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 string_t string_unescape(const char*in, int l)
136 const char*end = &in[l];
137 char*n = (char*)malloc(l);
146 if(s==end) syntaxerror("invalid \\ at end of string");
148 /* handle the various line endings (mac, dos, unix) */
161 case '\\': o[len++] = '\\';s++; break;
162 case '"': o[len++] = '"';s++; break;
163 case 'b': o[len++] = '\b';s++; break;
164 case 'f': o[len++] = '\f';s++; break;
165 case 'n': o[len++] = '\n';s++; break;
166 case 'r': o[len++] = '\r';s++; break;
167 case 't': o[len++] = '\t';s++; break;
168 case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': {
171 while(strchr("01234567", *s) && nr<3 && s<end) {
178 syntaxerror("octal number out of range (0-255): %d", num);
182 case 'x': case 'u': {
191 if(s==end) syntaxerror("invalid \\u or \\x at end of string");
194 if(s==end) syntaxerror("invalid \\u{ at end of string");
199 while(strchr("0123456789abcdefABCDEF", *s) && (bracket || nr < max) && s<end) {
201 if(*s>='0' && *s<='9') num |= *s - '0';
202 if(*s>='a' && *s<='f') num |= *s - 'a' + 10;
203 if(*s>='A' && *s<='F') num |= *s - 'A' + 10;
208 if(*s=='}' && s<end) {
211 syntaxerror("missing terminating '}'");
215 char*utf8 = getUTF8(num);
221 syntaxerror("byte out of range (0-255): %d", num);
227 syntaxerror("unknown escape sequence: \"\\%c\"", *s);
230 string_t out = string_new(n, len);
235 static void handleString(char*s, int len)
238 if(s[len-1]!='"') syntaxerror("String doesn't end with '\"'");
241 else if(s[0]=='\'') {
242 if(s[len-1]!='\'') syntaxerror("String doesn't end with '\"'");
245 else syntaxerror("String incorrectly terminated");
248 avm2_lval.str = string_unescape(s, len);
252 char start_of_expression;
254 static inline int mkid(int type)
256 char*s = malloc(yyleng+1);
257 memcpy(s, yytext, yyleng);
263 static inline int m(int type)
265 avm2_lval.token = type;
270 static char numberbuf[64];
271 static inline int handlenumber()
273 if(yyleng>sizeof(numberbuf)-1)
274 syntaxerror("decimal number overflow");
277 memcpy(s, yytext, yyleng);
282 for(t=0;t<yyleng;t++) {
285 syntaxerror("Invalid number");
287 } else if(!strchr("-0123456789", yytext[t])) {
288 syntaxerror("Invalid number");
292 avm2_lval.number_float = atof(s);
295 char l = (yytext[0]=='-');
297 char*max = l?"1073741824":"2147483647";
299 syntaxerror("integer overflow");
302 for(t=0;t<yyleng-l;t++) {
303 if(yytext[l+t]>max[t])
304 syntaxerror("integer overflow %s > %s", s+l,max);
305 else if(yytext[l+t]<max[t])
311 avm2_lval.number_int = v;
320 for(t=0;t<yyleng;t++) {
324 avm2_lval.number_uint = v;
334 void initialize_scanner();
335 #define YY_USER_INIT initialize_scanner();
337 #define c() {countlines(yytext, yyleng);}
344 NAME [a-zA-Z_][a-zA-Z0-9_\\]*
346 NUMBER [0-9]+(\.[0-9]*)?|-?\.[0-9]+
347 NUMBERWITHSIGN [+-]?({NUMBER})
349 STRING ["](\\[\x00-\xff]|[^\\"\n])*["]|['](\\[\x00-\xff]|[^\\'\n])*[']
351 MULTILINE_COMMENT [/][*]+([*][^/]|[^/*]|[^*][/]|[\x00-\x1f])*[*]+[/]
352 SINGLELINE_COMMENT \/\/[^\n]*\n
353 REGEXP [/]([^/\n]|\\[/])*[/][a-zA-Z]*
357 {SINGLELINE_COMMENT} {c(); /* single line comment */}
358 {MULTILINE_COMMENT} {c(); /* multi line comment */}
359 [/][*] {syntaxerror("syntax error: unterminated comment", yytext);}
361 ^include{S}+{STRING}{S}*/\n {c();handleInclude(yytext, yyleng, 1);}
362 ^include{S}+[^" \t\r\n][\x20-\xff]*{S}*/\n {c();handleInclude(yytext, yyleng, 0);}
363 {STRING} {c(); BEGIN(INITIAL);handleString(yytext, yyleng);return T_STRING;}
365 <BEGINNING,REGEXPOK>{
366 {REGEXP} {c(); BEGIN(INITIAL);return m(T_REGEXP);}
367 {NUMBERWITHSIGN} {c(); BEGIN(INITIAL);return handlenumber();}
370 \xef\xbb\xbf {/* utf 8 bom */}
373 {NUMBER} {c(); BEGIN(INITIAL);return handlenumber();}
375 3rr0r {/* for debugging: generates a tokenizer-level error */
376 syntaxerror("3rr0r");}
378 [&][&] {c();BEGIN(REGEXPOK);return m(T_ANDAND);}
379 [|][|] {c();BEGIN(REGEXPOK);return m(T_OROR);}
380 [!][=] {c();BEGIN(REGEXPOK);return m(T_NE);}
381 [=][=][=] {c();BEGIN(REGEXPOK);return m(T_EQEQEQ);}
382 [=][=] {c();BEGIN(REGEXPOK);return m(T_EQEQ);}
383 [>][=] {c();return m(T_GE);}
384 [<][=] {c();return m(T_LE);}
385 [-][-] {c();BEGIN(INITIAL);return m(T_MINUSMINUS);}
386 [+][+] {c();BEGIN(INITIAL);return m(T_PLUSPLUS);}
387 [+][=] {c();return m(T_PLUSBY);}
388 [-][=] {c();return m(T_MINUSBY);}
389 [/][=] {c();return m(T_DIVBY);}
390 [%][=] {c();return m(T_MODBY);}
391 [*][=] {c();return m(T_MULBY);}
392 [>][>][=] {c();return m(T_SHRBY);}
393 [<][<][=] {c();return m(T_SHLBY);}
394 [>][>][>][=] {c();return m(T_USHRBY);}
395 [<][<] {c();return m(T_SHL);}
396 [>][>][>] {c();return m(T_USHR);}
397 [>][>] {c();return m(T_SHR);}
398 \.\.\. {c();return m(T_DOTDOTDOT);}
399 \.\. {c();return m(T_DOTDOT);}
400 \. {c();return m('.');}
401 :: {c();return m(T_COLONCOLON);}
402 : {c();return m(':');}
403 implements {c();return m(KW_IMPLEMENTS);}
404 interface {c();return m(KW_INTERFACE);}
405 namespace {c();return m(KW_NAMESPACE);}
406 protected {c();return m(KW_PROTECTED);}
407 override {c();return m(KW_OVERRIDE);}
408 internal {c();return m(KW_INTERNAL);}
409 function {c();return m(KW_FUNCTION);}
410 package {c();return m(KW_PACKAGE);}
411 private {c();return m(KW_PRIVATE);}
412 Boolean {c();return m(KW_BOOLEAN);}
413 dynamic {c();return m(KW_DYNAMIC);}
414 extends {c();return m(KW_EXTENDS);}
415 delete {c();return m(KW_DELETE);}
416 return {c();return m(KW_RETURN);}
417 public {c();return m(KW_PUBLIC);}
418 native {c();return m(KW_NATIVE);}
419 static {c();return m(KW_STATIC);}
420 import {c();return m(KW_IMPORT);}
421 Number {c();return m(KW_NUMBER);}
422 while {c();return m(KW_WHILE);}
423 class {c();return m(KW_CLASS);}
424 const {c();return m(KW_CONST);}
425 final {c();return m(KW_FINAL);}
426 false {c();return m(KW_FALSE);}
427 break {c();return m(KW_BREAK);}
428 true {c();return m(KW_TRUE);}
429 uint {c();return m(KW_UINT);}
430 null {c();return m(KW_NULL);}
431 else {c();return m(KW_ELSE);}
432 use {c();return m(KW_USE);}
433 int {c();return m(KW_INT);}
434 new {c();return m(KW_NEW);}
435 get {c();return m(KW_GET);}
436 for {c();return m(KW_FOR);}
437 set {c();return m(KW_SET);}
438 var {c();return m(KW_VAR);}
439 is {c();return m(KW_IS) ;}
440 if {c();return m(KW_IF) ;}
441 as {c();return m(KW_AS);}
442 {NAME} {c();BEGIN(INITIAL);return mkid(T_IDENTIFIER);}
444 [+-\/*^~@$!%&\(=\[\]\{\}|?:;,<>] {c();BEGIN(REGEXPOK);return m(yytext[0]);}
445 [\)\]] {c();BEGIN(INITIAL);return m(yytext[0]);}
447 . {char c1=yytext[0];
452 char c = buf[t]=input();
453 if(c=='\n' || c==EOF) {
458 if(c1>='0' && c1<='9')
459 syntaxerror("syntax error: %s (identifiers must not start with a digit)");
461 syntaxerror("syntax error: %s", buf);
467 void*b = leave_file();
470 yy_delete_buffer(YY_CURRENT_BUFFER);
473 yy_delete_buffer(YY_CURRENT_BUFFER);
474 yy_switch_to_buffer(b);
485 static char mbuf[256];
486 char*token2string(enum yytokentype nr, YYSTYPE v)
488 if(nr==T_STRING) return "<string>";
489 else if(nr==T_INT) return "<int>";
490 else if(nr==T_UINT) return "<uint>";
491 else if(nr==T_BYTE) return "<byte>";
492 else if(nr==T_FLOAT) return "<float>";
493 else if(nr==T_REGEXP) return "REGEXP";
494 else if(nr==T_EOF) return "***END***";
495 else if(nr==T_GE) return ">=";
496 else if(nr==T_LE) return "<=";
497 else if(nr==T_MINUSMINUS) return "--";
498 else if(nr==T_PLUSPLUS) return "++";
499 else if(nr==KW_IMPLEMENTS) return "implements";
500 else if(nr==KW_INTERFACE) return "interface";
501 else if(nr==KW_NAMESPACE) return "namespace";
502 else if(nr==KW_PROTECTED) return "protected";
503 else if(nr==KW_OVERRIDE) return "override";
504 else if(nr==KW_INTERNAL) return "internal";
505 else if(nr==KW_FUNCTION) return "function";
506 else if(nr==KW_PACKAGE) return "package";
507 else if(nr==KW_PRIVATE) return "private";
508 else if(nr==KW_BOOLEAN) return "Boolean";
509 else if(nr==KW_DYNAMIC) return "dynamic";
510 else if(nr==KW_EXTENDS) return "extends";
511 else if(nr==KW_PUBLIC) return "public";
512 else if(nr==KW_NATIVE) return "native";
513 else if(nr==KW_STATIC) return "static";
514 else if(nr==KW_IMPORT) return "import";
515 else if(nr==KW_NUMBER) return "number";
516 else if(nr==KW_CLASS) return "class";
517 else if(nr==KW_CONST) return "const";
518 else if(nr==KW_FINAL) return "final";
519 else if(nr==KW_FALSE) return "False";
520 else if(nr==KW_TRUE) return "True";
521 else if(nr==KW_UINT) return "uint";
522 else if(nr==KW_NULL) return "null";
523 else if(nr==KW_ELSE) return "else";
524 else if(nr==KW_USE) return "use";
525 else if(nr==KW_INT) return "int";
526 else if(nr==KW_NEW) return "new";
527 else if(nr==KW_GET) return "get";
528 else if(nr==KW_FOR) return "for";
529 else if(nr==KW_SET) return "set";
530 else if(nr==KW_VAR) return "var";
531 else if(nr==KW_IS) return "is";
532 else if(nr==KW_AS) return "as";
533 else if(nr==T_IDENTIFIER) return "ID";
535 sprintf(mbuf, "%d", nr);
540 void initialize_scanner()