busybox/coreutils/printf.c

/* vi: set sw=4 ts=4: */
/* printf - format and print data

   Copyright 1999 Dave Cinege
   Portions copyright (C) 1990-1996 Free Software Foundation, Inc.

   Licensed under GPL v2 or later, see file LICENSE in this tarball for details.
*/

/* Usage: printf format [argument...]

   A front end to the printf function that lets it be used from the shell.

   Backslash escapes:

   \" = double quote
   \\ = backslash
   \a = alert (bell)
   \b = backspace
   \c = produce no further output
   \f = form feed
   \n = new line
   \r = carriage return
   \t = horizontal tab
   \v = vertical tab
   \0ooo = octal number (ooo is 0 to 3 digits)
   \xhhh = hexadecimal number (hhh is 1 to 3 digits)

   Additional directive:

   %b = print an argument string, interpreting backslash escapes

   The 'format' argument is re-used as many times as necessary
   to convert all of the given arguments.

   David MacKenzie <djm@gnu.ai.mit.edu>
*/

//   19990508 Busy Boxed! Dave Cinege

#include "libbb.h"

/* A note on bad input: neither bash 3.2 nor coreutils 6.10 stop on it.
 * They report it:
 *  bash: printf: XXX: invalid number
 *  printf: XXX: expected a numeric value
 *  bash: printf: 123XXX: invalid number
 *  printf: 123XXX: value not completely converted
 * but then they use 0 (or partially converted numeric prefix) as a value
 * and continue. They exit with 1 in this case.
 * Both accept insane field width/precision (e.g. %9999999999.9999999999d).
 * Both print error message and assume 0 if %*.*f width/precision is "bad"
 *  (but negative numbers are not "bad").
 * Both accept negative numbers for %u specifier.
 *
 * We try to be compatible.
 */

typedef void FAST_FUNC (*converter)(const char *arg, void *result);

static int multiconvert(const char *arg, void *result, converter convert)
{
	if (*arg == '"' || *arg == '\'') {
		arg = utoa((unsigned char)arg[1]);
	}
	errno = 0;
	convert(arg, result);
	if (errno) {
		bb_error_msg("%s: invalid number", arg);
		return 1;
	}
	return 0;
}

static void FAST_FUNC conv_strtoull(const char *arg, void *result)
{
	*(unsigned long long*)result = bb_strtoull(arg, NULL, 0);
	/* both coreutils 6.10 and bash 3.2:
	 * $ printf '%x\n' -2
	 * fffffffffffffffe
	 * Mimic that:
	 */
	if (errno) {
		*(unsigned long long*)result = bb_strtoll(arg, NULL, 0);
	}
}
static void FAST_FUNC conv_strtoll(const char *arg, void *result)
{
	*(long long*)result = bb_strtoll(arg, NULL, 0);
}
static void FAST_FUNC conv_strtod(const char *arg, void *result)
{
	char *end;
	/* Well, this one allows leading whitespace... so what? */
	/* What I like much less is that "-" accepted too! :( */
	*(double*)result = strtod(arg, &end);
	if (end[0]) {
		errno = ERANGE;
		*(double*)result = 0;
	}
}

/* Callers should check errno to detect errors */
static unsigned long long my_xstrtoull(const char *arg)
{
	unsigned long long result;
	if (multiconvert(arg, &result, conv_strtoull))
		result = 0;
	return result;
}
static long long my_xstrtoll(const char *arg)
{
	long long result;
	if (multiconvert(arg, &result, conv_strtoll))
		result = 0;
	return result;
}
static double my_xstrtod(const char *arg)
{
	double result;
	multiconvert(arg, &result, conv_strtod);
	return result;
}

static void print_esc_string(char *str)
{
	while (*str) {
		if (*str == '\\') {
			str++;
			bb_putchar(bb_process_escape_sequence((const char **)&str));
		} else {
			bb_putchar(*str);
			str++;
		}
	}
}

static void print_direc(char *format, unsigned fmt_length,
		int field_width, int precision,
		const char *argument)
{
	long long llv;
	double dv;
	char saved;
	char *have_prec, *have_width;

	saved = format[fmt_length];
	format[fmt_length] = '\0';

	have_prec = strstr(format, ".*");
	have_width = strchr(format, '*');
	if (have_width - 1 == have_prec)
		have_width = NULL;

	errno = 0;

	switch (format[fmt_length - 1]) {
	case 'c':
		printf(format, *argument);
		break;
	case 'd':
	case 'i':
		llv = my_xstrtoll(argument);
 print_long:
		if (!have_width) {
			if (!have_prec)
				printf(format, llv);
			else
				printf(format, precision, llv);
		} else {
			if (!have_prec)
				printf(format, field_width, llv);
			else
				printf(format, field_width, precision, llv);
		}
		break;
	case 'o':
	case 'u':
	case 'x':
	case 'X':
		llv = my_xstrtoull(argument);
		/* cheat: unsigned long and long have same width, so... */
		goto print_long;
	case 's':
		/* Are char* and long long the same? */
		if (sizeof(argument) == sizeof(llv)) {
			llv = (long long)(ptrdiff_t)argument;
			goto print_long;
		} else {
			/* Hope compiler will optimize it out by moving call
			 * instruction after the ifs... */
			if (!have_width) {
				if (!have_prec)
					printf(format, argument, /*unused:*/ argument, argument);
				else
					printf(format, precision, argument, /*unused:*/ argument);
			} else {
				if (!have_prec)
					printf(format, field_width, argument, /*unused:*/ argument);
				else
					printf(format, field_width, precision, argument);
			}
			break;
		}
	case 'f':
	case 'e':
	case 'E':
	case 'g':
	case 'G':
		dv = my_xstrtod(argument);
		if (!have_width) {
			if (!have_prec)
				printf(format, dv);
			else
				printf(format, precision, dv);
		} else {
			if (!have_prec)
				printf(format, field_width, dv);
			else
				printf(format, field_width, precision, dv);
		}
		break;
	} /* switch */

	format[fmt_length] = saved;
}

/* Handle params for "%*.*f". Negative numbers are ok (compat). */
static int get_width_prec(const char *str)
{
	int v = bb_strtoi(str, NULL, 10);
	if (errno) {
		bb_error_msg("%s: invalid number", str);
		v = 0;
	}
	return v;
}

/* Print the text in FORMAT, using ARGV for arguments to any '%' directives.
   Return advanced ARGV.  */
static char **print_formatted(char *f, char **argv, int *conv_err)
{
	char *direc_start;      /* Start of % directive.  */
	unsigned direc_length;  /* Length of % directive.  */
	int field_width;        /* Arg to first '*' */
	int precision;          /* Arg to second '*' */
	char **saved_argv = argv;

	for (; *f; ++f) {
		switch (*f) {
		case '%':
			direc_start = f++;
			direc_length = 1;
			field_width = precision = 0;
			if (*f == '%') {
				bb_putchar('%');
				break;
			}
			if (*f == 'b') {
				if (*argv) {
					print_esc_string(*argv);
					++argv;
				}
				break;
			}
			if (strchr("-+ #", *f)) {
				++f;
				++direc_length;
			}
			if (*f == '*') {
				++f;
				++direc_length;
				if (*argv)
					field_width = get_width_prec(*argv++);
			} else {
				while (isdigit(*f)) {
					++f;
					++direc_length;
				}
			}
			if (*f == '.') {
				++f;
				++direc_length;
				if (*f == '*') {
					++f;
					++direc_length;
					if (*argv)
						precision = get_width_prec(*argv++);
				} else {
					while (isdigit(*f)) {
						++f;
						++direc_length;
					}
				}
			}

			/* Remove "lLhz" size modifiers, repeatedly.
			 * bash does not like "%lld", but coreutils
			 * happily takes even "%Llllhhzhhzd"!
			 * We are permissive like coreutils */
			while ((*f | 0x20) == 'l' || *f == 'h' || *f == 'z') {
				overlapping_strcpy(f, f + 1);
			}
			/* Add "ll" if integer modifier, then print */
			{
				static const char format_chars[] ALIGN1 = "diouxXfeEgGcs";
				char *p = strchr(format_chars, *f);
				/* needed - try "printf %" without it */
				if (p == NULL) {
					bb_error_msg("%s: invalid format", direc_start);
					/* causes main() to exit with error */
					return saved_argv - 1;
				}
				++direc_length;
				if (p - format_chars <= 5) {
					/* it is one of "diouxX" */
					p = xmalloc(direc_length + 3);
					memcpy(p, direc_start, direc_length);
					p[direc_length + 1] = p[direc_length - 1];
					p[direc_length - 1] = 'l';
					p[direc_length] = 'l';
					//bb_error_msg("<%s>", p);
					direc_length += 2;
					direc_start = p;
				} else {
					p = NULL;
				}
				if (*argv) {
					print_direc(direc_start, direc_length, field_width,
								precision, *argv++);
				} else {
					print_direc(direc_start, direc_length, field_width,
								precision, "");
				}
				*conv_err |= errno;
				free(p);
			}
			break;
		case '\\':
			if (*++f == 'c') {
				return saved_argv; /* causes main() to exit */
			}
			bb_putchar(bb_process_escape_sequence((const char **)&f));
			f--;
			break;
		default:
			bb_putchar(*f);
		}
	}

	return argv;
}

int printf_main(int argc UNUSED_PARAM, char **argv)
{
	int conv_err;
	char *format;
	char **argv2;

	/* We must check that stdout is not closed.
	 * The reason for this is highly non-obvious.
	 * printf_main is used from shell.
	 * Shell must correctly handle 'printf "%s" foo'
	 * if stdout is closed. With stdio, output gets shoveled into
	 * stdout buffer, and even fflush cannot clear it out. It seems that
	 * even if libc receives EBADF on write attempts, it feels determined
	 * to output data no matter what. So it will try later,
	 * and possibly will clobber future output. Not good. */
// TODO: check fcntl() & O_ACCMODE == O_WRONLY or O_RDWR?
	if (fcntl(1, F_GETFL) == -1)
		return 1; /* match coreutils 6.10 (sans error msg to stderr) */
	//if (dup2(1, 1) != 1) - old way
	//	return 1;

	/* bash builtin errors out on "printf '-%s-\n' foo",
	 * coreutils-6.9 works. Both work with "printf -- '-%s-\n' foo".
	 * We will mimic coreutils. */
	if (argv[1] && argv[1][0] == '-' && argv[1][1] == '-' && !argv[1][2])
		argv++;
	if (!argv[1]) {
		if (ENABLE_ASH_BUILTIN_PRINTF
		 && applet_name[0] != 'p'
		) {
			bb_error_msg("usage: printf FORMAT [ARGUMENT...]");
			return 2; /* bash compat */
		}
		bb_show_usage();
	}

	format = argv[1];
	argv2 = argv + 2;

	conv_err = 0;
	do {
		argv = argv2;
		argv2 = print_formatted(format, argv, &conv_err);
	} while (argv2 > argv && *argv2);

	/* coreutils compat (bash doesn't do this):
	if (*argv)
		fprintf(stderr, "excess args ignored");
	*/

	return (argv2 < argv) /* if true, print_formatted errored out */
		|| conv_err; /* print_formatted saw invalid number */
}
1	/* vi: set sw=4 ts=4: */
2	/* printf - format and print data
3
4	Copyright 1999 Dave Cinege
5	Portions copyright (C) 1990-1996 Free Software Foundation, Inc.
6
7	Licensed under GPL v2 or later, see file LICENSE in this tarball for details.
8	*/
9
10	/* Usage: printf format [argument...]
11
12	A front end to the printf function that lets it be used from the shell.
13
14	Backslash escapes:
15
16	\" = double quote
17	\\ = backslash
18	\a = alert (bell)
19	\b = backspace
20	\c = produce no further output
21	\f = form feed
22	\n = new line
23	\r = carriage return
24	\t = horizontal tab
25	\v = vertical tab
26	\0ooo = octal number (ooo is 0 to 3 digits)
27	\xhhh = hexadecimal number (hhh is 1 to 3 digits)
28
29	Additional directive:
30
31	%b = print an argument string, interpreting backslash escapes
32
33	The 'format' argument is re-used as many times as necessary
34	to convert all of the given arguments.
35
36	David MacKenzie <djm@gnu.ai.mit.edu>
37	*/
38
39	// 19990508 Busy Boxed! Dave Cinege
40
41	#include "libbb.h"
42
43	/* A note on bad input: neither bash 3.2 nor coreutils 6.10 stop on it.
44	* They report it:
45	* bash: printf: XXX: invalid number
46	* printf: XXX: expected a numeric value
47	* bash: printf: 123XXX: invalid number
48	* printf: 123XXX: value not completely converted
49	* but then they use 0 (or partially converted numeric prefix) as a value
50	* and continue. They exit with 1 in this case.
51	* Both accept insane field width/precision (e.g. %9999999999.9999999999d).
52	* Both print error message and assume 0 if %.f width/precision is "bad"
53	* (but negative numbers are not "bad").
54	* Both accept negative numbers for %u specifier.
55	*
56	* We try to be compatible.
57	*/
58
59	typedef void FAST_FUNC (converter)(const char arg, void *result);
60
61	static int multiconvert(const char arg, void result, converter convert)
62	{
63	if (arg == '"' \|\| arg == '\'') {
64	arg = utoa((unsigned char)arg[1]);
65	}
66	errno = 0;
67	convert(arg, result);
68	if (errno) {
69	bb_error_msg("%s: invalid number", arg);
70	return 1;
71	}
72	return 0;
73	}
74
75	static void FAST_FUNC conv_strtoull(const char arg, void result)
76	{
77	(unsigned long long)result = bb_strtoull(arg, NULL, 0);
78	/* both coreutils 6.10 and bash 3.2:
79	* $ printf '%x\n' -2
80	* fffffffffffffffe
81	* Mimic that:
82	*/
83	if (errno) {
84	(unsigned long long)result = bb_strtoll(arg, NULL, 0);
85	}
86	}
87	static void FAST_FUNC conv_strtoll(const char arg, void result)
88	{
89	(long long)result = bb_strtoll(arg, NULL, 0);
90	}
91	static void FAST_FUNC conv_strtod(const char arg, void result)
92	{
93	char *end;
94	/* Well, this one allows leading whitespace... so what? */
95	/* What I like much less is that "-" accepted too! :( */
96	(double)result = strtod(arg, &end);
97	if (end[0]) {
98	errno = ERANGE;
99	(double)result = 0;
100	}
101	}
102
103	/* Callers should check errno to detect errors */
104	static unsigned long long my_xstrtoull(const char *arg)
105	{
106	unsigned long long result;
107	if (multiconvert(arg, &result, conv_strtoull))
108	result = 0;
109	return result;
110	}
111	static long long my_xstrtoll(const char *arg)
112	{
113	long long result;
114	if (multiconvert(arg, &result, conv_strtoll))
115	result = 0;
116	return result;
117	}
118	static double my_xstrtod(const char *arg)
119	{
120	double result;
121	multiconvert(arg, &result, conv_strtod);
122	return result;
123	}
124
125	static void print_esc_string(char *str)
126	{
127	while (*str) {
128	if (*str == '\\') {
129	str++;
130	bb_putchar(bb_process_escape_sequence((const char **)&str));
131	} else {
132	bb_putchar(*str);
133	str++;
134	}
135	}
136	}
137
138	static void print_direc(char *format, unsigned fmt_length,
139	int field_width, int precision,
140	const char *argument)
141	{
142	long long llv;
143	double dv;
144	char saved;
145	char have_prec, have_width;
146
147	saved = format[fmt_length];
148	format[fmt_length] = '\0';
149
150	have_prec = strstr(format, ".*");
151	have_width = strchr(format, '*');
152	if (have_width - 1 == have_prec)
153	have_width = NULL;
154
155	errno = 0;
156
157	switch (format[fmt_length - 1]) {
158	case 'c':
159	printf(format, *argument);
160	break;
161	case 'd':
162	case 'i':
163	llv = my_xstrtoll(argument);
164	print_long:
165	if (!have_width) {
166	if (!have_prec)
167	printf(format, llv);
168	else
169	printf(format, precision, llv);
170	} else {
171	if (!have_prec)
172	printf(format, field_width, llv);
173	else
174	printf(format, field_width, precision, llv);
175	}
176	break;
177	case 'o':
178	case 'u':
179	case 'x':
180	case 'X':
181	llv = my_xstrtoull(argument);
182	/* cheat: unsigned long and long have same width, so... */
183	goto print_long;
184	case 's':
185	/* Are char* and long long the same? */
186	if (sizeof(argument) == sizeof(llv)) {
187	llv = (long long)(ptrdiff_t)argument;
188	goto print_long;
189	} else {
190	/* Hope compiler will optimize it out by moving call
191	* instruction after the ifs... */
192	if (!have_width) {
193	if (!have_prec)
194	printf(format, argument, /unused:/ argument, argument);
195	else
196	printf(format, precision, argument, /unused:/ argument);
197	} else {
198	if (!have_prec)
199	printf(format, field_width, argument, /unused:/ argument);
200	else
201	printf(format, field_width, precision, argument);
202	}
203	break;
204	}
205	case 'f':
206	case 'e':
207	case 'E':
208	case 'g':
209	case 'G':
210	dv = my_xstrtod(argument);
211	if (!have_width) {
212	if (!have_prec)
213	printf(format, dv);
214	else
215	printf(format, precision, dv);
216	} else {
217	if (!have_prec)
218	printf(format, field_width, dv);
219	else
220	printf(format, field_width, precision, dv);
221	}
222	break;
223	} /* switch */
224
225	format[fmt_length] = saved;
226	}
227
228	/* Handle params for "%.f". Negative numbers are ok (compat). */
229	static int get_width_prec(const char *str)
230	{
231	int v = bb_strtoi(str, NULL, 10);
232	if (errno) {
233	bb_error_msg("%s: invalid number", str);
234	v = 0;
235	}
236	return v;
237	}
238
239	/* Print the text in FORMAT, using ARGV for arguments to any '%' directives.
240	Return advanced ARGV. */
241	static char *print_formatted(char f, char *argv, int conv_err)
242	{
243	char direc_start; / Start of % directive. */
244	unsigned direc_length; /* Length of % directive. */
245	int field_width; /* Arg to first '' /
246	int precision; /* Arg to second '' /
247	char **saved_argv = argv;
248
249	for (; *f; ++f) {
250	switch (*f) {
251	case '%':
252	direc_start = f++;
253	direc_length = 1;
254	field_width = precision = 0;
255	if (*f == '%') {
256	bb_putchar('%');
257	break;
258	}
259	if (*f == 'b') {
260	if (*argv) {
261	print_esc_string(*argv);
262	++argv;
263	}
264	break;
265	}
266	if (strchr("-+ #", *f)) {
267	++f;
268	++direc_length;
269	}
270	if (f == '') {
271	++f;
272	++direc_length;
273	if (*argv)
274	field_width = get_width_prec(*argv++);
275	} else {
276	while (isdigit(*f)) {
277	++f;
278	++direc_length;
279	}
280	}
281	if (*f == '.') {
282	++f;
283	++direc_length;
284	if (f == '') {
285	++f;
286	++direc_length;
287	if (*argv)
288	precision = get_width_prec(*argv++);
289	} else {
290	while (isdigit(*f)) {
291	++f;
292	++direc_length;
293	}
294	}
295	}
296
297	/* Remove "lLhz" size modifiers, repeatedly.
298	* bash does not like "%lld", but coreutils
299	* happily takes even "%Llllhhzhhzd"!
300	* We are permissive like coreutils */
301	while ((f \| 0x20) == 'l' \|\| f == 'h' \|\| *f == 'z') {
302	overlapping_strcpy(f, f + 1);
303	}
304	/* Add "ll" if integer modifier, then print */
305	{
306	static const char format_chars[] ALIGN1 = "diouxXfeEgGcs";
307	char p = strchr(format_chars, f);
308	/* needed - try "printf %" without it */
309	if (p == NULL) {
310	bb_error_msg("%s: invalid format", direc_start);
311	/* causes main() to exit with error */
312	return saved_argv - 1;
313	}
314	++direc_length;
315	if (p - format_chars <= 5) {
316	/* it is one of "diouxX" */
317	p = xmalloc(direc_length + 3);
318	memcpy(p, direc_start, direc_length);
319	p[direc_length + 1] = p[direc_length - 1];
320	p[direc_length - 1] = 'l';
321	p[direc_length] = 'l';
322	//bb_error_msg("<%s>", p);
323	direc_length += 2;
324	direc_start = p;
325	} else {
326	p = NULL;
327	}
328	if (*argv) {
329	print_direc(direc_start, direc_length, field_width,
330	precision, *argv++);
331	} else {
332	print_direc(direc_start, direc_length, field_width,
333	precision, "");
334	}
335	*conv_err \|= errno;
336	free(p);
337	}
338	break;
339	case '\\':
340	if (*++f == 'c') {
341	return saved_argv; /* causes main() to exit */
342	}
343	bb_putchar(bb_process_escape_sequence((const char **)&f));
344	f--;
345	break;
346	default:
347	bb_putchar(*f);
348	}
349	}
350
351	return argv;
352	}
353
354	int printf_main(int argc UNUSED_PARAM, char **argv)
355	{
356	int conv_err;
357	char *format;
358	char **argv2;
359
360	/* We must check that stdout is not closed.
361	* The reason for this is highly non-obvious.
362	* printf_main is used from shell.
363	* Shell must correctly handle 'printf "%s" foo'
364	* if stdout is closed. With stdio, output gets shoveled into
365	* stdout buffer, and even fflush cannot clear it out. It seems that
366	* even if libc receives EBADF on write attempts, it feels determined
367	* to output data no matter what. So it will try later,
368	* and possibly will clobber future output. Not good. */
369	// TODO: check fcntl() & O_ACCMODE == O_WRONLY or O_RDWR?
370	if (fcntl(1, F_GETFL) == -1)
371	return 1; /* match coreutils 6.10 (sans error msg to stderr) */
372	//if (dup2(1, 1) != 1) - old way
373	// return 1;
374
375	/* bash builtin errors out on "printf '-%s-\n' foo",
376	* coreutils-6.9 works. Both work with "printf -- '-%s-\n' foo".
377	* We will mimic coreutils. */
378	if (argv[1] && argv[1][0] == '-' && argv[1][1] == '-' && !argv[1][2])
379	argv++;
380	if (!argv[1]) {
381	if (ENABLE_ASH_BUILTIN_PRINTF
382	&& applet_name[0] != 'p'
383	) {
384	bb_error_msg("usage: printf FORMAT [ARGUMENT...]");
385	return 2; /* bash compat */
386	}
387	bb_show_usage();
388	}
389
390	format = argv[1];
391	argv2 = argv + 2;
392
393	conv_err = 0;
394	do {
395	argv = argv2;
396	argv2 = print_formatted(format, argv, &conv_err);
397	} while (argv2 > argv && *argv2);
398
399	/* coreutils compat (bash doesn't do this):
400	if (*argv)
401	fprintf(stderr, "excess args ignored");
402	*/
403
404	return (argv2 < argv) /* if true, print_formatted errored out */
405	\|\| conv_err; /* print_formatted saw invalid number */
406	}