1PRINTF(3) Linux Programmer's Manual PRINTF(3)
2
6 printf, fprintf, dprintf, sprintf, snprintf, vprintf, vfprintf, vd‐
7 printf, vsprintf, vsnprintf - formatted output conversion
8
10 #include <stdio.h>
11 int printf(const char *restrict format, ...);
13 int fprintf(FILE *restrict stream,
14 const char *restrict format, ...);
15 int dprintf(int fd,
16 const char *restrict format, ...);
17 int sprintf(char *restrict str,
18 const char *restrict format, ...);
19 int snprintf(char *restrict str, size_t size,
20 const char *restrict format, ...);
21 #include <stdarg.h>
23 int vprintf(const char *restrict format, va_list ap);
25 int vfprintf(FILE *restrict stream,
26 const char *restrict format, va_list ap);
27 int vdprintf(int fd,
28 const char *restrict format, va_list ap);
29 int vsprintf(char *restrict str,
30 const char *restrict format, va_list ap);
31 int vsnprintf(char *restrict str, size_t size,
32 const char *restrict format, va_list ap);
33 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
35 snprintf(), vsnprintf():
37 _XOPEN_SOURCE >= 500 || _ISOC99_SOURCE
38 || /* Glibc <= 2.19: */ _BSD_SOURCE
39 dprintf(), vdprintf():
41 Since glibc 2.10:
42 _POSIX_C_SOURCE >= 200809L
43 Before glibc 2.10:
44 _GNU_SOURCE
45
47 The functions in the printf() family produce output according to a for‐
48 mat as described below. The functions printf() and vprintf() write
49 output to stdout, the standard output stream; fprintf() and vfprintf()
50 write output to the given output stream; sprintf(), snprintf(),
51 vsprintf(), and vsnprintf() write to the character string str.
52 The function dprintf() is the same as fprintf() except that it outputs
54 to a file descriptor, fd, instead of to a stdio stream.
55 The functions snprintf() and vsnprintf() write at most size bytes (in‐
57 cluding the terminating null byte ('\0')) to str.
58 The functions vprintf(), vfprintf(), vdprintf(), vsprintf(), vs‐
60 nprintf() are equivalent to the functions printf(), fprintf(),
61 dprintf(), sprintf(), snprintf(), respectively, except that they are
62 called with a va_list instead of a variable number of arguments. These
63 functions do not call the va_end macro. Because they invoke the va_arg
64 macro, the value of ap is undefined after the call. See stdarg(3).
65 All of these functions write the output under the control of a format
67 string that specifies how subsequent arguments (or arguments accessed
68 via the variable-length argument facilities of stdarg(3)) are converted
69 for output.
70 C99 and POSIX.1-2001 specify that the results are undefined if a call
72 to sprintf(), snprintf(), vsprintf(), or vsnprintf() would cause copy‐
73 ing to take place between objects that overlap (e.g., if the target
74 string array and one of the supplied input arguments refer to the same
75 buffer). See NOTES.
76 Format of the format string
78 The format string is a character string, beginning and ending in its
79 initial shift state, if any. The format string is composed of zero or
80 more directives: ordinary characters (not %), which are copied un‐
81 changed to the output stream; and conversion specifications, each of
82 which results in fetching zero or more subsequent arguments. Each con‐
83 version specification is introduced by the character %, and ends with a
84 conversion specifier. In between there may be (in this order) zero or
85 more flags, an optional minimum field width, an optional precision and
86 an optional length modifier.
87 The overall syntax of a conversion specification is:
89 %[$][flags][width][.precision][length modifier]conversion
91 The arguments must correspond properly (after type promotion) with the
93 conversion specifier. By default, the arguments are used in the order
94 given, where each '*' (see Field width and Precision below) and each
95 conversion specifier asks for the next argument (and it is an error if
96 insufficiently many arguments are given). One can also specify explic‐
97 itly which argument is taken, at each place where an argument is re‐
98 quired, by writing "%m$" instead of '%' and "*m$" instead of '*', where
99 the decimal integer m denotes the position in the argument list of the
100 desired argument, indexed starting from 1. Thus,
101 printf("%*d", width, num);
103 and
105 printf("%2$*1$d", width, num);
107 are equivalent. The second style allows repeated references to the
109 same argument. The C99 standard does not include the style using '$',
110 which comes from the Single UNIX Specification. If the style using '$'
111 is used, it must be used throughout for all conversions taking an argu‐
112 ment and all width and precision arguments, but it may be mixed with
113 "%%" formats, which do not consume an argument. There may be no gaps
114 in the numbers of arguments specified using '$'; for example, if argu‐
115 ments 1 and 3 are specified, argument 2 must also be specified some‐
116 where in the format string.
117 For some numeric conversions a radix character ("decimal point") or
119 thousands' grouping character is used. The actual character used de‐
120 pends on the LC_NUMERIC part of the locale. (See setlocale(3).) The
121 POSIX locale uses '.' as radix character, and does not have a grouping
122 character. Thus,
123 printf("%'.2f", 1234567.89);
125 results in "1234567.89" in the POSIX locale, in "1234567,89" in the
127 nl_NL locale, and in "1.234.567,89" in the da_DK locale.
128 Flag characters
130 The character % is followed by zero or more of the following flags:
131 # The value should be converted to an "alternate form". For o
133 conversions, the first character of the output string is made
134 zero (by prefixing a 0 if it was not zero already). For x and X
135 conversions, a nonzero result has the string "0x" (or "0X" for X
136 conversions) prepended to it. For a, A, e, E, f, F, g, and G
137 conversions, the result will always contain a decimal point,
138 even if no digits follow it (normally, a decimal point appears
139 in the results of those conversions only if a digit follows).
140 For g and G conversions, trailing zeros are not removed from the
141 result as they would otherwise be. For other conversions, the
142 result is undefined.
143 0 The value should be zero padded. For d, i, o, u, x, X, a, A, e,
145 E, f, F, g, and G conversions, the converted value is padded on
146 the left with zeros rather than blanks. If the 0 and - flags
147 both appear, the 0 flag is ignored. If a precision is given
148 with a numeric conversion (d, i, o, u, x, and X), the 0 flag is
149 ignored. For other conversions, the behavior is undefined.
150 - The converted value is to be left adjusted on the field bound‐
152 ary. (The default is right justification.) The converted value
153 is padded on the right with blanks, rather than on the left with
154 blanks or zeros. A - overrides a 0 if both are given.
155 ' ' (a space) A blank should be left before a positive number (or
157 empty string) produced by a signed conversion.
158 + A sign (+ or -) should always be placed before a number produced
160 by a signed conversion. By default, a sign is used only for
161 negative numbers. A + overrides a space if both are used.
162 The five flag characters above are defined in the C99 standard. The
164 Single UNIX Specification specifies one further flag character.
165 ' For decimal conversion (i, d, u, f, F, g, G) the output is to be
167 grouped with thousands' grouping characters if the locale infor‐
168 mation indicates any. (See setlocale(3).) Note that many ver‐
169 sions of gcc(1) cannot parse this option and will issue a warn‐
170 ing. (SUSv2 did not include %'F, but SUSv3 added it.)
171 glibc 2.2 adds one further flag character.
173 I For decimal integer conversion (i, d, u) the output uses the lo‐
175 cale's alternative output digits, if any. For example, since
176 glibc 2.2.3 this will give Arabic-Indic digits in the Persian
177 ("fa_IR") locale.
178 Field width
180 An optional decimal digit string (with nonzero first digit) specifying
181 a minimum field width. If the converted value has fewer characters
182 than the field width, it will be padded with spaces on the left (or
183 right, if the left-adjustment flag has been given). Instead of a deci‐
184 mal digit string one may write "*" or "*m$" (for some decimal integer
185 m) to specify that the field width is given in the next argument, or in
186 the m-th argument, respectively, which must be of type int. A negative
187 field width is taken as a '-' flag followed by a positive field width.
188 In no case does a nonexistent or small field width cause truncation of
189 a field; if the result of a conversion is wider than the field width,
190 the field is expanded to contain the conversion result.
191 Precision
193 An optional precision, in the form of a period ('.') followed by an
194 optional decimal digit string. Instead of a decimal digit string one
195 may write "*" or "*m$" (for some decimal integer m) to specify that the
196 precision is given in the next argument, or in the m-th argument, re‐
197 spectively, which must be of type int. If the precision is given as
198 just '.', the precision is taken to be zero. A negative precision is
199 taken as if the precision were omitted. This gives the minimum number
200 of digits to appear for d, i, o, u, x, and X conversions, the number of
201 digits to appear after the radix character for a, A, e, E, f, and F
202 conversions, the maximum number of significant digits for g and G con‐
203 versions, or the maximum number of characters to be printed from a
204 string for s and S conversions.
205 Length modifier
207 Here, "integer conversion" stands for d, i, o, u, x, or X conversion.
208 hh A following integer conversion corresponds to a signed char or
210 unsigned char argument, or a following n conversion corresponds
211 to a pointer to a signed char argument.
212 h A following integer conversion corresponds to a short or un‐
214 signed short argument, or a following n conversion corresponds
215 to a pointer to a short argument.
216 l (ell) A following integer conversion corresponds to a long or
218 unsigned long argument, or a following n conversion corresponds
219 to a pointer to a long argument, or a following c conversion
220 corresponds to a wint_t argument, or a following s conversion
221 corresponds to a pointer to wchar_t argument.
222 ll (ell-ell). A following integer conversion corresponds to a long
224 long or unsigned long long argument, or a following n conversion
225 corresponds to a pointer to a long long argument.
226 q A synonym for ll. This is a nonstandard extension, derived from
228 BSD; avoid its use in new code.
229 L A following a, A, e, E, f, F, g, or G conversion corresponds to
231 a long double argument. (C99 allows %LF, but SUSv2 does not.)
232 j A following integer conversion corresponds to an intmax_t or
234 uintmax_t argument, or a following n conversion corresponds to a
235 pointer to an intmax_t argument.
236 z A following integer conversion corresponds to a size_t or
238 ssize_t argument, or a following n conversion corresponds to a
239 pointer to a size_t argument.
240 Z A nonstandard synonym for z that predates the appearance of z.
242 Do not use in new code.
243 t A following integer conversion corresponds to a ptrdiff_t argu‐
245 ment, or a following n conversion corresponds to a pointer to a
246 ptrdiff_t argument.
247 SUSv3 specifies all of the above, except for those modifiers explicitly
249 noted as being nonstandard extensions. SUSv2 specified only the length
250 modifiers h (in hd, hi, ho, hx, hX, hn) and l (in ld, li, lo, lx, lX,
251 ln, lc, ls) and L (in Le, LE, Lf, Lg, LG).
252 As a nonstandard extension, the GNU implementations treats ll and L as
254 synonyms, so that one can, for example, write llg (as a synonym for the
255 standards-compliant Lg) and Ld (as a synonym for the standards compli‐
256 ant lld). Such usage is nonportable.
257 Conversion specifiers
259 A character that specifies the type of conversion to be applied. The
260 conversion specifiers and their meanings are:
261 d, i The int argument is converted to signed decimal notation. The
263 precision, if any, gives the minimum number of digits that must
264 appear; if the converted value requires fewer digits, it is
265 padded on the left with zeros. The default precision is 1.
266 When 0 is printed with an explicit precision 0, the output is
267 empty.
268 o, u, x, X
270 The unsigned int argument is converted to unsigned octal (o),
271 unsigned decimal (u), or unsigned hexadecimal (x and X) nota‐
272 tion. The letters abcdef are used for x conversions; the let‐
273 ters ABCDEF are used for X conversions. The precision, if any,
274 gives the minimum number of digits that must appear; if the con‐
275 verted value requires fewer digits, it is padded on the left
276 with zeros. The default precision is 1. When 0 is printed with
277 an explicit precision 0, the output is empty.
278 e, E The double argument is rounded and converted in the style
280 [-]d.ddde±dd where there is one digit (which is nonzero if the
281 argument is nonzero) before the decimal-point character and the
282 number of digits after it is equal to the precision; if the pre‐
283 cision is missing, it is taken as 6; if the precision is zero,
284 no decimal-point character appears. An E conversion uses the
285 letter E (rather than e) to introduce the exponent. The expo‐
286 nent always contains at least two digits; if the value is zero,
287 the exponent is 00.
288 f, F The double argument is rounded and converted to decimal notation
290 in the style [-]ddd.ddd, where the number of digits after the
291 decimal-point character is equal to the precision specification.
292 If the precision is missing, it is taken as 6; if the precision
293 is explicitly zero, no decimal-point character appears. If a
294 decimal point appears, at least one digit appears before it.
295 (SUSv2 does not know about F and says that character string rep‐
297 resentations for infinity and NaN may be made available. SUSv3
298 adds a specification for F. The C99 standard specifies "[-]inf"
299 or "[-]infinity" for infinity, and a string starting with "nan"
300 for NaN, in the case of f conversion, and "[-]INF" or "[-]INFIN‐
301 ITY" or "NAN" in the case of F conversion.)
302 g, G The double argument is converted in style f or e (or F or E for
304 G conversions). The precision specifies the number of signifi‐
305 cant digits. If the precision is missing, 6 digits are given;
306 if the precision is zero, it is treated as 1. Style e is used
307 if the exponent from its conversion is less than -4 or greater
308 than or equal to the precision. Trailing zeros are removed from
309 the fractional part of the result; a decimal point appears only
310 if it is followed by at least one digit.
311 a, A (C99; not in SUSv2, but added in SUSv3) For a conversion, the
313 double argument is converted to hexadecimal notation (using the
314 letters abcdef) in the style [-]0xh.hhhhp±d; for A conversion
315 the prefix 0X, the letters ABCDEF, and the exponent separator P
316 is used. There is one hexadecimal digit before the decimal
317 point, and the number of digits after it is equal to the preci‐
318 sion. The default precision suffices for an exact representa‐
319 tion of the value if an exact representation in base 2 exists
320 and otherwise is sufficiently large to distinguish values of
321 type double. The digit before the decimal point is unspecified
322 for nonnormalized numbers, and nonzero but otherwise unspecified
323 for normalized numbers. The exponent always contains at least
324 one digit; if the value is zero, the exponent is 0.
325 c If no l modifier is present, the int argument is converted to an
327 unsigned char, and the resulting character is written. If an l
328 modifier is present, the wint_t (wide character) argument is
329 converted to a multibyte sequence by a call to the wcrtomb(3)
330 function, with a conversion state starting in the initial state,
331 and the resulting multibyte string is written.
332 s If no l modifier is present: the const char * argument is ex‐
334 pected to be a pointer to an array of character type (pointer to
335 a string). Characters from the array are written up to (but not
336 including) a terminating null byte ('\0'); if a precision is
337 specified, no more than the number specified are written. If a
338 precision is given, no null byte need be present; if the preci‐
339 sion is not specified, or is greater than the size of the array,
340 the array must contain a terminating null byte.
341 If an l modifier is present: the const wchar_t * argument is ex‐
343 pected to be a pointer to an array of wide characters. Wide
344 characters from the array are converted to multibyte characters
345 (each by a call to the wcrtomb(3) function, with a conversion
346 state starting in the initial state before the first wide char‐
347 acter), up to and including a terminating null wide character.
348 The resulting multibyte characters are written up to (but not
349 including) the terminating null byte. If a precision is speci‐
350 fied, no more bytes than the number specified are written, but
351 no partial multibyte characters are written. Note that the pre‐
352 cision determines the number of bytes written, not the number of
353 wide characters or screen positions. The array must contain a
354 terminating null wide character, unless a precision is given and
355 it is so small that the number of bytes written exceeds it be‐
356 fore the end of the array is reached.
357 C (Not in C99 or C11, but in SUSv2, SUSv3, and SUSv4.) Synonym
359 for lc. Don't use.
360 S (Not in C99 or C11, but in SUSv2, SUSv3, and SUSv4.) Synonym
362 for ls. Don't use.
363 p The void * pointer argument is printed in hexadecimal (as if by
365 %#x or %#lx).
366 n The number of characters written so far is stored into the inte‐
368 ger pointed to by the corresponding argument. That argument
369 shall be an int *, or variant whose size matches the (option‐
370 ally) supplied integer length modifier. No argument is con‐
371 verted. (This specifier is not supported by the bionic C li‐
372 brary.) The behavior is undefined if the conversion specifica‐
373 tion includes any flags, a field width, or a precision.
374 m (Glibc extension; supported by uClibc and musl.) Print output
376 of strerror(errno). No argument is required.
377 % A '%' is written. No argument is converted. The complete con‐
379 version specification is '%%'.
380
382 Upon successful return, these functions return the number of characters
383 printed (excluding the null byte used to end output to strings).
384 The functions snprintf() and vsnprintf() do not write more than size
386 bytes (including the terminating null byte ('\0')). If the output was
387 truncated due to this limit, then the return value is the number of
388 characters (excluding the terminating null byte) which would have been
389 written to the final string if enough space had been available. Thus,
390 a return value of size or more means that the output was truncated.
391 (See also below under NOTES.)
392 If an output error is encountered, a negative value is returned.
394
396 For an explanation of the terms used in this section, see at‐
397 tributes(7).
398 ┌─────────────────────────────────────┬───────────────┬────────────────┐
400 │Interface │ Attribute │ Value │
401 ├─────────────────────────────────────┼───────────────┼────────────────┤
402 │printf(), fprintf(), sprintf(), │ Thread safety │ MT-Safe locale │
403 │snprintf(), vprintf(), vfprintf(), │ │ │
404 │vsprintf(), vsnprintf() │ │ │
405 └─────────────────────────────────────┴───────────────┴────────────────┘
406
408 fprintf(), printf(), sprintf(), vprintf(), vfprintf(), vsprintf():
409 POSIX.1-2001, POSIX.1-2008, C89, C99.
410 snprintf(), vsnprintf(): POSIX.1-2001, POSIX.1-2008, C99.
412 The dprintf() and vdprintf() functions were originally GNU extensions
414 that were later standardized in POSIX.1-2008.
415 Concerning the return value of snprintf(), SUSv2 and C99 contradict
417 each other: when snprintf() is called with size=0 then SUSv2 stipulates
418 an unspecified return value less than 1, while C99 allows str to be
419 NULL in this case, and gives the return value (as always) as the number
420 of characters that would have been written in case the output string
421 has been large enough. POSIX.1-2001 and later align their specifica‐
422 tion of snprintf() with C99.
423 glibc 2.1 adds length modifiers hh, j, t, and z and conversion charac‐
425 ters a and A.
426 glibc 2.2 adds the conversion character F with C99 semantics, and the
428 flag character I.
429
431 Some programs imprudently rely on code such as the following
432 sprintf(buf, "%s some further text", buf);
434 to append text to buf. However, the standards explicitly note that the
436 results are undefined if source and destination buffers overlap when
437 calling sprintf(), snprintf(), vsprintf(), and vsnprintf(). Depending
438 on the version of gcc(1) used, and the compiler options employed, calls
439 such as the above will not produce the expected results.
440 The glibc implementation of the functions snprintf() and vsnprintf()
442 conforms to the C99 standard, that is, behaves as described above,
443 since glibc version 2.1. Until glibc 2.0.6, they would return -1 when
444 the output was truncated.
445
447 Because sprintf() and vsprintf() assume an arbitrarily long string,
448 callers must be careful not to overflow the actual space; this is often
449 impossible to assure. Note that the length of the strings produced is
450 locale-dependent and difficult to predict. Use snprintf() and vs‐
451 nprintf() instead (or asprintf(3) and vasprintf(3)).
452 Code such as printf(foo); often indicates a bug, since foo may contain
454 a % character. If foo comes from untrusted user input, it may contain
455 %n, causing the printf() call to write to memory and creating a securi‐
456 ty hole.
457
459 To print Pi to five decimal places:
460 #include <math.h>
462 #include <stdio.h>
463 fprintf(stdout, "pi = %.5f\n", 4 * atan(1.0));
464 To print a date and time in the form "Sunday, July 3, 10:02", where
466 weekday and month are pointers to strings:
467 #include <stdio.h>
469 fprintf(stdout, "%s, %s %d, %.2d:%.2d\n",
470 weekday, month, day, hour, min);
471 Many countries use the day-month-year order. Hence, an international‐
473 ized version must be able to print the arguments in an order specified
474 by the format:
475 #include <stdio.h>
477 fprintf(stdout, format,
478 weekday, month, day, hour, min);
479 where format depends on locale, and may permute the arguments. With
481 the value:
482 "%1$s, %3$d. %2$s, %4$d:%5$.2d\n"
484 one might obtain "Sonntag, 3. Juli, 10:02".
486 To allocate a sufficiently large string and print into it (code correct
488 for both glibc 2.0 and glibc 2.1):
489 #include <stdio.h>
491 #include <stdlib.h>
492 #include <stdarg.h>
493 char *
495 make_message(const char *fmt, ...)
496 {
497 int n = 0;
498 size_t size = 0;
499 char *p = NULL;
500 va_list ap;
501 /* Determine required size. */
503 va_start(ap, fmt);
505 n = vsnprintf(p, size, fmt, ap);
506 va_end(ap);
507 if (n < 0)
509 return NULL;
510 size = (size_t) n + 1; /* One extra byte for '\0' */
512 p = malloc(size);
513 if (p == NULL)
514 return NULL;
515 va_start(ap, fmt);
517 n = vsnprintf(p, size, fmt, ap);
518 va_end(ap);
519 if (n < 0) {
521 free(p);
522 return NULL;
523 }
524 return p;
526 }
527 If truncation occurs in glibc versions prior to 2.0.6, this is treated
529 as an error instead of being handled gracefully.
530
532 printf(1), asprintf(3), puts(3), scanf(3), setlocale(3), strfromd(3),
533 wcrtomb(3), wprintf(3), locale(5)
534
536 This page is part of release 5.12 of the Linux man-pages project. A
537 description of the project, information about reporting bugs, and the
538 latest version of this page, can be found at
539 https://www.kernel.org/doc/man-pages/.
540GNU 2021-03-22 PRINTF(3)