man3/scanf.3

   1 .\" Copyright (c) 1990, 1991 The Regents of the University of California.
   2 .\" All rights reserved.
   3 .\"
   4 .\" This code is derived from software contributed to Berkeley by
   5 .\" Chris Torek and the American National Standards Committee X3,
   6 .\" on Information Processing Systems.
   7 .\"
   8 .\" %%%LICENSE_START(BSD_4_CLAUSE_UCB)
   9 .\" Redistribution and use in source and binary forms, with or without
  10 .\" modification, are permitted provided that the following conditions
  11 .\" are met:
  12 .\" 1. Redistributions of source code must retain the above copyright
  13 .\"    notice, this list of conditions and the following disclaimer.
  14 .\" 2. Redistributions in binary form must reproduce the above copyright
  15 .\"    notice, this list of conditions and the following disclaimer in the
  16 .\"    documentation and/or other materials provided with the distribution.
  17 .\" 3. All advertising materials mentioning features or use of this software
  18 .\"    must display the following acknowledgement:
  19 .\"     This product includes software developed by the University of
  20 .\"     California, Berkeley and its contributors.
  21 .\" 4. Neither the name of the University nor the names of its contributors
  22 .\"    may be used to endorse or promote products derived from this software
  23 .\"    without specific prior written permission.
  24 .\"
  25 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  26 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  27 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  28 .\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  29 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  30 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  31 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  32 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  33 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  34 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  35 .\" SUCH DAMAGE.
  36 .\" %%%LICENSE_END
  37 .\"
  38 .\"     @(#)scanf.3     6.14 (Berkeley) 1/8/93
  39 .\"
  40 .\" Converted for Linux, Mon Nov 29 15:22:01 1993, faith@cs.unc.edu
  41 .\" modified to resemble the GNU libio setup used in the Linux libc
  42 .\" used in versions 4.x (x>4) and 5   Helmut.Geyer@iwr.uni-heidelberg.de
  43 .\" Modified, aeb, 970121
  44 .\" 2005-07-14, mtk, added description of %n$ form; various text
  45 .\"     incorporated from the GNU C library documentation ((C) The
  46 .\"     Free Software Foundation); other parts substantially rewritten.
  47 .\"
  48 .\" 2008-06-23, mtk
  49 .\"     Add ERRORS section.
  50 .\"     Document the 'a' and 'm' modifiers for dynamic string allocation.
  51 .\"
  52 .TH SCANF 3  2020-04-11 "GNU" "Linux Programmer's Manual"
  53 .SH NAME
  54 scanf, fscanf, sscanf, vscanf, vsscanf, vfscanf \- input format conversion
  55 .SH SYNOPSIS
  56 .nf
  57 .B #include <stdio.h>
  58 .PP
  59 .BI "int scanf(const char *" format ", ...);"
  60 .BI "int fscanf(FILE *" stream ", const char *" format ", ...);"
  61 .BI "int sscanf(const char *" str ", const char *" format ", ...);"
  62
  63 .B #include <stdarg.h>
  64 .PP
  65 .BI "int vscanf(const char *" format ", va_list " ap );
  66 .BI "int vsscanf(const char *" str ", const char *" format ", va_list " ap );
  67 .BI "int vfscanf(FILE *" stream ", const char *" format ", va_list " ap );
  68 .fi
  69 .PP
  70 .in -4n
  71 Feature Test Macro Requirements for glibc (see
  72 .BR feature_test_macros (7)):
  73 .in
  74 .ad l
  75 .PP
  76 .BR vscanf (),
  77 .BR vsscanf (),
  78 .BR vfscanf ():
  79 .RS 4
  80 _ISOC99_SOURCE || _POSIX_C_SOURCE\ >=\ 200112L
  81 .ad
  82 .RE
  83 .SH DESCRIPTION
  84 The
  85 .BR scanf ()
  86 family of functions scans input according to
  87 .I format
  88 as described below.
  89 This format may contain
  90 .IR "conversion specifications" ;
  91 the results from such conversions, if any,
  92 are stored in the locations pointed to by the
  93 .I pointer
  94 arguments that follow
  95 .IR format .
  96 Each
  97 .I pointer
  98 argument must be of a type that is appropriate for the value returned
  99 by the corresponding conversion specification.
 100 .PP
 101 If the number of conversion specifications in
 102 .I format
 103 exceeds the number of
 104 .I pointer
 105 arguments, the results are undefined.
 106 If the number of
 107 .I pointer
 108 arguments exceeds the number of conversion specifications, then the excess
 109 .I pointer
 110 arguments are evaluated, but are otherwise ignored.
 111 .PP
 112 The
 113 .BR scanf ()
 114 function reads input from the standard input stream
 115 .IR stdin ,
 116 .BR fscanf ()
 117 reads input from the stream pointer
 118 .IR stream ,
 119 and
 120 .BR sscanf ()
 121 reads its input from the character string pointed to by
 122 .IR str .
 123 .PP
 124 The
 125 .BR vfscanf ()
 126 function is analogous to
 127 .BR vfprintf (3)
 128 and reads input from the stream pointer
 129 .I stream
 130 using a variable argument list of pointers (see
 131 .BR stdarg (3).
 132 The
 133 .BR vscanf ()
 134 function scans a variable argument list from the standard input and the
 135 .BR vsscanf ()
 136 function scans it from a string; these are analogous to the
 137 .BR vprintf (3)
 138 and
 139 .BR vsprintf (3)
 140 functions respectively.
 141 .PP
 142 The
 143 .I format
 144 string consists of a sequence of
 145 .I directives
 146 which describe how to process the sequence of input characters.
 147 If processing of a directive fails, no further input is read, and
 148 .BR scanf ()
 149 returns.
 150 A "failure" can be either of the following:
 151 .IR "input failure" ,
 152 meaning that input characters were unavailable, or
 153 .IR "matching failure" ,
 154 meaning that the input was inappropriate (see below).
 155 .PP
 156 A directive is one of the following:
 157 .TP
 158 \(bu
 159 A sequence of white-space characters (space, tab, newline, etc.; see
 160 .BR isspace (3)).
 161 This directive matches any amount of white space,
 162 including none, in the input.
 163 .TP
 164 \(bu
 165 An ordinary character (i.e., one other than white space or \(aq%\(aq).
 166 This character must exactly match the next character of input.
 167 .TP
 168 \(bu
 169 A conversion specification,
 170 which commences with a \(aq%\(aq (percent) character.
 171 A sequence of characters from the input is converted according to
 172 this specification, and the result is placed in the corresponding
 173 .I pointer
 174 argument.
 175 If the next item of input does not match the conversion specification,
 176 the conversion fails\(emthis is a
 177 .IR "matching failure" .
 178 .PP
 179 Each
 180 .I conversion specification
 181 in
 182 .I format
 183 begins with either the character \(aq%\(aq or the character sequence
 184 "\fB%\fP\fIn\fP\fB$\fP"
 185 (see below for the distinction) followed by:
 186 .TP
 187 \(bu
 188 An optional \(aq*\(aq assignment-suppression character:
 189 .BR scanf ()
 190 reads input as directed by the conversion specification,
 191 but discards the input.
 192 No corresponding
 193 .I pointer
 194 argument is required, and this specification is not
 195 included in the count of successful assignments returned by
 196 .BR scanf ().
 197 .TP
 198 \(bu
 199 For decimal conversions, an optional quote character (\(aq).
 200 This specifies that the input number may include thousands'
 201 separators as defined by the
 202 .BR LC_NUMERIC
 203 category of the current locale.
 204 (See
 205 .BR setlocale (3).)
 206 The quote character may precede or follow the \(aq*\(aq
 207 assignment-suppression character.
 208 .TP
 209 \(bu
 210 An optional \(aqm\(aq character.
 211 This is used with string conversions
 212 .RI ( %s ,
 213 .IR %c ,
 214 .IR %[ ),
 215 and relieves the caller of the
 216 need to allocate a corresponding buffer to hold the input: instead,
 217 .BR scanf ()
 218 allocates a buffer of sufficient size,
 219 and assigns the address of this buffer to the corresponding
 220 .I pointer
 221 argument, which should be a pointer to a
 222 .I "char\ *"
 223 variable (this variable does not need to be initialized before the call).
 224 The caller should subsequently
 225 .BR free (3)
 226 this buffer when it is no longer required.
 227 .TP
 228 \(bu
 229 An optional decimal integer which specifies the
 230 .IR "maximum field width" .
 231 Reading of characters stops either when this maximum is reached or
 232 when a nonmatching character is found, whichever happens first.
 233 Most conversions discard initial white space characters (the exceptions
 234 are noted below),
 235 and these discarded characters don't count toward the maximum field width.
 236 String input conversions store a terminating null byte (\(aq\e0\(aq)
 237 to mark the end of the input;
 238 the maximum field width does not include this terminator.
 239 .TP
 240 \(bu
 241 An optional
 242 .IR "type modifier character" .
 243 For example, the
 244 .B l
 245 type modifier is used with integer conversions such as
 246 .B %d
 247 to specify that the corresponding
 248 .I pointer
 249 argument refers to a
 250 .I "long int"
 251 rather than a pointer to an
 252 .IR int .
 253 .TP
 254 \(bu
 255 A
 256 .I "conversion specifier"
 257 that specifies the type of input conversion to be performed.
 258 .PP
 259 The conversion specifications in
 260 .I format
 261 are of two forms, either beginning with \(aq%\(aq or beginning with
 262 "\fB%\fP\fIn\fP\fB$\fP".
 263 The two forms should not be mixed in the same
 264 .I format
 265 string, except that a string containing
 266 "\fB%\fP\fIn\fP\fB$\fP"
 267 specifications can include
 268 .B %%
 269 and
 270 .BR %* .
 271 If
 272 .I format
 273 contains \(aq%\(aq
 274 specifications, then these correspond in order with successive
 275 .I pointer
 276 arguments.
 277 In the
 278 "\fB%\fP\fIn\fP\fB$\fP"
 279 form (which is specified in POSIX.1-2001, but not C99),
 280 .I n
 281 is a decimal integer that specifies that the converted input should
 282 be placed in the location referred to by the
 283 .IR n -th
 284 .I pointer
 285 argument following
 286 .IR format .
 287 .SS Conversions
 288 The following
 289 .I "type modifier characters"
 290 can appear in a conversion specification:
 291 .TP
 292 .B h
 293 Indicates that the conversion will be one of
 294 \fBd\fP, \fBi\fP, \fBo\fP, \fBu\fP, \fBx\fP, \fBX\fP, or \fBn\fP
 295 and the next pointer is a pointer to a
 296 .I short int
 297 or
 298 .I unsigned short int
 299 (rather than
 300 .IR int ).
 301 .TP
 302 .B hh
 303 As for
 304 .BR h ,
 305 but the next pointer is a pointer to a
 306 .I signed char
 307 or
 308 .IR "unsigned char" .
 309 .TP
 310 .B j
 311 As for
 312 .BR h ,
 313 but the next pointer is a pointer to an
 314 .I intmax_t
 315 or a
 316 .IR uintmax_t .
 317 This modifier was introduced in C99.
 318 .TP
 319 .B l
 320 Indicates either that the conversion will be one of
 321 \fBd\fP, \fBi\fP, \fBo\fP, \fBu\fP, \fBx\fP, \fBX\fP, or \fBn\fP
 322 and the next pointer is a pointer to a
 323 .I long int
 324 or
 325 .I unsigned long int
 326 (rather than
 327 .IR int ),
 328 or that the conversion will be one of
 329 \fBe\fP, \fBf\fP, or \fBg\fP
 330 and the next pointer is a pointer to
 331 .I double
 332 (rather than
 333 .IR float ).
 334 Specifying two
 335 .B l
 336 characters is equivalent to
 337 .BR L .
 338 If used with
 339 .B %c
 340 or
 341 .BR %s ,
 342 the corresponding parameter is considered
 343 as a pointer to a wide character or wide-character string respectively.
 344 .\" This use of l was introduced in Amendment 1 to ISO C90.
 345 .TP
 346 .B L
 347 Indicates that the conversion will be either
 348 \fBe\fP, \fBf\fP, or \fBg\fP
 349 and the next pointer is a pointer to
 350 .I "long double"
 351 or the conversion will be
 352 \fBd\fP, \fBi\fP, \fBo\fP, \fBu\fP, or \fBx\fP
 353 and the next pointer is a pointer to
 354 .IR "long long" .
 355 .\" MTK, Jul 05: The following is no longer true for modern
 356 .\" ANSI C (i.e., C99):
 357 .\" (Note that long long is not an
 358 .\" ANSI C
 359 .\" type. Any program using this will not be portable to all
 360 .\" architectures).
 361 .TP
 362 .B q
 363 equivalent to
 364 .BR L .
 365 This specifier does not exist in ANSI C.
 366 .TP
 367 .B t
 368 As for
 369 .BR h ,
 370 but the next pointer is a pointer to a
 371 .IR ptrdiff_t .
 372 This modifier was introduced in C99.
 373 .TP
 374 .B z
 375 As for
 376 .BR h ,
 377 but the next pointer is a pointer to a
 378 .IR size_t .
 379 This modifier was introduced in C99.
 380 .PP
 381 The following
 382 .I "conversion specifiers"
 383 are available:
 384 .TP
 385 .B %
 386 Matches a literal \(aq%\(aq.
 387 That is,
 388 .B %\&%
 389 in the format string matches a
 390 single input \(aq%\(aq character.
 391 No conversion is done (but initial white space characters are discarded),
 392 and assignment does not occur.
 393 .TP
 394 .B d
 395 Matches an optionally signed decimal integer;
 396 the next pointer must be a pointer to
 397 .IR int .
 398 .\" .TP
 399 .\" .B D
 400 .\" Equivalent to
 401 .\" .IR ld ;
 402 .\" this exists only for backward compatibility.
 403 .\" (Note: thus only in libc4
 404 .\" In libc5 and glibc the
 405 .\" .B %D
 406 .\" is silently ignored, causing old programs to fail mysteriously.)
 407 .TP
 408 .B i
 409 Matches an optionally signed integer; the next pointer must be a pointer to
 410 .IR int .
 411 The integer is read in base 16 if it begins with
 412 .I 0x
 413 or
 414 .IR 0X ,
 415 in base 8 if it begins with
 416 .IR 0 ,
 417 and in base 10 otherwise.
 418 Only characters that correspond to the base are used.
 419 .TP
 420 .B o
 421 Matches an unsigned octal integer; the next pointer must be a pointer to
 422 .IR "unsigned int" .
 423 .TP
 424 .B u
 425 Matches an unsigned decimal integer; the next pointer must be a
 426 pointer to
 427 .IR "unsigned int" .
 428 .TP
 429 .B x
 430 Matches an unsigned hexadecimal integer; the next pointer must
 431 be a pointer to
 432 .IR "unsigned int" .
 433 .TP
 434 .B X
 435 Equivalent to
 436 .BR x .
 437 .TP
 438 .B f
 439 Matches an optionally signed floating-point number; the next pointer must
 440 be a pointer to
 441 .IR float .
 442 .TP
 443 .B e
 444 Equivalent to
 445 .BR f .
 446 .TP
 447 .B g
 448 Equivalent to
 449 .BR f .
 450 .TP
 451 .B E
 452 Equivalent to
 453 .BR f .
 454 .TP
 455 .B a
 456 (C99) Equivalent to
 457 .BR f .
 458 .TP
 459 .B s
 460 Matches a sequence of non-white-space characters;
 461 the next pointer must be a pointer to the initial element of a
 462 character array that is long enough to hold the input sequence and
 463 the terminating null byte (\(aq\e0\(aq), which is added automatically.
 464 The input string stops at white space or at the maximum field
 465 width, whichever occurs first.
 466 .TP
 467 .B c
 468 Matches a sequence of characters whose length is specified by the
 469 .I maximum field width
 470 (default 1); the next pointer must be a pointer to
 471 .IR char ,
 472 and there must be enough room for all the characters
 473 (no terminating null byte is added).
 474 The usual skip of leading white space is suppressed.
 475 To skip white space first, use an explicit space in the format.
 476 .TP
 477 .B \&[
 478 Matches a nonempty sequence of characters from the specified set of
 479 accepted characters; the next pointer must be a pointer to
 480 .IR char ,
 481 and there must be enough room for all the characters in the string, plus a
 482 terminating null byte.
 483 The usual skip of leading white space is suppressed.
 484 The string is to be made up of characters in (or not in) a particular set;
 485 the set is defined by the characters between the open bracket
 486 .B [
 487 character and a close bracket
 488 .B ]
 489 character.
 490 The set
 491 .I excludes
 492 those characters if the first character after the open bracket is a
 493 circumflex
 494 .RB ( ^ ).
 495 To include a close bracket in the set, make it the first character after
 496 the open bracket or the circumflex; any other position will end the set.
 497 The hyphen character
 498 .B \-
 499 is also special; when placed between two other characters, it adds all
 500 intervening characters to the set.
 501 To include a hyphen, make it the last
 502 character before the final close bracket.
 503 For instance,
 504 .B [^]0\-9\-]
 505 means
 506 the set "everything except close bracket, zero through nine, and hyphen".
 507 The string ends with the appearance of a character not in the (or, with a
 508 circumflex, in) set or when the field width runs out.
 509 .TP
 510 .B p
 511 Matches a pointer value (as printed by
 512 .B %p
 513 in
 514 .BR printf (3));
 515 the next pointer must be a pointer to a pointer to
 516 .IR void .
 517 .TP
 518 .B n
 519 Nothing is expected; instead, the number of characters consumed thus far
 520 from the input is stored through the next pointer, which must be a pointer
 521 to
 522 .IR int .
 523 This is
 524 .I not
 525 a conversion and does
 526 .I not
 527 increase the count returned by the function.
 528 The assignment can be suppressed with the
 529 .B *
 530 assignment-suppression character, but the effect on the
 531 return value is undefined.
 532 Therefore
 533 .B %*n
 534 conversions should not be used.
 535 .SH RETURN VALUE
 536 On success, these functions return the number of input items
 537 successfully matched and assigned;
 538 this can be fewer than provided for,
 539 or even zero, in the event of an early matching failure.
 540 .PP
 541 The value
 542 .B EOF
 543 is returned if the end of input is reached before either the first
 544 successful conversion or a matching failure occurs.
 545 .B EOF
 546 is also returned if a read error occurs,
 547 in which case the error indicator for the stream (see
 548 .BR ferror (3))
 549 is set, and
 550 .I errno
 551 is set to indicate the error.
 552 .SH ERRORS
 553 .TP
 554 .B EAGAIN
 555 The file descriptor underlying
 556 .I stream
 557 is marked nonblocking, and the read operation would block.
 558 .TP
 559 .B EBADF
 560 The file descriptor underlying
 561 .I stream
 562 is invalid, or not open for reading.
 563 .TP
 564 .B EILSEQ
 565 Input byte sequence does not form a valid character.
 566 .TP
 567 .B EINTR
 568 The read operation was interrupted by a signal; see
 569 .BR signal (7).
 570 .TP
 571 .B EINVAL
 572 Not enough arguments; or
 573 .I format
 574 is NULL.
 575 .TP
 576 .B ENOMEM
 577 Out of memory.
 578 .TP
 579 .B ERANGE
 580 The result of an integer conversion would exceed the size
 581 that can be stored in the corresponding integer type.
 582 .SH ATTRIBUTES
 583 For an explanation of the terms used in this section, see
 584 .BR attributes (7).
 585 .TS
 586 allbox;
 587 lbw20 lb lb
 588 l l l.
 589 Interface       Attribute       Value
 590 T{
 591 .BR scanf (),
 592 .BR fscanf (),
 593 .br
 594 .BR sscanf (),
 595 .BR vscanf (),
 596 .br
 597 .BR vsscanf (),
 598 .BR vfscanf ()
 599 T}      Thread safety   MT-Safe locale
 600 .TE
 601 .sp 1
 602 .SH CONFORMING TO
 603 The functions
 604 .BR fscanf (),
 605 .BR scanf (),
 606 and
 607 .BR sscanf ()
 608 conform to C89 and C99 and POSIX.1-2001.
 609 These standards do not specify the
 610 .B ERANGE
 611 error.
 612 .PP
 613 The
 614 .B q
 615 specifier is the 4.4BSD notation for
 616 .IR "long long" ,
 617 while
 618 .B ll
 619 or the usage of
 620 .B L
 621 in integer conversions is the GNU notation.
 622 .PP
 623 The Linux version of these functions is based on the
 624 .I GNU
 625 .I libio
 626 library.
 627 Take a look at the
 628 .I info
 629 documentation of
 630 .I GNU
 631 .I libc (glibc-1.08)
 632 for a more concise description.
 633 .SH NOTES
 634 .SS The 'a' assignment-allocation modifier
 635 Originally, the GNU C library supported dynamic allocation for string inputs
 636 (as a nonstandard extension) via the
 637 .B a
 638 character.
 639 (This feature is present at least as far back as glibc 2.0.)
 640 Thus, one could write the following to have
 641 .BR scanf ()
 642 allocate a buffer for an input string,
 643 with a pointer to that buffer being returned in
 644 .IR *buf :
 645 .PP
 646     char *buf;
 647     scanf("%as", &buf);
 648 .PP
 649 The use of the letter
 650 .B a
 651 for this purpose was problematic, since
 652 .B a
 653 is also specified by the ISO C standard as a synonym for
 654 .B f
 655 (floating-point input).
 656 POSIX.1-2008 instead specifies the
 657 .B m
 658 modifier for assignment allocation (as documented in DESCRIPTION, above).
 659 .PP
 660 Note that the
 661 .B a
 662 modifier is not available if the program is compiled with
 663 .I "gcc \-std=c99"
 664 or
 665 .IR "gcc \-D_ISOC99_SOURCE"
 666 (unless
 667 .B _GNU_SOURCE
 668 is also specified), in which case the
 669 .B a
 670 is interpreted as a specifier for floating-point numbers (see above).
 671 .PP
 672 Support for the
 673 .B m
 674 modifier was added to glibc starting with version 2.7,
 675 and new programs should use that modifier instead of
 676 .BR a .
 677 .PP
 678 As well as being standardized by POSIX, the
 679 .B m
 680 modifier has the following further advantages over
 681 the use of
 682 .BR a:
 683 .IP * 2
 684 It may also be applied to
 685 .B %c
 686 conversion specifiers (e.g.,
 687 .BR %3mc ).
 688 .IP *
 689 It avoids ambiguity with respect to the
 690 .B %a
 691 floating-point conversion specifier (and is unaffected by
 692 .IR "gcc \-std=c99"
 693 etc.).
 694 .SH BUGS
 695 All functions are fully C89 conformant, but provide the
 696 additional specifiers
 697 .B q
 698 and
 699 .B a
 700 as well as an additional behavior of the
 701 .B L
 702 and
 703 .B l
 704 specifiers.
 705 The latter may be considered to be a bug, as it changes the
 706 behavior of specifiers defined in C89.
 707 .PP
 708 Some combinations of the type modifiers and conversion
 709 specifiers defined by ANSI C do not make sense
 710 (e.g.,
 711 .BR "%Ld" ).
 712 While they may have a well-defined behavior on Linux, this need not
 713 to be so on other architectures.
 714 Therefore it usually is better to use
 715 modifiers that are not defined by ANSI C at all, that is, use
 716 .B q
 717 instead of
 718 .B L
 719 in combination with
 720 \fBd\fP, \fBi\fP, \fBo\fP, \fBu\fP, \fBx\fP, and \fBX\fP
 721 conversions or
 722 .BR ll .
 723 .PP
 724 The usage of
 725 .B q
 726 is not the same as on 4.4BSD,
 727 as it may be used in float conversions equivalently to
 728 .BR L .
 729 .SH EXAMPLE
 730 To use the dynamic allocation conversion specifier, specify
 731 .B m
 732 as a length modifier (thus
 733 .B %ms
 734 or
 735 \fB%m[\fP\fIrange\fP\fB]\fP).
 736 The caller must
 737 .BR free (3)
 738 the returned string, as in the following example:
 739 .PP
 740 .in +4n
 741 .EX
 742 char *p;
 743 int n;
 744
 745 errno = 0;
 746 n = scanf("%m[a\-z]", &p);
 747 if (n == 1) {
 748     printf("read: %s\en", p);
 749     free(p);
 750 } else if (errno != 0) {
 751     perror("scanf");
 752 } else {
 753     fprintf(stderr, "No matching characters\en");
 754 }
 755 .EE
 756 .in
 757 .PP
 758 As shown in the above example, it is necessary to call
 759 .BR free (3)
 760 only if the
 761 .BR scanf ()
 762 call successfully read a string.
 763 .SH SEE ALSO
 764 .BR getc (3),
 765 .BR printf (3),
 766 .BR setlocale (3),
 767 .BR strtod (3),
 768 .BR strtol (3),
 769 .BR strtoul (3)