1 /*-----------------------------------------------------------------------
2 SDCC.lex - lexical analyser for use with sdcc ( a freeware compiler for
3 8/16 bit microcontrollers)
4 Written by : Sandeep Dutta . sandeep.dutta@usa.net (1997)
6 This program is free software; you can redistribute it and/or modify it
7 under the terms of the GNU General Public License as published by the
8 Free Software Foundation; either version 2, or (at your option) any
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software
18 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
20 In other words, you are welcome to use, share and improve this program.
21 You are forbidden to forbid anyone else to use, share and improve
22 what you give them. Help stamp out software-hoarding!
23 -------------------------------------------------------------------------*/
42 char *stringLiteral();
45 extern int lineno, column;
46 extern char *filename ;
49 int process_pragma(char *);
55 #define YY_PROTO(proto) proto
57 #define YY_PROTO(proto) ()
62 int yywrap YY_PROTO((void))
67 static void yyunput (int, char *);
69 static void my_unput(char c)
71 yyunput(c, (yytext_ptr));
74 #define TKEYWORD(token) return (isTargetKeyword(yytext) ? token :\
79 extern int check_type ();
80 extern int isTargetKeyword ();
81 extern int checkCurrFile (char *);
82 extern int processPragma (char *);
83 extern int printListing (int );
84 struct optimize save_optimize ;
85 struct options save_options ;
92 asmp = asmbuff = realloc (asmbuff, INITIAL_INLINEASM);
93 asmbuffSize=INITIAL_INLINEASM;
99 yylval.yyinline = strdup (asmbuff);
104 if (asmp-asmbuff >= asmbuffSize-2) {
105 // increase the buffersize with 50%
106 int size=asmp-asmbuff;
107 asmbuffSize=asmbuffSize*3/2;
108 asmbuff = realloc (asmbuff, asmbuffSize);
115 if (asmp-asmbuff >= asmbuffSize-3) {
116 // increase the buffersize with 50%
117 int size=asmp-asmbuff;
118 asmbuffSize=asmbuffSize*3/2;
119 asmbuff = realloc (asmbuff, asmbuffSize);
124 "at" { count(); TKEYWORD(AT) ; }
125 "auto" { count(); return(AUTO); }
126 "bit" { count(); TKEYWORD(BIT) ; }
127 "break" { count(); return(BREAK); }
128 "case" { count(); return(CASE); }
129 "char" { count(); return(CHAR); }
130 "code" { count(); TKEYWORD(CODE); }
131 "const" { count(); return(CONST); }
132 "continue" { count(); return(CONTINUE); }
133 "critical" { count(); TKEYWORD(CRITICAL); }
134 "data" { count(); TKEYWORD(DATA); }
135 "default" { count(); return(DEFAULT); }
136 "do" { count(); return(DO); }
137 "double" { count(); werror(W_DOUBLE_UNSUPPORTED);return(FLOAT); }
138 "else" { count(); return(ELSE); }
139 "enum" { count(); return(ENUM); }
140 "extern" { count(); return(EXTERN); }
141 "far" { count(); TKEYWORD(XDATA); }
142 "eeprom" { count(); TKEYWORD(EEPROM); }
143 "float" { count(); return(FLOAT); }
144 "flash" { count(); TKEYWORD(CODE);}
145 "for" { count(); return(FOR); }
146 "goto" { count(); return(GOTO); }
147 "idata" { count(); TKEYWORD(IDATA);}
148 "if" { count(); return(IF); }
149 "int" { count(); return(INT); }
150 "interrupt" { count(); return(INTERRUPT);}
151 "nonbanked" { count(); TKEYWORD(NONBANKED);}
152 "banked" { count(); TKEYWORD(BANKED);}
153 "long" { count(); return(LONG); }
154 "near" { count(); TKEYWORD(DATA);}
155 "pdata" { count(); TKEYWORD(PDATA); }
156 "reentrant" { count(); TKEYWORD(REENTRANT);}
157 "register" { count(); return(REGISTER); }
158 "return" { count(); return(RETURN); }
159 "sfr" { count(); TKEYWORD(SFR) ; }
160 "sbit" { count(); TKEYWORD(SBIT) ; }
161 "short" { count(); return(SHORT); }
162 "signed" { count(); return(SIGNED); }
163 "sizeof" { count(); return(SIZEOF); }
164 "sram" { count(); TKEYWORD(XDATA);}
165 "static" { count(); return(STATIC); }
166 "struct" { count(); return(STRUCT); }
167 "switch" { count(); return(SWITCH); }
168 "typedef" { count(); return(TYPEDEF); }
169 "union" { count(); return(UNION); }
170 "unsigned" { count(); return(UNSIGNED); }
171 "void" { count(); return(VOID); }
172 "volatile" { count(); return(VOLATILE); }
173 "using" { count(); TKEYWORD(USING); }
174 "_naked" { count(); TKEYWORD(NAKED); }
175 "while" { count(); return(WHILE); }
176 "xdata" { count(); TKEYWORD(XDATA); }
177 "..." { count(); return(VAR_ARGS);}
178 "__typeof" { count(); return TYPEOF;}
179 "_JavaNative" { count(); TKEYWORD(JAVANATIVE);}
180 "_overlay" { count(); TKEYWORD(OVERLAY);}
181 {L}({L}|{D})* { count(); return(check_type()); }
182 0[xX]{H}+{IS}? { count(); yylval.val = constVal(yytext); return(CONSTANT); }
183 0{D}+{IS}? { count(); yylval.val = constVal(yytext); return(CONSTANT); }
184 {D}+{IS}? { count(); yylval.val = constVal(yytext); return(CONSTANT); }
185 '(\\.|[^\\'])+' { count();yylval.val = charVal (yytext); return(CONSTANT); /* ' make syntax highliter happy */}
186 {D}+{E}{FS}? { count(); yylval.val = constFloatVal(yytext);return(CONSTANT); }
187 {D}*"."{D}+({E})?{FS}? { count(); yylval.val = constFloatVal(yytext);return(CONSTANT); }
188 {D}+"."{D}*({E})?{FS}? { count(); yylval.val = constFloatVal(yytext);return(CONSTANT); }
189 \" { count(); yylval.val=strVal(stringLiteral()); return(STRING_LITERAL);}
190 ">>=" { count(); yylval.yyint = RIGHT_ASSIGN ; return(RIGHT_ASSIGN); }
191 "<<=" { count(); yylval.yyint = LEFT_ASSIGN ; return(LEFT_ASSIGN) ; }
192 "+=" { count(); yylval.yyint = ADD_ASSIGN ; return(ADD_ASSIGN) ; }
193 "-=" { count(); yylval.yyint = SUB_ASSIGN ; return(SUB_ASSIGN) ; }
194 "*=" { count(); yylval.yyint = MUL_ASSIGN ; return(MUL_ASSIGN) ; }
195 "/=" { count(); yylval.yyint = DIV_ASSIGN ; return(DIV_ASSIGN) ; }
196 "%=" { count(); yylval.yyint = MOD_ASSIGN ; return(MOD_ASSIGN) ; }
197 "&=" { count(); yylval.yyint = AND_ASSIGN ; return(AND_ASSIGN) ; }
198 "^=" { count(); yylval.yyint = XOR_ASSIGN ; return(XOR_ASSIGN) ; }
199 "|=" { count(); yylval.yyint = OR_ASSIGN ; return(OR_ASSIGN) ; }
200 ">>" { count(); return(RIGHT_OP); }
201 "<<" { count(); return(LEFT_OP); }
202 "++" { count(); return(INC_OP); }
203 "--" { count(); return(DEC_OP); }
204 "->" { count(); return(PTR_OP); }
205 "&&" { count(); return(AND_OP); }
206 "||" { count(); return(OR_OP); }
207 "<=" { count(); return(LE_OP); }
208 ">=" { count(); return(GE_OP); }
209 "==" { count(); return(EQ_OP); }
210 "!=" { count(); return(NE_OP); }
211 ";" { count(); return(';'); }
212 "{" { count(); NestLevel++ ; return('{'); }
213 "}" { count(); NestLevel--; return('}'); }
214 "," { count(); return(','); }
215 ":" { count(); return(':'); }
216 "=" { count(); return('='); }
217 "(" { count(); return('('); }
218 ")" { count(); return(')'); }
219 "[" { count(); return('['); }
220 "]" { count(); return(']'); }
221 "." { count(); return('.'); }
222 "&" { count(); return('&'); }
223 "!" { count(); return('!'); }
224 "~" { count(); return('~'); }
225 "-" { count(); return('-'); }
226 "+" { count(); return('+'); }
227 "*" { count(); return('*'); }
228 "/" { count(); return('/'); }
229 "%" { count(); return('%'); }
230 "<" { count(); return('<'); }
231 ">" { count(); return('>'); }
232 "^" { count(); return('^'); }
233 "|" { count(); return('|'); }
234 "?" { count(); return('?'); }
235 ^#line.*"\n" { count(); checkCurrFile(yytext); }
236 ^#pragma.*"\n" { count(); process_pragma(yytext); }
238 ^[^(]+"("[0-9]+") : error"[^\n]+ { werror(E_PRE_PROC_FAILED,yytext);count(); }
239 ^[^(]+"("[0-9]+") : warning"[^\n]+ { werror(W_PRE_PROC_WARNING,yytext);count(); }
242 [ \t\v\f] { count(); }
246 // that could have been removed by the preprocessor anyway
247 werror (W_STRAY_BACKSLASH, column);
254 int checkCurrFile ( char *s)
260 /* first check if this is a #line */
261 if ( strncmp(s,"#line",5) )
264 /* get to the line number */
271 sscanf(lineNum,"%d",&lNum);
273 /* now see if we have a file name */
274 while (*s != '\"' && *s)
277 /* if we don't have a filename then */
278 /* set the current line number to */
279 /* line number if printFlag is on */
281 lineno = mylineno = lNum ;
285 /* if we have a filename then check */
286 /* if it is "standard in" if yes then */
287 /* get the currentfile name info */
290 /* in c1mode fullSrcFileName is NULL */
291 if ( fullSrcFileName &&
292 strncmp(s,fullSrcFileName,strlen(fullSrcFileName)) == 0) {
293 lineno = mylineno = lNum;
294 currFname = fullSrcFileName ;
297 /* mark the end of the filename */
298 while (*s != '"') s++;
300 currFname = strdup (sb);
301 lineno = mylineno = lNum;
303 filename = currFname ;
313 for (i = 0; yytext[i] != '\0'; i++) {
314 if (yytext[i] == '\n') {
316 lineno = ++mylineno ;
319 if (yytext[i] == '\t')
320 column += 8 - (column % 8);
329 /* check if it is in the typedef table */
330 if (findSym(TypedefTab,NULL,yytext)) {
331 strncpyz(yylval.yychar,yytext, SDCC_NAME_MAX);
335 strncpyz (yylval.yychar,yytext, SDCC_NAME_MAX);
341 * Change by JTV 2001-05-19 to not concantenate strings
342 * to support ANSI hex and octal escape sequences in string liteals
345 char *stringLiteral()
347 #define STR_BUF_CHUNCK_LEN 1024
349 static struct dbuf_s dbuf;
353 dbuf_init(&dbuf, STR_BUF_CHUNCK_LEN);
355 dbuf_set_size(&dbuf, 0);
358 dbuf_append(&dbuf, "\"", 1);
359 /* put into the buffer till we hit the first \" */
361 while ((ch = input()) != 0) {
364 /* if it is a \ then escape char's are allowed */
367 /* \<newline> is a continuator */
374 dbuf_append(&dbuf, buf, 2); /* get the escape char, no further check */
376 break; /* carry on */
379 /* if new line we have a new line break, which is illegal */
380 werror(W_NEWLINE_IN_STRING);
381 dbuf_append(&dbuf, "\n", 1);
387 /* if this is a quote then we have work to do */
388 /* find the next non whitespace character */
389 /* if that is a double quote then carry on */
390 dbuf_append(&dbuf, "\"", 1); /* Pass end of this string or substring to evaluator */
391 while ((ch = input()) && (isspace(ch) || ch=='\\')) {
394 if ((ch = input()) != '\n') {
395 werror(W_STRAY_BACKSLASH, column);
421 dbuf_append(&dbuf, buf, 1); /* Put next substring introducer into output string */
426 return (char *)dbuf_c_str(&dbuf);
445 P_OVERLAY_ /* I had a strange conflict with P_OVERLAY while */
446 /* cross-compiling for MINGW32 with gcc 3.2 */
450 /* SAVE/RESTORE stack */
451 #define SAVE_RESTORE_SIZE 128
453 STACK_DCL(options_stack, struct options *, SAVE_RESTORE_SIZE)
454 STACK_DCL(optimize_stack, struct optimize *, SAVE_RESTORE_SIZE)
457 void doPragma (int op, char *cp)
462 struct options *optionsp;
463 struct optimize *optimizep;
465 optionsp = Safe_malloc(sizeof (struct options));
467 STACK_PUSH(options_stack, optionsp);
469 optimizep = Safe_malloc(sizeof (struct optimize));
470 *optimizep = optimize;
471 STACK_PUSH(optimize_stack, optimizep);
477 struct options *optionsp;
478 struct optimize *optimizep;
480 optionsp = STACK_POP(options_stack);
484 optimizep = STACK_POP(optimize_stack);
485 optimize = *optimizep;
486 Safe_free(optimizep);
491 optimize.loopInduction = 0 ;
495 optimize.loopInvariant = 0 ;
499 optimize.loopInduction = 1 ;
503 options.stackAuto = 1;
507 optimize.noJTabBoundary = 1;
511 optimize.global_cse = 0;
515 options.noOverlay = 1;
519 options.lessPedantic = 1;
525 /* append to the functions already listed
527 for (; options.calleeSaves[i] ;i++);
528 parseWithComma(&options.calleeSaves[i], Safe_strdup(cp));
533 parseWithComma(options.excludeRegs, Safe_strdup(cp));
541 optimize.noLoopReverse = 1;
549 int process_pragma(char *s)
551 #define NELEM(x) (sizeof (x) / sizeof (x)[0])
552 #define PRAGMA "#pragma"
554 static struct pragma_s {
559 { "RESTORE", P_RESTORE },
560 { "NOINDUCTION", P_NOINDUCTION },
561 { "NOINVARIANT", P_NOINVARIANT },
562 { "NOLOOPREVERSE", P_LOOPREV },
563 { "INDUCTION", P_INDUCTION },
564 { "STACKAUTO", P_STACKAUTO },
565 { "NOJTBOUND", P_NOJTBOUND },
566 { "NOGCSE", P_NOGCSE },
567 { "NOOVERLAY", P_NOOVERLAY },
568 { "CALLEE-SAVES", P_CALLEE_SAVES },
569 { "EXCLUDE", P_EXCLUDE },
571 { "OVERLAY", P_OVERLAY_ },
572 { "LESS_PEDANTIC", P_LESSPEDANTIC },
577 /* find the pragma */
578 while (strncmp(s, PRAGMA, (sizeof PRAGMA) - 1))
580 s += (sizeof PRAGMA) - 1;
582 /* look for the directive */
587 /* look for the end of the directive */
588 while ((!isspace(*s)) && (*s != '\n'))
591 /* First give the port a chance */
592 if (port->process_pragma && !port->process_pragma(cp))
595 for (i = 0; i < NELEM(pragma_tbl); i++) {
596 /* now compare and do what needs to be done */
597 size_t len = strlen(pragma_tbl[i].name);
599 if (strncmp(cp, pragma_tbl[i].name, len) == 0) {
600 doPragma(pragma_tbl[i].id, cp + len);
605 werror(W_UNKNOWN_PRAGMA,cp);
609 /* will return 1 if the string is a part
610 of a target specific keyword */
611 int isTargetKeyword(char *s)
615 if (port->keywords == NULL)
617 for ( i = 0 ; port->keywords[i] ; i++ ) {
618 if (strcmp(port->keywords[i],s) == 0)
625 extern int fatalError;
631 if (mylineno && filename) {
632 fprintf(stdout,"\n%s:%d: %s: token -> '%s' ; column %d\n",
637 // this comes from an empy file, no problem