xref: /freebsd-12.1/usr.bin/printf/printf.c (revision ad4e3b96)
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  * 3. All advertising materials mentioning features or use of this software
14  *    must display the following acknowledgement:
15  *	This product includes software developed by the University of
16  *	California, Berkeley and its contributors.
17  * 4. Neither the name of the University nor the names of its contributors
18  *    may be used to endorse or promote products derived from this software
19  *    without specific prior written permission.
20  *
21  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31  * SUCH DAMAGE.
32  */
33 
34 #if !defined(BUILTIN) && !defined(SHELL)
35 #ifndef lint
36 static char const copyright[] =
37 "@(#) Copyright (c) 1989, 1993\n\
38 	The Regents of the University of California.  All rights reserved.\n";
39 #endif /* not lint */
40 #endif
41 
42 #ifndef lint
43 #if 0
44 static char const sccsid[] = "@(#)printf.c	8.1 (Berkeley) 7/20/93";
45 #endif
46 static const char rcsid[] =
47   "$FreeBSD$";
48 #endif /* not lint */
49 
50 #include <sys/types.h>
51 
52 #include <err.h>
53 #include <errno.h>
54 #include <inttypes.h>
55 #include <limits.h>
56 #include <stdio.h>
57 #include <stdlib.h>
58 #include <string.h>
59 #include <unistd.h>
60 
61 #ifdef SHELL
62 #define main printfcmd
63 #include "bltin/bltin.h"
64 #include "memalloc.h"
65 #include "error.h"
66 #else
67 #define	warnx1(a, b, c)		warnx(a)
68 #define	warnx2(a, b, c)		warnx(a, b)
69 #define	warnx3(a, b, c)		warnx(a, b, c)
70 #endif
71 
72 #ifndef BUILTIN
73 #include <locale.h>
74 #endif
75 
76 #define PF(f, func) do { \
77 	char *b = NULL; \
78 	if (havewidth) \
79 		if (haveprec) \
80 			(void)asprintf(&b, f, fieldwidth, precision, func); \
81 		else \
82 			(void)asprintf(&b, f, fieldwidth, func); \
83 	else if (haveprec) \
84 		(void)asprintf(&b, f, precision, func); \
85 	else \
86 		(void)asprintf(&b, f, func); \
87 	if (b) { \
88 		(void)fputs(b, stdout); \
89 		free(b); \
90 	} \
91 } while (0)
92 
93 static int	 asciicode(void);
94 static char	*printf_doformat(char *, int *);
95 static int	 escape(char *, int, size_t *);
96 static int	 getchr(void);
97 static int	 getfloating(long double *, int);
98 static int	 getint(int *);
99 static int	 getnum(intmax_t *, uintmax_t *, int);
100 static const char
101 		*getstr(void);
102 static char	*mknum(char *, int);
103 static void	 usage(void);
104 
105 static char **gargv;
106 
107 int
108 #ifdef BUILTIN
109 progprintf(int argc, char *argv[])
110 #else
111 main(int argc, char *argv[])
112 #endif
113 {
114 	size_t len;
115 	int ch, chopped, end, rval;
116 	char *format, *fmt, *start;
117 
118 #if !defined(BUILTIN) && !defined(SHELL)
119 	(void) setlocale(LC_NUMERIC, "");
120 #endif
121 #ifdef SHELL
122 	optreset = 1; optind = 1; opterr = 0; /* initialize getopt */
123 #endif
124 	while ((ch = getopt(argc, argv, "")) != -1)
125 		switch (ch) {
126 		case '?':
127 		default:
128 			usage();
129 			return (1);
130 		}
131 	argc -= optind;
132 	argv += optind;
133 
134 	if (argc < 1) {
135 		usage();
136 		return (1);
137 	}
138 
139 #ifdef SHELL
140 	INTOFF;
141 #endif
142 	/*
143 	 * Basic algorithm is to scan the format string for conversion
144 	 * specifications -- once one is found, find out if the field
145 	 * width or precision is a '*'; if it is, gather up value.  Note,
146 	 * format strings are reused as necessary to use up the provided
147 	 * arguments, arguments of zero/null string are provided to use
148 	 * up the format string.
149 	 */
150 	fmt = format = *argv;
151 	chopped = escape(fmt, 1, &len);		/* backslash interpretation */
152 	rval = end = 0;
153 	gargv = ++argv;
154 	for (;;) {
155 		start = fmt;
156 		while (fmt < format + len) {
157 			if (fmt[0] == '%') {
158 				fwrite(start, 1, fmt - start, stdout);
159 				if (fmt[1] == '%') {
160 					/* %% prints a % */
161 					putchar('%');
162 					fmt += 2;
163 				} else {
164 					fmt = printf_doformat(fmt, &rval);
165 					if (fmt == NULL) {
166 #ifdef SHELL
167 						INTON;
168 #endif
169 						return (1);
170 					}
171 					end = 0;
172 				}
173 				start = fmt;
174 			} else
175 				fmt++;
176 		}
177 
178 		if (end == 1) {
179 			warnx1("missing format character", NULL, NULL);
180 #ifdef SHELL
181 			INTON;
182 #endif
183 			return (1);
184 		}
185 		fwrite(start, 1, fmt - start, stdout);
186 		if (chopped || !*gargv) {
187 #ifdef SHELL
188 			INTON;
189 #endif
190 			return (rval);
191 		}
192 		/* Restart at the beginning of the format string. */
193 		fmt = format;
194 		end = 1;
195 	}
196 	/* NOTREACHED */
197 }
198 
199 
200 static char *
201 printf_doformat(char *start, int *rval)
202 {
203 	static const char skip1[] = "#'-+ 0";
204 	static const char skip2[] = "0123456789";
205 	char *fmt;
206 	int fieldwidth, haveprec, havewidth, mod_ldbl, precision;
207 	char convch, nextch;
208 
209 	fmt = start + 1;
210 	/* skip to field width */
211 	fmt += strspn(fmt, skip1);
212 	if (*fmt == '*') {
213 		if (getint(&fieldwidth))
214 			return (NULL);
215 		havewidth = 1;
216 		++fmt;
217 	} else {
218 		havewidth = 0;
219 
220 		/* skip to possible '.', get following precision */
221 		fmt += strspn(fmt, skip2);
222 	}
223 	if (*fmt == '.') {
224 		/* precision present? */
225 		++fmt;
226 		if (*fmt == '*') {
227 			if (getint(&precision))
228 				return (NULL);
229 			haveprec = 1;
230 			++fmt;
231 		} else {
232 			haveprec = 0;
233 
234 			/* skip to conversion char */
235 			fmt += strspn(fmt, skip2);
236 		}
237 	} else
238 		haveprec = 0;
239 	if (!*fmt) {
240 		warnx1("missing format character", NULL, NULL);
241 		return (NULL);
242 	}
243 
244 	/*
245 	 * Look for a length modifier.  POSIX doesn't have these, so
246 	 * we only support them for floating-point conversions, which
247 	 * are extensions.  This is useful because the L modifier can
248 	 * be used to gain extra range and precision, while omitting
249 	 * it is more likely to produce consistent results on different
250 	 * architectures.  This is not so important for integers
251 	 * because overflow is the only bad thing that can happen to
252 	 * them, but consider the command  printf %a 1.1
253 	 */
254 	if (*fmt == 'L') {
255 		mod_ldbl = 1;
256 		fmt++;
257 		if (!strchr("aAeEfFgG", *fmt)) {
258 			warnx2("bad modifier L for %%%c", *fmt, NULL);
259 			return (NULL);
260 		}
261 	} else {
262 		mod_ldbl = 0;
263 	}
264 
265 	convch = *fmt;
266 	nextch = *++fmt;
267 	*fmt = '\0';
268 	switch (convch) {
269 	case 'b': {
270 		size_t len;
271 		char *p;
272 		int getout;
273 
274 #ifdef SHELL
275 		p = savestr(getstr());
276 #else
277 		p = strdup(getstr());
278 #endif
279 		if (p == NULL) {
280 			warnx2("%s", strerror(ENOMEM), NULL);
281 			return (NULL);
282 		}
283 		getout = escape(p, 0, &len);
284 		*(fmt - 1) = 's';
285 		PF(start, p);
286 		*(fmt - 1) = 'b';
287 #ifdef SHELL
288 		ckfree(p);
289 #else
290 		free(p);
291 #endif
292 		if (getout)
293 			return (fmt);
294 		break;
295 	}
296 	case 'c': {
297 		char p;
298 
299 		p = getchr();
300 		PF(start, p);
301 		break;
302 	}
303 	case 's': {
304 		const char *p;
305 
306 		p = getstr();
307 		PF(start, p);
308 		break;
309 	}
310 	case 'd': case 'i': case 'o': case 'u': case 'x': case 'X': {
311 		char *f;
312 		intmax_t val;
313 		uintmax_t uval;
314 		int signedconv;
315 
316 		signedconv = (convch == 'd' || convch == 'i');
317 		if ((f = mknum(start, convch)) == NULL)
318 			return (NULL);
319 		if (getnum(&val, &uval, signedconv))
320 			*rval = 1;
321 		if (signedconv)
322 			PF(f, val);
323 		else
324 			PF(f, uval);
325 		break;
326 	}
327 	case 'e': case 'E':
328 	case 'f': case 'F':
329 	case 'g': case 'G':
330 	case 'a': case 'A': {
331 		long double p;
332 
333 		if (getfloating(&p, mod_ldbl))
334 			*rval = 1;
335 		if (mod_ldbl)
336 			PF(start, p);
337 		else
338 			PF(start, (double)p);
339 		break;
340 	}
341 	default:
342 		warnx2("illegal format character %c", convch, NULL);
343 		return (NULL);
344 	}
345 	*fmt = nextch;
346 	return (fmt);
347 }
348 
349 static char *
350 mknum(char *str, int ch)
351 {
352 	static char *copy;
353 	static size_t copy_size;
354 	char *newcopy;
355 	size_t len, newlen;
356 
357 	len = strlen(str) + 2;
358 	if (len > copy_size) {
359 		newlen = ((len + 1023) >> 10) << 10;
360 #ifdef SHELL
361 		if ((newcopy = ckrealloc(copy, newlen)) == NULL)
362 #else
363 		if ((newcopy = realloc(copy, newlen)) == NULL)
364 #endif
365 		{
366 			warnx2("%s", strerror(ENOMEM), NULL);
367 			return (NULL);
368 		}
369 		copy = newcopy;
370 		copy_size = newlen;
371 	}
372 
373 	memmove(copy, str, len - 3);
374 	copy[len - 3] = 'j';
375 	copy[len - 2] = ch;
376 	copy[len - 1] = '\0';
377 	return (copy);
378 }
379 
380 static int
381 escape(char *fmt, int percent, size_t *len)
382 {
383 	char *save, *store;
384 	int value, c;
385 
386 	for (save = store = fmt; (c = *fmt); ++fmt, ++store) {
387 		if (c != '\\') {
388 			*store = c;
389 			continue;
390 		}
391 		switch (*++fmt) {
392 		case '\0':		/* EOS, user error */
393 			*store = '\\';
394 			*++store = '\0';
395 			*len = store - save;
396 			return (0);
397 		case '\\':		/* backslash */
398 		case '\'':		/* single quote */
399 			*store = *fmt;
400 			break;
401 		case 'a':		/* bell/alert */
402 			*store = '\a';
403 			break;
404 		case 'b':		/* backspace */
405 			*store = '\b';
406 			break;
407 		case 'c':
408 			*store = '\0';
409 			*len = store - save;
410 			return (1);
411 		case 'f':		/* form-feed */
412 			*store = '\f';
413 			break;
414 		case 'n':		/* newline */
415 			*store = '\n';
416 			break;
417 		case 'r':		/* carriage-return */
418 			*store = '\r';
419 			break;
420 		case 't':		/* horizontal tab */
421 			*store = '\t';
422 			break;
423 		case 'v':		/* vertical tab */
424 			*store = '\v';
425 			break;
426 					/* octal constant */
427 		case '0': case '1': case '2': case '3':
428 		case '4': case '5': case '6': case '7':
429 			c = (!percent && *fmt == '0') ? 4 : 3;
430 			for (value = 0;
431 			    c-- && *fmt >= '0' && *fmt <= '7'; ++fmt) {
432 				value <<= 3;
433 				value += *fmt - '0';
434 			}
435 			--fmt;
436 			if (percent && value == '%') {
437 				*store++ = '%';
438 				*store = '%';
439 			} else
440 				*store = value;
441 			break;
442 		default:
443 			*store = *fmt;
444 			break;
445 		}
446 	}
447 	*store = '\0';
448 	*len = store - save;
449 	return (0);
450 }
451 
452 static int
453 getchr(void)
454 {
455 	if (!*gargv)
456 		return ('\0');
457 	return ((int)**gargv++);
458 }
459 
460 static const char *
461 getstr(void)
462 {
463 	if (!*gargv)
464 		return ("");
465 	return (*gargv++);
466 }
467 
468 static int
469 getint(int *ip)
470 {
471 	intmax_t val;
472 	uintmax_t uval;
473 	int rval;
474 
475 	if (getnum(&val, &uval, 1))
476 		return (1);
477 	rval = 0;
478 	if (val < INT_MIN || val > INT_MAX) {
479 		warnx3("%s: %s", *gargv, strerror(ERANGE));
480 		rval = 1;
481 	}
482 	*ip = (int)val;
483 	return (rval);
484 }
485 
486 static int
487 getnum(intmax_t *ip, uintmax_t *uip, int signedconv)
488 {
489 	char *ep;
490 	int rval;
491 
492 	if (!*gargv) {
493 		*ip = 0;
494 		return (0);
495 	}
496 	if (**gargv == '"' || **gargv == '\'') {
497 		if (signedconv)
498 			*ip = asciicode();
499 		else
500 			*uip = asciicode();
501 		return (0);
502 	}
503 	rval = 0;
504 	errno = 0;
505 	if (signedconv)
506 		*ip = strtoimax(*gargv, &ep, 0);
507 	else
508 		*uip = strtoumax(*gargv, &ep, 0);
509 	if (ep == *gargv) {
510 		warnx2("%s: expected numeric value", *gargv, NULL);
511 		rval = 1;
512 	}
513 	else if (*ep != '\0') {
514 		warnx2("%s: not completely converted", *gargv, NULL);
515 		rval = 1;
516 	}
517 	if (errno == ERANGE) {
518 		warnx3("%s: %s", *gargv, strerror(ERANGE));
519 		rval = 1;
520 	}
521 	++gargv;
522 	return (rval);
523 }
524 
525 static int
526 getfloating(long double *dp, int mod_ldbl)
527 {
528 	char *ep;
529 	int rval;
530 
531 	if (!*gargv) {
532 		*dp = 0.0;
533 		return (0);
534 	}
535 	if (**gargv == '"' || **gargv == '\'') {
536 		*dp = asciicode();
537 		return (0);
538 	}
539 	rval = 0;
540 	errno = 0;
541 	if (mod_ldbl)
542 		*dp = strtold(*gargv, &ep);
543 	else
544 		*dp = strtod(*gargv, &ep);
545 	if (ep == *gargv) {
546 		warnx2("%s: expected numeric value", *gargv, NULL);
547 		rval = 1;
548 	} else if (*ep != '\0') {
549 		warnx2("%s: not completely converted", *gargv, NULL);
550 		rval = 1;
551 	}
552 	if (errno == ERANGE) {
553 		warnx3("%s: %s", *gargv, strerror(ERANGE));
554 		rval = 1;
555 	}
556 	++gargv;
557 	return (rval);
558 }
559 
560 static int
561 asciicode(void)
562 {
563 	int ch;
564 
565 	ch = **gargv;
566 	if (ch == '\'' || ch == '"')
567 		ch = (*gargv)[1];
568 	++gargv;
569 	return (ch);
570 }
571 
572 static void
573 usage(void)
574 {
575 	(void)fprintf(stderr, "usage: printf format [arguments ...]\n");
576 }
577