xref: /openbsd-src/usr.bin/printf/printf.c (revision d874cce4b1d9fe6b41c9e4f2117a77d8a4a37b92) 
1 /*	$OpenBSD: printf.c,v 1.13 2008/06/19 16:24:00 millert Exp $	*/
2 
3 /*
4  * Copyright (c) 1989 The Regents of the University of California.
5  * 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 #ifndef lint
33 #if !defined(SHELL) && !defined(BUILTIN)
34 char copyright[] =
35 "@(#) Copyright (c) 1989 The Regents of the University of California.\n\
36  All rights reserved.\n";
37 #endif
38 #endif /* not lint */
39 
40 #ifndef lint
41 /*static char sccsid[] = "from: @(#)printf.c	5.9 (Berkeley) 6/1/90";*/
42 static char rcsid[] = "$OpenBSD: printf.c,v 1.13 2008/06/19 16:24:00 millert Exp $";
43 #endif /* not lint */
44 
45 #include <ctype.h>
46 #include <stdio.h>
47 #include <stdlib.h>
48 #include <string.h>
49 #include <limits.h>
50 #include <locale.h>
51 #include <errno.h>
52 #include <err.h>
53 
54 static int	 print_escape_str(const char *);
55 static int	 print_escape(const char *);
56 
57 static int	 getchr(void);
58 static double	 getdouble(void);
59 static int	 getint(void);
60 static long	 getlong(void);
61 static unsigned long getulong(void);
62 static char	*getstr(void);
63 static char	*mklong(const char *, int);
64 static void      check_conversion(const char *, const char *);
65 static void	 usage(void);
66 
67 static int	rval;
68 static char  **gargv;
69 
70 #define isodigit(c)	((c) >= '0' && (c) <= '7')
71 #define octtobin(c)	((c) - '0')
72 #define hextobin(c)	((c) >= 'A' && (c) <= 'F' ? c - 'A' + 10 : (c) >= 'a' && (c) <= 'f' ? c - 'a' + 10 : c - '0')
73 
74 #ifdef SHELL
75 #define main printfcmd
76 #include "../../bin/sh/bltin/bltin.h"
77 #include <stdarg.h>
78 
79 static void
80 warnx(const char *fmt, ...)
81 {
82 
83 	char buf[64];
84 	va_list ap;
85 
86 	va_start(ap, fmt);
87 	vsnprintf(buf, sizeof buf, fmt, ap);
88 	va_end(ap);
89 
90 	error(buf);
91 }
92 #endif /* SHELL */
93 
94 #define PF(f, func) { \
95 	if (fieldwidth) \
96 		if (precision) \
97 			(void)printf(f, fieldwidth, precision, func); \
98 		else \
99 			(void)printf(f, fieldwidth, func); \
100 	else if (precision) \
101 		(void)printf(f, precision, func); \
102 	else \
103 		(void)printf(f, func); \
104 }
105 
106 int
107 #ifdef BUILTIN
108 progprintf(int argc, char *argv[])
109 #else
110 main(int argc, char *argv[])
111 #endif
112 {
113 	char *fmt, *start;
114 	int fieldwidth, precision;
115 	char convch, nextch;
116 	char *format;
117 
118 #if !defined(SHELL) && !defined(BUILTIN)
119 	setlocale (LC_ALL, "");
120 #endif
121 
122 	/* Need to accept/ignore "--" option. */
123 	if (argc > 1 && strcmp(argv[1], "--") == 0) {
124 		argc--;
125 		argv++;
126 	}
127 
128 	if (argc < 2) {
129 		usage();
130 		return (1);
131 	}
132 
133 	format = *++argv;
134 	gargv = ++argv;
135 
136 #define SKIP1	"#-+ 0"
137 #define SKIP2	"*0123456789"
138 	do {
139 		/*
140 		 * Basic algorithm is to scan the format string for conversion
141 		 * specifications -- once one is found, find out if the field
142 		 * width or precision is a '*'; if it is, gather up value.
143 		 * Note, format strings are reused as necessary to use up the
144 		 * provided arguments, arguments of zero/null string are
145 		 * provided to use up the format string.
146 		 */
147 
148 		/* find next format specification */
149 		for (fmt = format; *fmt; fmt++) {
150 			switch (*fmt) {
151 			case '%':
152 				start = fmt++;
153 
154 				if (*fmt == '%') {
155 					putchar ('%');
156 					break;
157 				} else if (*fmt == 'b') {
158 					char *p = getstr();
159 					if (print_escape_str(p)) {
160 						return (rval);
161 					}
162 					break;
163 				}
164 
165 				/* skip to field width */
166 				for (; strchr(SKIP1, *fmt); ++fmt) ;
167 				fieldwidth = *fmt == '*' ? getint() : 0;
168 
169 				/* skip to possible '.', get following precision */
170 				for (; strchr(SKIP2, *fmt); ++fmt) ;
171 				if (*fmt == '.')
172 					++fmt;
173 				precision = *fmt == '*' ? getint() : 0;
174 
175 				for (; strchr(SKIP2, *fmt); ++fmt) ;
176 				if (!*fmt) {
177 					warnx ("missing format character");
178 					return(1);
179 				}
180 
181 				convch = *fmt;
182 				nextch = *(fmt + 1);
183 				*(fmt + 1) = '\0';
184 				switch(convch) {
185 				case 'c': {
186 					char p = getchr();
187 					PF(start, p);
188 					break;
189 				}
190 				case 's': {
191 					char *p = getstr();
192 					PF(start, p);
193 					break;
194 				}
195 				case 'd':
196 				case 'i': {
197 					long p;
198 					char *f = mklong(start, convch);
199 					if (!f) {
200 						warnx("out of memory");
201 						return (1);
202 					}
203 					p = getlong();
204 					PF(f, p);
205 					break;
206 				}
207 				case 'o':
208 				case 'u':
209 				case 'x':
210 				case 'X': {
211 					unsigned long p;
212 					char *f = mklong(start, convch);
213 					if (!f) {
214 						warnx("out of memory");
215 						return (1);
216 					}
217 					p = getulong();
218 					PF(f, p);
219 					break;
220 				}
221 				case 'e':
222 				case 'E':
223 				case 'f':
224 				case 'g':
225 				case 'G': {
226 					double p = getdouble();
227 					PF(start, p);
228 					break;
229 				}
230 				default:
231 					warnx ("%s: invalid directive", start);
232 					return(1);
233 				}
234 				*(fmt + 1) = nextch;
235 				break;
236 
237 			case '\\':
238 				fmt += print_escape(fmt);
239 				break;
240 
241 			default:
242 				putchar (*fmt);
243 				break;
244 			}
245 		}
246 	} while (gargv > argv && *gargv);
247 
248 	return (rval);
249 }
250 
251 
252 /*
253  * Print SysV echo(1) style escape string
254  *	Halts processing string and returns 1 if a \c escape is encountered.
255  */
256 static int
257 print_escape_str(const char *str)
258 {
259 	int value;
260 	int c;
261 
262 	while (*str) {
263 		if (*str == '\\') {
264 			str++;
265 			/*
266 			 * %b string octal constants are not like those in C.
267 			 * They start with a \0, and are followed by 0, 1, 2,
268 			 * or 3 octal digits.
269 			 */
270 			if (*str == '0') {
271 				str++;
272 				for (c = 3, value = 0; c-- && isodigit(*str); str++) {
273 					value <<= 3;
274 					value += octtobin(*str);
275 				}
276 				putchar (value);
277 				str--;
278 			} else if (*str == 'c') {
279 				return 1;
280 			} else {
281 				str--;
282 				str += print_escape(str);
283 			}
284 		} else {
285 			putchar (*str);
286 		}
287 		str++;
288 	}
289 
290 	return 0;
291 }
292 
293 /*
294  * Print "standard" escape characters
295  */
296 static int
297 print_escape(const char *str)
298 {
299 	const char *start = str;
300 	int value;
301 	int c;
302 
303 	str++;
304 
305 	switch (*str) {
306 	case '0': case '1': case '2': case '3':
307 	case '4': case '5': case '6': case '7':
308 		for (c = 3, value = 0; c-- && isodigit(*str); str++) {
309 			value <<= 3;
310 			value += octtobin(*str);
311 		}
312 		putchar(value);
313 		return str - start - 1;
314 		/* NOTREACHED */
315 
316 	case 'x':
317 		str++;
318 		for (value = 0; isxdigit(*str); str++) {
319 			value <<= 4;
320 			value += hextobin(*str);
321 		}
322 		if (value > UCHAR_MAX) {
323 			warnx ("escape sequence out of range for character");
324 			rval = 1;
325 		}
326 		putchar (value);
327 		return str - start - 1;
328 		/* NOTREACHED */
329 
330 	case '\\':			/* backslash */
331 		putchar('\\');
332 		break;
333 
334 	case '\'':			/* single quote */
335 		putchar('\'');
336 		break;
337 
338 	case '"':			/* double quote */
339 		putchar('"');
340 		break;
341 
342 	case 'a':			/* alert */
343 		putchar('\a');
344 		break;
345 
346 	case 'b':			/* backspace */
347 		putchar('\b');
348 		break;
349 
350 	case 'e':			/* escape */
351 #ifdef __GNUC__
352 		putchar('\e');
353 #else
354 		putchar(033);
355 #endif
356 		break;
357 
358 	case 'f':			/* form-feed */
359 		putchar('\f');
360 		break;
361 
362 	case 'n':			/* newline */
363 		putchar('\n');
364 		break;
365 
366 	case 'r':			/* carriage-return */
367 		putchar('\r');
368 		break;
369 
370 	case 't':			/* tab */
371 		putchar('\t');
372 		break;
373 
374 	case 'v':			/* vertical-tab */
375 		putchar('\v');
376 		break;
377 
378 	default:
379 		putchar(*str);
380 		warnx("unknown escape sequence `\\%c'", *str);
381 		rval = 1;
382 	}
383 
384 	return 1;
385 }
386 
387 static char *
388 mklong(const char *str, int ch)
389 {
390 	static char *copy;
391 	static int copysize;
392 	int len;
393 
394 	len = strlen(str) + 2;
395 	if (copysize < len) {
396 		char *newcopy;
397 		copysize = len + 256;
398 
399 		newcopy = realloc(copy, copysize);
400 		if (newcopy == NULL) {
401 			copysize = 0;
402 			free(copy);
403 			copy = NULL;
404 			return (NULL);
405 		}
406 		copy = newcopy;
407 	}
408 	(void) memmove(copy, str, len - 3);
409 	copy[len - 3] = 'l';
410 	copy[len - 2] = ch;
411 	copy[len - 1] = '\0';
412 	return (copy);
413 }
414 
415 static int
416 getchr(void)
417 {
418 	if (!*gargv)
419 		return((int)'\0');
420 	return((int)**gargv++);
421 }
422 
423 static char *
424 getstr(void)
425 {
426 	if (!*gargv)
427 		return("");
428 	return(*gargv++);
429 }
430 
431 static char *number = "+-.0123456789";
432 static int
433 getint(void)
434 {
435 	if (!*gargv)
436 		return(0);
437 
438 	if (strchr(number, **gargv))
439 		return(atoi(*gargv++));
440 
441 	return 0;
442 }
443 
444 static long
445 getlong(void)
446 {
447 	long val;
448 	char *ep;
449 
450 	if (!*gargv)
451 		return(0L);
452 
453 	if (**gargv == '\"' || **gargv == '\'')
454 		return (long) *((*gargv++)+1);
455 
456 	errno = 0;
457 	val = strtol (*gargv, &ep, 0);
458 	check_conversion(*gargv++, ep);
459 	return val;
460 }
461 
462 static unsigned long
463 getulong(void)
464 {
465 	unsigned long val;
466 	char *ep;
467 
468 	if (!*gargv)
469 		return(0UL);
470 
471 	if (**gargv == '\"' || **gargv == '\'')
472 		return (unsigned long) *((*gargv++)+1);
473 
474 	errno = 0;
475 	val = strtoul (*gargv, &ep, 0);
476 	check_conversion(*gargv++, ep);
477 	return val;
478 }
479 
480 static double
481 getdouble(void)
482 {
483 	double val;
484 	char *ep;
485 
486 	if (!*gargv)
487 		return(0.0);
488 
489 	if (**gargv == '\"' || **gargv == '\'')
490 		return (double) *((*gargv++)+1);
491 
492 	errno = 0;
493 	val = strtod (*gargv, &ep);
494 	check_conversion(*gargv++, ep);
495 	return val;
496 }
497 
498 static void
499 check_conversion(const char *s, const char *ep)
500 {
501 	if (*ep) {
502 		if (ep == s)
503 			warnx ("%s: expected numeric value", s);
504 		else
505 			warnx ("%s: not completely converted", s);
506 		rval = 1;
507 	} else if (errno == ERANGE) {
508 		warnx ("%s: %s", s, strerror(ERANGE));
509 		rval = 1;
510 	}
511 }
512 
513 static void
514 usage(void)
515 {
516 	(void)fprintf(stderr, "usage: printf format [arg ...]\n");
517 }
518