1PRINTF(3) Linux Programmer's Manual PRINTF(3)
2
6 printf, fprintf, sprintf, snprintf, vprintf, vfprintf, vsprintf,
7 vsnprintf - formatted output conversion
8
10 #include <stdio.h>
11 int printf(const char *format, ...);
13 int fprintf(FILE *stream, const char *format, ...);
14 int sprintf(char *str, const char *format, ...);
15 int snprintf(char *str, size_t size, const char *format, ...);
16 #include <stdarg.h>
18 int vprintf(const char *format, va_list ap);
20 int vfprintf(FILE *stream, const char *format, va_list ap);
21 int vsprintf(char *str, const char *format, va_list ap);
22 int vsnprintf(char *str, size_t size, const char *format, va_list ap);
23 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
25 snprintf(), vsnprintf():
27 _BSD_SOURCE || _XOPEN_SOURCE >= 500 || _ISOC99_SOURCE ||
28 _POSIX_C_SOURCE >= 200112L;
29 or cc -std=c99
30
32 The functions in the printf() family produce output according to a for‐
33 mat as described below. The functions printf() and vprintf() write
34 output to stdout, the standard output stream; fprintf() and vfprintf()
35 write output to the given output stream; sprintf(), snprintf(),
36 vsprintf() and vsnprintf() write to the character string str.
37 The functions snprintf() and vsnprintf() write at most size bytes
39 (including the terminating null byte ('\0')) to str.
40 The functions vprintf(), vfprintf(), vsprintf(), vsnprintf() are equiv‐
42 alent to the functions printf(), fprintf(), sprintf(), snprintf(),
43 respectively, except that they are called with a va_list instead of a
44 variable number of arguments. These functions do not call the va_end
45 macro. Because they invoke the va_arg macro, the value of ap is unde‐
46 fined after the call. See stdarg(3).
47 These eight functions write the output under the control of a format
49 string that specifies how subsequent arguments (or arguments accessed
50 via the variable-length argument facilities of stdarg(3)) are converted
51 for output.
52 C99 and POSIX.1-2001 specify that the results are undefined if a call
54 to sprintf(), snprintf(), vsprintf(), or vsnprintf() would cause copy‐
55 ing to take place between objects that overlap (e.g., if the target
56 string array and one of the supplied input arguments refer to the same
57 buffer). See NOTES.
58 Return value
60 Upon successful return, these functions return the number of characters
61 printed (excluding the null byte used to end output to strings).
62 The functions snprintf() and vsnprintf() do not write more than size
64 bytes (including the terminating null byte ('\0')). If the output was
65 truncated due to this limit then the return value is the number of
66 characters (excluding the terminating null byte) which would have been
67 written to the final string if enough space had been available. Thus,
68 a return value of size or more means that the output was truncated.
69 (See also below under NOTES.)
70 If an output error is encountered, a negative value is returned.
72 Format of the format string
74 The format string is a character string, beginning and ending in its
75 initial shift state, if any. The format string is composed of zero or
76 more directives: ordinary characters (not %), which are copied
77 unchanged to the output stream; and conversion specifications, each of
78 which results in fetching zero or more subsequent arguments. Each con‐
79 version specification is introduced by the character %, and ends with a
80 conversion specifier. In between there may be (in this order) zero or
81 more flags, an optional minimum field width, an optional precision and
82 an optional length modifier.
83 The arguments must correspond properly (after type promotion) with the
85 conversion specifier. By default, the arguments are used in the order
86 given, where each '*' and each conversion specifier asks for the next
87 argument (and it is an error if insufficiently many arguments are
88 given). One can also specify explicitly which argument is taken, at
89 each place where an argument is required, by writing "%m$" instead of
90 '%' and "*m$" instead of '*', where the decimal integer m denotes the
91 position in the argument list of the desired argument, indexed starting
92 from 1. Thus,
93 printf("%*d", width, num);
95 and
97 printf("%2$*1$d", width, num);
99 are equivalent. The second style allows repeated references to the
101 same argument. The C99 standard does not include the style using '$',
102 which comes from the Single UNIX Specification. If the style using '$'
103 is used, it must be used throughout for all conversions taking an argu‐
104 ment and all width and precision arguments, but it may be mixed with
105 "%%" formats which do not consume an argument. There may be no gaps in
106 the numbers of arguments specified using '$'; for example, if arguments
107 1 and 3 are specified, argument 2 must also be specified somewhere in
108 the format string.
109 For some numeric conversions a radix character ("decimal point") or
111 thousands' grouping character is used. The actual character used
112 depends on the LC_NUMERIC part of the locale. The POSIX locale uses
113 '.' as radix character, and does not have a grouping character. Thus,
114 printf("%'.2f", 1234567.89);
116 results in "1234567.89" in the POSIX locale, in "1234567,89" in the
118 nl_NL locale, and in "1.234.567,89" in the da_DK locale.
119 The flag characters
121 The character % is followed by zero or more of the following flags:
122 # The value should be converted to an "alternate form". For o
124 conversions, the first character of the output string is made
125 zero (by prefixing a 0 if it was not zero already). For x and X
126 conversions, a nonzero result has the string "0x" (or "0X" for X
127 conversions) prepended to it. For a, A, e, E, f, F, g, and G
128 conversions, the result will always contain a decimal point,
129 even if no digits follow it (normally, a decimal point appears
130 in the results of those conversions only if a digit follows).
131 For g and G conversions, trailing zeros are not removed from the
132 result as they would otherwise be. For other conversions, the
133 result is undefined.
134 0 The value should be zero padded. For d, i, o, u, x, X, a, A, e,
136 E, f, F, g, and G conversions, the converted value is padded on
137 the left with zeros rather than blanks. If the 0 and - flags
138 both appear, the 0 flag is ignored. If a precision is given
139 with a numeric conversion (d, i, o, u, x, and X), the 0 flag is
140 ignored. For other conversions, the behavior is undefined.
141 - The converted value is to be left adjusted on the field bound‐
143 ary. (The default is right justification.) Except for n con‐
144 versions, the converted value is padded on the right with
145 blanks, rather than on the left with blanks or zeros. A - over‐
146 rides a 0 if both are given.
147 ' ' (a space) A blank should be left before a positive number (or
149 empty string) produced by a signed conversion.
150 + A sign (+ or -) should always be placed before a number produced
152 by a signed conversion. By default a sign is used only for neg‐
153 ative numbers. A + overrides a space if both are used.
154 The five flag characters above are defined in the C standard. The
156 SUSv2 specifies one further flag character.
157 ' For decimal conversion (i, d, u, f, F, g, G) the output is to be
159 grouped with thousands' grouping characters if the locale infor‐
160 mation indicates any. Note that many versions of gcc(1) cannot
161 parse this option and will issue a warning. SUSv2 does not
162 include %'F.
163 glibc 2.2 adds one further flag character.
165 I For decimal integer conversion (i, d, u) the output uses the
167 locale's alternative output digits, if any. For example, since
168 glibc 2.2.3 this will give Arabic-Indic digits in the Persian
169 ("fa_IR") locale.
170 The field width
172 An optional decimal digit string (with nonzero first digit) specifying
173 a minimum field width. If the converted value has fewer characters
174 than the field width, it will be padded with spaces on the left (or
175 right, if the left-adjustment flag has been given). Instead of a deci‐
176 mal digit string one may write "*" or "*m$" (for some decimal integer
177 m) to specify that the field width is given in the next argument, or in
178 the m-th argument, respectively, which must be of type int. A negative
179 field width is taken as a '-' flag followed by a positive field width.
180 In no case does a nonexistent or small field width cause truncation of
181 a field; if the result of a conversion is wider than the field width,
182 the field is expanded to contain the conversion result.
183 The precision
185 An optional precision, in the form of a period ('.') followed by an
186 optional decimal digit string. Instead of a decimal digit string one
187 may write "*" or "*m$" (for some decimal integer m) to specify that the
188 precision is given in the next argument, or in the m-th argument,
189 respectively, which must be of type int. If the precision is given as
190 just '.', or the precision is negative, the precision is taken to be
191 zero. This gives the minimum number of digits to appear for d, i, o,
192 u, x, and X conversions, the number of digits to appear after the radix
193 character for a, A, e, E, f, and F conversions, the maximum number of
194 significant digits for g and G conversions, or the maximum number of
195 characters to be printed from a string for s and S conversions.
196 The length modifier
198 Here, "integer conversion" stands for d, i, o, u, x, or X conversion.
199 hh A following integer conversion corresponds to a signed char or
201 unsigned char argument, or a following n conversion corresponds
202 to a pointer to a signed char argument.
203 h A following integer conversion corresponds to a short int or
205 unsigned short int argument, or a following n conversion corre‐
206 sponds to a pointer to a short int argument.
207 l (ell) A following integer conversion corresponds to a long int
209 or unsigned long int argument, or a following n conversion cor‐
210 responds to a pointer to a long int argument, or a following c
211 conversion corresponds to a wint_t argument, or a following s
212 conversion corresponds to a pointer to wchar_t argument.
213 ll (ell-ell). A following integer conversion corresponds to a long
215 long int or unsigned long long int argument, or a following n
216 conversion corresponds to a pointer to a long long int argument.
217 L A following a, A, e, E, f, F, g, or G conversion corresponds to
219 a long double argument. (C99 allows %LF, but SUSv2 does not.)
220 q ("quad". 4.4BSD and Linux libc5 only. Don't use.) This is a
222 synonym for ll.
223 j A following integer conversion corresponds to an intmax_t or
225 uintmax_t argument.
226 z A following integer conversion corresponds to a size_t or
228 ssize_t argument. (Linux libc5 has Z with this meaning. Don't
229 use it.)
230 t A following integer conversion corresponds to a ptrdiff_t argu‐
232 ment.
233 The SUSv2 knows about only the length modifiers h (in hd, hi, ho, hx,
235 hX, hn) and l (in ld, li, lo, lx, lX, ln, lc, ls) and L (in Le, LE, Lf,
236 Lg, LG).
237 The conversion specifier
239 A character that specifies the type of conversion to be applied. The
240 conversion specifiers and their meanings are:
241 d, i The int argument is converted to signed decimal notation. The
243 precision, if any, gives the minimum number of digits that must
244 appear; if the converted value requires fewer digits, it is
245 padded on the left with zeros. The default precision is 1.
246 When 0 is printed with an explicit precision 0, the output is
247 empty.
248 o, u, x, X
250 The unsigned int argument is converted to unsigned octal (o),
251 unsigned decimal (u), or unsigned hexadecimal (x and X) nota‐
252 tion. The letters abcdef are used for x conversions; the let‐
253 ters ABCDEF are used for X conversions. The precision, if any,
254 gives the minimum number of digits that must appear; if the con‐
255 verted value requires fewer digits, it is padded on the left
256 with zeros. The default precision is 1. When 0 is printed with
257 an explicit precision 0, the output is empty.
258 e, E The double argument is rounded and converted in the style
260 [-]d.ddde±dd where there is one digit before the decimal-point
261 character and the number of digits after it is equal to the pre‐
262 cision; if the precision is missing, it is taken as 6; if the
263 precision is zero, no decimal-point character appears. An E
264 conversion uses the letter E (rather than e) to introduce the
265 exponent. The exponent always contains at least two digits; if
266 the value is zero, the exponent is 00.
267 f, F The double argument is rounded and converted to decimal notation
269 in the style [-]ddd.ddd, where the number of digits after the
270 decimal-point character is equal to the precision specification.
271 If the precision is missing, it is taken as 6; if the precision
272 is explicitly zero, no decimal-point character appears. If a
273 decimal point appears, at least one digit appears before it.
274 (The SUSv2 does not know about F and says that character string
276 representations for infinity and NaN may be made available. The
277 C99 standard specifies "[-]inf" or "[-]infinity" for infinity,
278 and a string starting with "nan" for NaN, in the case of f con‐
279 version, and "[-]INF" or "[-]INFINITY" or "NAN*" in the case of
280 F conversion.)
281 g, G The double argument is converted in style f or e (or F or E for
283 G conversions). The precision specifies the number of signifi‐
284 cant digits. If the precision is missing, 6 digits are given;
285 if the precision is zero, it is treated as 1. Style e is used
286 if the exponent from its conversion is less than -4 or greater
287 than or equal to the precision. Trailing zeros are removed from
288 the fractional part of the result; a decimal point appears only
289 if it is followed by at least one digit.
290 a, A (C99; not in SUSv2) For a conversion, the double argument is
292 converted to hexadecimal notation (using the letters abcdef) in
293 the style [-]0xh.hhhhp±; for A conversion the prefix 0X, the
294 letters ABCDEF, and the exponent separator P is used. There is
295 one hexadecimal digit before the decimal point, and the number
296 of digits after it is equal to the precision. The default pre‐
297 cision suffices for an exact representation of the value if an
298 exact representation in base 2 exists and otherwise is suffi‐
299 ciently large to distinguish values of type double. The digit
300 before the decimal point is unspecified for nonnormalized num‐
301 bers, and nonzero but otherwise unspecified for normalized num‐
302 bers.
303 c If no l modifier is present, the int argument is converted to an
305 unsigned char, and the resulting character is written. If an l
306 modifier is present, the wint_t (wide character) argument is
307 converted to a multibyte sequence by a call to the wcrtomb(3)
308 function, with a conversion state starting in the initial state,
309 and the resulting multibyte string is written.
310 s If no l modifier is present: The const char * argument is
312 expected to be a pointer to an array of character type (pointer
313 to a string). Characters from the array are written up to (but
314 not including) a terminating null byte ('\0'); if a precision is
315 specified, no more than the number specified are written. If a
316 precision is given, no null byte need be present; if the preci‐
317 sion is not specified, or is greater than the size of the array,
318 the array must contain a terminating null byte.
319 If an l modifier is present: The const wchar_t * argument is
321 expected to be a pointer to an array of wide characters. Wide
322 characters from the array are converted to multibyte characters
323 (each by a call to the wcrtomb(3) function, with a conversion
324 state starting in the initial state before the first wide char‐
325 acter), up to and including a terminating null wide character.
326 The resulting multibyte characters are written up to (but not
327 including) the terminating null byte. If a precision is speci‐
328 fied, no more bytes than the number specified are written, but
329 no partial multibyte characters are written. Note that the pre‐
330 cision determines the number of bytes written, not the number of
331 wide characters or screen positions. The array must contain a
332 terminating null wide character, unless a precision is given and
333 it is so small that the number of bytes written exceeds it
334 before the end of the array is reached.
335 C (Not in C99, but in SUSv2.) Synonym for lc. Don't use.
337 S (Not in C99, but in SUSv2.) Synonym for ls. Don't use.
339 p The void * pointer argument is printed in hexadecimal (as if by
341 %#x or %#lx).
342 n The number of characters written so far is stored into the inte‐
344 ger indicated by the int * (or variant) pointer argument. No
345 argument is converted.
346 m (Glibc extension.) Print output of strerror(errno). No argu‐
348 ment is required.
349 % A '%' is written. No argument is converted. The complete con‐
351 version specification is '%%'.
352
354 The fprintf(), printf(), sprintf(), vprintf(), vfprintf(), and
355 vsprintf() functions conform to C89 and C99. The snprintf() and
356 vsnprintf() functions conform to C99.
357 Concerning the return value of snprintf(), SUSv2 and C99 contradict
359 each other: when snprintf() is called with size=0 then SUSv2 stipulates
360 an unspecified return value less than 1, while C99 allows str to be
361 NULL in this case, and gives the return value (as always) as the number
362 of characters that would have been written in case the output string
363 has been large enough.
364 Linux libc4 knows about the five C standard flags. It knows about the
366 length modifiers h, l, L, and the conversions c, d, e, E, f, F, g, G,
367 i, n, o, p, s, u, x, and X, where F is a synonym for f. Additionally,
368 it accepts D, O, and U as synonyms for ld, lo, and lu. (This is bad,
369 and caused serious bugs later, when support for %D disappeared.) No
370 locale-dependent radix character, no thousands' separator, no NaN or
371 infinity, no "%m$" and "*m$".
372 Linux libc5 knows about the five C standard flags and the ' flag,
374 locale, "%m$" and "*m$". It knows about the length modifiers h, l, L,
375 Z, and q, but accepts L and q both for long double and for long long
376 int (this is a bug). It no longer recognizes F, D, O, and U, but adds
377 the conversion character m, which outputs strerror(errno).
378 glibc 2.0 adds conversion characters C and S.
380 glibc 2.1 adds length modifiers hh, j, t, and z and conversion charac‐
382 ters a and A.
383 glibc 2.2 adds the conversion character F with C99 semantics, and the
385 flag character I.
386
388 Some programs imprudently rely on code such as the following
389 sprintf(buf, "%s some further text", buf);
391 to append text to buf. However, the standards explicitly note that the
393 results are undefined if source and destination buffers overlap when
394 calling sprintf(), snprintf(), vsprintf(), and vsnprintf(). Depending
395 on the version of gcc(1) used, and the compiler options employed, calls
396 such as the above will not produce the expected results.
397 The glibc implementation of the functions snprintf() and vsnprintf()
399 conforms to the C99 standard, that is, behaves as described above,
400 since glibc version 2.1. Until glibc 2.0.6 they would return -1 when
401 the output was truncated.
402
404 Because sprintf() and vsprintf() assume an arbitrarily long string,
405 callers must be careful not to overflow the actual space; this is often
406 impossible to assure. Note that the length of the strings produced is
407 locale-dependent and difficult to predict. Use snprintf() and
408 vsnprintf() instead (or asprintf(3) and vasprintf(3)).
409 Linux libc4.[45] does not have a snprintf(), but provides a libbsd that
411 contains an snprintf() equivalent to sprintf(), that is, one that
412 ignores the size argument. Thus, the use of snprintf() with early
413 libc4 leads to serious security problems.
414 Code such as printf(foo); often indicates a bug, since foo may contain
416 a % character. If foo comes from untrusted user input, it may contain
417 %n, causing the printf() call to write to memory and creating a secu‐
418 rity hole.
419
421 To print Pi to five decimal places:
422 #include <math.h>
424 #include <stdio.h>
425 fprintf(stdout, "pi = %.5f\n", 4 * atan(1.0));
426 To print a date and time in the form "Sunday, July 3, 10:02", where
428 weekday and month are pointers to strings:
429 #include <stdio.h>
431 fprintf(stdout, "%s, %s %d, %.2d:%.2d\n",
432 weekday, month, day, hour, min);
433 Many countries use the day-month-year order. Hence, an international‐
435 ized version must be able to print the arguments in an order specified
436 by the format:
437 #include <stdio.h>
439 fprintf(stdout, format,
440 weekday, month, day, hour, min);
441 where format depends on locale, and may permute the arguments. With
443 the value:
444 "%1$s, %3$d. %2$s, %4$d:%5$.2d\n"
446 one might obtain "Sonntag, 3. Juli, 10:02".
448 To allocate a sufficiently large string and print into it (code correct
450 for both glibc 2.0 and glibc 2.1):
451 If truncation occurs in glibc versions prior to 2.0.6, this is treated
453 as an error instead of being handled gracefully.
454 #include <stdio.h>
456 #include <stdlib.h>
457 #include <stdarg.h>
458 char *
460 make_message(const char *fmt, ...)
461 {
462 int n;
463 int size = 100; /* Guess we need no more than 100 bytes */
464 char *p, *np;
465 va_list ap;
466 if ((p = malloc(size)) == NULL)
468 return NULL;
469 while (1) {
471 /* Try to print in the allocated space */
473 va_start(ap, fmt);
475 n = vsnprintf(p, size, fmt, ap);
476 va_end(ap);
477 /* Check error code */
479 if (n < 0)
481 return NULL;
482 /* If that worked, return the string */
484 if (n < size)
486 return p;
487 /* Else try again with more space */
489 size = n + 1; /* Precisely what is needed */
491 if ((np = realloc (p, size)) == NULL) {
494 free(p);
495 return NULL;
496 } else {
497 p = np;
498 }
499 }
500 }
501
503 printf(1), asprintf(3), dprintf(3), scanf(3), setlocale(3), wcrtomb(3),
504 wprintf(3), locale(5)
505
507 This page is part of release 3.53 of the Linux man-pages project. A
508 description of the project, and information about reporting bugs, can
509 be found at http://www.kernel.org/doc/man-pages/.
510GNU 2013-03-05 PRINTF(3)