3 Routines for compiling Flash2 AVM2 ABC Actionscript
5 Extension module for the rfxswf library.
6 Part of the swftools package.
8 Copyright (c) 2008 Matthias Kramm <kramm@quiss.org>
10 This program is free software; you can redistribute it and/or modify
11 it under the terms of the GNU General Public License as published by
12 the Free Software Foundation; either version 2 of the License, or
13 (at your option) any later version.
15 This program is distributed in the hope that it will be useful,
16 but WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 GNU General Public License for more details.
20 You should have received a copy of the GNU General Public License
21 along with this program; if not, write to the Free Software
22 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
31 #include "tokenizer.h"
34 static void countlines(char*text, int len) {
46 static int verbose = 1;
47 static void dbg(const char*format, ...)
54 va_start(arglist, format);
55 vsprintf(buf, format, arglist);
58 while(l && buf[l-1]=='\n') {
62 printf("(tokenizer) ");
67 void syntaxerror(const char*format, ...)
74 va_start(arglist, format);
75 vsprintf(buf, format, arglist);
77 fprintf(stderr, "%s:%d:%d: error: %s\n", current_filename_short, current_line, current_column, buf);
81 void warning(const char*format, ...)
88 va_start(arglist, format);
89 vsprintf(buf, format, arglist);
91 fprintf(stderr, "%s:%d:%d: warning: %s\n", current_filename_short, current_line, current_column, buf);
96 #ifndef YY_CURRENT_BUFFER
97 #define YY_CURRENT_BUFFER yy_current_buffer
100 void handleInclude(char*text, int len, char quotes)
104 char*p1 = strchr(text, '"');
105 char*p2 = strrchr(text, '"');
106 if(!p1 || !p2 || p1==p2) {
107 syntaxerror("Invalid include in line %d\n", current_line);
110 filename = strdup(p1+1);
114 while(!strchr(" \n\r\t", text[i1])) i1++;
116 while(strchr(" \n\r\t", text[i1])) i1++;
117 while(strchr(" \n\r\t", text[i2-1])) i2--;
118 if(i2!=len) text[i2]=0;
119 filename = strdup(&text[i1]);
122 char*fullfilename = enter_file(filename, YY_CURRENT_BUFFER);
123 yyin = fopen(fullfilename, "rb");
125 syntaxerror("Couldn't open include file \"%s\"\n", fullfilename);
128 yy_switch_to_buffer(yy_create_buffer( yyin, YY_BUF_SIZE ) );
129 //BEGIN(INITIAL); keep context
132 static int do_unescape(const char*s, const char*end, char*n)
138 if(o) o[len] = *s;len++;
143 if(s==end) syntaxerror("invalid \\ at end of string");
145 /* handle the various line endings (mac, dos, unix) */
158 case '\\': if(o) o[len] = '\\';s++;len++; break;
159 case '"': if(o) o[len] = '"';s++;len++; break;
160 case 'b': if(o) o[len] = '\b';s++;len++; break;
161 case 'f': if(o) o[len] = '\f';s++;len++; break;
162 case 'n': if(o) o[len] = '\n';s++;len++; break;
163 case 'r': if(o) o[len] = '\r';s++;len++; break;
164 case 't': if(o) o[len] = '\t';s++;len++; break;
165 case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': {
168 while(strchr("01234567", *s) && nr<3 && s<end) {
175 syntaxerror("octal number out of range (0-255): %d", num);
176 if(o) o[len] = num;len++;
179 case 'x': case 'u': {
188 if(s==end) syntaxerror("invalid \\u or \\x at end of string");
191 if(s==end) syntaxerror("invalid \\u{ at end of string");
196 while(strchr("0123456789abcdefABCDEF", *s) && (bracket || nr < max) && s<end) {
198 if(*s>='0' && *s<='9') num |= *s - '0';
199 if(*s>='a' && *s<='f') num |= *s - 'a' + 10;
200 if(*s>='A' && *s<='F') num |= *s - 'A' + 10;
205 if(*s=='}' && s<end) {
208 syntaxerror("missing terminating '}'");
212 char*utf8 = getUTF8(num);
214 if(o) o[len] = *utf8;utf8++;len++;
218 syntaxerror("byte out of range (0-255): %d", num);
219 if(o) o[len] = num;len++;
224 syntaxerror("unknown escape sequence: \"\\%c\"", *s);
231 static string_t string_unescape(const char*in, int l)
234 const char*end = &in[l];
236 int len = do_unescape(s, end, 0);
237 char*n = (char*)malloc(len+1);
238 do_unescape(s, end, n);
239 string_t out = string_new(n, len);
243 static void handleString(char*s, int len)
246 if(s[len-1]!='"') syntaxerror("String doesn't end with '\"'");
249 else if(s[0]=='\'') {
250 if(s[len-1]!='\'') syntaxerror("String doesn't end with '\"'");
253 else syntaxerror("String incorrectly terminated");
256 avm2_lval.str = string_unescape(s, len);
260 char start_of_expression;
262 static inline int mkid(int type)
264 char*s = malloc(yyleng+1);
265 memcpy(s, yytext, yyleng);
271 static inline int m(int type)
273 avm2_lval.token = type;
278 static char numberbuf[64];
281 if(yyleng>sizeof(numberbuf)-1)
282 syntaxerror("decimal number overflow");
284 memcpy(s, yytext, yyleng);
289 static inline int setint(int v)
291 avm2_lval.number_int = v;
299 static inline int setuint(unsigned int v)
301 avm2_lval.number_uint = v;
310 static inline int handlefloat()
313 avm2_lval.number_float = atof(s);
317 static inline int handleint()
320 char l = (yytext[0]=='-');
322 char*max = l?"1073741824":"2147483647";
324 syntaxerror("integer overflow");
327 for(t=0;t<yyleng-l;t++) {
328 if(yytext[l+t]>max[t])
329 syntaxerror("integer overflow %s > %s", s+l,max);
330 else if(yytext[l+t]<max[t])
340 for(t=0;t<yyleng;t++) {
348 static inline int handlehex()
350 char l = (yytext[0]=='-')+2;
353 syntaxerror("integer overflow");
356 for(t=l;t<yyleng;t++) {
361 else if(c>='a' && c<='f' ||
365 if(l && v>1073741824)
366 syntaxerror("signed integer overflow");
367 if(!l && v>2147483647)
368 syntaxerror("unsigned integer overflow");
371 return setint(-(int)v);
377 void handleLabel(char*text, int len)
380 for(t=len-1;t>=0;--t) {
385 char*s = malloc(t+1);
386 memcpy(s, yytext, t);
391 void initialize_scanner();
392 #define YY_USER_INIT initialize_scanner();
394 #define c() {countlines(yytext, yyleng);}
396 //Boolean {c();return m(KW_BOOLEAN);}
397 //int {c();return m(KW_INT);}
398 //uint {c();return m(KW_UINT);}
399 //Number {c();return m(KW_NUMBER);}
407 NAME [a-zA-Z_][a-zA-Z0-9_\\]*
410 HEXINT 0x[a-zA-Z0-9]+
412 FLOAT [0-9]+(\.[0-9]*)?|\.[0-9]+
414 HEXWITHSIGN [+-]?({HEXINT})
415 INTWITHSIGN [+-]?({INT})
416 FLOATWITHSIGN [+-]?({FLOAT})
418 STRING ["](\\[\x00-\xff]|[^\\"\n])*["]|['](\\[\x00-\xff]|[^\\'\n])*[']
420 MULTILINE_COMMENT [/][*]+([*][^/]|[^/*]|[^*][/]|[\x00-\x1f])*[*]+[/]
421 SINGLELINE_COMMENT \/\/[^\n]*\n
422 REGEXP [/]([^/\n]|\\[/])*[/][a-zA-Z]*
426 {SINGLELINE_COMMENT} {c(); /* single line comment */}
427 {MULTILINE_COMMENT} {c(); /* multi line comment */}
428 [/][*] {syntaxerror("syntax error: unterminated comment", yytext);}
430 ^include{S}+{STRING}{S}*/\n {c();handleInclude(yytext, yyleng, 1);}
431 ^include{S}+[^" \t\r\n][\x20-\xff]*{S}*/\n {c();handleInclude(yytext, yyleng, 0);}
432 {STRING} {c(); BEGIN(INITIAL);handleString(yytext, yyleng);return T_STRING;}
434 <BEGINNING,REGEXPOK>{
435 {REGEXP} {c(); BEGIN(INITIAL);return m(T_REGEXP);}
436 {HEXWITHSIGN} {c(); BEGIN(INITIAL);return handlehex();}
437 {INTWITHSIGN} {c(); BEGIN(INITIAL);return handleint();}
438 {FLOATWITHSIGN} {c(); BEGIN(INITIAL);return handlefloat();}
441 \xef\xbb\xbf {/* utf 8 bom */}
444 {HEXINT} {c(); BEGIN(INITIAL);return handlehex();}
445 {INT} {c(); BEGIN(INITIAL);return handleint();}
446 {FLOAT} {c(); BEGIN(INITIAL);return handlefloat();}
448 3rr0r {/* for debugging: generates a tokenizer-level error */
449 syntaxerror("3rr0r");}
451 {NAME}{S}*:{S}*for/{_} {c();handleLabel(yytext, yyleng-3);return T_FOR;}
452 {NAME}{S}*:{S}*do/{_} {c();handleLabel(yytext, yyleng-2);return T_DO;}
453 {NAME}{S}*:{S}*while/{_} {c();handleLabel(yytext, yyleng-5);return T_WHILE;}
454 {NAME}{S}*:{S}*switch/{_} {c();handleLabel(yytext, yyleng-6);return T_SWITCH;}
455 for {c();avm2_lval.id="";return T_FOR;}
456 do {c();avm2_lval.id="";return T_DO;}
457 while {c();avm2_lval.id="";return T_WHILE;}
458 switch {c();avm2_lval.id="";return T_SWITCH;}
460 [&][&] {c();BEGIN(REGEXPOK);return m(T_ANDAND);}
461 [|][|] {c();BEGIN(REGEXPOK);return m(T_OROR);}
462 [!][=] {c();BEGIN(REGEXPOK);return m(T_NE);}
463 [!][=][=] {c();BEGIN(REGEXPOK);return m(T_NEE);}
464 [=][=][=] {c();BEGIN(REGEXPOK);return m(T_EQEQEQ);}
465 [=][=] {c();BEGIN(REGEXPOK);return m(T_EQEQ);}
466 [>][=] {c();return m(T_GE);}
467 [<][=] {c();return m(T_LE);}
468 [-][-] {c();BEGIN(INITIAL);return m(T_MINUSMINUS);}
469 [+][+] {c();BEGIN(INITIAL);return m(T_PLUSPLUS);}
470 [+][=] {c();return m(T_PLUSBY);}
471 [-][=] {c();return m(T_MINUSBY);}
472 [/][=] {c();return m(T_DIVBY);}
473 [%][=] {c();return m(T_MODBY);}
474 [*][=] {c();return m(T_MULBY);}
475 [>][>][=] {c();return m(T_SHRBY);}
476 [<][<][=] {c();return m(T_SHLBY);}
477 [>][>][>][=] {c();return m(T_USHRBY);}
478 [<][<] {c();return m(T_SHL);}
479 [>][>][>] {c();return m(T_USHR);}
480 [>][>] {c();return m(T_SHR);}
481 \.\.\. {c();return m(T_DOTDOTDOT);}
482 \.\. {c();return m(T_DOTDOT);}
483 \. {c();return m('.');}
484 :: {c();return m(T_COLONCOLON);}
485 : {c();return m(':');}
486 instanceof {c();return m(KW_INSTANCEOF);}
487 implements {c();return m(KW_IMPLEMENTS);}
488 interface {c();return m(KW_INTERFACE);}
489 namespace {c();return m(KW_NAMESPACE);}
490 protected {c();return m(KW_PROTECTED);}
491 undefined {c();return m(KW_UNDEFINED);}
492 continue {c();return m(KW_CONTINUE);}
493 override {c();return m(KW_OVERRIDE);}
494 internal {c();return m(KW_INTERNAL);}
495 function {c();return m(KW_FUNCTION);}
496 default {c();return m(KW_DEFAULT);}
497 package {c();return m(KW_PACKAGE);}
498 private {c();return m(KW_PRIVATE);}
499 dynamic {c();return m(KW_DYNAMIC);}
500 extends {c();return m(KW_EXTENDS);}
501 delete {c();return m(KW_DELETE);}
502 return {c();return m(KW_RETURN);}
503 public {c();return m(KW_PUBLIC);}
504 native {c();return m(KW_NATIVE);}
505 static {c();return m(KW_STATIC);}
506 import {c();return m(KW_IMPORT);}
507 typeof {c();return m(KW_TYPEOF);}
508 class {c();return m(KW_CLASS);}
509 const {c();return m(KW_CONST);}
510 catch {c();return m(KW_CATCH);}
511 final {c();return m(KW_FINAL);}
512 false {c();return m(KW_FALSE);}
513 break {c();return m(KW_BREAK);}
514 super {c();return m(KW_SUPER);}
515 each {c();return m(KW_EACH);}
516 void {c();return m(KW_VOID);}
517 true {c();return m(KW_TRUE);}
518 null {c();return m(KW_NULL);}
519 else {c();return m(KW_ELSE);}
520 case {c();return m(KW_CASE);}
521 use {c();return m(KW_USE);}
522 new {c();return m(KW_NEW);}
523 get {c();return m(KW_GET);}
524 set {c();return m(KW_SET);}
525 var {c();return m(KW_VAR);}
526 try {c();return m(KW_TRY);}
527 is {c();return m(KW_IS) ;}
528 in {c();return m(KW_IN) ;}
529 if {c();return m(KW_IF) ;}
530 as {c();return m(KW_AS);}
531 {NAME} {c();BEGIN(INITIAL);return mkid(T_IDENTIFIER);}
533 [+-\/*^~@$!%&\(=\[\]\{\}|?:;,<>] {c();BEGIN(REGEXPOK);return m(yytext[0]);}
534 [\)\]] {c();BEGIN(INITIAL);return m(yytext[0]);}
536 . {char c1=yytext[0];
541 char c = buf[t]=input();
542 if(c=='\n' || c==EOF) {
547 if(c1>='0' && c1<='9')
548 syntaxerror("syntax error: %s (identifiers must not start with a digit)");
550 syntaxerror("syntax error: %s", buf);
556 void*b = leave_file();
559 yy_delete_buffer(YY_CURRENT_BUFFER);
562 yy_delete_buffer(YY_CURRENT_BUFFER);
563 yy_switch_to_buffer(b);
574 static char mbuf[256];
575 char*token2string(enum yytokentype nr, YYSTYPE v)
577 if(nr==T_STRING) return "<string>";
578 else if(nr==T_INT) return "<int>";
579 else if(nr==T_UINT) return "<uint>";
580 else if(nr==T_BYTE) return "<byte>";
581 else if(nr==T_FLOAT) return "<float>";
582 else if(nr==T_REGEXP) return "REGEXP";
583 else if(nr==T_EOF) return "***END***";
584 else if(nr==T_GE) return ">=";
585 else if(nr==T_LE) return "<=";
586 else if(nr==T_MINUSMINUS) return "--";
587 else if(nr==T_PLUSPLUS) return "++";
588 else if(nr==KW_IMPLEMENTS) return "implements";
589 else if(nr==KW_INTERFACE) return "interface";
590 else if(nr==KW_NAMESPACE) return "namespace";
591 else if(nr==KW_PROTECTED) return "protected";
592 else if(nr==KW_OVERRIDE) return "override";
593 else if(nr==KW_INTERNAL) return "internal";
594 else if(nr==KW_FUNCTION) return "function";
595 else if(nr==KW_PACKAGE) return "package";
596 else if(nr==KW_PRIVATE) return "private";
597 else if(nr==KW_BOOLEAN) return "Boolean";
598 else if(nr==KW_DYNAMIC) return "dynamic";
599 else if(nr==KW_EXTENDS) return "extends";
600 else if(nr==KW_PUBLIC) return "public";
601 else if(nr==KW_NATIVE) return "native";
602 else if(nr==KW_STATIC) return "static";
603 else if(nr==KW_IMPORT) return "import";
604 else if(nr==KW_NUMBER) return "number";
605 else if(nr==KW_CLASS) return "class";
606 else if(nr==KW_CONST) return "const";
607 else if(nr==KW_FINAL) return "final";
608 else if(nr==KW_FALSE) return "False";
609 else if(nr==KW_TRUE) return "True";
610 else if(nr==KW_UINT) return "uint";
611 else if(nr==KW_NULL) return "null";
612 else if(nr==KW_ELSE) return "else";
613 else if(nr==KW_USE) return "use";
614 else if(nr==KW_INT) return "int";
615 else if(nr==KW_NEW) return "new";
616 else if(nr==KW_GET) return "get";
617 else if(nr==KW_SET) return "set";
618 else if(nr==KW_VAR) return "var";
619 else if(nr==KW_IS) return "is";
620 else if(nr==KW_AS) return "as";
621 else if(nr==T_IDENTIFIER) return "ID";
623 sprintf(mbuf, "%d", nr);
628 void initialize_scanner()