* scripttempl/elf.sc: Use PROVIDE to define etext, edata, and end.
[external/binutils.git] / ld / ldlex.l
1 %{
2
3 /* Copyright (C) 1991, 1992, 1993, 1994 Free Software Foundation, Inc.
4
5 This file is part of GLD, the Gnu Linker.
6
7 GLD is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2, or (at your option)
10 any later version.
11
12 GLD is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License
18 along with GLD; see the file COPYING.  If not, write to
19 the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.  */
20
21 /*
22 This was written by steve chamberlain
23                     sac@cygnus.com
24 */
25
26
27 #include <ansidecl.h>
28 #include <stdio.h>
29 /* start-sanitize-mpw */
30 #ifdef MPW
31 /* Prevent enum redefinition problems. */
32 #define TRUE_FALSE_ALREADY_DEFINED
33 #endif /* MPW */
34 /* end-sanitize-mpw */
35 #include "bfd.h"
36 #include "sysdep.h"
37 #include "ld.h"
38 #include "ldgram.h"
39 #include "ldmisc.h"
40 #include "ldexp.h"
41 #include "ldlang.h"
42 #include "ldfile.h"
43 #include "ldlex.h"
44 #include "ldmain.h"
45
46 /* The type of top-level parser input.
47    yylex and yyparse (indirectly) both check this.  */
48 input_type parser_input;
49
50 /* Radix to use for bfd_scan_vma -- 0 (default to base 10) or 16.  */
51 int hex_mode;
52
53 /* Line number in the current input file.
54    (FIXME Actually, it doesn't appear to get reset for each file?)  */
55 unsigned int lineno = 1;
56
57 /* Support for flex reading from more than one input file (stream).
58    `include_stack' is flex's input state for each open file;
59    `file_name_stack' is the file names.
60
61    If `include_stack_ptr' is 0, we haven't started reading anything yet.
62    Otherwise, stack elements 0 through `include_stack_ptr - 1' are valid.  */
63
64 #undef YY_INPUT
65 #define YY_INPUT(buf,result,max_size) yy_input(buf, &result, max_size)
66
67 #define MAX_INCLUDE_DEPTH 10
68 static YY_BUFFER_STATE include_stack[MAX_INCLUDE_DEPTH];
69 static char *file_name_stack[MAX_INCLUDE_DEPTH];
70 static unsigned int include_stack_ptr = 0;
71
72 static YY_BUFFER_STATE yy_create_string_buffer PARAMS ((const char *string,
73                                                         size_t size));
74 static void yy_input PARAMS ((char *, int *result, int max_size));
75
76 static void comment PARAMS ((void));
77 static void lex_warn_invalid PARAMS ((char *where, char *what));
78
79 /* STATES 
80         EXPRESSION      definitely in an expression
81         SCRIPT          definitely in a script
82         BOTH            either EXPRESSION or SCRIPT
83         DEFSYMEXP       in an argument to -defsym
84         MRI             in an MRI script
85 */
86 #define RTOKEN(x)  {  yylval.token = x; return x; }
87
88 /* Some versions of flex want this.  */
89 #ifndef yywrap
90 int yywrap () { return 1; }
91 #endif
92 %}
93
94 %a 4000
95 %o 5000
96
97 CMDFILENAMECHAR   [_a-zA-Z0-9\/\.\\_\+\$\:\[\]\\\,\=\&\!\<\>\-\~]
98 CMDFILENAMECHAR1  [_a-zA-Z0-9\/\.\\_\+\$\:\[\]\\\,\=\&\!\<\>\~]
99 FILENAMECHAR1   [_a-zA-Z\/\.\\\$\_\~]
100 SYMBOLCHARN     [_a-zA-Z\/\.\\0-9]
101 FILENAMECHAR    [_a-zA-Z0-9\/\.\-\_\+\=\$\:\[\]\\\,\~]
102 FILENAME        {FILENAMECHAR}+
103 WHITE           [ \t\n\r]+ 
104
105 NOCFILENAMECHAR [_a-zA-Z0-9\/\.\-\_\+\$\:\[\]\\\~]
106
107
108 %s SCRIPT
109 %s EXPRESSION
110 %s BOTH
111 %s DEFSYMEXP
112 %s MRI
113 %%
114
115   if (parser_input != input_selected)
116     {
117       /* The first token of the input determines the initial parser state.  */
118       input_type t = parser_input;
119       parser_input = input_selected;
120       switch (t)
121         {
122         case input_script: return INPUT_SCRIPT; break;
123         case input_mri_script: return INPUT_MRI_SCRIPT; break;
124         case input_defsym: return INPUT_DEFSYM; break;
125         default: abort ();
126         }
127     }
128
129 <BOTH,SCRIPT,EXPRESSION>"/*"    { comment(); }
130
131
132 <DEFSYMEXP>"-"                  { RTOKEN('-');}
133 <DEFSYMEXP>"+"                  { RTOKEN('+');}
134 <DEFSYMEXP>{FILENAMECHAR1}{SYMBOLCHARN}*   { yylval.name = buystring(yytext); return NAME; }
135 <DEFSYMEXP>"="                  { RTOKEN('='); }
136
137 <MRI,EXPRESSION>"$"([0-9A-Fa-f])+ {
138                                 yylval.integer = bfd_scan_vma (yytext+1, 0,16);
139                                 return INT;
140                         }
141
142 <MRI,EXPRESSION>([0-9A-Fa-f])+(H|X|B|O|D) {
143                                    int ibase ;
144                                    switch (yytext[yyleng-1]) {
145                                     case 'X': 
146                                     case 'H':
147                                      ibase = 16;
148                                      break;
149                                     case 'O':
150                                      ibase = 8;
151                                      break;
152                                     case 'B':
153                                      ibase = 2;
154                                      break;
155                                     default:
156                                      ibase = 10;
157                                    }
158                                    yylval.integer = bfd_scan_vma (yytext+1, 0,
159                                                                   ibase);
160                                    return INT;
161                                  }
162 <SCRIPT,DEFSYMEXP,MRI,BOTH,EXPRESSION>"$"?"0x"?([0-9A-Fa-f])+(M|K|m|k)? {
163                                   yylval.integer = bfd_scan_vma (yytext, 0,
164                                                                  hex_mode);
165                                   if (yytext[yyleng-1]=='M'
166                                         || yytext[yyleng-1] == 'm') {
167                                       yylval.integer *= 1024*1024;
168                                     }   
169                                   if (yytext[yyleng-1]=='K' 
170                                 || yytext[yyleng-1]=='k') {
171                                       yylval.integer *= 1024;
172                                     }           
173                                   return INT;
174                                 }
175 <BOTH,SCRIPT,EXPRESSION>"]"             { RTOKEN(']');}
176 <BOTH,SCRIPT,EXPRESSION>"["             { RTOKEN('[');}
177 <BOTH,SCRIPT,EXPRESSION>"<<="   { RTOKEN(LSHIFTEQ);}
178 <BOTH,SCRIPT,EXPRESSION>">>="   { RTOKEN(RSHIFTEQ);}
179 <BOTH,SCRIPT,EXPRESSION>"||"    { RTOKEN(OROR);}
180 <BOTH,SCRIPT,EXPRESSION>"=="    { RTOKEN(EQ);}
181 <BOTH,SCRIPT,EXPRESSION>"!="    { RTOKEN(NE);}
182 <BOTH,SCRIPT,EXPRESSION>">="    { RTOKEN(GE);}
183 <BOTH,SCRIPT,EXPRESSION>"<="    { RTOKEN(LE);}
184 <BOTH,SCRIPT,EXPRESSION>"<<"    { RTOKEN(LSHIFT);}
185 <BOTH,SCRIPT,EXPRESSION>">>"    { RTOKEN(RSHIFT);}
186 <BOTH,SCRIPT,EXPRESSION>"+="    { RTOKEN(PLUSEQ);}
187 <BOTH,SCRIPT,EXPRESSION>"-="    { RTOKEN(MINUSEQ);}
188 <BOTH,SCRIPT,EXPRESSION>"*="    { RTOKEN(MULTEQ);}
189 <BOTH,SCRIPT,EXPRESSION>"/="    { RTOKEN(DIVEQ);}
190 <BOTH,SCRIPT,EXPRESSION>"&="    { RTOKEN(ANDEQ);}
191 <BOTH,SCRIPT,EXPRESSION>"|="    { RTOKEN(OREQ);}
192 <BOTH,SCRIPT,EXPRESSION>"&&"    { RTOKEN(ANDAND);}
193 <BOTH,SCRIPT,EXPRESSION>">"             { RTOKEN('>');}
194 <MRI,BOTH,SCRIPT,EXPRESSION>","         { RTOKEN(',');}
195 <BOTH,SCRIPT,EXPRESSION>"&"             { RTOKEN('&');}
196 <BOTH,SCRIPT,EXPRESSION>"|"             { RTOKEN('|');}
197 <BOTH,SCRIPT,EXPRESSION>"~"             { RTOKEN('~');}
198 <BOTH,SCRIPT,EXPRESSION>"!"             { RTOKEN('!');}
199 <BOTH,SCRIPT,EXPRESSION>"?"             { RTOKEN('?');}
200 <BOTH,SCRIPT,EXPRESSION>"*"             { RTOKEN('*');}
201 <BOTH,SCRIPT,EXPRESSION>"+"             { RTOKEN('+');}
202 <BOTH,SCRIPT,EXPRESSION>"-"             { RTOKEN('-');}
203 <BOTH,SCRIPT,EXPRESSION>"/"             { RTOKEN('/');}
204 <BOTH,SCRIPT,EXPRESSION>"%"             { RTOKEN('%');}
205 <BOTH,SCRIPT,EXPRESSION>"<"             { RTOKEN('<');}
206 <MRI,BOTH,SCRIPT,EXPRESSION>"="          { RTOKEN('=');}
207 <BOTH,SCRIPT,EXPRESSION>"}"                     { RTOKEN('}') ; }
208 <BOTH,SCRIPT,EXPRESSION>"{"                     { RTOKEN('{'); }
209 <BOTH,SCRIPT,EXPRESSION>")"                     { RTOKEN(')');}
210 <BOTH,SCRIPT,EXPRESSION>"("                     { RTOKEN('(');}
211 <BOTH,SCRIPT,EXPRESSION>":"             { RTOKEN(':'); }
212 <BOTH,SCRIPT,EXPRESSION>";"             { RTOKEN(';');}
213 <BOTH,SCRIPT>"MEMORY"           { RTOKEN(MEMORY);}
214 <BOTH,SCRIPT>"ORIGIN"           { RTOKEN(ORIGIN);}
215 <BOTH,SCRIPT>"BLOCK"                    { RTOKEN(BLOCK);}
216 <BOTH,SCRIPT>"LENGTH"           { RTOKEN(LENGTH);}
217 <EXPRESSION,BOTH,SCRIPT>"ALIGN"                 { RTOKEN(ALIGN_K);}
218 <EXPRESSION,BOTH,SCRIPT>"ADDR"                  { RTOKEN(ADDR);}
219 <BOTH,SCRIPT>"ENTRY"                    { RTOKEN(ENTRY);}
220 <EXPRESSION,BOTH,SCRIPT>"NEXT"                  { RTOKEN(NEXT);}
221 <EXPRESSION,BOTH,SCRIPT>"sizeof_headers"        { RTOKEN(SIZEOF_HEADERS);}
222 <EXPRESSION,BOTH,SCRIPT>"SIZEOF_HEADERS"        { RTOKEN(SIZEOF_HEADERS);}
223 <BOTH,SCRIPT>"MAP"                      { RTOKEN(MAP);}
224 <EXPRESSION,BOTH,SCRIPT>"SIZEOF"                { RTOKEN(SIZEOF);}
225 <BOTH,SCRIPT>"TARGET"           { RTOKEN(TARGET_K);}
226 <BOTH,SCRIPT>"SEARCH_DIR"               { RTOKEN(SEARCH_DIR);}
227 <BOTH,SCRIPT>"OUTPUT"           { RTOKEN(OUTPUT);}
228 <BOTH,SCRIPT>"INPUT"                    { RTOKEN(INPUT);}
229 <EXPRESSION,BOTH,SCRIPT>"DEFINED"               { RTOKEN(DEFINED);}
230 <BOTH,SCRIPT>"CREATE_OBJECT_SYMBOLS"    { RTOKEN(CREATE_OBJECT_SYMBOLS);}
231 <BOTH,SCRIPT>"CONSTRUCTORS"             { RTOKEN( CONSTRUCTORS);}
232 <BOTH,SCRIPT>"FORCE_COMMON_ALLOCATION" { RTOKEN(FORCE_COMMON_ALLOCATION);}
233 <BOTH,SCRIPT>"SECTIONS"         { RTOKEN(SECTIONS);}
234 <BOTH,SCRIPT>"FILL"                     { RTOKEN(FILL);}
235 <BOTH,SCRIPT>"STARTUP"          { RTOKEN(STARTUP);}
236 <BOTH,SCRIPT>"OUTPUT_FORMAT"            { RTOKEN(OUTPUT_FORMAT);}
237 <BOTH,SCRIPT>"OUTPUT_ARCH"              { RTOKEN( OUTPUT_ARCH);}
238 <BOTH,SCRIPT>"HLL"                      { RTOKEN(HLL);}
239 <BOTH,SCRIPT>"SYSLIB"           { RTOKEN(SYSLIB);}
240 <BOTH,SCRIPT>"FLOAT"                    { RTOKEN(FLOAT);}
241 <BOTH,SCRIPT>"QUAD"                     { RTOKEN( QUAD);}
242 <BOTH,SCRIPT>"LONG"                     { RTOKEN( LONG);}
243 <BOTH,SCRIPT>"SHORT"                    { RTOKEN( SHORT);}
244 <BOTH,SCRIPT>"BYTE"                     { RTOKEN( BYTE);}
245 <BOTH,SCRIPT>"NOFLOAT"          { RTOKEN(NOFLOAT);}
246 <EXPRESSION,BOTH,SCRIPT>"NOLOAD"                { RTOKEN(NOLOAD);}
247 <BOTH,SCRIPT>"DSECT"                    { RTOKEN(DSECT);}
248 <BOTH,SCRIPT>"COPY"                     { RTOKEN(COPY);}
249 <BOTH,SCRIPT>"INFO"                     { RTOKEN(INFO);}
250 <BOTH,SCRIPT>"OVERLAY"          { RTOKEN(OVERLAY);}
251 <BOTH,SCRIPT>"o"                        { RTOKEN(ORIGIN);}
252 <BOTH,SCRIPT>"org"                      { RTOKEN(ORIGIN);}
253 <BOTH,SCRIPT>"l"                        { RTOKEN( LENGTH);}
254 <BOTH,SCRIPT>"len"                      { RTOKEN( LENGTH);}
255 <BOTH,SCRIPT>"INCLUDE"                  { RTOKEN(INCLUDE);}
256 <EXPRESSION,BOTH,SCRIPT>"AT"                    { RTOKEN(AT);}
257 <BOTH,SCRIPT>"PROVIDE"                  { RTOKEN(PROVIDE); }
258 <MRI>"\n"                       { ++ lineno;  RTOKEN(NEWLINE); }
259 <MRI>"\r"                       { ++ lineno;  RTOKEN(NEWLINE); }
260 <MRI>"*".*                      { /* Mri comment line */ }
261 <MRI>"END"                      { RTOKEN(ENDWORD); }
262 <MRI>"ALIGNMOD"         { RTOKEN(ALIGNMOD);}
263 <MRI>"ALIGN"            { RTOKEN(ALIGN_K);}
264
265 <MRI>"CHIP"                     { RTOKEN(CHIP); }
266 <MRI>"BASE"                     { RTOKEN(BASE); }
267 <MRI>"ALIAS"                     { RTOKEN(ALIAS); }
268 <MRI>"TRUNCATE"                     { RTOKEN(TRUNCATE); }
269 <MRI>"LOAD"                     { RTOKEN(LOAD); }
270 <MRI>"PUBLIC"                   { RTOKEN(PUBLIC); }
271 <MRI>"ORDER"                    { RTOKEN(ORDER); }
272 <MRI>"NAME"                     { RTOKEN(NAMEWORD); }
273 <MRI>"FORMAT"                   { RTOKEN(FORMAT); }
274 <MRI>"LIST".*                   { RTOKEN(LIST); /* LIST and ignore to end of line */ }
275 <MRI>"SECT"                     { RTOKEN(SECT); }
276 <EXPRESSION,BOTH,SCRIPT,MRI>"ABSOLUTE"                  { RTOKEN(ABSOLUTE); }
277 <MRI>"end"                      { RTOKEN(ENDWORD); }
278 <MRI>"chip"                     { RTOKEN(CHIP); }
279 <MRI>"load"                     { RTOKEN(LOAD); }
280 <MRI>"order"                    { RTOKEN(ORDER); }
281 <MRI>"name"                     { RTOKEN(NAMEWORD); }
282 <MRI>"format"                   { RTOKEN(FORMAT); }
283 <MRI>"list".*                   { RTOKEN(LIST); /* LIST and ignore to end of line */ }
284 <MRI>"sect"                     { RTOKEN(SECT); }
285 <EXPRESSION,BOTH,SCRIPT,MRI>"absolute"                  { RTOKEN(ABSOLUTE); }
286
287 <MRI>{FILENAMECHAR1}{NOCFILENAMECHAR}*  {
288 /* Filename without commas, needed to parse mri stuff */
289                                  yylval.name = buystring(yytext); 
290                                   return NAME;
291                                 }
292
293
294 <BOTH,EXPRESSION>{FILENAMECHAR1}{FILENAMECHAR}* {
295                                  yylval.name = buystring(yytext); 
296                                   return NAME;
297                                 }
298 <SCRIPT>{FILENAMECHAR}* { yylval.name = buystring(yytext); 
299                                   return NAME;
300                                 }
301
302 <EXPRESSION,BOTH,SCRIPT>"\""[^\"]*"\"" {
303                                         /* No matter the state, quotes
304                                            give what's inside */
305                                         yylval.name = buystring(yytext+1);
306                                         yylval.name[yyleng-2] = 0;
307                                         return NAME;
308                                 }
309 <BOTH,SCRIPT,EXPRESSION>"\n"            { lineno++;}
310 <BOTH,SCRIPT,EXPRESSION>"\r"            { lineno++;}
311 <MRI,BOTH,SCRIPT,EXPRESSION>[ \t]
312
313 <<EOF>> {
314   include_stack_ptr--;
315     
316   if (include_stack_ptr == 0) 
317   {
318     yyterminate();
319   }
320   else 
321   {
322     yy_switch_to_buffer(include_stack[include_stack_ptr]);
323
324   }
325   BEGIN(SCRIPT);
326   ldfile_input_filename = file_name_stack[include_stack_ptr-1];
327
328   return END;
329 }
330
331 <SCRIPT,MRI>.           lex_warn_invalid(" in script", yytext);
332 <EXPRESSION,DEFSYMEXP,BOTH>.    lex_warn_invalid(" in expression", yytext);
333     
334 %%
335 \f
336
337 /* Switch flex to reading script file NAME, open on FILE,
338    saving the current input info on the include stack.  */
339
340 void
341 lex_push_file (file, name)
342      FILE *file;
343      char *name;
344 {
345   if (include_stack_ptr >= MAX_INCLUDE_DEPTH) 
346     {
347       einfo("%F:includes nested too deeply\n");
348     }
349   file_name_stack[include_stack_ptr] = name;
350   include_stack[include_stack_ptr] = YY_CURRENT_BUFFER;
351
352   include_stack_ptr++;
353   yyin = file;
354   yy_switch_to_buffer(yy_create_buffer(yyin, YY_BUF_SIZE));
355   BEGIN (SCRIPT);
356 }
357
358 /* Return a newly created flex input buffer containing STRING,
359    which is SIZE bytes long.  */
360
361 static YY_BUFFER_STATE 
362 yy_create_string_buffer (string, size)
363      CONST char *string;
364      size_t size;
365 {
366   YY_BUFFER_STATE b;
367
368   /* Calls to m-alloc get turned by sed into xm-alloc.  */
369   b = (YY_BUFFER_STATE) malloc (sizeof (struct yy_buffer_state));
370   b->yy_input_file = 0;
371   b->yy_buf_size = size;
372
373   /* yy_ch_buf has to be 2 characters longer than the size given because
374      we need to put in 2 end-of-buffer characters.  */
375   b->yy_ch_buf = (YY_CHAR *) malloc ((unsigned) (b->yy_buf_size + 3));
376
377   b->yy_ch_buf[0] = '\n';
378   strcpy (b->yy_ch_buf+1, string);
379   b->yy_ch_buf[size+1] = YY_END_OF_BUFFER_CHAR;
380   b->yy_ch_buf[size+2] = YY_END_OF_BUFFER_CHAR;
381   b->yy_n_chars = size+1;
382   b->yy_buf_pos = &b->yy_ch_buf[1];
383   b->yy_eof_status = EOF_NOT_SEEN;
384
385   return b;
386 }
387
388 /* Switch flex to reading from STRING, saving the current input info
389    on the include stack.  */
390
391 void
392 lex_redirect (string)
393      CONST char *string;
394 {
395   YY_BUFFER_STATE tmp;
396
397   yy_init = 0;
398   if (include_stack_ptr >= MAX_INCLUDE_DEPTH) 
399     {
400       einfo("%F: macros nested too deeply\n");
401     }
402   file_name_stack[include_stack_ptr] = "redirect";
403   include_stack[include_stack_ptr] = YY_CURRENT_BUFFER;
404   include_stack_ptr++;
405   tmp = yy_create_string_buffer (string, strlen (string));
406   yy_switch_to_buffer (tmp);
407   BEGIN (SCRIPT);
408 }
409 \f
410 /* Functions to switch to a different flex start condition,
411    saving the current start condition on `state_stack'.  */
412
413 static int state_stack[MAX_INCLUDE_DEPTH * 2];
414 static int *state_stack_p = state_stack;
415
416 void
417 ldlex_script ()
418 {
419   *(state_stack_p)++ = yy_start;
420   BEGIN (SCRIPT);
421 }
422
423 void
424 ldlex_mri_script ()
425 {
426   *(state_stack_p)++ = yy_start;
427   BEGIN (MRI);
428 }
429
430 void
431 ldlex_defsym ()
432 {
433   *(state_stack_p)++ = yy_start;
434   BEGIN (DEFSYMEXP);
435 }
436            
437 void
438 ldlex_expression ()
439 {
440   *(state_stack_p)++ = yy_start;
441   BEGIN (EXPRESSION);
442 }
443
444 void
445 ldlex_both ()
446 {
447   *(state_stack_p)++ = yy_start;
448   BEGIN (BOTH);
449 }
450
451 void
452 ldlex_popstate ()
453 {
454   yy_start = *(--state_stack_p);
455 }
456 \f
457
458 /* Place up to MAX_SIZE characters in BUF and return in *RESULT
459    either the number of characters read, or 0 to indicate EOF.  */
460
461 static void
462 yy_input (buf, result, max_size)
463      char *buf;
464      int *result;
465      int max_size;
466 {
467   *result = 0; 
468   if (yy_current_buffer->yy_input_file)
469     {
470       if (yyin)
471         {
472           *result = read (fileno (yyin), (char *) buf, max_size);
473           if (*result < 0) 
474             einfo ("%F%P: read in flex scanner failed");
475         }
476     }
477 }
478
479 /* Eat the rest of a C-style comment.  */
480
481 static void
482 comment ()
483 {
484   int c;
485
486   while (1)
487   {
488     c = input();
489     while (c != '*' && c != EOF) 
490     {
491       if (c == '\n' || c == '\r')
492         lineno++;
493       c = input();
494     }
495
496     if (c == '*')
497     {
498       c = input();
499       while (c == '*')
500        c = input();
501       if (c == '/')
502        break;                   /* found the end */
503     }
504
505     if (c == '\n' || c == '\r')
506       lineno++;
507
508     if (c == EOF)
509     {
510       einfo( "%F%P: EOF in comment\n");
511       break;
512     }
513   }
514 }
515
516 /* Warn the user about a garbage character WHAT in the input
517    in context WHERE.  */
518
519 static void
520 lex_warn_invalid (where, what)
521      char *where, *what;
522 {
523   fprintf(stderr,
524           "%s: ignoring invalid character `%s'%s\n",
525           program_name, what, where);
526 }