xref: /freebsd-12.1/usr.bin/printf/printf.c (revision 81bd5041)
1 /*-
2  * Copyright (c) 1989, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  * 4. Neither the name of the University nor the names of its contributors
14  *    may be used to endorse or promote products derived from this software
15  *    without specific prior written permission.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 #ifndef SHELL
31 #ifndef lint
32 static char const copyright[] =
33 "@(#) Copyright (c) 1989, 1993\n\
34 	The Regents of the University of California.  All rights reserved.\n";
35 #endif /* not lint */
36 #endif
37 
38 #ifndef lint
39 #if 0
40 static char const sccsid[] = "@(#)printf.c	8.1 (Berkeley) 7/20/93";
41 #endif
42 static const char rcsid[] =
43   "$FreeBSD$";
44 #endif /* not lint */
45 
46 #include <sys/types.h>
47 
48 #include <err.h>
49 #include <errno.h>
50 #include <inttypes.h>
51 #include <limits.h>
52 #include <locale.h>
53 #include <stdio.h>
54 #include <stdlib.h>
55 #include <string.h>
56 #include <unistd.h>
57 
58 #ifdef SHELL
59 #define main printfcmd
60 #include "bltin/bltin.h"
61 #include "memalloc.h"
62 #include "error.h"
63 #else
64 #define	warnx1(a, b, c)		warnx(a)
65 #define	warnx2(a, b, c)		warnx(a, b)
66 #define	warnx3(a, b, c)		warnx(a, b, c)
67 #endif
68 
69 #define PF(f, func) do { \
70 	char *b = NULL; \
71 	if (havewidth) \
72 		if (haveprec) \
73 			(void)asprintf(&b, f, fieldwidth, precision, func); \
74 		else \
75 			(void)asprintf(&b, f, fieldwidth, func); \
76 	else if (haveprec) \
77 		(void)asprintf(&b, f, precision, func); \
78 	else \
79 		(void)asprintf(&b, f, func); \
80 	if (b) { \
81 		(void)fputs(b, stdout); \
82 		free(b); \
83 	} \
84 } while (0)
85 
86 static int	 asciicode(void);
87 static char	*printf_doformat(char *, int *);
88 static int	 escape(char *, int, size_t *);
89 static int	 getchr(void);
90 static int	 getfloating(long double *, int);
91 static int	 getint(int *);
92 static int	 getnum(intmax_t *, uintmax_t *, int);
93 static const char
94 		*getstr(void);
95 static char	*mknum(char *, char);
96 static void	 usage(void);
97 
98 static char **gargv;
99 
100 int
101 main(int argc, char *argv[])
102 {
103 	size_t len;
104 	int ch, chopped, end, rval;
105 	char *format, *fmt, *start;
106 
107 #ifndef SHELL
108 	(void) setlocale(LC_ALL, "");
109 #endif
110 #ifdef SHELL
111 	optreset = 1; optind = 1; opterr = 0; /* initialize getopt */
112 #endif
113 	while ((ch = getopt(argc, argv, "")) != -1)
114 		switch (ch) {
115 		case '?':
116 		default:
117 			usage();
118 			return (1);
119 		}
120 	argc -= optind;
121 	argv += optind;
122 
123 	if (argc < 1) {
124 		usage();
125 		return (1);
126 	}
127 
128 #ifdef SHELL
129 	INTOFF;
130 #endif
131 	/*
132 	 * Basic algorithm is to scan the format string for conversion
133 	 * specifications -- once one is found, find out if the field
134 	 * width or precision is a '*'; if it is, gather up value.  Note,
135 	 * format strings are reused as necessary to use up the provided
136 	 * arguments, arguments of zero/null string are provided to use
137 	 * up the format string.
138 	 */
139 	fmt = format = *argv;
140 	chopped = escape(fmt, 1, &len);		/* backslash interpretation */
141 	rval = end = 0;
142 	gargv = ++argv;
143 	for (;;) {
144 		start = fmt;
145 		while (fmt < format + len) {
146 			if (fmt[0] == '%') {
147 				fwrite(start, 1, fmt - start, stdout);
148 				if (fmt[1] == '%') {
149 					/* %% prints a % */
150 					putchar('%');
151 					fmt += 2;
152 				} else {
153 					fmt = printf_doformat(fmt, &rval);
154 					if (fmt == NULL) {
155 #ifdef SHELL
156 						INTON;
157 #endif
158 						return (1);
159 					}
160 					end = 0;
161 				}
162 				start = fmt;
163 			} else
164 				fmt++;
165 		}
166 
167 		if (end == 1) {
168 			warnx1("missing format character", NULL, NULL);
169 #ifdef SHELL
170 			INTON;
171 #endif
172 			return (1);
173 		}
174 		fwrite(start, 1, fmt - start, stdout);
175 		if (chopped || !*gargv) {
176 #ifdef SHELL
177 			INTON;
178 #endif
179 			return (rval);
180 		}
181 		/* Restart at the beginning of the format string. */
182 		fmt = format;
183 		end = 1;
184 	}
185 	/* NOTREACHED */
186 }
187 
188 
189 static char *
190 printf_doformat(char *start, int *rval)
191 {
192 	static const char skip1[] = "#'-+ 0";
193 	static const char skip2[] = "0123456789";
194 	char *fmt;
195 	int fieldwidth, haveprec, havewidth, mod_ldbl, precision;
196 	char convch, nextch;
197 
198 	fmt = start + 1;
199 	/* skip to field width */
200 	fmt += strspn(fmt, skip1);
201 	if (*fmt == '*') {
202 		if (getint(&fieldwidth))
203 			return (NULL);
204 		havewidth = 1;
205 		++fmt;
206 	} else {
207 		havewidth = 0;
208 
209 		/* skip to possible '.', get following precision */
210 		fmt += strspn(fmt, skip2);
211 	}
212 	if (*fmt == '.') {
213 		/* precision present? */
214 		++fmt;
215 		if (*fmt == '*') {
216 			if (getint(&precision))
217 				return (NULL);
218 			haveprec = 1;
219 			++fmt;
220 		} else {
221 			haveprec = 0;
222 
223 			/* skip to conversion char */
224 			fmt += strspn(fmt, skip2);
225 		}
226 	} else
227 		haveprec = 0;
228 	if (!*fmt) {
229 		warnx1("missing format character", NULL, NULL);
230 		return (NULL);
231 	}
232 
233 	/*
234 	 * Look for a length modifier.  POSIX doesn't have these, so
235 	 * we only support them for floating-point conversions, which
236 	 * are extensions.  This is useful because the L modifier can
237 	 * be used to gain extra range and precision, while omitting
238 	 * it is more likely to produce consistent results on different
239 	 * architectures.  This is not so important for integers
240 	 * because overflow is the only bad thing that can happen to
241 	 * them, but consider the command  printf %a 1.1
242 	 */
243 	if (*fmt == 'L') {
244 		mod_ldbl = 1;
245 		fmt++;
246 		if (!strchr("aAeEfFgG", *fmt)) {
247 			warnx2("bad modifier L for %%%c", *fmt, NULL);
248 			return (NULL);
249 		}
250 	} else {
251 		mod_ldbl = 0;
252 	}
253 
254 	convch = *fmt;
255 	nextch = *++fmt;
256 	*fmt = '\0';
257 	switch (convch) {
258 	case 'b': {
259 		size_t len;
260 		char *p;
261 		int getout;
262 
263 #ifdef SHELL
264 		p = savestr(getstr());
265 #else
266 		p = strdup(getstr());
267 #endif
268 		if (p == NULL) {
269 			warnx2("%s", strerror(ENOMEM), NULL);
270 			return (NULL);
271 		}
272 		getout = escape(p, 0, &len);
273 		*(fmt - 1) = 's';
274 		PF(start, p);
275 		*(fmt - 1) = 'b';
276 #ifdef SHELL
277 		ckfree(p);
278 #else
279 		free(p);
280 #endif
281 		if (getout)
282 			return (fmt);
283 		break;
284 	}
285 	case 'c': {
286 		char p;
287 
288 		p = getchr();
289 		PF(start, p);
290 		break;
291 	}
292 	case 's': {
293 		const char *p;
294 
295 		p = getstr();
296 		PF(start, p);
297 		break;
298 	}
299 	case 'd': case 'i': case 'o': case 'u': case 'x': case 'X': {
300 		char *f;
301 		intmax_t val;
302 		uintmax_t uval;
303 		int signedconv;
304 
305 		signedconv = (convch == 'd' || convch == 'i');
306 		if ((f = mknum(start, convch)) == NULL)
307 			return (NULL);
308 		if (getnum(&val, &uval, signedconv))
309 			*rval = 1;
310 		if (signedconv)
311 			PF(f, val);
312 		else
313 			PF(f, uval);
314 		break;
315 	}
316 	case 'e': case 'E':
317 	case 'f': case 'F':
318 	case 'g': case 'G':
319 	case 'a': case 'A': {
320 		long double p;
321 
322 		if (getfloating(&p, mod_ldbl))
323 			*rval = 1;
324 		if (mod_ldbl)
325 			PF(start, p);
326 		else
327 			PF(start, (double)p);
328 		break;
329 	}
330 	default:
331 		warnx2("illegal format character %c", convch, NULL);
332 		return (NULL);
333 	}
334 	*fmt = nextch;
335 	return (fmt);
336 }
337 
338 static char *
339 mknum(char *str, char ch)
340 {
341 	static char *copy;
342 	static size_t copy_size;
343 	char *newcopy;
344 	size_t len, newlen;
345 
346 	len = strlen(str) + 2;
347 	if (len > copy_size) {
348 		newlen = ((len + 1023) >> 10) << 10;
349 #ifdef SHELL
350 		if ((newcopy = ckrealloc(copy, newlen)) == NULL)
351 #else
352 		if ((newcopy = realloc(copy, newlen)) == NULL)
353 #endif
354 		{
355 			warnx2("%s", strerror(ENOMEM), NULL);
356 			return (NULL);
357 		}
358 		copy = newcopy;
359 		copy_size = newlen;
360 	}
361 
362 	memmove(copy, str, len - 3);
363 	copy[len - 3] = 'j';
364 	copy[len - 2] = ch;
365 	copy[len - 1] = '\0';
366 	return (copy);
367 }
368 
369 static int
370 escape(char *fmt, int percent, size_t *len)
371 {
372 	char *save, *store;
373 	int value, c;
374 
375 	for (save = store = fmt; (c = *fmt); ++fmt, ++store) {
376 		if (c != '\\') {
377 			*store = c;
378 			continue;
379 		}
380 		switch (*++fmt) {
381 		case '\0':		/* EOS, user error */
382 			*store = '\\';
383 			*++store = '\0';
384 			*len = store - save;
385 			return (0);
386 		case '\\':		/* backslash */
387 		case '\'':		/* single quote */
388 			*store = *fmt;
389 			break;
390 		case 'a':		/* bell/alert */
391 			*store = '\a';
392 			break;
393 		case 'b':		/* backspace */
394 			*store = '\b';
395 			break;
396 		case 'c':
397 			*store = '\0';
398 			*len = store - save;
399 			return (1);
400 		case 'f':		/* form-feed */
401 			*store = '\f';
402 			break;
403 		case 'n':		/* newline */
404 			*store = '\n';
405 			break;
406 		case 'r':		/* carriage-return */
407 			*store = '\r';
408 			break;
409 		case 't':		/* horizontal tab */
410 			*store = '\t';
411 			break;
412 		case 'v':		/* vertical tab */
413 			*store = '\v';
414 			break;
415 					/* octal constant */
416 		case '0': case '1': case '2': case '3':
417 		case '4': case '5': case '6': case '7':
418 			c = (!percent && *fmt == '0') ? 4 : 3;
419 			for (value = 0;
420 			    c-- && *fmt >= '0' && *fmt <= '7'; ++fmt) {
421 				value <<= 3;
422 				value += *fmt - '0';
423 			}
424 			--fmt;
425 			if (percent && value == '%') {
426 				*store++ = '%';
427 				*store = '%';
428 			} else
429 				*store = value;
430 			break;
431 		default:
432 			*store = *fmt;
433 			break;
434 		}
435 	}
436 	*store = '\0';
437 	*len = store - save;
438 	return (0);
439 }
440 
441 static int
442 getchr(void)
443 {
444 	if (!*gargv)
445 		return ('\0');
446 	return ((int)**gargv++);
447 }
448 
449 static const char *
450 getstr(void)
451 {
452 	if (!*gargv)
453 		return ("");
454 	return (*gargv++);
455 }
456 
457 static int
458 getint(int *ip)
459 {
460 	intmax_t val;
461 	uintmax_t uval;
462 	int rval;
463 
464 	if (getnum(&val, &uval, 1))
465 		return (1);
466 	rval = 0;
467 	if (val < INT_MIN || val > INT_MAX) {
468 		warnx3("%s: %s", *gargv, strerror(ERANGE));
469 		rval = 1;
470 	}
471 	*ip = (int)val;
472 	return (rval);
473 }
474 
475 static int
476 getnum(intmax_t *ip, uintmax_t *uip, int signedconv)
477 {
478 	char *ep;
479 	int rval;
480 
481 	if (!*gargv) {
482 		*ip = 0;
483 		return (0);
484 	}
485 	if (**gargv == '"' || **gargv == '\'') {
486 		if (signedconv)
487 			*ip = asciicode();
488 		else
489 			*uip = asciicode();
490 		return (0);
491 	}
492 	rval = 0;
493 	errno = 0;
494 	if (signedconv)
495 		*ip = strtoimax(*gargv, &ep, 0);
496 	else
497 		*uip = strtoumax(*gargv, &ep, 0);
498 	if (ep == *gargv) {
499 		warnx2("%s: expected numeric value", *gargv, NULL);
500 		rval = 1;
501 	}
502 	else if (*ep != '\0') {
503 		warnx2("%s: not completely converted", *gargv, NULL);
504 		rval = 1;
505 	}
506 	if (errno == ERANGE) {
507 		warnx3("%s: %s", *gargv, strerror(ERANGE));
508 		rval = 1;
509 	}
510 	++gargv;
511 	return (rval);
512 }
513 
514 static int
515 getfloating(long double *dp, int mod_ldbl)
516 {
517 	char *ep;
518 	int rval;
519 
520 	if (!*gargv) {
521 		*dp = 0.0;
522 		return (0);
523 	}
524 	if (**gargv == '"' || **gargv == '\'') {
525 		*dp = asciicode();
526 		return (0);
527 	}
528 	rval = 0;
529 	errno = 0;
530 	if (mod_ldbl)
531 		*dp = strtold(*gargv, &ep);
532 	else
533 		*dp = strtod(*gargv, &ep);
534 	if (ep == *gargv) {
535 		warnx2("%s: expected numeric value", *gargv, NULL);
536 		rval = 1;
537 	} else if (*ep != '\0') {
538 		warnx2("%s: not completely converted", *gargv, NULL);
539 		rval = 1;
540 	}
541 	if (errno == ERANGE) {
542 		warnx3("%s: %s", *gargv, strerror(ERANGE));
543 		rval = 1;
544 	}
545 	++gargv;
546 	return (rval);
547 }
548 
549 static int
550 asciicode(void)
551 {
552 	int ch;
553 
554 	ch = **gargv;
555 	if (ch == '\'' || ch == '"')
556 		ch = (*gargv)[1];
557 	++gargv;
558 	return (ch);
559 }
560 
561 static void
562 usage(void)
563 {
564 	(void)fprintf(stderr, "usage: printf format [arguments ...]\n");
565 }
566