Home | History | Annotate | Line # | Download | only in hexdump
parse.c revision 1.23 
      1 /*	$NetBSD: parse.c,v 1.23 2006/10/09 11:37:31 dsl Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1989, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  * 3. Neither the name of the University nor the names of its contributors
     16  *    may be used to endorse or promote products derived from this software
     17  *    without specific prior written permission.
     18  *
     19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     29  * SUCH DAMAGE.
     30  */
     31 
     32 #if HAVE_NBTOOL_CONFIG_H
     33 #include "nbtool_config.h"
     34 #endif
     35 
     36 #include <sys/cdefs.h>
     37 #if !defined(lint)
     38 #if 0
     39 static char sccsid[] = "@(#)parse.c	8.1 (Berkeley) 6/6/93";
     40 #else
     41 __RCSID("$NetBSD: parse.c,v 1.23 2006/10/09 11:37:31 dsl Exp $");
     42 #endif
     43 #endif /* not lint */
     44 
     45 #include <sys/types.h>
     46 #include <sys/file.h>
     47 
     48 #include <ctype.h>
     49 #include <err.h>
     50 #include <errno.h>
     51 #include <fcntl.h>
     52 #include <stdio.h>
     53 #include <stdlib.h>
     54 #include <string.h>
     55 #include <util.h>
     56 
     57 #include "hexdump.h"
     58 
     59 FU *endfu;					/* format at end-of-data */
     60 
     61 void
     62 addfile(char *name)
     63 {
     64 	char *p;
     65 	FILE *fp;
     66 	int ch;
     67 	char buf[2048 + 1];
     68 
     69 	if ((fp = fopen(name, "r")) == NULL)
     70 		err(1, "fopen %s", name);
     71 	while (fgets(buf, sizeof(buf), fp)) {
     72 		if (!(p = strchr(buf, '\n'))) {
     73 			warnx("line too long.");
     74 			while ((ch = getchar()) != '\n' && ch != EOF);
     75 			continue;
     76 		}
     77 		*p = '\0';
     78 		for (p = buf; *p && isspace((unsigned char)*p); ++p);
     79 		if (!*p || *p == '#')
     80 			continue;
     81 		add(p);
     82 	}
     83 	(void)fclose(fp);
     84 }
     85 
     86 void
     87 add(const char *fmt)
     88 {
     89 	const char *p;
     90 	static FS **nextfs;
     91 	FS *tfs;
     92 	FU *tfu, **nextfu;
     93 	const char *savep;
     94 
     95 	/* start new linked list of format units */
     96 	tfs = ecalloc(1, sizeof(FS));
     97 	if (!fshead)
     98 		fshead = tfs;
     99 	else
    100 		*nextfs = tfs;
    101 	nextfs = &tfs->nextfs;
    102 	nextfu = &tfs->nextfu;
    103 
    104 	/* take the format string and break it up into format units */
    105 	for (p = fmt;;) {
    106 		/* skip leading white space */
    107 		for (; isspace((unsigned char)*p); ++p);
    108 		if (!*p)
    109 			break;
    110 
    111 		/* allocate a new format unit and link it in */
    112 		tfu = ecalloc(1, sizeof(FU));
    113 		*nextfu = tfu;
    114 		nextfu = &tfu->nextfu;
    115 		tfu->reps = 1;
    116 
    117 		/* if leading digit, repetition count */
    118 		if (isdigit((unsigned char)*p)) {
    119 			for (savep = p; isdigit((unsigned char)*p); ++p);
    120 			if (!isspace((unsigned char)*p) && *p != '/')
    121 				badfmt(fmt);
    122 			/* may overwrite either white space or slash */
    123 			tfu->reps = atoi(savep);
    124 			tfu->flags = F_SETREP;
    125 			/* skip trailing white space */
    126 			for (++p; isspace((unsigned char)*p); ++p);
    127 		}
    128 
    129 		/* skip slash and trailing white space */
    130 		if (*p == '/')
    131 			while (isspace((unsigned char)*++p));
    132 
    133 		/* byte count */
    134 		if (isdigit((unsigned char)*p)) {
    135 			for (savep = p; isdigit((unsigned char)*p); ++p);
    136 			if (!isspace((unsigned char)*p))
    137 				badfmt(fmt);
    138 			tfu->bcnt = atoi(savep);
    139 			/* skip trailing white space */
    140 			for (++p; isspace((unsigned char)*p); ++p);
    141 		}
    142 
    143 		/* format */
    144 		if (*p != '"')
    145 			badfmt(fmt);
    146 		for (savep = ++p; *p != '"';)
    147 			if (*p++ == 0)
    148 				badfmt(fmt);
    149 		tfu->fmt = emalloc(p - savep + 1);
    150 		(void) strncpy(tfu->fmt, savep, p - savep);
    151 		tfu->fmt[p - savep] = '\0';
    152 		escape(tfu->fmt);
    153 		p++;
    154 	}
    155 }
    156 
    157 static const char *spec = ".#-+ 0123456789";
    158 
    159 int
    160 size(FS *fs)
    161 {
    162 	FU *fu;
    163 	int bcnt, cursize;
    164 	char *fmt;
    165 	int prec;
    166 
    167 	/* figure out the data block size needed for each format unit */
    168 	for (cursize = 0, fu = fs->nextfu; fu; fu = fu->nextfu) {
    169 		if (fu->bcnt) {
    170 			cursize += fu->bcnt * fu->reps;
    171 			continue;
    172 		}
    173 		for (bcnt = prec = 0, fmt = fu->fmt; *fmt; ++fmt) {
    174 			if (*fmt != '%')
    175 				continue;
    176 			/*
    177 			 * skip any special chars -- save precision in
    178 			 * case it's a %s format.
    179 			 */
    180 			while (strchr(spec + 1, *++fmt));
    181 			if (*fmt == '.' && isdigit((unsigned char)*++fmt)) {
    182 				prec = atoi(fmt);
    183 				while (isdigit((unsigned char)*++fmt));
    184 			}
    185 			switch(*fmt) {
    186 			case 'c':
    187 				bcnt += 1;
    188 				break;
    189 			case 'd': case 'i': case 'o': case 'u':
    190 			case 'x': case 'X':
    191 				bcnt += 4;
    192 				break;
    193 			case 'e': case 'E': case 'f': case 'g': case 'G':
    194 				bcnt += 8;
    195 				break;
    196 			case 's':
    197 				bcnt += prec;
    198 				break;
    199 			case '_':
    200 				switch(*++fmt) {
    201 				case 'c': case 'p': case 'u':
    202 					bcnt += 1;
    203 					break;
    204 				}
    205 			}
    206 		}
    207 		cursize += bcnt * fu->reps;
    208 	}
    209 	return (cursize);
    210 }
    211 
    212 void
    213 rewrite(FS *fs)
    214 {
    215 	enum { NOTOKAY, USEBCNT, USEPREC } sokay;
    216 	PR *pr, **nextpr;
    217 	FU *fu;
    218 	char *p1, *p2;
    219 	char savech, *fmtp, cs[3];
    220 	int nconv, prec;
    221 
    222 	prec = 0;
    223 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
    224 		/*
    225 		 * Break each format unit into print units; each conversion
    226 		 * character gets its own.
    227 		 */
    228 		nextpr = &fu->nextpr;
    229 		for (nconv = 0, fmtp = fu->fmt; *fmtp; nextpr = &pr->nextpr) {
    230 			pr = emalloc(sizeof *pr);
    231 			memset(pr, 0, sizeof *pr);
    232 			*nextpr = pr;
    233 
    234 			/* Skip preceding text and up to the next % sign. */
    235 			for (p1 = fmtp; *p1 && *p1 != '%'; ++p1);
    236 
    237 			/* Only text in the string. */
    238 			if (!*p1) {
    239 				pr->fmt = fmtp;
    240 				pr->flags = F_TEXT;
    241 				break;
    242 			}
    243 
    244 			/*
    245 			 * Get precision for %s -- if have a byte count, don't
    246 			 * need it.
    247 			 */
    248 			if (fu->bcnt) {
    249 				sokay = USEBCNT;
    250 				/* Skip to conversion character. */
    251 				for (++p1; *p1 && strchr(spec, *p1); ++p1);
    252 			} else {
    253 				/* Skip any special chars, field width. */
    254 				while (*++p1 && strchr(spec + 1, *p1));
    255 				if (*p1 == '.' &&
    256 				    isdigit((unsigned char)*++p1)) {
    257 					sokay = USEPREC;
    258 					prec = atoi(p1);
    259 					while (isdigit((unsigned char)*++p1))
    260 						continue;
    261 				} else
    262 					sokay = NOTOKAY;
    263 			}
    264 
    265 			p2 = *p1 ? p1 + 1 : p1;	/* Set end pointer. */
    266 			cs[0] = *p1;		/* Set conversion string. */
    267 			cs[1] = '\0';
    268 
    269 			/*
    270 			 * Figure out the byte count for each conversion;
    271 			 * rewrite the format as necessary, set up blank-
    272 			 * padding for end of data.
    273 			 */
    274 			switch(cs[0]) {
    275 			case 'c':
    276 				pr->flags = F_CHAR;
    277 				switch(fu->bcnt) {
    278 				case 0: case 1:
    279 					pr->bcnt = 1;
    280 					break;
    281 				default:
    282 					p1[1] = '\0';
    283 					badcnt(p1);
    284 				}
    285 				break;
    286 			case 'd': case 'i':
    287 				pr->flags = F_INT;
    288 				goto isint;
    289 			case 'o': case 'u': case 'x': case 'X':
    290 				pr->flags = F_UINT;
    291 isint:				cs[2] = '\0';
    292 				cs[1] = cs[0];
    293 				cs[0] = 'q';
    294 				switch(fu->bcnt) {
    295 				case 0: case 4:
    296 					pr->bcnt = 4;
    297 					break;
    298 				case 1:
    299 					pr->bcnt = 1;
    300 					break;
    301 				case 2:
    302 					pr->bcnt = 2;
    303 					break;
    304 				case 8:
    305 					pr->bcnt = 8;
    306 					break;
    307 				default:
    308 					p1[1] = '\0';
    309 					badcnt(p1);
    310 				}
    311 				break;
    312 			case 'e': case 'E': case 'f': case 'g': case 'G':
    313 				pr->flags = F_DBL;
    314 				switch(fu->bcnt) {
    315 				case 0: case 8:
    316 					pr->bcnt = 8;
    317 					break;
    318 				case 4:
    319 					pr->bcnt = 4;
    320 					break;
    321 				default:
    322 					p1[1] = '\0';
    323 					badcnt(p1);
    324 				}
    325 				break;
    326 			case 's':
    327 				pr->flags = F_STR;
    328 				switch(sokay) {
    329 				case NOTOKAY:
    330 					badsfmt();
    331 				case USEBCNT:
    332 					pr->bcnt = fu->bcnt;
    333 					break;
    334 				case USEPREC:
    335 					pr->bcnt = prec;
    336 					break;
    337 				}
    338 				break;
    339 			case '_':
    340 				++p2;
    341 				switch(p1[1]) {
    342 				case 'A':
    343 					endfu = fu;
    344 					fu->flags |= F_IGNORE;
    345 					/* FALLTHROUGH */
    346 				case 'a':
    347 					pr->flags = F_ADDRESS;
    348 					++p2;
    349 					switch(p1[2]) {
    350 					case 'd': case 'o': case'x':
    351 						cs[0] = 'q';
    352 						cs[1] = p1[2];
    353 						cs[2] = '\0';
    354 						break;
    355 					default:
    356 						p1[3] = '\0';
    357 						badconv(p1);
    358 					}
    359 					break;
    360 				case 'c':
    361 					pr->flags = F_C;
    362 					/* cs[0] = 'c';	set in conv_c */
    363 					goto isint2;
    364 				case 'p':
    365 					pr->flags = F_P;
    366 					cs[0] = 'c';
    367 					goto isint2;
    368 				case 'u':
    369 					pr->flags = F_U;
    370 					/* cs[0] = 'c';	set in conv_u */
    371 isint2:					switch(fu->bcnt) {
    372 					case 0: case 1:
    373 						pr->bcnt = 1;
    374 						break;
    375 					default:
    376 						p1[2] = '\0';
    377 						badcnt(p1);
    378 					}
    379 					break;
    380 				default:
    381 					p1[2] = '\0';
    382 					badconv(p1);
    383 				}
    384 				break;
    385 			default:
    386 				p1[1] = '\0';
    387 				badconv(p1);
    388 			}
    389 
    390 			/*
    391 			 * Copy to PR format string, set conversion character
    392 			 * pointer, update original.
    393 			 */
    394 			savech = *p2;
    395 			p1[0] = '\0';
    396 			pr->fmt = emalloc(strlen(fmtp) + strlen(cs) + 1);
    397 			(void)strcpy(pr->fmt, fmtp);
    398 			(void)strcat(pr->fmt, cs);
    399 			*p2 = savech;
    400 			pr->cchar = pr->fmt + (p1 - fmtp);
    401 			fmtp = p2;
    402 
    403 			/* Only one conversion character if byte count. */
    404 			if (!(pr->flags&F_ADDRESS) && fu->bcnt && nconv++)
    405 				errx(1,
    406 			    "byte count with multiple conversion characters");
    407 		}
    408 		/*
    409 		 * If format unit byte count not specified, figure it out
    410 		 * so can adjust rep count later.
    411 		 */
    412 		if (!fu->bcnt)
    413 			for (pr = fu->nextpr; pr; pr = pr->nextpr)
    414 				fu->bcnt += pr->bcnt;
    415 	}
    416 	/*
    417 	 * If the format string interprets any data at all, and it's
    418 	 * not the same as the blocksize, and its last format unit
    419 	 * interprets any data at all, and has no iteration count,
    420 	 * repeat it as necessary.
    421 	 *
    422 	 * If, rep count is greater than 1, no trailing whitespace
    423 	 * gets output from the last iteration of the format unit.
    424 	 */
    425 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
    426 		if (!fu->nextfu && fs->bcnt < blocksize &&
    427 		    !(fu->flags&F_SETREP) && fu->bcnt)
    428 			fu->reps += (blocksize - fs->bcnt) / fu->bcnt;
    429 		if (fu->reps > 1) {
    430 			if (!fu->nextpr)
    431 				break;
    432 			for (pr = fu->nextpr;; pr = pr->nextpr)
    433 				if (!pr->nextpr)
    434 					break;
    435 			for (p1 = pr->fmt, p2 = NULL; *p1; ++p1)
    436 				p2 = isspace((unsigned char)*p1) ? p1 : NULL;
    437 			if (p2)
    438 				pr->nospace = p2;
    439 		}
    440 	}
    441 #ifdef DEBUG
    442 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
    443 		(void)printf("fmt:");
    444 		for (pr = fu->nextpr; pr; pr = pr->nextpr)
    445 			(void)printf(" {%s}", pr->fmt);
    446 		(void)printf("\n");
    447 	}
    448 #endif
    449 }
    450 
    451 void
    452 escape(char *p1)
    453 {
    454 	char *p2;
    455 
    456 	/* alphabetic escape sequences have to be done in place */
    457 	for (p2 = p1;; ++p1, ++p2) {
    458 		if (!*p1) {
    459 			*p2 = *p1;
    460 			break;
    461 		}
    462 		if (*p1 == '\\')
    463 			switch(*++p1) {
    464 			case '\0':
    465 				*p2 = '\\';
    466 				*++p2 = '\0';
    467 				return;	/* incomplete escape sequence */
    468 			case 'a':
    469 			     /* *p2 = '\a'; */
    470 				*p2 = '\007';
    471 				break;
    472 			case 'b':
    473 				*p2 = '\b';
    474 				break;
    475 			case 'f':
    476 				*p2 = '\f';
    477 				break;
    478 			case 'n':
    479 				*p2 = '\n';
    480 				break;
    481 			case 'r':
    482 				*p2 = '\r';
    483 				break;
    484 			case 't':
    485 				*p2 = '\t';
    486 				break;
    487 			case 'v':
    488 				*p2 = '\v';
    489 				break;
    490 			default:
    491 				*p2 = *p1;
    492 				break;
    493 			}
    494 		else
    495 			*p2 = *p1;
    496 	}
    497 }
    498 
    499 void
    500 badcnt(char *s)
    501 {
    502 	errx(1, "%s: bad byte count", s);
    503 }
    504 
    505 void
    506 badsfmt(void)
    507 {
    508 	errx(1, "%%s: requires a precision or a byte count");
    509 }
    510 
    511 void
    512 badfmt(const char *fmt)
    513 {
    514 	errx(1, "\"%s\": bad format", fmt);
    515 }
    516 
    517 void
    518 badconv(char *ch)
    519 {
    520 	errx(1, "%%%s: bad conversion character", ch);
    521 }
    522