1 .\" Copyright (c) 1999 Andries Brouwer (aeb@cwi.nl)
3 .\" Earlier versions of this page influenced the present text.
4 .\" It was derived from a Berkeley page with version
5 .\" @(#)printf.3 6.14 (Berkeley) 7/30/91
6 .\" converted for Linux by faith@cs.unc.edu, updated by
7 .\" Helmut.Geyer@iwr.uni-heidelberg.de, agulbra@troll.no and Bruno Haible.
9 .\" %%%LICENSE_START(GPLv2+_DOC_FULL)
10 .\" This is free documentation; you can redistribute it and/or
11 .\" modify it under the terms of the GNU General Public License as
12 .\" published by the Free Software Foundation; either version 2 of
13 .\" the License, or (at your option) any later version.
15 .\" The GNU General Public License's references to "object code"
16 .\" and "executables" are to be interpreted as the output of any
17 .\" document formatting or typesetting system, including
18 .\" intermediate and printed output.
20 .\" This manual is distributed in the hope that it will be useful,
21 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
22 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23 .\" GNU General Public License for more details.
25 .\" You should have received a copy of the GNU General Public
26 .\" License along with this manual; if not, see
27 .\" <http://www.gnu.org/licenses/>.
30 .\" 1999-11-25 aeb - Rewritten, using SUSv2 and C99.
31 .\" 2000-07-26 jsm28@hermes.cam.ac.uk - three small fixes
32 .\" 2000-10-16 jsm28@hermes.cam.ac.uk - more fixes
34 .TH PRINTF 3 2013-03-05 "GNU" "Linux Programmer's Manual"
36 printf, fprintf, sprintf, snprintf, vprintf, vfprintf, vsprintf,
37 vsnprintf \- formatted output conversion
41 .BI "int printf(const char *" format ", ...);"
43 .BI "int fprintf(FILE *" stream ", const char *" format ", ...);"
45 .BI "int sprintf(char *" str ", const char *" format ", ...);"
47 .BI "int snprintf(char *" str ", size_t " size ", const char *" format ", ...);"
49 .B #include <stdarg.h>
51 .BI "int vprintf(const char *" format ", va_list " ap );
53 .BI "int vfprintf(FILE *" stream ", const char *" format ", va_list " ap );
55 .BI "int vsprintf(char *" str ", const char *" format ", va_list " ap );
57 .BI "int vsnprintf(char *" str ", size_t " size ", const char *" format \
61 Feature Test Macro Requirements for glibc (see
62 .BR feature_test_macros (7)):
69 _BSD_SOURCE || _XOPEN_SOURCE\ >=\ 500 || _ISOC99_SOURCE ||
70 _POSIX_C_SOURCE\ >=\ 200112L;
79 family produce output according to a
88 the standard output stream;
92 write output to the given output
99 write to the character string
108 bytes (including the terminating null byte (\(aq\e0\(aq)) to
116 are equivalent to the functions
121 respectively, except that they are called with a
123 instead of a variable number of arguments.
124 These functions do not call the
127 Because they invoke the
131 is undefined after the call.
135 These eight functions write the output under the control of a
137 string that specifies how subsequent arguments (or arguments accessed via
138 the variable-length argument facilities of
140 are converted for output.
142 C99 and POSIX.1-2001 specify that the results are undefined if a call to
148 would cause copying to take place between objects that overlap
149 (e.g., if the target string array and one of the supplied input arguments
150 refer to the same buffer).
153 Upon successful return, these functions return the number of characters
154 printed (excluding the null byte used to end output to strings).
160 do not write more than
162 bytes (including the terminating null byte (\(aq\e0\(aq)).
163 If the output was truncated due to this limit then the return value
164 is the number of characters (excluding the terminating null byte)
165 which would have been written to the final string if enough space
167 Thus, a return value of
169 or more means that the output was truncated.
170 (See also below under NOTES.)
172 If an output error is encountered, a negative value is returned.
173 .SS Format of the format string
174 The format string is a character string, beginning and ending
175 in its initial shift state, if any.
176 The format string is composed of zero or more directives: ordinary
179 which are copied unchanged to the output stream;
180 and conversion specifications, each of which results in fetching zero or
181 more subsequent arguments.
182 Each conversion specification is introduced by
186 .IR "conversion specifier" .
187 In between there may be (in this order) zero or more
194 .IR "length modifier" .
196 The arguments must correspond properly (after type promotion) with the
197 conversion specifier.
198 By default, the arguments are used in the order
199 given, where each \(aq*\(aq and each conversion specifier asks for the next
200 argument (and it is an error if insufficiently many arguments are given).
201 One can also specify explicitly which argument is taken,
202 at each place where an argument is required, by writing "%m$" instead
203 of \(aq%\(aq and "*m$" instead of \(aq*\(aq,
204 where the decimal integer m denotes
205 the position in the argument list of the desired argument, indexed starting
211 printf("%*d", width, num);
219 printf("%2$*1$d", width, num);
224 The second style allows repeated references to the
226 The C99 standard does not include the style using \(aq$\(aq,
227 which comes from the Single UNIX Specification.
229 \(aq$\(aq is used, it must be used throughout for all conversions taking an
230 argument and all width and precision arguments, but it may be mixed
231 with "%%" formats which do not consume an argument.
233 gaps in the numbers of arguments specified using \(aq$\(aq; for example, if
234 arguments 1 and 3 are specified, argument 2 must also be specified
235 somewhere in the format string.
237 For some numeric conversions a radix character ("decimal point") or
238 thousands' grouping character is used.
239 The actual character used
244 uses \(aq.\(aq as radix character, and does not have a grouping character.
249 printf("%\(aq.2f", 1234567.89);
253 results in "1234567.89" in the POSIX locale, in "1234567,89" in the
254 nl_NL locale, and in "1.234.567,89" in the da_DK locale.
255 .SS The flag characters
256 The character % is followed by zero or more of the following flags:
259 The value should be converted to an "alternate form".
262 conversions, the first character of the output string is made zero
263 (by prefixing a 0 if it was not zero already).
268 conversions, a nonzero result has the string "0x" (or "0X" for
270 conversions) prepended to it.
281 conversions, the result will always contain a decimal point, even if no
282 digits follow it (normally, a decimal point appears in the results of those
283 conversions only if a digit follows).
288 conversions, trailing zeros are not removed from the result as they would
290 For other conversions, the result is undefined.
293 The value should be zero padded.
310 conversions, the converted value is padded on the left with zeros rather
316 flags both appear, the
319 If a precision is given with a numeric conversion
330 For other conversions, the behavior is undefined.
333 The converted value is to be left adjusted on the field boundary.
334 (The default is right justification.)
337 conversions, the converted value is padded on the right with blanks, rather
338 than on the left with blanks or zeros.
346 (a space) A blank should be left before a positive number
347 (or empty string) produced by a signed conversion.
350 A sign (+ or \-) should always be placed before a number produced by a signed
352 By default a sign is used only for negative numbers.
355 overrides a space if both are used.
357 The five flag characters above are defined in the C standard.
358 The SUSv2 specifies one further flag character.
361 For decimal conversion
369 the output is to be grouped with thousands' grouping characters
370 if the locale information indicates any.
371 Note that many versions of
373 cannot parse this option and will issue a warning.
375 include \fI%\(aqF\fP.
377 glibc 2.2 adds one further flag character.
380 For decimal integer conversion
384 the output uses the locale's alternative output digits, if any.
385 For example, since glibc 2.2.3 this will give Arabic-Indic digits
386 in the Persian ("fa_IR") locale.
387 .\" outdigits keyword in locale file
389 An optional decimal digit string (with nonzero first digit) specifying
390 a minimum field width.
391 If the converted value has fewer characters
392 than the field width, it will be padded with spaces on the left
393 (or right, if the left-adjustment flag has been given).
394 Instead of a decimal digit string one may write "*" or "*m$"
395 (for some decimal integer \fIm\fP) to specify that the field width
396 is given in the next argument, or in the \fIm\fP-th argument, respectively,
397 which must be of type
399 A negative field width is taken as a \(aq\-\(aq flag followed by a
400 positive field width.
401 In no case does a nonexistent or small field width cause truncation of a
402 field; if the result of a conversion is wider than the field width, the
403 field is expanded to contain the conversion result.
405 An optional precision, in the form of a period (\(aq.\(aq) followed by an
406 optional decimal digit string.
407 Instead of a decimal digit string one may write "*" or "*m$"
408 (for some decimal integer m) to specify that the precision
409 is given in the next argument, or in the m-th argument, respectively,
410 which must be of type
412 If the precision is given as just \(aq.\(aq, or the precision is negative,
413 the precision is taken to be zero.
414 This gives the minimum number of digits to appear for
422 conversions, the number of digits to appear after the radix character for
430 conversions, the maximum number of significant digits for
434 conversions, or the maximum number of characters to be printed from a
440 .SS The length modifier
441 Here, "integer conversion" stands for
452 A following integer conversion corresponds to a
456 argument, or a following
458 conversion corresponds to a pointer to a
463 A following integer conversion corresponds to a
466 .I unsigned short int
467 argument, or a following
469 conversion corresponds to a pointer to a
474 (ell) A following integer conversion corresponds to a
478 argument, or a following
480 conversion corresponds to a pointer to a
482 argument, or a following
484 conversion corresponds to a
486 argument, or a following
488 conversion corresponds to a pointer to
494 A following integer conversion corresponds to a
497 .I unsigned long long int
498 argument, or a following
500 conversion corresponds to a pointer to a
515 conversion corresponds to a
518 (C99 allows %LF, but SUSv2 does not.)
521 ("quad". 4.4BSD and Linux libc5 only.
523 This is a synonym for
527 A following integer conversion corresponds to an
534 A following integer conversion corresponds to a
545 A following integer conversion corresponds to a
549 The SUSv2 only knows about the length modifiers
577 .SS The conversion specifier
578 A character that specifies the type of conversion to be applied.
579 The conversion specifiers and their meanings are:
584 argument is converted to signed decimal notation.
585 The precision, if any, gives the minimum number of digits
586 that must appear; if the converted value requires fewer digits, it is
587 padded on the left with zeros.
588 The default precision is 1.
589 When 0 is printed with an explicit precision 0, the output is empty.
591 .BR o ", " u ", " x ", " X
594 argument is converted to unsigned octal
598 or unsigned hexadecimal
607 conversions; the letters
612 The precision, if any, gives the minimum number of digits
613 that must appear; if the converted value requires fewer digits, it is
614 padded on the left with zeros.
615 The default precision is 1.
616 When 0 is printed with an explicit precision 0, the output is empty.
621 argument is rounded and converted in the style
622 .RB [\-]d \&. ddd e \(+-dd
623 where there is one digit before the decimal-point character and the number
624 of digits after it is equal to the precision; if the precision is missing,
625 it is taken as 6; if the precision is zero, no decimal-point character
629 conversion uses the letter
633 to introduce the exponent.
634 The exponent always contains at least two
635 digits; if the value is zero, the exponent is 00.
640 argument is rounded and converted to decimal notation in the style
642 where the number of digits after the decimal-point character is equal to
643 the precision specification.
644 If the precision is missing, it is taken as
645 6; if the precision is explicitly zero, no decimal-point character appears.
646 If a decimal point appears, at least one digit appears before it.
648 (The SUSv2 does not know about
650 and says that character string representations for infinity and NaN
651 may be made available.
652 The C99 standard specifies "[\-]inf" or "[\-]infinity"
653 for infinity, and a string starting with "nan" for NaN, in the case of
655 conversion, and "[\-]INF" or "[\-]INFINITY" or "NAN*" in the case of
662 argument is converted in style
673 The precision specifies the number of significant digits.
674 If the precision is missing, 6 digits are given; if the precision is zero,
678 is used if the exponent from its conversion is less than \-4 or greater
679 than or equal to the precision.
680 Trailing zeros are removed from the
681 fractional part of the result; a decimal point appears only if it is
682 followed by at least one digit.
685 (C99; not in SUSv2) For
689 argument is converted to hexadecimal notation (using the letters abcdef)
691 .RB [\-] 0x h \&. hhhh p \(+-;
694 conversion the prefix
696 the letters ABCDEF, and the exponent separator
699 There is one hexadecimal digit before the decimal point,
700 and the number of digits after it is equal to the precision.
701 The default precision suffices for an exact representation of the value
702 if an exact representation in base 2 exists
703 and otherwise is sufficiently large to distinguish values of type
705 The digit before the decimal point is unspecified for nonnormalized
706 numbers, and nonzero but otherwise unspecified for normalized numbers.
711 modifier is present, the
713 argument is converted to an
714 .IR "unsigned char" ,
715 and the resulting character is written.
718 modifier is present, the
720 (wide character) argument is converted to a multibyte sequence by a call
723 function, with a conversion state starting in the initial state, and the
724 resulting multibyte string is written.
729 modifier is present: The
731 argument is expected to be a pointer to an array of character type (pointer
733 Characters from the array are written up to (but not
734 including) a terminating null byte (\(aq\\0\(aq);
735 if a precision is specified, no more than the number specified
737 If a precision is given, no null byte need be present;
738 if the precision is not specified, or is greater than the size of the
739 array, the array must contain a terminating null byte.
743 modifier is present: The
745 argument is expected to be a pointer to an array of wide characters.
746 Wide characters from the array are converted to multibyte characters
747 (each by a call to the
749 function, with a conversion state starting in the initial state before
750 the first wide character), up to and including a terminating null
752 The resulting multibyte characters are written up to
753 (but not including) the terminating null byte.
755 specified, no more bytes than the number specified are written, but
756 no partial multibyte characters are written.
757 Note that the precision
758 determines the number of
760 written, not the number of
763 .IR "screen positions" .
764 The array must contain a terminating null wide character, unless a
765 precision is given and it is so small that the number of bytes written
766 exceeds it before the end of the array is reached.
769 (Not in C99, but in SUSv2.)
775 (Not in C99, but in SUSv2.)
783 pointer argument is printed in hexadecimal (as if by
789 The number of characters written so far is stored into the integer
792 (or variant) pointer argument.
793 No argument is converted.
798 .IR strerror(errno) .
799 No argument is required.
802 A \(aq%\(aq is written.
803 No argument is converted.
804 The complete conversion
805 specification is \(aq%%\(aq.
815 functions conform to C89 and C99.
820 functions conform to C99.
822 Concerning the return value of
824 SUSv2 and C99 contradict each other: when
828 then SUSv2 stipulates an unspecified return value less than 1,
831 to be NULL in this case, and gives the return value (as always)
832 as the number of characters that would have been written in case
833 the output string has been large enough.
835 Linux libc4 knows about the five C standard flags.
836 It knows about the length modifiers \fBh\fP, \fBl\fP, \fBL\fP,
838 \fBc\fP, \fBd\fP, \fBe\fP, \fBE\fP, \fBf\fP, \fBF\fP,
839 \fBg\fP, \fBG\fP, \fBi\fP, \fBn\fP, \fBo\fP, \fBp\fP,
840 \fBs\fP, \fBu\fP, \fBx\fP, and \fBX\fP,
841 where \fBF\fP is a synonym for \fBf\fP.
842 Additionally, it accepts \fBD\fP, \fBO\fP, and \fBU\fP as synonyms
843 for \fBld\fP, \fBlo\fP, and \fBlu\fP.
844 (This is bad, and caused serious bugs later, when
845 support for \fB%D\fP disappeared.)
846 No locale-dependent radix character,
847 no thousands' separator, no NaN or infinity, no "%m$" and "*m$".
849 Linux libc5 knows about the five C standard flags and the \(aq flag,
850 locale, "%m$" and "*m$".
851 It knows about the length modifiers \fBh\fP, \fBl\fP, \fBL\fP,
852 \fBZ\fP, and \fBq\fP, but accepts \fBL\fP and \fBq\fP
853 both for \fIlong double\fP and for \fIlong long int\fP (this is a bug).
854 It no longer recognizes \fBF\fP, \fBD\fP, \fBO\fP, and \fBU\fP,
855 but adds the conversion character
858 .IR strerror(errno) .
860 glibc 2.0 adds conversion characters \fBC\fP and \fBS\fP.
862 glibc 2.1 adds length modifiers \fBhh\fP, \fBj\fP, \fBt\fP, and \fBz\fP
863 and conversion characters \fBa\fP and \fBA\fP.
865 glibc 2.2 adds the conversion character \fBF\fP with C99 semantics,
866 and the flag character \fBI\fP.
868 Some programs imprudently rely on code such as the following
870 sprintf(buf, "%s some further text", buf);
874 However, the standards explicitly note that the results are undefined
875 if source and destination buffers overlap when calling
881 .\" http://sourceware.org/bugzilla/show_bug.cgi?id=7075
882 Depending on the version of
884 used, and the compiler options employed, calls such as the above will
886 produce the expected results.
888 The glibc implementation of the functions
892 conforms to the C99 standard, that is, behaves as described above,
893 since glibc version 2.1.
894 Until glibc 2.0.6 they would return \-1
895 when the output was truncated.
897 .\" UNIX V7 defines the three routines
901 .\" and has the flag \-, the width or precision *, the length modifier l,
902 .\" and the conversions doxfegcsu, and also D,O,U,X as synonyms for ld,lo,lu,lx.
903 .\" This is still true for 2.9.1BSD, but 2.10BSD has the flags
904 .\" #, + and <space> and no longer mentions D,O,U,X.
909 .\" and warns not to use D,O,U,X.
910 .\" 4.3BSD Reno has the flag 0, the length modifiers h and L,
911 .\" and the conversions n, p, E, G, X (with current meaning)
912 .\" and deprecates D,O,U.
913 .\" 4.4BSD introduces the functions
916 .\" .BR vsnprintf (),
917 .\" and the length modifier q.
918 .\" FreeBSD also has functions
921 .\" .BR vasprintf (),
922 .\" that allocate a buffer large enough for
924 .\" In glibc there are functions
928 .\" that print to a file descriptor instead of a stream.
934 assume an arbitrarily long string, callers must be careful not to overflow
935 the actual space; this is often impossible to assure.
937 of the strings produced is locale-dependent and difficult to predict.
947 Linux libc4.[45] does not have a
949 but provides a libbsd that contains an
953 that is, one that ignores the
958 with early libc4 leads to serious security problems.
962 often indicates a bug, since
964 may contain a % character.
967 comes from untrusted user input, it may contain \fB%n\fP, causing the
969 call to write to memory and creating a security hole.
971 .\" Some floating-point conversions under early libc4
972 .\" caused memory leaks.
976 to five decimal places:
982 fprintf(stdout, "pi = %.5f\en", 4 * atan(1.0));
986 To print a date and time in the form "Sunday, July 3, 10:02",
991 are pointers to strings:
996 fprintf(stdout, "%s, %s %d, %.2d:%.2d\en",
997 weekday, month, day, hour, min);
1001 Many countries use the day-month-year order.
1002 Hence, an internationalized version must be able to print
1003 the arguments in an order specified by the format:
1008 fprintf(stdout, format,
1009 weekday, month, day, hour, min);
1015 depends on locale, and may permute the arguments.
1020 "%1$s, %3$d. %2$s, %4$d:%5$.2d\en"
1024 one might obtain "Sonntag, 3. Juli, 10:02".
1026 To allocate a sufficiently large string and print into it
1027 (code correct for both glibc 2.0 and glibc 2.1):
1029 If truncation occurs in glibc versions prior to 2.0.6, this is treated as an
1030 error instead of being handled gracefully.
1038 make_message(const char *fmt, ...)
1041 int size = 100; /* Guess we need no more than 100 bytes */
1045 if ((p = malloc(size)) == NULL)
1050 /* Try to print in the allocated space */
1053 n = vsnprintf(p, size, fmt, ap);
1056 /* Check error code */
1061 /* If that worked, return the string */
1066 /* Else try again with more space */
1068 size = n + 1; /* Precisely what is needed */
1071 if ((np = realloc (p, size)) == NULL) {