]> git.ipfire.org Git - thirdparty/binutils-gdb.git/blob - gdb/ada-lex.l
* doublest.h (DOUBLEST_PRINT_FORMAT): Remove % from string.
[thirdparty/binutils-gdb.git] / gdb / ada-lex.l
1 /* FLEX lexer for Ada expressions, for GDB.
2 Copyright (C) 1994, 1997, 1998, 2000, 2001, 2002, 2003, 2007, 2008
3 Free Software Foundation, Inc.
4
5 This file is part of GDB.
6
7 This program 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 of the License, or
10 (at your option) any later version.
11
12 This program 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 this program; if not, write to the Free Software
19 Foundation, Inc., 51 Franklin Street, Fifth Floor,
20 Boston, MA 02110-1301, USA. */
21
22 /*----------------------------------------------------------------------*/
23
24 /* The converted version of this file is to be included in ada-exp.y, */
25 /* the Ada parser for gdb. The function yylex obtains characters from */
26 /* the global pointer lexptr. It returns a syntactic category for */
27 /* each successive token and places a semantic value into yylval */
28 /* (ada-lval), defined by the parser. */
29
30 DIG [0-9]
31 NUM10 ({DIG}({DIG}|_)*)
32 HEXDIG [0-9a-f]
33 NUM16 ({HEXDIG}({HEXDIG}|_)*)
34 OCTDIG [0-7]
35 LETTER [a-z_]
36 ID ({LETTER}({LETTER}|{DIG})*|"<"{LETTER}({LETTER}|{DIG})*">")
37 WHITE [ \t\n]
38 TICK ("'"{WHITE}*)
39 GRAPHIC [a-z0-9 #&'()*+,-./:;<>=_|!$%?@\[\]\\^`{}~]
40 OPER ([-+*/=<>&]|"<="|">="|"**"|"/="|"and"|"or"|"xor"|"not"|"mod"|"rem"|"abs")
41
42 EXP (e[+-]{NUM10})
43 POSEXP (e"+"?{NUM10})
44
45 %{
46
47 #define NUMERAL_WIDTH 256
48 #define LONGEST_SIGN ((ULONGEST) 1 << (sizeof(LONGEST) * HOST_CHAR_BIT - 1))
49
50 /* Temporary staging for numeric literals. */
51 static char numbuf[NUMERAL_WIDTH];
52 static void canonicalizeNumeral (char *s1, const char *);
53 static struct stoken processString (const char*, int);
54 static int processInt (const char *, const char *, const char *);
55 static int processReal (const char *);
56 static struct stoken processId (const char *, int);
57 static int processAttribute (const char *);
58 static int find_dot_all (const char *);
59
60 #undef YY_DECL
61 #define YY_DECL static int yylex ( void )
62
63 #undef YY_INPUT
64 #define YY_INPUT(BUF, RESULT, MAX_SIZE) \
65 if ( *lexptr == '\000' ) \
66 (RESULT) = YY_NULL; \
67 else \
68 { \
69 *(BUF) = *lexptr; \
70 (RESULT) = 1; \
71 lexptr += 1; \
72 }
73
74 static int find_dot_all (const char *);
75
76 %}
77
78 %option case-insensitive interactive nodefault
79
80 %s BEFORE_QUAL_QUOTE
81
82 %%
83
84 {WHITE} { }
85
86 "--".* { yyterminate(); }
87
88 {NUM10}{POSEXP} {
89 canonicalizeNumeral (numbuf, yytext);
90 return processInt (NULL, numbuf, strrchr(numbuf, 'e')+1);
91 }
92
93 {NUM10} {
94 canonicalizeNumeral (numbuf, yytext);
95 return processInt (NULL, numbuf, NULL);
96 }
97
98 {NUM10}"#"{HEXDIG}({HEXDIG}|_)*"#"{POSEXP} {
99 canonicalizeNumeral (numbuf, yytext);
100 return processInt (numbuf,
101 strchr (numbuf, '#') + 1,
102 strrchr(numbuf, '#') + 1);
103 }
104
105 {NUM10}"#"{HEXDIG}({HEXDIG}|_)*"#" {
106 canonicalizeNumeral (numbuf, yytext);
107 return processInt (numbuf, strchr (numbuf, '#') + 1, NULL);
108 }
109
110 "0x"{HEXDIG}+ {
111 canonicalizeNumeral (numbuf, yytext+2);
112 return processInt ("16#", numbuf, NULL);
113 }
114
115
116 {NUM10}"."{NUM10}{EXP} {
117 canonicalizeNumeral (numbuf, yytext);
118 return processReal (numbuf);
119 }
120
121 {NUM10}"."{NUM10} {
122 canonicalizeNumeral (numbuf, yytext);
123 return processReal (numbuf);
124 }
125
126 {NUM10}"#"{NUM16}"."{NUM16}"#"{EXP} {
127 error (_("Based real literals not implemented yet."));
128 }
129
130 {NUM10}"#"{NUM16}"."{NUM16}"#" {
131 error (_("Based real literals not implemented yet."));
132 }
133
134 <INITIAL>"'"({GRAPHIC}|\")"'" {
135 yylval.typed_val.type = type_char ();
136 yylval.typed_val.val = yytext[1];
137 return CHARLIT;
138 }
139
140 <INITIAL>"'[\""{HEXDIG}{2}"\"]'" {
141 int v;
142 yylval.typed_val.type = type_char ();
143 sscanf (yytext+3, "%2x", &v);
144 yylval.typed_val.val = v;
145 return CHARLIT;
146 }
147
148 \"({GRAPHIC}|"[\""({HEXDIG}{2}|\")"\"]")*\" {
149 yylval.sval = processString (yytext+1, yyleng-2);
150 return STRING;
151 }
152
153 \" {
154 error (_("ill-formed or non-terminated string literal"));
155 }
156
157
158 if {
159 while (*lexptr != 'i' && *lexptr != 'I')
160 lexptr -= 1;
161 yyrestart(NULL);
162 return 0;
163 }
164
165 /* ADA KEYWORDS */
166
167 abs { return ABS; }
168 and { return _AND_; }
169 else { return ELSE; }
170 in { return IN; }
171 mod { return MOD; }
172 new { return NEW; }
173 not { return NOT; }
174 null { return NULL_PTR; }
175 or { return OR; }
176 others { return OTHERS; }
177 rem { return REM; }
178 then { return THEN; }
179 xor { return XOR; }
180
181 /* ATTRIBUTES */
182
183 {TICK}[a-zA-Z][a-zA-Z]+ { return processAttribute (yytext+1); }
184
185 /* PUNCTUATION */
186
187 "=>" { return ARROW; }
188 ".." { return DOTDOT; }
189 "**" { return STARSTAR; }
190 ":=" { return ASSIGN; }
191 "/=" { return NOTEQUAL; }
192 "<=" { return LEQ; }
193 ">=" { return GEQ; }
194
195 <BEFORE_QUAL_QUOTE>"'" { BEGIN INITIAL; return '\''; }
196
197 [-&*+./:<>=|;\[\]] { return yytext[0]; }
198
199 "," { if (paren_depth == 0 && comma_terminates)
200 {
201 lexptr -= 1;
202 yyrestart(NULL);
203 return 0;
204 }
205 else
206 return ',';
207 }
208
209 "(" { paren_depth += 1; return '('; }
210 ")" { if (paren_depth == 0)
211 {
212 lexptr -= 1;
213 yyrestart(NULL);
214 return 0;
215 }
216 else
217 {
218 paren_depth -= 1;
219 return ')';
220 }
221 }
222
223 "."{WHITE}*all { return DOT_ALL; }
224
225 "."{WHITE}*{ID} {
226 yylval.sval = processId (yytext+1, yyleng-1);
227 return DOT_ID;
228 }
229
230 {ID}({WHITE}*"."{WHITE}*({ID}|\"{OPER}\"))*(" "*"'")? {
231 int all_posn = find_dot_all (yytext);
232
233 if (all_posn == -1 && yytext[yyleng-1] == '\'')
234 {
235 BEGIN BEFORE_QUAL_QUOTE;
236 yyless (yyleng-1);
237 }
238 else if (all_posn >= 0)
239 yyless (all_posn);
240 yylval.sval = processId (yytext, yyleng);
241 return NAME;
242 }
243
244
245 /* GDB EXPRESSION CONSTRUCTS */
246
247 "'"[^']+"'"{WHITE}*:: {
248 yyless (yyleng - 2);
249 yylval.sval = processId (yytext, yyleng);
250 return NAME;
251 }
252
253 "::" { return COLONCOLON; }
254
255 [{}@] { return yytext[0]; }
256
257 /* REGISTERS AND GDB CONVENIENCE VARIABLES */
258
259 "$"({LETTER}|{DIG}|"$")* {
260 yylval.sval.ptr = yytext;
261 yylval.sval.length = yyleng;
262 return SPECIAL_VARIABLE;
263 }
264
265 /* CATCH-ALL ERROR CASE */
266
267 . { error (_("Invalid character '%s' in expression."), yytext); }
268 %%
269
270 #include <ctype.h>
271 #include "gdb_string.h"
272
273 /* Initialize the lexer for processing new expression. */
274
275 void
276 lexer_init (FILE *inp)
277 {
278 BEGIN INITIAL;
279 yyrestart (inp);
280 }
281
282
283 /* Copy S2 to S1, removing all underscores, and downcasing all letters. */
284
285 static void
286 canonicalizeNumeral (char *s1, const char *s2)
287 {
288 for (; *s2 != '\000'; s2 += 1)
289 {
290 if (*s2 != '_')
291 {
292 *s1 = tolower(*s2);
293 s1 += 1;
294 }
295 }
296 s1[0] = '\000';
297 }
298
299 /* Interprets the prefix of NUM that consists of digits of the given BASE
300 as an integer of that BASE, with the string EXP as an exponent.
301 Puts value in yylval, and returns INT, if the string is valid. Causes
302 an error if the number is improperly formated. BASE, if NULL, defaults
303 to "10", and EXP to "1". The EXP does not contain a leading 'e' or 'E'.
304 */
305
306 static int
307 processInt (const char *base0, const char *num0, const char *exp0)
308 {
309 ULONGEST result;
310 long exp;
311 int base;
312
313 char *trailer;
314
315 if (base0 == NULL)
316 base = 10;
317 else
318 {
319 base = strtol (base0, (char **) NULL, 10);
320 if (base < 2 || base > 16)
321 error (_("Invalid base: %d."), base);
322 }
323
324 if (exp0 == NULL)
325 exp = 0;
326 else
327 exp = strtol(exp0, (char **) NULL, 10);
328
329 errno = 0;
330 result = strtoulst (num0, (const char **) &trailer, base);
331 if (errno == ERANGE)
332 error (_("Integer literal out of range"));
333 if (isxdigit(*trailer))
334 error (_("Invalid digit `%c' in based literal"), *trailer);
335
336 while (exp > 0)
337 {
338 if (result > (ULONG_MAX / base))
339 error (_("Integer literal out of range"));
340 result *= base;
341 exp -= 1;
342 }
343
344 if ((result >> (gdbarch_int_bit (current_gdbarch)-1)) == 0)
345 yylval.typed_val.type = type_int ();
346 else if ((result >> (gdbarch_long_bit (current_gdbarch)-1)) == 0)
347 yylval.typed_val.type = type_long ();
348 else if (((result >> (gdbarch_long_bit (current_gdbarch)-1)) >> 1) == 0)
349 {
350 /* We have a number representable as an unsigned integer quantity.
351 For consistency with the C treatment, we will treat it as an
352 anonymous modular (unsigned) quantity. Alas, the types are such
353 that we need to store .val as a signed quantity. Sorry
354 for the mess, but C doesn't officially guarantee that a simple
355 assignment does the trick (no, it doesn't; read the reference manual).
356 */
357 yylval.typed_val.type = builtin_type_unsigned_long;
358 if (result & LONGEST_SIGN)
359 yylval.typed_val.val =
360 (LONGEST) (result & ~LONGEST_SIGN)
361 - (LONGEST_SIGN>>1) - (LONGEST_SIGN>>1);
362 else
363 yylval.typed_val.val = (LONGEST) result;
364 return INT;
365 }
366 else
367 yylval.typed_val.type = type_long_long ();
368
369 yylval.typed_val.val = (LONGEST) result;
370 return INT;
371 }
372
373 static int
374 processReal (const char *num0)
375 {
376 sscanf (num0, "%" DOUBLEST_SCAN_FORMAT, &yylval.typed_val_float.dval);
377
378 yylval.typed_val_float.type = type_float ();
379 if (sizeof(DOUBLEST) >= gdbarch_double_bit (current_gdbarch)
380 / TARGET_CHAR_BIT)
381 yylval.typed_val_float.type = type_double ();
382 if (sizeof(DOUBLEST) >= gdbarch_long_double_bit (current_gdbarch)
383 / TARGET_CHAR_BIT)
384 yylval.typed_val_float.type = type_long_double ();
385
386 return FLOAT;
387 }
388
389
390 /* Store a canonicalized version of NAME0[0..LEN-1] in yylval.ssym. The
391 resulting string is valid until the next call to ada_parse. It differs
392 from NAME0 in that:
393 + Characters between '...' or <...> are transfered verbatim to
394 yylval.ssym.
395 + <, >, and trailing "'" characters in quoted sequences are removed
396 (a leading quote is preserved to indicate that the name is not to be
397 GNAT-encoded).
398 + Unquoted whitespace is removed.
399 + Unquoted alphabetic characters are mapped to lower case.
400 Result is returned as a struct stoken, but for convenience, the string
401 is also null-terminated. Result string valid until the next call of
402 ada_parse.
403 */
404 static struct stoken
405 processId (const char *name0, int len)
406 {
407 char *name = obstack_alloc (&temp_parse_space, len + 11);
408 int i0, i;
409 struct stoken result;
410
411 while (len > 0 && isspace (name0[len-1]))
412 len -= 1;
413 i = i0 = 0;
414 while (i0 < len)
415 {
416 if (isalnum (name0[i0]))
417 {
418 name[i] = tolower (name0[i0]);
419 i += 1; i0 += 1;
420 }
421 else switch (name0[i0])
422 {
423 default:
424 name[i] = name0[i0];
425 i += 1; i0 += 1;
426 break;
427 case ' ': case '\t':
428 i0 += 1;
429 break;
430 case '\'':
431 do
432 {
433 name[i] = name0[i0];
434 i += 1; i0 += 1;
435 }
436 while (i0 < len && name0[i0] != '\'');
437 i0 += 1;
438 break;
439 case '<':
440 i0 += 1;
441 while (i0 < len && name0[i0] != '>')
442 {
443 name[i] = name0[i0];
444 i += 1; i0 += 1;
445 }
446 i0 += 1;
447 break;
448 }
449 }
450 name[i] = '\000';
451
452 result.ptr = name;
453 result.length = i;
454 return result;
455 }
456
457 /* Return TEXT[0..LEN-1], a string literal without surrounding quotes,
458 with special hex character notations replaced with characters.
459 Result valid until the next call to ada_parse. */
460
461 static struct stoken
462 processString (const char *text, int len)
463 {
464 const char *p;
465 char *q;
466 const char *lim = text + len;
467 struct stoken result;
468
469 q = result.ptr = obstack_alloc (&temp_parse_space, len);
470 p = text;
471 while (p < lim)
472 {
473 if (p[0] == '[' && p[1] == '"' && p+2 < lim)
474 {
475 if (p[2] == '"') /* "...["""]... */
476 {
477 *q = '"';
478 p += 4;
479 }
480 else
481 {
482 int chr;
483 sscanf (p+2, "%2x", &chr);
484 *q = (char) chr;
485 p += 5;
486 }
487 }
488 else
489 *q = *p;
490 q += 1;
491 p += 1;
492 }
493 result.length = q - result.ptr;
494 return result;
495 }
496
497 /* Returns the position within STR of the '.' in a
498 '.{WHITE}*all' component of a dotted name, or -1 if there is none.
499 Note: we actually don't need this routine, since 'all' can never be an
500 Ada identifier. Thus, looking up foo.all or foo.all.x as a name
501 must fail, and will eventually be interpreted as (foo).all or
502 (foo).all.x. However, this does avoid an extraneous lookup. */
503
504 static int
505 find_dot_all (const char *str)
506 {
507 int i;
508 for (i = 0; str[i] != '\000'; i += 1)
509 {
510 if (str[i] == '.')
511 {
512 int i0 = i;
513 do
514 i += 1;
515 while (isspace (str[i]));
516 if (strncmp (str+i, "all", 3) == 0
517 && ! isalnum (str[i+3]) && str[i+3] != '_')
518 return i0;
519 }
520 }
521 return -1;
522 }
523
524 /* Returns non-zero iff string SUBSEQ matches a subsequence of STR, ignoring
525 case. */
526
527 static int
528 subseqMatch (const char *subseq, const char *str)
529 {
530 if (subseq[0] == '\0')
531 return 1;
532 else if (str[0] == '\0')
533 return 0;
534 else if (tolower (subseq[0]) == tolower (str[0]))
535 return subseqMatch (subseq+1, str+1) || subseqMatch (subseq, str+1);
536 else
537 return subseqMatch (subseq, str+1);
538 }
539
540
541 static struct { const char *name; int code; }
542 attributes[] = {
543 { "address", TICK_ADDRESS },
544 { "unchecked_access", TICK_ACCESS },
545 { "unrestricted_access", TICK_ACCESS },
546 { "access", TICK_ACCESS },
547 { "first", TICK_FIRST },
548 { "last", TICK_LAST },
549 { "length", TICK_LENGTH },
550 { "max", TICK_MAX },
551 { "min", TICK_MIN },
552 { "modulus", TICK_MODULUS },
553 { "pos", TICK_POS },
554 { "range", TICK_RANGE },
555 { "size", TICK_SIZE },
556 { "tag", TICK_TAG },
557 { "val", TICK_VAL },
558 { NULL, -1 }
559 };
560
561 /* Return the syntactic code corresponding to the attribute name or
562 abbreviation STR. */
563
564 static int
565 processAttribute (const char *str)
566 {
567 int i, k;
568
569 for (i = 0; attributes[i].code != -1; i += 1)
570 if (strcasecmp (str, attributes[i].name) == 0)
571 return attributes[i].code;
572
573 for (i = 0, k = -1; attributes[i].code != -1; i += 1)
574 if (subseqMatch (str, attributes[i].name))
575 {
576 if (k == -1)
577 k = i;
578 else
579 error (_("ambiguous attribute name: `%s'"), str);
580 }
581 if (k == -1)
582 error (_("unrecognized attribute: `%s'"), str);
583
584 return attributes[k].code;
585 }
586
587 int
588 yywrap(void)
589 {
590 return 1;
591 }
592
593 /* Dummy definition to suppress warnings about unused static definitions. */
594 typedef void (*dummy_function) ();
595 dummy_function ada_flex_use[] =
596 {
597 (dummy_function) yyunput
598 };