tokenizer_is_namespace(s)) 
-        return T_NAMESPACE;
-    else
-        return T_IDENTIFIER;
+    return T_IDENTIFIER;
 }
+static int tokenerror();
 
 
 //Boolean                      {c();return m(KW_BOOLEAN);}
 //int                          {c();return m(KW_INT);}
 //uint                         {c();return m(KW_UINT);}
 //Number                       {c();return m(KW_NUMBER);}
+//XMLCOMMENT  <!--([^->]|(-/[^-])|(--/[^>]))*-->
+
+//{XMLCOMMENT}                 
 
 %}
 
 %s REGEXPOK
 %s BEGINNING
+%s DEFAULT
+%x XMLTEXT
+%x XML
+
+X1 parsing identifiers with a non unicode lexer is a knightmare we have to skip all possible
+X2 combinations of byte order markers or utf8 space chars and i dont quite like the fact that
+X3 lex doesnt support proper comments in this section either...
+X4 {NAME_HEAD}{NAME_TAIL} 
 
-NAME	 [a-zA-Z_][a-zA-Z0-9_\\]*
-_        [^a-zA-Z0-9_\\]
+NAME_NOC2EF  [a-zA-Z_\x80-\xc1\xc3-\xee\xf0-\xff]
+NAME_EF      [\xef][a-zA-Z0-9_\\\x80-\xba\xbc-\xff]
+NAME_C2      [\xc2][a-zA-Z0-9_\\\x80-\x9f\xa1-\xff]
+NAME_EFBB    [\xef][\xbb][a-zA-Z0-9_\\\x80-\xbe\xc0-\xff]
+NAME_TAIL    [a-zA-Z_0-9\\\x80-\xff]*
+NAME_HEAD    (({NAME_NOC2EF})|({NAME_EF})|({NAME_C2})|({NAME_EFBB}))
+NAME	     {NAME_HEAD}{NAME_TAIL} 
+
+_            [^a-zA-Z0-9_\\\x80-\xff]
 
 HEXINT    0x[a-zA-Z0-9]+
 HEXFLOAT  0x[a-zA-Z0-9]*\.[a-zA-Z0-9]*
@@ -512,9 +522,14 @@ HEXFLOATWITHSIGN [+-]?({HEXFLOAT})
 INTWITHSIGN [+-]?({INT})
 FLOATWITHSIGN [+-]?({FLOAT})
 
-CDATA    <!\[CDATA\[([^]]|\][^]]|\]\][^>])*\]*\]\]\>
+CDATA       <!\[CDATA\[([^]]|\][^]]|\]\][^>])*\]*\]\]\>
+XMLCOMMENT  <!--([^->]|[-]+[^>-]|>)*-*-->
+XML         <[^>]+{S}>
+XMLID       [A-Za-z0-9_\x80-\xff]+([:][A-Za-z0-9_\x80-\xff]+)?
+XMLSTRING   ["][^"]*["]
+
 STRING   ["](\\[\x00-\xff]|[^\\"\n])*["]|['](\\[\x00-\xff]|[^\\'\n])*[']
-S 	 [ \n\r\t]
+S 	 ([ \n\r\t\xa0]|[\xc2][\xa0])
 MULTILINE_COMMENT [/][*]+([*][^/]|[^/*]|[^*][/]|[\x00-\x1f])*[*]+[/]
 SINGLELINE_COMMENT \/\/[^\n\r]*[\n\r]
 REGEXP   [/]([^/\n]|\\[/])*[/][a-zA-Z]*
@@ -526,41 +541,68 @@ REGEXP   [/]([^/\n]|\\[/])*[/][a-zA-Z]*
 [/][*]                       {syntaxerror("syntax error: unterminated comment", yytext);}
 
 ^include{S}+{STRING}{S}*/\n    {l();handleInclude(yytext, yyleng, 1);}
-^include{S}+[^" \t\r\n][\x20-\xff]*{S}*/\n    {l();handleInclude(yytext, yyleng, 0);}
-{STRING}                     {l(); BEGIN(INITIAL);handleString(yytext, yyleng);return T_STRING;}
-{CDATA}                      {l(); BEGIN(INITIAL);handleCData(yytext, yyleng);return T_STRING;}
+^include{S}+[^" \t\xa0\r\n][\x20-\xff]*{S}*/\n    {l();handleInclude(yytext, yyleng, 0);}
+{STRING}                     {l(); BEGIN(DEFAULT);handleString(yytext, yyleng);return T_STRING;}
+{CDATA}                      {l(); BEGIN(DEFAULT);handleCData(yytext, yyleng);return T_STRING;}
+
+<DEFAULT,BEGINNING,REGEXPOK>{
+{XMLCOMMENT}                 {l(); BEGIN(DEFAULT);handleRaw(yytext, yyleng);return T_STRING;}
+}
+
+<XML>{
+{XMLSTRING}                  {l(); handleRaw(yytext, yyleng);return T_STRING;}
+[{]                          {c(); BEGIN(REGEXPOK);return m('{');}
+[<]                          {c(); return m('<');}
+[/]                          {c(); return m('/');}
+[>]                          {c(); return m('>');}
+[=]                          {c(); return m('=');}
+{XMLID}                      {c(); handleRaw(yytext, yyleng);return T_IDENTIFIER;}
+{S}                          {l();}
+<<EOF>>                      {syntaxerror("unexpected end of file");}
+}
+
+<XMLTEXT>{
+[^<>{]+                      {l(); handleRaw(yytext, yyleng);return T_STRING;}
+[{]                          {c(); BEGIN(REGEXPOK);return m('{');}
+[<]                          {c(); BEGIN(XML);return m('<');}
+[>]                          {c(); return m('>');}
+{XMLCOMMENT}                 {l(); handleRaw(yytext, yyleng);return T_STRING;}
+{CDATA}                      {l(); handleRaw(yytext, yyleng);return T_STRING;}
+<<EOF>>                      {syntaxerror("unexpected end of file");}
+}
 
 <BEGINNING,REGEXPOK>{
-{REGEXP}                     {c(); BEGIN(INITIAL);return handleregexp();} 
-{HEXWITHSIGN}/{_}            {c(); BEGIN(INITIAL);return handlehex();}
-{HEXFLOATWITHSIGN}/{_}       {c(); BEGIN(INITIAL);return handlehexfloat();}
-{INTWITHSIGN}/{_}            {c(); BEGIN(INITIAL);return handleint();}
-{FLOATWITHSIGN}/{_}          {c(); BEGIN(INITIAL);return handlefloat();}
+{REGEXP}                     {c(); BEGIN(DEFAULT);return handleregexp();} 
+{HEXWITHSIGN}/{_}            {c(); BEGIN(DEFAULT);return handlehex();}
+{HEXFLOATWITHSIGN}/{_}       {c(); BEGIN(DEFAULT);return handlehexfloat();}
+{INTWITHSIGN}/{_}            {c(); BEGIN(DEFAULT);return handleint();}
+{FLOATWITHSIGN}/{_}          {c(); BEGIN(DEFAULT);return handlefloat();}
 }
 
 <REGEXPOK>[\{]               {c(); BEGIN(REGEXPOK);return m(T_DICTSTART);}
-[\{]                         {c(); BEGIN(INITIAL); return m('{');}
+[\{]                         {c(); BEGIN(DEFAULT); return m('{');}
 
-\xef\xbb\xbf                 {/* utf 8 bom */}
+\xef\xbb\xbf                 {/* utf 8 bom (0xfeff) */}
 {S}                          {l();}
 
-{HEXINT}/{_}                 {c(); BEGIN(INITIAL);return handlehex();}
-{HEXFLOAT}/{_}               {c(); BEGIN(INITIAL);return handlehexfloat();}
-{INT}/{_}                    {c(); BEGIN(INITIAL);return handleint();}
-{FLOAT}/{_}                  {c(); BEGIN(INITIAL);return handlefloat();}
-NaN                          {c(); BEGIN(INITIAL);return m(KW_NAN);}
+{HEXINT}/{_}                 {c(); BEGIN(DEFAULT);return handlehex();}
+{HEXFLOAT}/{_}               {c(); BEGIN(DEFAULT);return handlehexfloat();}
+{INT}/{_}                    {c(); BEGIN(DEFAULT);return handleint();}
+{FLOAT}/{_}                  {c(); BEGIN(DEFAULT);return handlefloat();}
+NaN                          {c(); BEGIN(DEFAULT);return m(KW_NAN);}
 
 3rr0r                        {/* for debugging: generates a tokenizer-level error */
                               syntaxerror("3rr0r");}
 
-{NAME}{S}*:{S}*for/{_}       {l();BEGIN(INITIAL);handleLabel(yytext, yyleng-3);return T_FOR;}
-{NAME}{S}*:{S}*do/{_}        {l();BEGIN(INITIAL);handleLabel(yytext, yyleng-2);return T_DO;}
-{NAME}{S}*:{S}*while/{_}     {l();BEGIN(INITIAL);handleLabel(yytext, yyleng-5);return T_WHILE;}
-{NAME}{S}*:{S}*switch/{_}    {l();BEGIN(INITIAL);handleLabel(yytext, yyleng-6);return T_SWITCH;}
-for                          {c();BEGIN(INITIAL);a3_lval.id="";return T_FOR;}
-do                           {c();BEGIN(INITIAL);a3_lval.id="";return T_DO;}
-while                        {c();BEGIN(INITIAL);a3_lval.id="";return T_WHILE;}
-switch                       {c();BEGIN(INITIAL);a3_lval.id="";return T_SWITCH;}
+{NAME}{S}*:{S}*for/{_}       {l();BEGIN(DEFAULT);handleLabel(yytext, yyleng-3);return T_FOR;}
+{NAME}{S}*:{S}*do/{_}        {l();BEGIN(DEFAULT);handleLabel(yytext, yyleng-2);return T_DO;}
+{NAME}{S}*:{S}*while/{_}     {l();BEGIN(DEFAULT);handleLabel(yytext, yyleng-5);return T_WHILE;}
+{NAME}{S}*:{S}*switch/{_}    {l();BEGIN(DEFAULT);handleLabel(yytext, yyleng-6);return T_SWITCH;}
+default{S}xml                {l();BEGIN(DEFAULT);return m(KW_DEFAULT_XML);}
+for                          {c();BEGIN(DEFAULT);a3_lval.id="";return T_FOR;}
+do                           {c();BEGIN(DEFAULT);a3_lval.id="";return T_DO;}
+while                        {c();BEGIN(DEFAULT);a3_lval.id="";return T_WHILE;}
+switch                       {c();BEGIN(DEFAULT);a3_lval.id="";return T_SWITCH;}
 
 [&][&]                       {c();BEGIN(REGEXPOK);return m(T_ANDAND);}
 [|][|]                       {c();BEGIN(REGEXPOK);return m(T_OROR);}
@@ -570,8 +612,8 @@ switch                       {c();BEGIN(INITIAL);a3_lval.id="";return T_SWITCH;}
 [=][=]                       {c();BEGIN(REGEXPOK);return m(T_EQEQ);}
 [>][=]                       {c();BEGIN(REGEXPOK);return m(T_GE);}
 [<][=]                       {c();BEGIN(REGEXPOK);return m(T_LE);}
-[-][-]                       {c();BEGIN(INITIAL);return m(T_MINUSMINUS);}
-[+][+]                       {c();BEGIN(INITIAL);return m(T_PLUSPLUS);}
+[-][-]                       {c();BEGIN(DEFAULT);return m(T_MINUSMINUS);}
+[+][+]                       {c();BEGIN(DEFAULT);return m(T_PLUSPLUS);}
 [+][=]                       {c();BEGIN(REGEXPOK);return m(T_PLUSBY);}
 [\^][=]                      {c();BEGIN(REGEXPOK);return m(T_XORBY);}
 [-][=]                       {c();BEGIN(REGEXPOK);return m(T_MINUSBY);}
@@ -579,6 +621,7 @@ switch                       {c();BEGIN(INITIAL);a3_lval.id="";return T_SWITCH;}
 [%][=]                       {c();BEGIN(REGEXPOK);return m(T_MODBY);}
 [*][=]                       {c();BEGIN(REGEXPOK);return m(T_MULBY);}
 [|][=]                       {c();BEGIN(REGEXPOK);return m(T_ORBY);}
+[&][=]                       {c();BEGIN(REGEXPOK);return m(T_ANDBY);}
 [>][>][=]                    {c();BEGIN(REGEXPOK);return m(T_SHRBY);}
 [<][<][=]                    {c();BEGIN(REGEXPOK);return m(T_SHLBY);}
 [>][>][>][=]                 {c();BEGIN(REGEXPOK);return m(T_USHRBY);}
@@ -592,78 +635,62 @@ switch                       {c();BEGIN(INITIAL);a3_lval.id="";return T_SWITCH;}
 :                            {c();BEGIN(REGEXPOK);return m(':');}
 instanceof                   {c();BEGIN(REGEXPOK);return m(KW_INSTANCEOF);}
 implements                   {c();BEGIN(REGEXPOK);return m(KW_IMPLEMENTS);}
-interface                    {c();BEGIN(INITIAL);return m(KW_INTERFACE);}
-namespace                    {c();BEGIN(INITIAL);return m(KW_NAMESPACE);}
-protected                    {c();BEGIN(INITIAL);return m(KW_PROTECTED);}
-undefined                    {c();BEGIN(INITIAL);return m(KW_UNDEFINED);}
-continue                     {c();BEGIN(INITIAL);return m(KW_CONTINUE);}
-override                     {c();BEGIN(INITIAL);return m(KW_OVERRIDE);}
-internal                     {c();BEGIN(INITIAL);return m(KW_INTERNAL);}
-function                     {c();BEGIN(INITIAL);return m(KW_FUNCTION);}
-finally                      {c();BEGIN(INITIAL);return m(KW_FINALLY);}
-default                      {c();BEGIN(INITIAL);return m(KW_DEFAULT);}
-package                      {c();BEGIN(INITIAL);return m(KW_PACKAGE);}
-private                      {c();BEGIN(INITIAL);return m(KW_PRIVATE);}
-dynamic                      {c();BEGIN(INITIAL);return m(KW_DYNAMIC);}
-extends                      {c();BEGIN(INITIAL);return m(KW_EXTENDS);}
+interface                    {c();BEGIN(DEFAULT);return m(KW_INTERFACE);}
+protected                    {c();BEGIN(DEFAULT);return m(KW_PROTECTED);}
+namespace                    {c();BEGIN(DEFAULT);return m(KW_NAMESPACE);}
+undefined                    {c();BEGIN(DEFAULT);return m(KW_UNDEFINED);}
+arguments                    {c();BEGIN(DEFAULT);return m(KW_ARGUMENTS);}
+continue                     {c();BEGIN(DEFAULT);return m(KW_CONTINUE);}
+override                     {c();BEGIN(DEFAULT);return m(KW_OVERRIDE);}
+internal                     {c();BEGIN(DEFAULT);return m(KW_INTERNAL);}
+function                     {c();BEGIN(DEFAULT);return m(KW_FUNCTION);}
+finally                      {c();BEGIN(DEFAULT);return m(KW_FINALLY);}
+default                      {c();BEGIN(DEFAULT);return m(KW_DEFAULT);}
+package                      {c();BEGIN(DEFAULT);return m(KW_PACKAGE);}
+private                      {c();BEGIN(DEFAULT);return m(KW_PRIVATE);}
+dynamic                      {c();BEGIN(DEFAULT);return m(KW_DYNAMIC);}
+extends                      {c();BEGIN(DEFAULT);return m(KW_EXTENDS);}
 delete                       {c();BEGIN(REGEXPOK);return m(KW_DELETE);}
 return                       {c();BEGIN(REGEXPOK);return m(KW_RETURN);}
-public                       {c();BEGIN(INITIAL);return m(KW_PUBLIC);}
-native                       {c();BEGIN(INITIAL);return m(KW_NATIVE);}
-static                       {c();BEGIN(INITIAL);return m(KW_STATIC);}
+public                       {c();BEGIN(DEFAULT);return m(KW_PUBLIC);}
+native                       {c();BEGIN(DEFAULT);return m(KW_NATIVE);}
+static                       {c();BEGIN(DEFAULT);return m(KW_STATIC);}
 import                       {c();BEGIN(REGEXPOK);return m(KW_IMPORT);}
 typeof                       {c();BEGIN(REGEXPOK);return m(KW_TYPEOF);}
 throw                        {c();BEGIN(REGEXPOK);return m(KW_THROW);}
-class                        {c();BEGIN(INITIAL);return m(KW_CLASS);}
-const                        {c();BEGIN(INITIAL);return m(KW_CONST);}
-catch                        {c();BEGIN(INITIAL);return m(KW_CATCH);}
-final                        {c();BEGIN(INITIAL);return m(KW_FINAL);}
-false                        {c();BEGIN(INITIAL);return m(KW_FALSE);}
-break                        {c();BEGIN(INITIAL);return m(KW_BREAK);}
-super                        {c();BEGIN(INITIAL);return m(KW_SUPER);}
-each                         {c();BEGIN(INITIAL);return m(KW_EACH);}
-void                         {c();BEGIN(INITIAL);return m(KW_VOID);}
-true                         {c();BEGIN(INITIAL);return m(KW_TRUE);}
-null                         {c();BEGIN(INITIAL);return m(KW_NULL);}
-else                         {c();BEGIN(INITIAL);return m(KW_ELSE);}
+class                        {c();BEGIN(DEFAULT);return m(KW_CLASS);}
+const                        {c();BEGIN(DEFAULT);return m(KW_CONST);}
+catch                        {c();BEGIN(DEFAULT);return m(KW_CATCH);}
+final                        {c();BEGIN(DEFAULT);return m(KW_FINAL);}
+false                        {c();BEGIN(DEFAULT);return m(KW_FALSE);}
+break                        {c();BEGIN(DEFAULT);return m(KW_BREAK);}
+super                        {c();BEGIN(DEFAULT);return m(KW_SUPER);}
+each                         {c();BEGIN(DEFAULT);return m(KW_EACH);}
+void                         {c();BEGIN(DEFAULT);return m(KW_VOID);}
+true                         {c();BEGIN(DEFAULT);return m(KW_TRUE);}
+null                         {c();BEGIN(DEFAULT);return m(KW_NULL);}
+else                         {c();BEGIN(DEFAULT);return m(KW_ELSE);}
 case                         {c();BEGIN(REGEXPOK);return m(KW_CASE);}
 with                         {c();BEGIN(REGEXPOK);return m(KW_WITH);}
 use                          {c();BEGIN(REGEXPOK);return m(KW_USE);}
 new                          {c();BEGIN(REGEXPOK);return m(KW_NEW);}
-get                          {c();BEGIN(INITIAL);return m(KW_GET);}
-set                          {c();BEGIN(INITIAL);return m(KW_SET);}
-var                          {c();BEGIN(INITIAL);return m(KW_VAR);}
-try                          {c();BEGIN(INITIAL);return m(KW_TRY);}
+get                          {c();BEGIN(DEFAULT);return m(KW_GET);}
+set                          {c();BEGIN(DEFAULT);return m(KW_SET);}
+var                          {c();BEGIN(DEFAULT);return m(KW_VAR);}
+try                          {c();BEGIN(DEFAULT);return m(KW_TRY);}
 is                           {c();BEGIN(REGEXPOK);return m(KW_IS) ;}
 in                           {c();BEGIN(REGEXPOK);return m(KW_IN) ;}
-if                           {c();BEGIN(INITIAL);return m(KW_IF) ;}
+if                           {c();BEGIN(DEFAULT);return m(KW_IF) ;}
 as                           {c();BEGIN(REGEXPOK);return m(KW_AS);}
-$?{NAME}                       {c();BEGIN(INITIAL);return handleIdentifier();}
+$?{NAME}                       {c();BEGIN(DEFAULT);return handleIdentifier();}
 
-[\]\}*]                       {c();BEGIN(INITIAL);return m(yytext[0]);}
+[\]\}*]                       {c();BEGIN(DEFAULT);return m(yytext[0]);}
 [+-\/^~@$!%&\(=\[|?:;,<>]   {c();BEGIN(REGEXPOK);return m(yytext[0]);}
-[\)\]]                           {c();BEGIN(INITIAL);return m(yytext[0]);}
-
-.		             {/* ERROR */
-                              char c1=yytext[0];
-                              char buf[128];
-                              buf[0] = yytext[0];
-                              int t;
-                              for(t=1;t<128;t++) {
-		                  char c = buf[t]=input();
-		                  if(c=='\n' || c==EOF)  {
-                                      buf[t] = 0;
-		                      break;
-                                  }
-		              }
-			      if(c1>='0' && c1<='9')
-			          syntaxerror("syntax error: %s (identifiers must not start with a digit)");
-                              else
-		                  syntaxerror("syntax error: %s", buf);
-		              printf("\n");
-			      exit(1);
-		              yyterminate();
-		             }
+[\)\]]                           {c();BEGIN(DEFAULT);return m(yytext[0]);}
+
+<DEFAULT,BEGINNING,REGEXPOK,XML,XMLTEXT>{
+.		             {tokenerror();}
+}
 <<EOF>>		             {l();
                               void*b = leave_file();
 			      if (!b) {
@@ -683,6 +710,29 @@ int yywrap()
     return 1;
 }
 
+static int tokenerror()
+{
+    char c1=yytext[0];
+    char buf[128];
+    buf[0] = yytext[0];
+    int t;
+    for(t=1;t<128;t++) {
+        char c = buf[t]=input();
+        if(c=='\n' || c==EOF)  {
+            buf[t] = 0;
+            break;
+        }
+    }
+    if(c1>='0' && c1<='9')
+        syntaxerror("syntax error: %s (identifiers must not start with a digit)");
+    else
+        syntaxerror("syntax error [state=%d]: %s", (yy_start-1)/2, buf);
+    printf("\n");
+    exit(1);
+    yyterminate();
+}
+
+
 static char mbuf[256];
 char*token2string(enum yytokentype nr, YYSTYPE v)
 {
@@ -750,6 +800,27 @@ char*token2string(enum yytokentype nr, YYSTYPE v)
     }
 }
 
+void tokenizer_begin_xml()
+{
+    dbg("begin reading xml");
+    BEGIN(XML);
+}
+void tokenizer_begin_xmltext()
+{
+    dbg("begin reading xml text");
+    BEGIN(XMLTEXT);
+}
+void tokenizer_end_xmltext()
+{
+    dbg("end reading xml text");
+    BEGIN(XML);
+}
+void tokenizer_end_xml()
+{
+    dbg("end reading xml");
+    BEGIN(DEFAULT);
+}
+
 void initialize_scanner()
 {
     BEGIN(BEGINNING);