Line data Source code
1 : /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
2 : /*
3 : * This file is part of the LibreOffice project.
4 : *
5 : * This Source Code Form is subject to the terms of the Mozilla Public
6 : * License, v. 2.0. If a copy of the MPL was not distributed with this
7 : * file, You can obtain one at http://mozilla.org/MPL/2.0/.
8 : *
9 : * This file incorporates work covered by the following license notice:
10 : *
11 : * Licensed to the Apache Software Foundation (ASF) under one or more
12 : * contributor license agreements. See the NOTICE file distributed
13 : * with this work for additional information regarding copyright
14 : * ownership. The ASF licenses this file to you under the Apache
15 : * License, Version 2.0 (the "License"); you may not use this file
16 : * except in compliance with the License. You may obtain a copy of
17 : * the License at http://www.apache.org/licenses/LICENSE-2.0 .
18 : */
19 :
20 : #include <sal/types.h>
21 : #include <stdio.h>
22 : #include <ctype.h>
23 : #include <string.h>
24 : #include "cppdef.h"
25 : #include "cpp.h"
26 :
27 : /*ER evaluate macros to pDefOut */
28 :
29 : /*
30 : * skipnl() skips over input text to the end of the line.
31 : * skipws() skips over "whitespace" (spaces or tabs), but
32 : * not skip over the end of the line. It skips over
33 : * TOK_SEP, however (though that shouldn't happen).
34 : * scanid() reads the next token (C identifier) into token[].
35 : * The caller has already read the first character of
36 : * the identifier. Unlike macroid(), the token is
37 : * never expanded.
38 : * macroid() reads the next token (C identifier) into token[].
39 : * If it is a #defined macro, it is expanded, and
40 : * macroid() returns TRUE, otherwise, FALSE.
41 : * catenate() Does the dirty work of token concatenation, TRUE if it did.
42 : * scanstring() Reads a string from the input stream, calling
43 : * a user-supplied function for each character.
44 : * This function may be output() to write the
45 : * string to the output file, or save() to save
46 : * the string in the work buffer.
47 : * scannumber() Reads a C numeric constant from the input stream,
48 : * calling the user-supplied function for each
49 : * character. (output() or save() as noted above.)
50 : * save() Save one character in the work[] buffer.
51 : * savestring() Saves a string in malloc() memory.
52 : * getfile() Initialize a new FILEINFO structure, called when
53 : * #include opens a new file, or a macro is to be
54 : * expanded.
55 : * getmem() Get a specified number of bytes from malloc memory.
56 : * output() Write one character to stdout (calling PUTCHAR) --
57 : * implemented as a function so its address may be
58 : * passed to scanstring() and scannumber().
59 : * lookid() Scans the next token (identifier) from the input
60 : * stream. Looks for it in the #defined symbol table.
61 : * Returns a pointer to the definition, if found, or NULL
62 : * if not present. The identifier is stored in token[].
63 : * defnedel() Define enter/delete subroutine. Updates the
64 : * symbol table.
65 : * get() Read the next byte from the current input stream,
66 : * handling end of (macro/file) input and embedded
67 : * comments appropriately. Note that the global
68 : * instring is -- essentially -- a parameter to get().
69 : * cget() Like get(), but skip over TOK_SEP.
70 : * unget() Push last gotten character back on the input stream.
71 : * cerror(), cwarn(), cfatal(), cierror(), ciwarn()
72 : * These routines format an print messages to the user.
73 : * cerror & cwarn take a format and a single string argument.
74 : * cierror & ciwarn take a format and a single int (char) argument.
75 : * cfatal takes a format and a single string argument.
76 : */
77 :
78 : /*
79 : * This table must be rewritten for a non-Ascii machine.
80 : *
81 : * Note that several "non-visible" characters have special meaning:
82 : * Hex 1D DEF_MAGIC -- a flag to prevent #define recursion.
83 : * Hex 1E TOK_SEP -- a delimiter for token concatenation
84 : * Hex 1F COM_SEP -- a zero-width whitespace for comment concatenation
85 : */
86 : #if TOK_SEP != 0x1E || COM_SEP != 0x1F || DEF_MAGIC != 0x1D
87 : << error type table is not correct >>
88 : #endif
89 :
90 : #define DOL LET
91 :
92 :
93 : char type[256] = { /* Character type codes Hex */
94 : END, 000, 000, 000, 000, 000, 000, 000, /* 00 */
95 : 000, SPA, 000, 000, 000, 000, 000, 000, /* 08 */
96 : 000, 000, 000, 000, 000, 000, 000, 000, /* 10 */
97 : 000, 000, 000, 000, 000, LET, 000, SPA, /* 18 */
98 : SPA,OP_NOT, QUO, 000, DOL,OP_MOD,OP_AND, QUO, /* 20 !"#$%&' */
99 : OP_LPA,OP_RPA,OP_MUL,OP_ADD, 000,OP_SUB, DOT,OP_DIV, /* 28 ()*+,-./ */
100 : DIG, DIG, DIG, DIG, DIG, DIG, DIG, DIG, /* 30 01234567 */
101 : DIG, DIG,OP_COL, 000, OP_LT, OP_EQ, OP_GT,OP_QUE, /* 38 89:;<=>? */
102 : 000, LET, LET, LET, LET, LET, LET, LET, /* 40 @ABCDEFG */
103 : LET, LET, LET, LET, LET, LET, LET, LET, /* 48 HIJKLMNO */
104 : LET, LET, LET, LET, LET, LET, LET, LET, /* 50 PQRSTUVW */
105 : LET, LET, LET, 000, BSH, 000,OP_XOR, LET, /* 58 XYZ[\]^_ */
106 : 000, LET, LET, LET, LET, LET, LET, LET, /* 60 `abcdefg */
107 : LET, LET, LET, LET, LET, LET, LET, LET, /* 68 hijklmno */
108 : LET, LET, LET, LET, LET, LET, LET, LET, /* 70 pqrstuvw */
109 : LET, LET, LET, 000, OP_OR, 000,OP_NOT, 000, /* 78 xyz{|}~ */
110 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
111 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
112 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
113 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
114 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
115 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
116 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
117 : 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
118 : };
119 :
120 :
121 : /*
122 : * C P P S y m b o l T a b l e s
123 : */
124 :
125 : /*
126 : * SBSIZE defines the number of hash-table slots for the symbol table.
127 : * It must be a power of 2.
128 : */
129 : #ifndef SBSIZE
130 : #define SBSIZE 64
131 : #endif
132 : #define SBMASK (SBSIZE - 1)
133 : #if (SBSIZE ^ SBMASK) != ((SBSIZE * 2) - 1)
134 : << error, SBSIZE must be a power of 2 >>
135 : #endif
136 :
137 :
138 : static DEFBUF *symtab[SBSIZE]; /* Symbol table queue headers */
139 :
140 363 : void InitCpp6()
141 : {
142 : int i;
143 23595 : for( i = 0; i < SBSIZE; i++ )
144 23232 : symtab[ i ] = NULL;
145 363 : }
146 :
147 :
148 :
149 : /*
150 : * Skip to the end of the current input line.
151 : */
152 11979924 : void skipnl()
153 : {
154 : int c;
155 :
156 : do
157 : { /* Skip to newline */
158 11979924 : c = get();
159 : }
160 11979924 : while (c != '\n' && c != EOF_CHAR);
161 216106 : }
162 :
163 : /*
164 : * Skip over whitespace
165 : */
166 11408643 : int skipws()
167 : {
168 : int c;
169 :
170 : do { /* Skip whitespace */
171 11408643 : c = get();
172 11408643 : } while (type[c] == SPA);
173 2255914 : return (c);
174 : }
175 :
176 : /*
177 : * Get the next token (an id) into the token buffer.
178 : * Note: this code is duplicated in lookid().
179 : * Change one, change both.
180 : */
181 1150173 : void scanid(int c)
182 : {
183 : char* bp;
184 :
185 1150173 : if (c == DEF_MAGIC) /* Eat the magic token */
186 0 : c = get(); /* undefiner. */
187 1150173 : bp = token;
188 : do
189 : {
190 10156627 : if (bp < &token[IDMAX]) /* token dim is IDMAX+1 */
191 10156627 : *bp++ = (char)c;
192 10156627 : c = get();
193 : }
194 10156627 : while (type[c] == LET || type[c] == DIG);
195 1150173 : unget();
196 1150173 : *bp = EOS;
197 1150173 : }
198 :
199 : /*
200 : * If c is a letter, scan the id. if it's #defined, expand it and scan
201 : * the next character and try again.
202 : *
203 : * Else, return the character. If type[c] is a LET, the token is in token.
204 : */
205 497065 : int macroid(int c)
206 : {
207 : DEFBUF* dp;
208 :
209 497065 : if (infile != NULL && infile->fp != NULL)
210 281040 : recursion = 0;
211 1075246 : while (type[c] == LET && (dp = lookid(c)) != NULL)
212 : {
213 81116 : expand(dp);
214 81116 : c = get();
215 : }
216 497065 : return (c);
217 : }
218 :
219 : /*
220 : * A token was just read (via macroid).
221 : * If the next character is TOK_SEP, concatenate the next token
222 : * return TRUE -- which should recall macroid after refreshing
223 : * macroid's argument. If it is not TOK_SEP, unget() the character
224 : * and return FALSE.
225 : */
226 87744 : int catenate()
227 : {
228 : int c;
229 : char* token1;
230 :
231 87744 : if (get() != TOK_SEP) /* Token concatenation */
232 : {
233 87744 : unget();
234 87744 : return (FALSE);
235 : }
236 : else
237 : {
238 0 : token1 = savestring(token); /* Save first token */
239 0 : c = macroid(get()); /* Scan next token */
240 0 : switch(type[c]) /* What was it? */
241 : {
242 : case LET: /* An identifier, ... */
243 0 : if (strlen(token1) + strlen(token) >= NWORK)
244 0 : cfatal("work buffer overflow doing %s #", token1);
245 0 : sprintf(work, "%s%s", token1, token);
246 0 : break;
247 :
248 : case DIG: /* A digit string */
249 0 : strcpy(work, token1);
250 0 : workp = work + strlen(work);
251 : do
252 : {
253 0 : save(c);
254 : }
255 0 : while ((c = get()) != TOK_SEP);
256 : /*
257 : * The trailing TOK_SEP is no longer needed.
258 : */
259 0 : save(EOS);
260 0 : break;
261 :
262 : default: /* An error, ... */
263 0 : if (isprint(c))
264 0 : cierror("Strange character '%c' after #", c);
265 : else
266 0 : cierror("Strange character (%d.) after #", c);
267 0 : strcpy(work, token1);
268 0 : unget();
269 0 : break;
270 : }
271 : /*
272 : * work has the concatenated token and token1 has
273 : * the first token (no longer needed). Unget the
274 : * new (concatenated) token after freeing token1.
275 : * Finally, setup to read the new token.
276 : */
277 0 : free(token1); /* Free up memory */
278 0 : ungetstring(work); /* Unget the new thing, */
279 0 : return (TRUE);
280 : }
281 : }
282 :
283 : /*
284 : * Scan off a string. Warning if terminated by newline or EOF.
285 : * outfun() outputs the character -- to a buffer if in a macro.
286 : * TRUE if ok, FALSE if error.
287 : */
288 84111 : int scanstring(int delim,
289 : #ifndef _NO_PROTO
290 : void (*outfun)( int ) /* BP */ /* Output function */
291 : #else
292 : void (*outfun)() /* BP */
293 : #endif
294 : )
295 : {
296 : int c;
297 :
298 84111 : instring = TRUE; /* Don't strip comments */
299 84111 : (*outfun)(delim);
300 1911544 : while ((c = get()) != delim &&
301 1743322 : c != '\n' &&
302 : c != EOF_CHAR)
303 : {
304 1743322 : if (c != DEF_MAGIC)
305 1743322 : (*outfun)(c);
306 1743322 : if (c == '\\')
307 642 : (*outfun)(get());
308 : }
309 84111 : instring = FALSE;
310 84111 : if (c == delim)
311 : {
312 84111 : (*outfun)(c);
313 84111 : return (TRUE);
314 : }
315 : else
316 : {
317 0 : cerror("Unterminated string", NULLST);
318 0 : unget();
319 0 : return (FALSE);
320 : }
321 : }
322 :
323 : /*
324 : * Process a number. We know that c is from 0 to 9 or dot.
325 : * Algorithm from Dave Conroy's Decus C.
326 : */
327 489020 : void scannumber(int c,
328 : #ifndef _NO_PROTO
329 : void (*outfun)( int ) /* BP */ /* Output/store func */
330 : #else
331 : void (*outfun)() /* BP */
332 : #endif
333 : )
334 : {
335 : int radix; /* 8, 10, or 16 */
336 : int expseen; /* 'e' seen in floater */
337 : int signseen; /* '+' or '-' seen */
338 : int octal89; /* For bad octal test */
339 : int dotflag; /* TRUE if '.' was seen */
340 :
341 489020 : expseen = FALSE; /* No exponent seen yet */
342 489020 : signseen = TRUE; /* No +/- allowed yet */
343 489020 : octal89 = FALSE; /* No bad octal yet */
344 489020 : radix = 10; /* Assume decimal */
345 489020 : if ((dotflag = (c == '.')) != FALSE) /* . something? */
346 : {
347 0 : (*outfun)('.'); /* Always out the dot */
348 0 : if (type[(c = get())] != DIG) /* If not a float numb, */
349 : {
350 0 : unget(); /* Rescan strange char */
351 489020 : return; /* All done for now */
352 : }
353 : } /* End of float test */
354 489020 : else if (c == '0') /* Octal or hex? */
355 : {
356 17511 : (*outfun)(c); /* Stuff initial zero */
357 17511 : radix = 8; /* Assume it's octal */
358 17511 : c = get(); /* Look for an 'x' */
359 17511 : if (c == 'x' || c == 'X') /* Did we get one? */
360 : {
361 11084 : radix = 16; /* Remember new radix */
362 11084 : (*outfun)(c); /* Stuff the 'x' */
363 11084 : c = get(); /* Get next character */
364 : }
365 : }
366 : for (;;) /* Process curr. char. */
367 : {
368 : /*
369 : * Note that this algorithm accepts "012e4" and "03.4"
370 : * as legitimate floating-point numbers.
371 : */
372 1825753 : if (radix != 16 && (c == 'e' || c == 'E'))
373 : {
374 0 : if (expseen) /* Already saw 'E'? */
375 0 : break; /* Exit loop, bad nbr. */
376 0 : expseen = TRUE; /* Set exponent seen */
377 0 : signseen = FALSE; /* We can read '+' now */
378 0 : radix = 10; /* Decimal exponent */
379 : }
380 1825753 : else if (radix != 16 && c == '.')
381 : {
382 0 : if (dotflag) /* Saw dot already? */
383 0 : break; /* Exit loop, two dots */
384 0 : dotflag = TRUE; /* Remember the dot */
385 0 : radix = 10; /* Decimal fraction */
386 : }
387 1825753 : else if (c == '+' || c == '-') /* 1.0e+10 */
388 : {
389 4177 : if (signseen) /* Sign in wrong place? */
390 4177 : break; /* Exit loop, not nbr. */
391 : /* signseen = TRUE; */ /* Remember we saw it */
392 : }
393 : else /* Check the digit */
394 : {
395 1821576 : switch (c)
396 : {
397 : case '8': case '9': /* Sometimes wrong */
398 178981 : octal89 = TRUE; /* Do check later */
399 : case '0': case '1': case '2': case '3':
400 : case '4': case '5': case '6': case '7':
401 1328983 : break; /* Always ok */
402 :
403 : case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
404 : case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
405 7750 : if (radix == 16) /* Alpha's are ok only */
406 7750 : break; /* if reading hex. */
407 : default: /* At number end */
408 484843 : goto done; /* Break from for loop */
409 : } /* End of switch */
410 : } /* End general case */
411 1336733 : (*outfun)(c); /* Accept the character */
412 1336733 : signseen = TRUE; /* Don't read sign now */
413 1336733 : c = get(); /* Read another char */
414 1336733 : } /* End of scan loop */
415 : /*
416 : * When we break out of the scan loop, c contains the first
417 : * character (maybe) not in the number. If the number is an
418 : * integer, allow a trailing 'L' for long and/or a trailing 'U'
419 : * for unsigned. If not those, push the trailing character back
420 : * on the input stream. Floating point numbers accept a trailing
421 : * 'L' for "long double".
422 : */
423 : done:
424 489020 : if (dotflag || expseen) /* Floating point? */
425 : {
426 0 : if (c == 'l' || c == 'L')
427 : {
428 0 : (*outfun)(c);
429 0 : get(); /* Ungotten later */
430 : }
431 : }
432 : else /* Else it's an integer */
433 : {
434 : /*
435 : * We know that dotflag and expseen are both zero, now:
436 : * dotflag signals "saw 'L'", and
437 : * expseen signals "saw 'U'".
438 : */
439 : for (;;)
440 : {
441 494868 : switch (c)
442 : {
443 : case 'l':
444 : case 'L':
445 2924 : if (dotflag)
446 0 : goto nomore;
447 2924 : dotflag = TRUE;
448 2924 : break;
449 :
450 : case 'u':
451 : case 'U':
452 2924 : if (expseen)
453 0 : goto nomore;
454 2924 : expseen = TRUE;
455 2924 : break;
456 :
457 : default:
458 489020 : goto nomore;
459 : }
460 5848 : (*outfun)(c); /* Got 'L' or 'U'. */
461 5848 : c = get(); /* Look at next, too. */
462 5848 : }
463 : }
464 : nomore:
465 489020 : unget(); /* Not part of a number */
466 489020 : if (octal89 && radix == 8)
467 0 : cwarn("Illegal digit in octal number", NULLST);
468 : }
469 :
470 10581181 : void save(int c)
471 : {
472 10581181 : if (workp >= &work[NWORK])
473 : {
474 0 : work[NWORK-1] = '\0';
475 0 : cfatal("Work buffer overflow: %s", work);
476 : }
477 : else
478 10581181 : *workp++ = (char)c;
479 10581181 : }
480 :
481 : /*
482 : * Store a string into free memory.
483 : */
484 592232 : char* savestring(char* text)
485 : {
486 : char* result;
487 :
488 592232 : result = getmem(strlen(text) + 1);
489 592232 : strcpy(result, text);
490 592232 : return (result);
491 : }
492 :
493 : /*
494 : * Common FILEINFO buffer initialization for a new file or macro.
495 : */
496 85682 : FILEINFO* getfile(int bufsize, char* name)
497 : {
498 : FILEINFO* file;
499 : int size;
500 :
501 85682 : size = strlen(name); /* File/macro name */
502 85682 : file = (FILEINFO*) getmem(sizeof (FILEINFO) + bufsize + size);
503 85682 : file->parent = infile; /* Chain files together */
504 85682 : file->fp = NULL; /* No file yet */
505 85682 : file->filename = savestring(name); /* Save file/macro name */
506 85682 : file->progname = NULL; /* No #line seen yet */
507 85682 : file->unrecur = 0; /* No macro fixup */
508 85682 : file->bptr = file->buffer; /* Initialize line ptr */
509 85682 : file->buffer[0] = EOS; /* Force first read */
510 85682 : file->line = 0; /* (Not used just yet) */
511 85682 : if (infile != NULL) /* If #include file */
512 85319 : infile->line = line; /* Save current line */
513 85682 : infile = file; /* New current file */
514 85682 : line = 1; /* Note first line */
515 85682 : return (file); /* All done. */
516 : }
517 :
518 : /*
519 : * Get a block of free memory.
520 : */
521 1173647 : char* getmem(int size)
522 : {
523 : char* result;
524 :
525 1173647 : if ((result = malloc((unsigned) size)) == NULL)
526 0 : cfatal("Out of memory", NULLST);
527 1173647 : return (result);
528 : }
529 :
530 : /*
531 : * Look for the next token in the symbol table. Returns token in "token".
532 : * If found, returns the table pointer; Else returns NULL.
533 : */
534 672521 : DEFBUF* lookid(int c)
535 : {
536 : int nhash;
537 : DEFBUF* dp;
538 : char* np;
539 672521 : int temp = 0;
540 : int isrecurse; /* For #define foo foo */
541 :
542 672521 : np = token;
543 672521 : nhash = 0;
544 672521 : if (0 != (isrecurse = (c == DEF_MAGIC)))/* If recursive macro */
545 0 : c = get(); /* hack, skip DEF_MAGIC */
546 : do
547 : {
548 11542720 : if (np < &token[IDMAX]) /* token dim is IDMAX+1 */
549 : {
550 11542720 : *np++ = (char)c; /* Store token byte */
551 11542720 : nhash += c; /* Update hash value */
552 : }
553 11542720 : c = get(); /* And get another byte */
554 : }
555 11542720 : while (type[c] == LET || type[c] == DIG);
556 672521 : unget(); /* Rescan terminator */
557 672521 : *np = EOS; /* Terminate token */
558 672521 : if (isrecurse) /* Recursive definition */
559 0 : return (NULL); /* undefined just now */
560 672521 : nhash += (np - token); /* Fix hash value */
561 672521 : dp = symtab[nhash & SBMASK]; /* Starting bucket */
562 13300555 : while (dp != (DEFBUF*) NULL) /* Search symbol table */
563 : {
564 12713807 : if (dp->hash == nhash && /* Fast precheck */
565 511487 : (temp = strcmp(dp->name, token)) >= 0)
566 : {
567 246807 : break;
568 : }
569 11955513 : dp = dp->link; /* Nope, try next one */
570 : }
571 672521 : return ((temp == 0) ? dp : NULL);
572 : }
573 :
574 : /*
575 : * Enter this name in the lookup table (delete = FALSE)
576 : * or delete this name (delete = TRUE).
577 : * Returns a pointer to the define block (delete = FALSE)
578 : * Returns NULL if the symbol wasn't defined (delete = TRUE).
579 : */
580 495399 : DEFBUF* defendel(char* name, int delete)
581 : {
582 : DEFBUF* dp;
583 : DEFBUF** prevp;
584 : char* np;
585 : int nhash;
586 : int temp;
587 : int size;
588 :
589 10605803 : for (nhash = 0, np = name; *np != EOS;)
590 9615005 : nhash += *np++;
591 495399 : size = (np - name);
592 495399 : nhash += size;
593 495399 : prevp = &symtab[nhash & SBMASK];
594 8873249 : while ((dp = *prevp) != (DEFBUF*) NULL)
595 : {
596 8390921 : if (dp->hash == nhash &&
597 360099 : (temp = strcmp(dp->name, name)) >= 0)
598 : {
599 148371 : if (temp > 0)
600 148342 : dp = NULL; /* Not found */
601 : else
602 : {
603 29 : *prevp = dp->link; /* Found, unlink and */
604 29 : if (dp->repl != NULL) /* Free the replacement */
605 29 : free(dp->repl); /* if any, and then */
606 29 : free((char*) dp); /* Free the symbol */
607 29 : dp = NULL;
608 : }
609 148371 : break;
610 : }
611 7882451 : prevp = &dp->link;
612 : }
613 495399 : if (!delete)
614 : {
615 495370 : dp = (DEFBUF*) getmem(sizeof (DEFBUF) + size);
616 495370 : dp->link = *prevp;
617 495370 : *prevp = dp;
618 495370 : dp->hash = nhash;
619 495370 : dp->repl = NULL;
620 495370 : dp->nargs = 0;
621 495370 : strcpy(dp->name, name);
622 : }
623 495399 : return (dp);
624 : }
625 :
626 : #if OSL_DEBUG_LEVEL > 1
627 :
628 : void dumpdef(char* why)
629 : {
630 : DEFBUF* dp;
631 : DEFBUF** syp;
632 : FILE* pRememberOut = NULL;
633 :
634 : if ( bDumpDefs ) /*ER */
635 : {
636 : pRememberOut = pCppOut;
637 : pCppOut = pDefOut;
638 : }
639 : fprintf( pCppOut, "CPP symbol table dump %s\n", why);
640 : for (syp = symtab; syp < &symtab[SBSIZE]; syp++)
641 : {
642 : if ((dp = *syp) != (DEFBUF*) NULL)
643 : {
644 : fprintf( pCppOut, "symtab[%" SAL_PRI_PTRDIFFT "d]\n", (syp - symtab));
645 : do
646 : {
647 : dumpadef((char*) NULL, dp);
648 : }
649 : while ((dp = dp->link) != (DEFBUF*) NULL);
650 : }
651 : }
652 : if ( bDumpDefs )
653 : {
654 : fprintf( pCppOut, "\n");
655 : pCppOut = pRememberOut;
656 : }
657 : }
658 :
659 : void dumpadef(char* why, DEFBUF* dp)
660 : {
661 : char* cp;
662 : int c;
663 : FILE* pRememberOut = NULL;
664 :
665 : /*ER dump #define's to pDefOut */
666 : if ( bDumpDefs )
667 : {
668 : pRememberOut = pCppOut;
669 : pCppOut = pDefOut;
670 : }
671 : fprintf( pCppOut, " \"%s\" [%d]", dp->name, dp->nargs);
672 : if (why != NULL)
673 : fprintf( pCppOut, " (%s)", why);
674 : if (dp->repl != NULL)
675 : {
676 : fprintf( pCppOut, " => ");
677 : for (cp = dp->repl; (c = *cp++ & 0xFF) != EOS;)
678 : {
679 : #ifdef SOLAR
680 : if (c == DEL)
681 : {
682 : c = *cp++ & 0xFF;
683 : if( c == EOS ) break;
684 : fprintf( pCppOut, "<%%%d>", c - MAC_PARM);
685 : }
686 : #else
687 : if (c >= MAC_PARM && c <= (MAC_PARM + PAR_MAC))
688 : fprintf( pCppOut, "<%%%d>", c - MAC_PARM);
689 : #endif
690 : else if (isprint(c) || c == '\n' || c == '\t')
691 : PUTCHAR(c);
692 : else if (c < ' ')
693 : fprintf( pCppOut, "<^%c>", c + '@');
694 : else
695 : fprintf( pCppOut, "<\\0%o>", c);
696 : }
697 : /*ER evaluate macros to pDefOut */
698 : #ifdef EVALDEFS
699 : if ( bDumpDefs && !bIsInEval && dp->nargs <= 0 )
700 : {
701 : FILEINFO* infileSave = infile;
702 : char* tokenSave = savestring( token );
703 : char* workSave = savestring( work );
704 : int lineSave = line;
705 : int wronglineSave = wrongline;
706 : int recursionSave = recursion;
707 : FILEINFO* file;
708 : EVALTYPE valEval;
709 :
710 : bIsInEval = 1;
711 : infile = NULL; /* start from scrap */
712 : line = 0;
713 : wrongline = 0;
714 : *token = EOS;
715 : *work = EOS;
716 : recursion = 0;
717 : file = getfile( strlen( dp->repl ), dp->name );
718 : strcpy( file->buffer, dp->repl );
719 : fprintf( pCppOut, " ===> ");
720 : nEvalOff = 0;
721 : cppmain(); /* get() frees also *file */
722 : valEval = 0;
723 : if ( 0 == evaluate( EvalBuf, &valEval ) )
724 : {
725 : #ifdef EVALFLOATS
726 : if ( valEval != (EVALTYPE)((long)valEval ) )
727 : fprintf( pCppOut, " ==eval=> %f", valEval );
728 : else
729 : #endif
730 : fprintf( pCppOut, " ==eval=> %ld", (long)valEval );
731 : }
732 : recursion = recursionSave;
733 : wrongline = wronglineSave;
734 : line = lineSave;
735 : strcpy( work, workSave );
736 : free( workSave );
737 : strcpy( token, tokenSave );
738 : free( tokenSave );
739 : infile = infileSave;
740 : bIsInEval = 0;
741 : }
742 : #endif
743 : }
744 : else
745 : {
746 : fprintf( pCppOut, ", no replacement.");
747 : }
748 : PUTCHAR('\n');
749 : if ( bDumpDefs )
750 : pCppOut = pRememberOut;
751 : }
752 : #endif
753 :
754 : /*
755 : * G E T
756 : */
757 :
758 : /*
759 : * Return the next character from a macro or the current file.
760 : * Handle end of file from #include files.
761 : */
762 65390813 : int get()
763 : {
764 : int c;
765 : FILEINFO* file;
766 : int popped; /* Recursion fixup */
767 :
768 65390813 : popped = 0;
769 : get_from_file:
770 65476132 : if ((file = infile) == NULL)
771 0 : return (EOF_CHAR);
772 : newline:
773 :
774 : /*
775 : * Read a character from the current input line or macro.
776 : * At EOS, either finish the current macro (freeing temp.
777 : * storage) or read another line from the current input file.
778 : * At EOF, exit the current file (#include) or, at EOF from
779 : * the cpp input file, return EOF_CHAR to finish processing.
780 : */
781 66743242 : if ((c = *file->bptr++ & 0xFF) == EOS)
782 : {
783 : /*
784 : * Nothing in current line or macro. Get next line (if
785 : * input from a file), or do end of file/macro processing.
786 : * In the latter case, jump back to restart from the top.
787 : */
788 1310953 : if (file->fp == NULL) /* NULL if macro */
789 : {
790 81116 : popped++;
791 81116 : recursion -= file->unrecur;
792 81116 : if (recursion < 0)
793 0 : recursion = 0;
794 81116 : infile = file->parent; /* Unwind file chain */
795 : }
796 : else /* Else get from a file */
797 : {
798 1229837 : if ((file->bptr = fgets(file->buffer, NBUFF, file->fp)) != NULL)
799 : {
800 : #if OSL_DEBUG_LEVEL > 1
801 : if (debug > 1) /* Dump it to stdout */
802 : {
803 : fprintf( pCppOut, "\n#line %d (%s), %s",
804 : line, file->filename, file->buffer);
805 : }
806 : #endif
807 1225271 : goto newline; /* process the line */
808 : }
809 : else
810 : {
811 4566 : if( file->fp != stdin )
812 4566 : fclose(file->fp); /* Close finished file */
813 4566 : if ((infile = file->parent) != NULL)
814 : {
815 : /*
816 : * There is an "ungotten" newline in the current
817 : * infile buffer (set there by doinclude() in
818 : * cpp1.c). Thus, we know that the mainline code
819 : * is skipping over blank lines and will do a
820 : * #line at its convenience.
821 : */
822 4203 : wrongline = TRUE; /* Need a #line now */
823 : }
824 : }
825 : }
826 : /*
827 : * Free up space used by the (finished) file or macro and
828 : * restart input from the parent file/macro, if any.
829 : */
830 85682 : free(file->filename); /* Free name and */
831 85682 : if (file->progname != NULL) /* if a #line was seen, */
832 0 : free(file->progname); /* free it, too. */
833 85682 : free((char*) file); /* Free file space */
834 85682 : if (infile == NULL) /* If at end of file */
835 363 : return (EOF_CHAR); /* Return end of file */
836 85319 : line = infile->line; /* Reset line number */
837 85319 : goto get_from_file; /* Get from the top. */
838 : }
839 : /*
840 : * Common processing for the new character.
841 : */
842 65432289 : if (c == DEF_MAGIC && file->fp != NULL) /* Don't allow delete */
843 0 : goto newline; /* from a file */
844 65432289 : if (file->parent != NULL) /* Macro or #include */
845 : {
846 62727130 : if (popped != 0)
847 61695 : file->parent->unrecur += popped;
848 : else
849 : {
850 62665435 : recursion -= file->parent->unrecur;
851 62665435 : if (recursion < 0)
852 555 : recursion = 0;
853 62665435 : file->parent->unrecur = 0;
854 : }
855 : }
856 : #if (HOST == SYS_UNIX)
857 65432289 : if (c == '\r')
858 0 : return get(); /* DOS fuck */
859 : #endif
860 65432289 : if (c == '\n') /* Maintain current */
861 1825642 : ++line; /* line counter */
862 65432289 : if (instring) /* Strings just return */
863 12612732 : return (c); /* the character. */
864 52819557 : else if (c == '/') /* Comment? */
865 : {
866 161028 : instring = TRUE; /* So get() won't loop */
867 :
868 161028 : c = get();
869 161028 : if ((c != '*') && (c != '/')) /* Next byte '*'? */
870 : {
871 1111 : instring = FALSE; /* Nope, no comment */
872 1111 : unget(); /* Push the char. back */
873 1111 : return ('/'); /* Return the slash */
874 : }
875 159917 : if (keepcomments) /* If writing comments */
876 : {
877 0 : PUTCHAR('/'); /* Write out the */
878 : /* initializer */
879 0 : if( '*' == c )
880 0 : PUTCHAR('*');
881 : else
882 0 : PUTCHAR('/');
883 : }
884 159917 : if( '*' == c )
885 : {
886 : for (;;) /* Eat a comment */
887 : {
888 5304153 : c = get();
889 : test:
890 5425129 : if (keepcomments && c != EOF_CHAR)
891 0 : cput(c);
892 5425129 : switch (c)
893 : {
894 : case EOF_CHAR:
895 0 : cerror("EOF in comment", NULLST);
896 0 : return (EOF_CHAR);
897 :
898 : case '/':
899 42446 : if ((c = get()) != '*') /* Don't let comments */
900 42446 : goto test; /* Nest. */
901 : #ifdef STRICT_COMMENTS
902 : cwarn("Nested comments", NULLST);
903 : #endif
904 : /* Fall into * stuff */
905 : case '*':
906 102069 : if ((c = get()) != '/') /* If comment doesn't */
907 78530 : goto test; /* end, look at next */
908 23539 : instring = FALSE; /* End of comment, */
909 23539 : if (keepcomments) /* Put out the comment */
910 : {
911 0 : cput(c); /* terminator, too */
912 : }
913 : /*
914 : * A comment is syntactically "whitespace" --
915 : * however, there are certain strange sequences
916 : * such as
917 : * #define foo(x) (something)
918 : * foo|* comment *|(123)
919 : * these are '/' ^ ^
920 : * where just returning space (or COM_SEP) will cause
921 : * problems. This can be "fixed" by overwriting the
922 : * '/' in the input line buffer with ' ' (or COM_SEP)
923 : * but that may mess up an error message.
924 : * So, we peek ahead -- if the next character is
925 : * "whitespace" we just get another character, if not,
926 : * we modify the buffer. All in the name of purity.
927 : */
928 23539 : if (*file->bptr == '\n' || type[*file->bptr & 0xFF] == SPA)
929 : goto newline;
930 25 : return ((file->bptr[-1] = ' '));
931 :
932 : case '\n': /* we'll need a #line */
933 85419 : if (!keepcomments)
934 85419 : wrongline = TRUE; /* later... */
935 : default: /* Anything else is */
936 5280614 : break; /* Just a character */
937 : } /* End switch */
938 5280614 : } /* End comment loop */
939 : }
940 : else /* c++ comment */
941 : {
942 : for (;;) /* Eat a comment */
943 : {
944 5111099 : c = get();
945 5111099 : if (keepcomments && c != EOF_CHAR)
946 0 : cput(c);
947 5111099 : if( EOF_CHAR == c )
948 0 : return (EOF_CHAR);
949 5111099 : else if( '\n' == c )
950 : {
951 136378 : instring = FALSE; /* End of comment, */
952 136378 : return( c );
953 : }
954 4974721 : }
955 : }
956 : } /* End if in comment */
957 52658529 : else if (!inmacro && c == '\\') /* If backslash, peek */
958 : {
959 18325 : if ((c = get()) == '\n') /* for a <nl>. If so, */
960 : {
961 18325 : wrongline = TRUE;
962 18325 : goto newline;
963 : }
964 : else /* Backslash anything */
965 : {
966 0 : unget(); /* Get it later */
967 0 : return ('\\'); /* Return the backslash */
968 : }
969 : }
970 52640204 : else if (c == '\f' || c == VT) /* Form Feed, Vertical */
971 : {
972 0 : c = ' '; /* Tab are whitespace */
973 : }
974 52640204 : else if (c == 0xef) /* eat up UTF-8 BOM */
975 : {
976 0 : if((c = get()) == 0xbb)
977 : {
978 0 : if((c = get()) == 0xbf)
979 : {
980 0 : c = get();
981 0 : return c;
982 : }
983 : else
984 : {
985 0 : unget();
986 0 : unget();
987 0 : return 0xef;
988 : }
989 : }
990 : else
991 : {
992 0 : unget();
993 0 : return 0xef;
994 : }
995 : }
996 52640204 : return (c); /* Just return the char */
997 : }
998 :
999 : /*
1000 : * Backup the pointer to reread the last character. Fatal error
1001 : * (code bug) if we backup too far. unget() may be called,
1002 : * without problems, at end of file. Only one character may
1003 : * be ungotten. If you need to unget more, call ungetstring().
1004 : */
1005 2981527 : void unget()
1006 : {
1007 : FILEINFO* file;
1008 :
1009 2981527 : if ((file = infile) == NULL)
1010 2981527 : return; /* Unget after EOF */
1011 2981527 : if (--file->bptr < file->buffer)
1012 0 : cfatal("Too much pushback", NULLST);
1013 2981527 : if (*file->bptr == '\n') /* Ungetting a newline? */
1014 585438 : --line; /* Unget the line number, too */
1015 : }
1016 :
1017 : /*
1018 : * Push a string back on the input stream. This is done by treating
1019 : * the text as if it were a macro.
1020 : */
1021 0 : void ungetstring(char* text)
1022 : {
1023 : FILEINFO* file;
1024 0 : file = getfile(strlen(text) + 1, "");
1025 0 : strcpy(file->buffer, text);
1026 0 : }
1027 :
1028 : /*
1029 : * Get one character, absorb "funny space" after comments or
1030 : * token concatenation
1031 : */
1032 18725 : int cget()
1033 : {
1034 : int c;
1035 :
1036 : do
1037 : {
1038 18725 : c = get();
1039 : }
1040 18725 : while (c == TOK_SEP);
1041 18725 : return (c);
1042 : }
1043 :
1044 : /*
1045 : * Error messages and other hacks. The first byte of severity
1046 : * is 'S' for string arguments and 'I' for int arguments. This
1047 : * is needed for portability with machines that have int's that
1048 : * are shorter than char *'s.
1049 : */
1050 :
1051 : /*
1052 : * Print filenames, macro names, and line numbers for error messages.
1053 : */
1054 0 : static void domsg(char* severity, char* format, void* arg)
1055 : {
1056 : char* tp;
1057 : FILEINFO* file;
1058 :
1059 0 : fprintf(stderr, "%sline %d, %s: ", MSG_PREFIX, line, &severity[1]);
1060 0 : if (*severity == 'S')
1061 0 : fprintf(stderr, format, (char*)arg);
1062 : else
1063 0 : fprintf(stderr, format, *((int*)arg) );
1064 0 : putc('\n', stderr);
1065 0 : if ((file = infile) == NULL)
1066 0 : return; /* At end of file */
1067 0 : if (file->fp != NULL)
1068 : {
1069 0 : tp = file->buffer; /* Print current file */
1070 0 : fprintf(stderr, "%s", tp); /* name, making sure */
1071 0 : if (tp[strlen(tp) - 1] != '\n') /* there's a newline */
1072 0 : putc('\n', stderr);
1073 : }
1074 0 : while ((file = file->parent) != NULL) /* Print #includes, too */
1075 : {
1076 0 : if (file->fp == NULL)
1077 0 : fprintf(stderr, "from macro %s\n", file->filename);
1078 : else
1079 : {
1080 0 : tp = file->buffer;
1081 0 : fprintf(stderr, "from file %s, line %d:\n%s",
1082 0 : (file->progname != NULL)
1083 : ? file->progname : file->filename,
1084 : file->line, tp);
1085 0 : if (tp[strlen(tp) - 1] != '\n')
1086 0 : putc('\n', stderr);
1087 : }
1088 : }
1089 : }
1090 :
1091 : /*
1092 : * Print a normal error message, string argument.
1093 : */
1094 0 : void cerror(char* format, char* sarg)
1095 : {
1096 0 : domsg("SError", format, sarg);
1097 0 : errors++;
1098 0 : }
1099 :
1100 : /*
1101 : * Print a normal error message, numeric argument.
1102 : */
1103 0 : void cierror(char* format, int narg)
1104 : {
1105 0 : domsg("IError", format, &narg);
1106 0 : errors++;
1107 0 : }
1108 :
1109 : /*
1110 : * A real disaster
1111 : */
1112 0 : void cfatal(char* format, char* sarg)
1113 : {
1114 0 : domsg("SFatal error", format, sarg);
1115 0 : exit(IO_ERROR);
1116 : }
1117 :
1118 : /*
1119 : * A non-fatal error, string argument.
1120 : */
1121 0 : void cwarn(char* format, char* sarg)
1122 : {
1123 0 : domsg("SWarning", format, sarg);
1124 0 : }
1125 :
1126 : /*
1127 : * A non-fatal error, numeric argument.
1128 : */
1129 0 : void ciwarn(char* format, int narg)
1130 : {
1131 0 : domsg("IWarning", format, &narg);
1132 0 : }
1133 :
1134 : /* vim:set shiftwidth=4 softtabstop=4 expandtab: */
|
