xref: /openbsd-src/usr.bin/hexdump/parse.c (revision f2da64fbbbf1b03f09f390ab01267c93dfd77c4c)
1 /*	$OpenBSD: parse.c,v 1.22 2016/09/04 16:41:43 tb Exp $	*/
2 /*	$NetBSD: parse.c,v 1.12 2001/12/07 13:37:39 bjh21 Exp $	*/
3 
4 /*
5  * Copyright (c) 1989, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. Neither the name of the University nor the names of its contributors
17  *    may be used to endorse or promote products derived from this software
18  *    without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  */
32 
33 #include <sys/types.h>
34 #include <sys/file.h>
35 
36 #include <ctype.h>
37 #include <err.h>
38 #include <errno.h>
39 #include <fcntl.h>
40 #include <stdio.h>
41 #include <stdlib.h>
42 #include <string.h>
43 
44 #include "hexdump.h"
45 
46 FU *endfu;					/* format at end-of-data */
47 
48 static __dead void	 badcnt(char *);
49 static __dead void	 badconv(char *);
50 static __dead void	 badfmt(const char *);
51 static __dead void	 badsfmt(void);
52 static void		 escape(char *);
53 
54 void
55 addfile(char *name)
56 {
57 	FILE *fp;
58 	size_t len;
59 	char *buf, *lbuf, *p;
60 
61 	if ((fp = fopen(name, "r")) == NULL)
62 		err(1, "fopen %s", name);
63 
64 	lbuf = NULL;
65 	while ((buf = fgetln(fp, &len))) {
66 		if (buf[len - 1] == '\n')
67 			buf[len - 1] = '\0';
68 		else {
69 			/* EOF without EOL, copy and add the NUL */
70 			if ((lbuf = malloc(len + 1)) == NULL)
71 				err(1, NULL);
72 			memcpy(lbuf, buf, len);
73 			lbuf[len] = '\0';
74 			buf = lbuf;
75 		}
76 		for (p = buf; isspace((unsigned char)*p); ++p);
77 		if (!*p || *p == '#')
78 			continue;
79 		add(p);
80 	}
81 	free(lbuf);
82 	(void)fclose(fp);
83 }
84 
85 void
86 add(const char *fmt)
87 {
88 	const char *p;
89 	static FS **nextfs;
90 	FS *tfs;
91 	FU *tfu, **nextfu;
92 	const char *savep;
93 
94 	/* start new linked list of format units */
95 	if ((tfs = calloc(1, sizeof(FS))) == NULL)
96 		err(1, NULL);
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 		if ((tfu = calloc(1, sizeof(FU))) == NULL)
113 			err(1, NULL);
114 		*nextfu = tfu;
115 		nextfu = &tfu->nextfu;
116 		tfu->reps = 1;
117 
118 		/* if leading digit, repetition count */
119 		if (isdigit((unsigned char)*p)) {
120 			for (savep = p; isdigit((unsigned char)*p); ++p);
121 			if (!isspace((unsigned char)*p) && *p != '/')
122 				badfmt(fmt);
123 			/* may overwrite either white space or slash */
124 			tfu->reps = atoi(savep);
125 			tfu->flags = F_SETREP;
126 			/* skip trailing white space */
127 			for (++p; isspace((unsigned char)*p); ++p);
128 		}
129 
130 		/* skip slash and trailing white space */
131 		if (*p == '/')
132 			while (isspace((unsigned char)*++p));
133 
134 		/* byte count */
135 		if (isdigit((unsigned char)*p)) {
136 			for (savep = p; isdigit((unsigned char)*p); ++p);
137 			if (!isspace((unsigned char)*p))
138 				badfmt(fmt);
139 			tfu->bcnt = atoi(savep);
140 			/* skip trailing white space */
141 			for (++p; isspace((unsigned char)*p); ++p);
142 		}
143 
144 		/* format */
145 		if (*p != '"')
146 			badfmt(fmt);
147 		for (savep = ++p; *p != '"';)
148 			if (*p++ == 0)
149 				badfmt(fmt);
150 		if ((tfu->fmt = strndup(savep, p - savep)) == NULL)
151 			err(1, NULL);
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 (*++fmt && 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[4];
220 	int nconv, prec;
221 
222 	nextpr = NULL;
223 	prec = 0;
224 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
225 		/*
226 		 * Break each format unit into print units; each conversion
227 		 * character gets its own.
228 		 */
229 		for (nconv = 0, fmtp = fu->fmt; *fmtp; nextpr = &pr->nextpr) {
230 			if ((pr = calloc(1, sizeof(PR))) == NULL)
231 				err(1, NULL);
232 			if (!fu->nextpr)
233 				fu->nextpr = pr;
234 			else
235 				*nextpr = pr;
236 
237 			/* Skip preceding text and up to the next % sign. */
238 			for (p1 = fmtp; *p1 && *p1 != '%'; ++p1);
239 
240 			/* Only text in the string. */
241 			if (!*p1) {
242 				pr->fmt = fmtp;
243 				pr->flags = F_TEXT;
244 				break;
245 			}
246 
247 			/*
248 			 * Get precision for %s -- if have a byte count, don't
249 			 * need it.
250 			 */
251 			if (fu->bcnt) {
252 				sokay = USEBCNT;
253 				/* Skip to conversion character. */
254 				for (++p1; *p1 && strchr(spec, *p1); ++p1);
255 			} else {
256 				/* Skip any special chars, field width. */
257 				while (*++p1 && strchr(spec + 1, *p1));
258 				if (*p1 == '.' &&
259 				    isdigit((unsigned char)*++p1)) {
260 					sokay = USEPREC;
261 					prec = atoi(p1);
262 					while (isdigit((unsigned char)*++p1))
263 						continue;
264 				} else
265 					sokay = NOTOKAY;
266 			}
267 
268 			p2 = *p1 ? p1 + 1 : p1;	/* Set end pointer. */
269 			cs[0] = *p1;		/* Set conversion string. */
270 			cs[1] = '\0';
271 
272 			/*
273 			 * Figure out the byte count for each conversion;
274 			 * rewrite the format as necessary, set up blank-
275 			 * padding for end of data.
276 			 */
277 			switch(cs[0]) {
278 			case 'c':
279 				pr->flags = F_CHAR;
280 				switch(fu->bcnt) {
281 				case 0: case 1:
282 					pr->bcnt = 1;
283 					break;
284 				default:
285 					p1[1] = '\0';
286 					badcnt(p1);
287 				}
288 				break;
289 			case 'd': case 'i':
290 			case 'o': case 'u': case 'x': case 'X':
291 				if (cs[0] == 'd' || cs[0] == 'i')
292 					pr->flags = F_INT;
293 				else
294 					pr->flags = F_UINT;
295 
296 				cs[3] = '\0';
297 				cs[2] = cs[0];
298 				cs[1] = 'l';
299 				cs[0] = 'l';
300 				switch(fu->bcnt) {
301 				case 0: case 4:
302 					pr->bcnt = 4;
303 					break;
304 				case 1:
305 					pr->bcnt = 1;
306 					break;
307 				case 2:
308 					pr->bcnt = 2;
309 					break;
310 				case 8:
311 					pr->bcnt = 8;
312 					break;
313 				default:
314 					p1[1] = '\0';
315 					badcnt(p1);
316 				}
317 				break;
318 			case 'e': case 'E': case 'f': case 'g': case 'G':
319 				pr->flags = F_DBL;
320 				switch(fu->bcnt) {
321 				case 0: case 8:
322 					pr->bcnt = 8;
323 					break;
324 				case 4:
325 					pr->bcnt = 4;
326 					break;
327 				default:
328 					p1[1] = '\0';
329 					badcnt(p1);
330 				}
331 				break;
332 			case 's':
333 				pr->flags = F_STR;
334 				switch(sokay) {
335 				case NOTOKAY:
336 					badsfmt();
337 				case USEBCNT:
338 					pr->bcnt = fu->bcnt;
339 					break;
340 				case USEPREC:
341 					pr->bcnt = prec;
342 					break;
343 				}
344 				break;
345 			case '_':
346 				++p2;
347 				switch(p1[1]) {
348 				case 'A':
349 					endfu = fu;
350 					fu->flags |= F_IGNORE;
351 					/* FALLTHROUGH */
352 				case 'a':
353 					pr->flags = F_ADDRESS;
354 					++p2;
355 					switch(p1[2]) {
356 					case 'd': case 'o': case'x':
357 						cs[0] = 'l';
358 						cs[1] = 'l';
359 						cs[2] = p1[2];
360 						cs[3] = '\0';
361 						break;
362 					default:
363 						if (p1[2])
364 							p1[3] = '\0';
365 						badconv(p1);
366 					}
367 					break;
368 				case 'c':
369 				case 'p':
370 				case 'u':
371 					if (p1[1] == 'c') {
372 						pr->flags = F_C;
373 						/* cs[0] = 'c';	set in conv_c */
374 					} else if (p1[1] == 'p') {
375 						pr->flags = F_P;
376 						cs[0] = 'c';
377 					} else {
378 						pr->flags = F_U;
379 						/* cs[0] = 'c';	set in conv_u */
380 					}
381 
382 					switch(fu->bcnt) {
383 					case 0: case 1:
384 						pr->bcnt = 1;
385 						break;
386 					default:
387 						p1[2] = '\0';
388 						badcnt(p1);
389 					}
390 					break;
391 				default:
392 					if (p1[1])
393 						p1[2] = '\0';
394 					badconv(p1);
395 				}
396 				break;
397 			default:
398 				if (cs[0])
399 					p1[1] = '\0';
400 				badconv(p1);
401 			}
402 
403 			/*
404 			 * Copy to PR format string, set conversion character
405 			 * pointer, update original.
406 			 */
407 			savech = *p2;
408 			p1[0] = '\0';
409 			if (asprintf(&pr->fmt, "%s%s", fmtp, cs) == -1)
410 				err(1, NULL);
411 			*p2 = savech;
412 			pr->cchar = pr->fmt + (p1 - fmtp);
413 			fmtp = p2;
414 
415 			/* Only one conversion character if byte count. */
416 			if (!(pr->flags&F_ADDRESS) && fu->bcnt && nconv++)
417 				errx(1,
418 			    "byte count with multiple conversion characters");
419 		}
420 		/*
421 		 * If format unit byte count not specified, figure it out
422 		 * so can adjust rep count later.
423 		 */
424 		if (!fu->bcnt)
425 			for (pr = fu->nextpr; pr; pr = pr->nextpr)
426 				fu->bcnt += pr->bcnt;
427 	}
428 	/*
429 	 * If the format string interprets any data at all, and it's
430 	 * not the same as the blocksize, and its last format unit
431 	 * interprets any data at all, and has no iteration count,
432 	 * repeat it as necessary.
433 	 *
434 	 * If, rep count is greater than 1, no trailing whitespace
435 	 * gets output from the last iteration of the format unit.
436 	 */
437 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
438 		if (!fu->nextfu && fs->bcnt < blocksize &&
439 		    !(fu->flags&F_SETREP) && fu->bcnt)
440 			fu->reps += (blocksize - fs->bcnt) / fu->bcnt;
441 		if (fu->reps > 1) {
442 			if (!fu->nextpr)
443 				break;
444 			for (pr = fu->nextpr;; pr = pr->nextpr)
445 				if (!pr->nextpr)
446 					break;
447 			for (p1 = pr->fmt, p2 = NULL; *p1; ++p1)
448 				p2 = isspace((unsigned char)*p1) ? p1 : NULL;
449 			if (p2)
450 				pr->nospace = p2;
451 		}
452 	}
453 #ifdef DEBUG
454 	for (fu = fs->nextfu; fu; fu = fu->nextfu) {
455 		(void)printf("fmt:");
456 		for (pr = fu->nextpr; pr; pr = pr->nextpr)
457 			(void)printf(" {%s}", pr->fmt);
458 		(void)printf("\n");
459 	}
460 #endif
461 }
462 
463 static void
464 escape(char *p1)
465 {
466 	char *p2;
467 
468 	/* alphabetic escape sequences have to be done in place */
469 	for (p2 = p1;; ++p1, ++p2) {
470 		if (!*p1) {
471 			*p2 = *p1;
472 			break;
473 		}
474 		if (*p1 == '\\') {
475 			switch(*++p1) {
476 			case '\0':
477 				*p2++ = '\\';
478 				*p2 = '\0';
479 				return;	/* incomplete escape sequence */
480 			case 'a':
481 			     /* *p2 = '\a'; */
482 				*p2 = '\007';
483 				break;
484 			case 'b':
485 				*p2 = '\b';
486 				break;
487 			case 'f':
488 				*p2 = '\f';
489 				break;
490 			case 'n':
491 				*p2 = '\n';
492 				break;
493 			case 'r':
494 				*p2 = '\r';
495 				break;
496 			case 't':
497 				*p2 = '\t';
498 				break;
499 			case 'v':
500 				*p2 = '\v';
501 				break;
502 			default:
503 				*p2 = *p1;
504 				break;
505 			}
506 		} else
507 			*p2 = *p1;
508 	}
509 }
510 
511 static __dead void
512 badcnt(char *s)
513 {
514 	errx(1, "%s: bad byte count", s);
515 }
516 
517 static __dead void
518 badsfmt(void)
519 {
520 	errx(1, "%%s: requires a precision or a byte count");
521 }
522 
523 static __dead void
524 badfmt(const char *fmt)
525 {
526 	errx(1, "\"%s\": bad format", fmt);
527 }
528 
529 static __dead void
530 badconv(char *ch)
531 {
532 	errx(1, "%%%s: bad conversion character", ch);
533 }
534