unplugged-vendor/prebuilts/xcc/linux-x86/xtensa/RI-2018.0-linux/XtensaTools/info/libc.info-1

8503 lines
296 KiB
Plaintext
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

This is
/home/xpgcust/tree/RI-2019.1/ib/p4root/Xtensa/SWConfig/../Target-libs/newlib/newlib/libc/libc.info,
produced by makeinfo version 4.8 from
/home/xpgcust/tree/RI-2019.1/ib/p4root/Xtensa/SWConfig/../Target-libs/newlib/newlib/libc/libc.texinfo.
10/2018
This file documents the ANSI C library, version 2.20.0 for Xtensa(R)
Tools Version 14.2.
Copyright (C) 1992, 1993, 1994-2012 Red Hat, Inc. Copyright (C)
1999-2012 Tensilica, Inc.
`libc' includes software developed by the University of California,
Berkeley and its contributors.
libc includes software developed by Martin Jackson, Graham Haley and
Steve Chamberlain of Tadpole Technology and released to Cygnus.
libc uses floating-point conversion software developed at AT&T, which
includes this copyright information:
The author of this software is David M. Gay.
Copyright (c) 1991 by AT&T.
Permission to use, copy, modify, and distribute this software for any
purpose without fee is hereby granted, provided that this entire notice
is included in all copies of any software which is or includes a copy
or modification of this software and in all copies of the supporting
documentation for such software.
THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR
IMPLIED WARRANTY. IN PARTICULAR, NEITHER THE AUTHOR NOR AT&T MAKES ANY
REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE MERCHANTABILITY
OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE.
Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.
Permission is granted to copy and distribute modified versions of
this manual under the conditions for verbatim copying, subject to the
terms of the GNU General Public License, which includes the provision
that the entire resulting derived work is distributed under the terms
of a permission notice identical to this one.
Permission is granted to copy and distribute translations of this
manual into another language, under the above conditions for modified
versions.
This publication is provided "AS IS." Tensilica, Inc. (hereafter
"Tensilica") does not make any warranty of any kind, either expressed
or implied, including, but not limited to, the implied warranties of
merchantability and fitness for a particular purpose. Information in
this document is provided solely to enable system and software
developers to use Tensilica(R) processors. Unless specifically set
forth herein, there are no express or implied patent, copyright or any
other intellectual property rights or licenses granted hereunder to
design or fabricate Tensilica integrated circuits or integrated
circuits based on the information in this document. Tensilica does not
warrant that the contents of this publication, whether individually or
as one or more groups, meets your requirements or that the publication
is error-free. This publication could include technical inaccuracies
or typographical errors. Changes may be made to the information
herein, and these changes may be incorporated in new editions of this
publication.
The following terms are trademarks or registered trademarks of
Tensilica, Inc.: FLIX, OSKit, Sea of Processors, Tensilica, Vectra,
Xplorer, XPRES, and Xtensa. All other trademarks and registered
trademarks are the property of their respective companies.

File: libc.info, Node: Top, Next: Revisions, Up: (dir)
The Red Hat newlib C Library
****************************
* Menu:
* Revisions::
* Introduction::
* Stdlib::
* Ctype::
* Stdio::
* Strings::
* Wchar strings::
* Signals::
* Timefns::
* Locale::
* Reentrancy::
* Misc::
* Xtensa::
* Syscalls::
* Arglists::
* Library Index::

File: libc.info, Node: Revisions, Next: Introduction, Prev: Top, Up: Top
Changes from Previous Versions
******************************
Tensilica newlib C library has been upgraded from version 1.19.0 to
version 2.0.0.

File: libc.info, Node: Introduction, Next: Stdlib, Prev: Revisions, Up: Top
1 Introduction
**************
This reference manual describes the functions provided by the Red Hat
"newlib" version of the standard ANSI C library. This document is not
intended as an overview or a tutorial for the C library. Each library
function is listed with a synopsis of its use, a brief description,
return values (including error handling), and portability issues.
Some of the library functions depend on support from the underlying
operating system and may not be available on every platform. For
embedded systems in particular, many of these underlying operating
system services may not be available or may not be fully functional.
The specific operating system subroutines required for a particular
library function are listed in the "Portability" section of the
function description. *Note Syscalls::, for a description of the
relevant operating system calls.

File: libc.info, Node: Stdlib, Next: Ctype, Prev: Introduction, Up: Top
2 Standard Utility Functions (`stdlib.h')
*****************************************
This chapter groups utility functions useful in a variety of programs.
The corresponding declarations are in the header file `stdlib.h'.
* Menu:
* _Exit:: End program execution without cleaning up
* a64l:: String to long long
* abort:: Abnormal termination of a program
* abs:: Integer absolute value (magnitude)
* assert:: Macro for Debugging Diagnostics
* atexit:: Request execution of functions at program exit
* atof:: String to double or float
* atoi:: String to integer
* atoll:: String to long long
* bsearch:: Binary search
* calloc:: Allocate space for arrays
* div:: Divide two integers
* ecvtbuf:: Double or float to string of digits
* ecvt:: Double or float to string of digits (malloc result)
* __env_lock:: Lock environment list for getenv and setenv
* gvcvt:: Format double or float as string
* exit:: End program execution
* getenv:: Look up environment variable
* labs:: Long integer absolute value (magnitude)
* ldiv:: Divide two long integers
* llabs:: Long long integer absolute value (magnitude)
* lldiv:: Divide two long long integers
* malloc:: Allocate and manage memory (malloc, realloc, free)
* mallinfo:: Get information about allocated memory
* __malloc_lock:: Lock memory pool for malloc and free
* mbsrtowcs:: Convert a character string to a wide-character string
* mbstowcs:: Minimal multibyte string to wide string converter
* mblen:: Minimal multibyte length
* mbtowc:: Minimal multibyte to wide character converter
* on_exit:: Request execution of functions at program exit
* qsort:: Array sort
* rand:: Pseudo-random numbers
* rand48:: Uniformly distributed pseudo-random numbers
* strtod:: String to double or float
* strtol:: String to long
* strtoll:: String to long long
* strtoul:: String to unsigned long
* strtoull:: String to unsigned long long
* wcsrtombs:: Convert a wide-character string to a character string
* wcstod:: Wide string to double or float
* wcstol:: Wide string to long
* wcstoll:: Wide string to long long
* wcstoul:: Wide string to unsigned long
* wcstoull:: Wide string to unsigned long long
* system:: Execute command string
* wcstombs:: Minimal wide string to multibyte string converter
* wctomb:: Minimal wide character to multibyte converter

File: libc.info, Node: _Exit, Next: a64l, Up: Stdlib
2.1 `_Exit'--end program execution with no cleanup processing
=============================================================
*Synopsis*
#include <stdlib.h>
void _Exit(int CODE);
*Description*
Use `_Exit' to return control from a program to the host operating
environment. Use the argument CODE to pass an exit status to the
operating environment: two particular values, `EXIT_SUCCESS' and
`EXIT_FAILURE', are defined in ``stdlib.h'' to indicate success or
failure in a portable fashion.
`_Exit' differs from `exit' in that it does not run any
application-defined cleanup functions registered with `atexit' and it
does not clean up files and streams. It is identical to `_exit'.
*Returns*
`_Exit' does not return to its caller.
*Portability*
`_Exit' is defined by the C99 standard.
Supporting OS subroutines required: `_exit'.

File: libc.info, Node: a64l, Next: abort, Prev: _Exit, Up: Stdlib
2.2 `a64l', `l64a'--convert between radix-64 ASCII string and long
==================================================================
*Synopsis*
#include <stdlib.h>
long a64l(const char *INPUT);
char *l64a(long INPUT);
*Description*
Conversion is performed between long and radix-64 characters. The
`l64a' routine transforms up to 32 bits of input value starting from
least significant bits to the most significant bits. The input value
is split up into a maximum of 5 groups of 6 bits and possibly one group
of 2 bits (bits 31 and 30).
Each group of 6 bits forms a value from 0-63 which is translated into
a character as follows:
* 0 = '.'
* 1 = '/'
* 2-11 = '0' to '9'
* 12-37 = 'A' to 'Z'
* 38-63 = 'a' to 'z'
When the remaining bits are zero or all bits have been translated, a
null terminator is appended to the string. An input value of 0 results
in the empty string.
The `a64l' function performs the reverse translation. Each
character is used to generate a 6-bit value for up to 30 bits and then
a 2-bit value to complete a 32-bit result. The null terminator means
that the remaining digits are 0. An empty input string or NULL string
results in 0L. An invalid string results in undefined behavior. If
the size of a long is greater than 32 bits, the result is sign-extended.
*Returns*
`l64a' returns a null-terminated string of 0 to 6 characters. `a64l'
returns the 32-bit translated value from the input character string.
*Portability*
`l64a' and `a64l' are non-ANSI and are defined by the Single Unix
Specification.
Supporting OS subroutines required: None.

File: libc.info, Node: abort, Next: abs, Prev: a64l, Up: Stdlib
2.3 `abort'--abnormal termination of a program
==============================================
*Synopsis*
#include <stdlib.h>
void abort(void);
*Description*
Use `abort' to signal that your program has detected a condition it
cannot deal with. Normally, `abort' ends your program's execution.
Before terminating your program, `abort' raises the exception
`SIGABRT' (using ``raise(SIGABRT)''). If you have used `signal' to
register an exception handler for this condition, that handler has the
opportunity to retain control, thereby avoiding program termination.
In this implementation, `abort' does not perform any stream- or
file-related cleanup (the host environment may do so; if not, you can
arrange for your program to do its own cleanup with a `SIGABRT'
exception handler).
*Returns*
`abort' does not return to its caller.
*Portability*
ANSI C requires `abort'.
Supporting OS subroutines required: `_exit' and optionally, `write'.

File: libc.info, Node: abs, Next: assert, Prev: abort, Up: Stdlib
2.4 `abs'--integer absolute value (magnitude)
=============================================
*Synopsis*
#include <stdlib.h>
int abs(int I);
*Description*
`abs' returns the absolute value of I (also called the magnitude of I).
That is, if I is negative, the result is the opposite of I, but if I
is nonnegative the result is I.
The similar function `labs' uses and returns `long' rather than
`int' values.
*Returns*
The result is a nonnegative integer.
*Portability*
`abs' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: assert, Next: atexit, Prev: abs, Up: Stdlib
2.5 `assert'--macro for debugging diagnostics
=============================================
*Synopsis*
#include <assert.h>
void assert(int EXPRESSION);
*Description*
Use this macro to embed debuggging diagnostic statements in your
programs. The argument EXPRESSION should be an expression which
evaluates to true (nonzero) when your program is working as you
intended.
When EXPRESSION evaluates to false (zero), `assert' calls `abort',
after first printing a message showing what failed and where:
Assertion failed: EXPRESSION, file FILENAME, line LINENO, function: FUNC
If the name of the current function is not known (for example, when
using a C89 compiler that does not understand __func__), the function
location is omitted.
The macro is defined to permit you to turn off all uses of `assert'
at compile time by defining `NDEBUG' as a preprocessor variable. If
you do this, the `assert' macro expands to
(void(0))
*Returns*
`assert' does not return a value.
*Portability*
The `assert' macro is required by ANSI, as is the behavior when
`NDEBUG' is defined.
Supporting OS subroutines required (only if enabled): `close',
`fstat', `getpid', `isatty', `kill', `lseek', `read', `sbrk', `write'.

File: libc.info, Node: atexit, Next: atof, Prev: assert, Up: Stdlib
2.6 `atexit'--request execution of functions at program exit
============================================================
*Synopsis*
#include <stdlib.h>
int atexit (void (*FUNCTION)(void));
*Description*
You can use `atexit' to enroll functions in a list of functions that
will be called when your program terminates normally. The argument is
a pointer to a user-defined function (which must not require arguments
and must not return a result).
The functions are kept in a LIFO stack; that is, the last function
enrolled by `atexit' will be the first to execute when your program
exits.
There is no built-in limit to the number of functions you can enroll
in this list; however, after every group of 32 functions is enrolled,
`atexit' will call `malloc' to get space for the next part of the list.
The initial list of 32 functions is statically allocated, so you can
always count on at least that many slots available.
*Returns*
`atexit' returns `0' if it succeeds in enrolling your function, `-1' if
it fails (possible only if no space was available for `malloc' to
extend the list of functions).
*Portability*
`atexit' is required by the ANSI standard, which also specifies that
implementations must support enrolling at least 32 functions.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: atof, Next: atoi, Prev: atexit, Up: Stdlib
2.7 `atof', `atoff'--string to double or float
==============================================
*Synopsis*
#include <stdlib.h>
double atof(const char *S);
float atoff(const char *S);
*Description*
`atof' converts the initial portion of a string to a `double'. `atoff'
converts the initial portion of a string to a `float'.
The functions parse the character string S, locating a substring
which can be converted to a floating-point value. The substring must
match the format:
[+|-]DIGITS[.][DIGITS][(e|E)[+|-]DIGITS]
The substring converted is the longest initial fragment of S that
has the expected format, beginning with the first non-whitespace
character. The substring is empty if `str' is empty, consists entirely
of whitespace, or if the first non-whitespace character is something
other than `+', `-', `.', or a digit.
`atof(S)' is implemented as `strtod(S, NULL)'. `atoff(S)' is
implemented as `strtof(S, NULL)'.
*Returns*
`atof' returns the converted substring value, if any, as a `double'; or
`0.0', if no conversion could be performed. If the correct value is
out of the range of representable values, plus or minus `HUGE_VAL' is
returned, and `ERANGE' is stored in `errno'. If the correct value
would cause underflow, `0.0' is returned and `ERANGE' is stored in
`errno'.
`atoff' obeys the same rules as `atof', except that it returns a
`float'.
*Portability*
`atof' is ANSI C. `atof', `atoi', and `atol' are subsumed by `strod'
and `strol', but are used extensively in existing code. These functions
are less reliable, but may be faster if the argument is verified to be
in a valid range.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: atoi, Next: atoll, Prev: atof, Up: Stdlib
2.8 `atoi', `atol'--string to integer
=====================================
*Synopsis*
#include <stdlib.h>
int atoi(const char *S);
long atol(const char *S);
int _atoi_r(struct _reent *PTR, const char *S);
long _atol_r(struct _reent *PTR, const char *S);
*Description*
`atoi' converts the initial portion of a string to an `int'. `atol'
converts the initial portion of a string to a `long'.
`atoi(s)' is implemented as `(int)strtol(s, NULL, 10).' `atol(s)' is
implemented as `strtol(s, NULL, 10).'
`_atoi_r' and `_atol_r' are reentrant versions of `atoi' and `atol'
respectively, passing the reentrancy struct pointer.
*Returns*
The functions return the converted value, if any. If no conversion was
made, `0' is returned.
*Portability*
`atoi', `atol' are ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: atoll, Next: bsearch, Prev: atoi, Up: Stdlib
2.9 `atoll'--convert a string to a long long integer
====================================================
*Synopsis*
#include <stdlib.h>
long long atoll(const char *STR);
long long _atoll_r(struct _reent *PTR, const char *STR);
*Description*
The function `atoll' converts the initial portion of the string pointed
to by `*STR' to a type `long long'. A call to atoll(str) in this
implementation is equivalent to strtoll(str, (char **)NULL, 10)
including behavior on error.
The alternate function `_atoll_r' is a reentrant version. The extra
argument REENT is a pointer to a reentrancy structure.
*Returns*
The converted value.
*Portability*
`atoll' is ISO 9899 (C99) and POSIX 1003.1-2001 compatable.
No supporting OS subroutines are required.

File: libc.info, Node: bsearch, Next: calloc, Prev: atoll, Up: Stdlib
2.10 `bsearch'--binary search
=============================
*Synopsis*
#include <stdlib.h>
void *bsearch(const void *KEY, const void *BASE,
size_t NMEMB, size_t SIZE,
int (*COMPAR)(const void *, const void *));
*Description*
`bsearch' searches an array beginning at BASE for any element that
matches KEY, using binary search. NMEMB is the element count of the
array; SIZE is the size of each element.
The array must be sorted in ascending order with respect to the
comparison function COMPAR (which you supply as the last argument of
`bsearch').
You must define the comparison function `(*COMPAR)' to have two
arguments; its result must be negative if the first argument is less
than the second, zero if the two arguments match, and positive if the
first argument is greater than the second (where "less than" and
"greater than" refer to whatever arbitrary ordering is appropriate).
*Returns*
Returns a pointer to an element of ARRAY that matches KEY. If more
than one matching element is available, the result may point to any of
them.
*Portability*
`bsearch' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: calloc, Next: div, Prev: bsearch, Up: Stdlib
2.11 `calloc'--allocate space for arrays
========================================
*Synopsis*
#include <stdlib.h>
void *calloc(size_t N, size_t S);
void *_calloc_r(void *REENT, size_t N, size_t S);
*Description*
Use `calloc' to request a block of memory sufficient to hold an array
of N elements, each of which has size S.
The memory allocated by `calloc' comes out of the same memory pool
used by `malloc', but the memory block is initialized to all zero
bytes. (To avoid the overhead of initializing the space, use `malloc'
instead.)
The alternate function `_calloc_r' is reentrant. The extra argument
REENT is a pointer to a reentrancy structure.
*Returns*
If successful, a pointer to the newly allocated space.
If unsuccessful, `NULL'.
*Portability*
`calloc' is ANSI.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: div, Next: ecvtbuf, Prev: calloc, Up: Stdlib
2.12 `div'--divide two integers
===============================
*Synopsis*
#include <stdlib.h>
div_t div(int N, int D);
*Description*
Divide N/D, returning quotient and remainder as two integers in a
structure `div_t'.
*Returns*
The result is represented with the structure
typedef struct
{
int quot;
int rem;
} div_t;
where the `quot' field represents the quotient, and `rem' the
remainder. For nonzero D, if ``R = div(N,D);'' then N equals ``R.rem +
D*R.quot''.
To divide `long' rather than `int' values, use the similar function
`ldiv'.
*Portability*
`div' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: ecvt, Next: __env_lock, Prev: ecvtbuf, Up: Stdlib
2.13 `ecvt', `ecvtf', `fcvt', `fcvtf'--double or float to string
================================================================
*Synopsis*
#include <stdlib.h>
char *ecvt(double VAL, int CHARS, int *DECPT, int *SGN);
char *ecvtf(float VAL, int CHARS, int *DECPT, int *SGN);
char *fcvt(double VAL, int DECIMALS,
int *DECPT, int *SGN);
char *fcvtf(float VAL, int DECIMALS,
int *DECPT, int *SGN);
*Description*
`ecvt' and `fcvt' produce (null-terminated) strings of digits
representating the `double' number VAL. `ecvtf' and `fcvtf' produce
the corresponding character representations of `float' numbers.
(The `stdlib' functions `ecvtbuf' and `fcvtbuf' are reentrant
versions of `ecvt' and `fcvt'.)
The only difference between `ecvt' and `fcvt' is the interpretation
of the second argument (CHARS or DECIMALS). For `ecvt', the second
argument CHARS specifies the total number of characters to write (which
is also the number of significant digits in the formatted string, since
these two functions write only digits). For `fcvt', the second
argument DECIMALS specifies the number of characters to write after the
decimal point; all digits for the integer part of VAL are always
included.
Since `ecvt' and `fcvt' write only digits in the output string, they
record the location of the decimal point in `*DECPT', and the sign of
the number in `*SGN'. After formatting a number, `*DECPT' contains the
number of digits to the left of the decimal point. `*SGN' contains `0'
if the number is positive, and `1' if it is negative.
*Returns*
All four functions return a pointer to the new string containing a
character representation of VAL.
*Portability*
None of these functions are ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: gvcvt, Next: exit, Prev: __env_lock, Up: Stdlib
2.14 `gvcvt', `gcvtf'--format double or float as string
=======================================================
*Synopsis*
#include <stdlib.h>
char *gcvt(double VAL, int PRECISION, char *BUF);
char *gcvtf(float VAL, int PRECISION, char *BUF);
*Description*
`gcvt' writes a fully formatted number as a null-terminated string in
the buffer `*BUF'. `gdvtf' produces corresponding character
representations of `float' numbers.
`gcvt' uses the same rules as the `printf' format
``%.PRECISIONg''--only negative values are signed (with ``-''), and
either exponential or ordinary decimal-fraction format is chosen
depending on the number of significant digits (specified by PRECISION).
*Returns*
The result is a pointer to the formatted representation of VAL (the
same as the argument BUF).
*Portability*
Neither function is ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: ecvtbuf, Next: ecvt, Prev: div, Up: Stdlib
2.15 `ecvtbuf', `fcvtbuf'--double or float to string
====================================================
*Synopsis*
#include <stdio.h>
char *ecvtbuf(double VAL, int CHARS, int *DECPT,
int *SGN, char *BUF);
char *fcvtbuf(double VAL, int DECIMALS, int *DECPT,
int *SGN, char *BUF);
*Description*
`ecvtbuf' and `fcvtbuf' produce (null-terminated) strings of digits
representating the `double' number VAL.
The only difference between `ecvtbuf' and `fcvtbuf' is the
interpretation of the second argument (CHARS or DECIMALS). For
`ecvtbuf', the second argument CHARS specifies the total number of
characters to write (which is also the number of significant digits in
the formatted string, since these two functions write only digits). For
`fcvtbuf', the second argument DECIMALS specifies the number of
characters to write after the decimal point; all digits for the integer
part of VAL are always included.
Since `ecvtbuf' and `fcvtbuf' write only digits in the output
string, they record the location of the decimal point in `*DECPT', and
the sign of the number in `*SGN'. After formatting a number, `*DECPT'
contains the number of digits to the left of the decimal point. `*SGN'
contains `0' if the number is positive, and `1' if it is negative. For
both functions, you supply a pointer BUF to an area of memory to hold
the converted string.
*Returns*
Both functions return a pointer to BUF, the string containing a
character representation of VAL.
*Portability*
Neither function is ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: __env_lock, Next: gvcvt, Prev: ecvt, Up: Stdlib
2.16 `__env_lock', `__env_unlock'--lock environ variable
========================================================
*Synopsis*
#include <envlock.h>
void __env_lock (struct _reent *REENT);
void __env_unlock (struct _reent *REENT);
*Description*
The `setenv' family of routines call these functions when they need to
modify the environ variable. The version of these routines supplied in
the library use the lock API defined in sys/lock.h. If multiple
threads of execution can call `setenv', or if `setenv' can be called
reentrantly, then you need to define your own versions of these
functions in order to safely lock the memory pool during a call. If
you do not, the memory pool may become corrupted.
A call to `setenv' may call `__env_lock' recursively; that is, the
sequence of calls may go `__env_lock', `__env_lock', `__env_unlock',
`__env_unlock'. Any implementation of these routines must be careful
to avoid causing a thread to wait for a lock that it already holds.

File: libc.info, Node: exit, Next: getenv, Prev: gvcvt, Up: Stdlib
2.17 `exit'--end program execution
==================================
*Synopsis*
#include <stdlib.h>
void exit(int CODE);
*Description*
Use `exit' to return control from a program to the host operating
environment. Use the argument CODE to pass an exit status to the
operating environment: two particular values, `EXIT_SUCCESS' and
`EXIT_FAILURE', are defined in ``stdlib.h'' to indicate success or
failure in a portable fashion.
`exit' does two kinds of cleanup before ending execution of your
program. First, it calls all application-defined cleanup functions you
have enrolled with `atexit'. Second, files and streams are cleaned up:
any pending output is delivered to the host system, each open file or
stream is closed, and files created by `tmpfile' are deleted.
*Returns*
`exit' does not return to its caller.
*Portability*
ANSI C requires `exit', and specifies that `EXIT_SUCCESS' and
`EXIT_FAILURE' must be defined.
Supporting OS subroutines required: `_exit'.

File: libc.info, Node: getenv, Next: labs, Prev: exit, Up: Stdlib
2.18 `getenv'--look up environment variable
===========================================
*Synopsis*
#include <stdlib.h>
char *getenv(const char *NAME);
*Description*
`getenv' searches the list of environment variable names and values
(using the global pointer "`char **environ'") for a variable whose name
matches the string at NAME. If a variable name matches, `getenv'
returns a pointer to the associated value.
*Returns*
A pointer to the (string) value of the environment variable, or `NULL'
if there is no such environment variable.
*Portability*
`getenv' is ANSI, but the rules for properly forming names of
environment variables vary from one system to another.
`getenv' requires a global pointer `environ'.

File: libc.info, Node: labs, Next: ldiv, Prev: getenv, Up: Stdlib
2.19 `labs'--long integer absolute value
========================================
*Synopsis*
#include <stdlib.h>
long labs(long I);
*Description*
`labs' returns the absolute value of I (also called the magnitude of
I). That is, if I is negative, the result is the opposite of I, but if
I is nonnegative the result is I.
The similar function `abs' uses and returns `int' rather than `long'
values.
*Returns*
The result is a nonnegative long integer.
*Portability*
`labs' is ANSI.
No supporting OS subroutine calls are required.

File: libc.info, Node: ldiv, Next: llabs, Prev: labs, Up: Stdlib
2.20 `ldiv'--divide two long integers
=====================================
*Synopsis*
#include <stdlib.h>
ldiv_t ldiv(long N, long D);
*Description*
Divide N/D, returning quotient and remainder as two long integers in a
structure `ldiv_t'.
*Returns*
The result is represented with the structure
typedef struct
{
long quot;
long rem;
} ldiv_t;
where the `quot' field represents the quotient, and `rem' the
remainder. For nonzero D, if ``R = ldiv(N,D);'' then N equals ``R.rem
+ D*R.quot''.
To divide `int' rather than `long' values, use the similar function
`div'.
*Portability*
`ldiv' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: llabs, Next: lldiv, Prev: ldiv, Up: Stdlib
2.21 `llabs'--compute the absolute value of an long long integer.
=================================================================
*Synopsis*
#include <stdlib.h>
long long llabs(long long j);
*Description*
The `llabs' function computes the absolute value of the long long
integer argument J (also called the magnitude of J).
The similar function `labs' uses and returns `long' rather than
`long long' values.
*Returns*
A nonnegative long long integer.
*Portability*
`llabs' is ISO 9899 (C99) compatable.
No supporting OS subroutines are required.

File: libc.info, Node: lldiv, Next: malloc, Prev: llabs, Up: Stdlib
2.22 `lldiv'--divide two long long integers
===========================================
*Synopsis*
#include <stdlib.h>
lldiv_t lldiv(long long N, long long D);
*Description*
Divide N/D, returning quotient and remainder as two long long integers
in a structure `lldiv_t'.
*Returns*
The result is represented with the structure
typedef struct
{
long long quot;
long long rem;
} lldiv_t;
where the `quot' field represents the quotient, and `rem' the
remainder. For nonzero D, if ``R = ldiv(N,D);'' then N equals ``R.rem
+ D*R.quot''.
To divide `long' rather than `long long' values, use the similar
function `ldiv'.
*Portability*
`lldiv' is ISO 9899 (C99) compatable.
No supporting OS subroutines are required.

File: libc.info, Node: malloc, Next: mallinfo, Prev: lldiv, Up: Stdlib
2.23 `malloc', `realloc', `free'--manage memory
===============================================
*Synopsis*
#include <stdlib.h>
void *malloc(size_t NBYTES);
void *realloc(void *APTR, size_t NBYTES);
void *reallocf(void *APTR, size_t NBYTES);
void free(void *APTR);
void *memalign(size_t ALIGN, size_t NBYTES);
size_t malloc_usable_size(void *APTR);
void *_malloc_r(void *REENT, size_t NBYTES);
void *_realloc_r(void *REENT,
void *APTR, size_t NBYTES);
void *_reallocf_r(void *REENT,
void *APTR, size_t NBYTES);
void _free_r(void *REENT, void *APTR);
void *_memalign_r(void *REENT,
size_t ALIGN, size_t NBYTES);
size_t _malloc_usable_size_r(void *REENT, void *APTR);
*Description*
These functions manage a pool of system memory.
Use `malloc' to request allocation of an object with at least NBYTES
bytes of storage available. If the space is available, `malloc'
returns a pointer to a newly allocated block as its result.
If you already have a block of storage allocated by `malloc', but
you no longer need all the space allocated to it, you can make it
smaller by calling `realloc' with both the object pointer and the new
desired size as arguments. `realloc' guarantees that the contents of
the smaller object match the beginning of the original object.
Similarly, if you need more space for an object, use `realloc' to
request the larger size; again, `realloc' guarantees that the beginning
of the new, larger object matches the contents of the original object.
When you no longer need an object originally allocated by `malloc'
or `realloc' (or the related function `calloc'), return it to the
memory storage pool by calling `free' with the address of the object as
the argument. You can also use `realloc' for this purpose by calling
it with `0' as the NBYTES argument.
The `reallocf' function behaves just like `realloc' except if the
function is required to allocate new storage and this fails. In this
case `reallocf' will free the original object passed in whereas
`realloc' will not.
The `memalign' function returns a block of size NBYTES aligned to a
ALIGN boundary. The ALIGN argument must be a power of two.
The `malloc_usable_size' function takes a pointer to a block
allocated by `malloc'. It returns the amount of space that is
available in the block. This may or may not be more than the size
requested from `malloc', due to alignment or minimum size constraints.
The alternate functions `_malloc_r', `_realloc_r', `_reallocf_r',
`_free_r', `_memalign_r', and `_malloc_usable_size_r' are reentrant
versions. The extra argument REENT is a pointer to a reentrancy
structure.
If you have multiple threads of execution which may call any of these
routines, or if any of these routines may be called reentrantly, then
you must provide implementations of the `__malloc_lock' and
`__malloc_unlock' functions for your system. See the documentation for
those functions.
These functions operate by calling the function `_sbrk_r' or `sbrk',
which allocates space. You may need to provide one of these functions
for your system. `_sbrk_r' is called with a positive value to allocate
more space, and with a negative value to release previously allocated
space if it is no longer required. *Note Stubs::.
*Returns*
`malloc' returns a pointer to the newly allocated space, if successful;
otherwise it returns `NULL'. If your application needs to generate
empty objects, you may use `malloc(0)' for this purpose.
`realloc' returns a pointer to the new block of memory, or `NULL' if
a new block could not be allocated. `NULL' is also the result when you
use ``realloc(APTR,0)'' (which has the same effect as ``free(APTR)'').
You should always check the result of `realloc'; successful
reallocation is not guaranteed even when you request a smaller object.
`free' does not return a result.
`memalign' returns a pointer to the newly allocated space.
`malloc_usable_size' returns the usable size.
*Portability*
`malloc', `realloc', and `free' are specified by the ANSI C standard,
but other conforming implementations of `malloc' may behave differently
when NBYTES is zero.
`memalign' is part of SVR4.
`malloc_usable_size' is not portable.
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: mallinfo, Next: __malloc_lock, Prev: malloc, Up: Stdlib
2.24 `mallinfo', `malloc_stats', `mallopt'--malloc support
==========================================================
*Synopsis*
#include <malloc.h>
struct mallinfo mallinfo(void);
void malloc_stats(void);
int mallopt(int PARAMETER, VALUE);
struct mallinfo _mallinfo_r(void *REENT);
void _malloc_stats_r(void *REENT);
int _mallopt_r(void *REENT, int PARAMETER, VALUE);
*Description*
`mallinfo' returns a structure describing the current state of memory
allocation. The structure is defined in malloc.h. The following
fields are defined: `arena' is the total amount of space in the heap;
`ordblks' is the number of chunks which are not in use; `uordblks' is
the total amount of space allocated by `malloc'; `fordblks' is the
total amount of space not in use; `keepcost' is the size of the top
most memory block.
`malloc_stats' print some statistics about memory allocation on
standard error.
`mallopt' takes a parameter and a value. The parameters are defined
in malloc.h, and may be one of the following: `M_TRIM_THRESHOLD' sets
the maximum amount of unused space in the top most block before
releasing it back to the system in `free' (the space is released by
calling `_sbrk_r' with a negative argument); `M_TOP_PAD' is the amount
of padding to allocate whenever `_sbrk_r' is called to allocate more
space.
The alternate functions `_mallinfo_r', `_malloc_stats_r', and
`_mallopt_r' are reentrant versions. The extra argument REENT is a
pointer to a reentrancy structure.
*Returns*
`mallinfo' returns a mallinfo structure. The structure is defined in
malloc.h.
`malloc_stats' does not return a result.
`mallopt' returns zero if the parameter could not be set, or
non-zero if it could be set.
*Portability*
`mallinfo' and `mallopt' are provided by SVR4, but `mallopt' takes
different parameters on different systems. `malloc_stats' is not
portable.

File: libc.info, Node: __malloc_lock, Next: mbsrtowcs, Prev: mallinfo, Up: Stdlib
2.25 `__malloc_lock', `__malloc_unlock'--lock malloc pool
=========================================================
*Synopsis*
#include <malloc.h>
void __malloc_lock (struct _reent *REENT);
void __malloc_unlock (struct _reent *REENT);
*Description*
The `malloc' family of routines call these functions when they need to
lock the memory pool. The version of these routines supplied in the
library use the lock API defined in sys/lock.h. If multiple threads of
execution can call `malloc', or if `malloc' can be called reentrantly,
then you need to define your own versions of these functions in order
to safely lock the memory pool during a call. If you do not, the
memory pool may become corrupted.
A call to `malloc' may call `__malloc_lock' recursively; that is,
the sequence of calls may go `__malloc_lock', `__malloc_lock',
`__malloc_unlock', `__malloc_unlock'. Any implementation of these
routines must be careful to avoid causing a thread to wait for a lock
that it already holds.

File: libc.info, Node: mblen, Next: mbtowc, Prev: mbstowcs, Up: Stdlib
2.26 `mblen'--minimal multibyte length function
===============================================
*Synopsis*
#include <stdlib.h>
int mblen(const char *S, size_t N);
*Description*
When _MB_CAPABLE is not defined, this is a minimal ANSI-conforming
implementation of `mblen'. In this case, the only "multi-byte
character sequences" recognized are single bytes, and thus `1' is
returned unless S is the null pointer or has a length of 0 or is the
empty string.
When _MB_CAPABLE is defined, this routine calls `_mbtowc_r' to
perform the conversion, passing a state variable to allow state
dependent decoding. The result is based on the locale setting which may
be restricted to a defined set of locales.
*Returns*
This implementation of `mblen' returns `0' if S is `NULL' or the empty
string; it returns `1' if not _MB_CAPABLE or the character is a
single-byte character; it returns `-1' if the multi-byte character is
invalid; otherwise it returns the number of bytes in the multibyte
character.
*Portability*
`mblen' is required in the ANSI C standard. However, the precise
effects vary with the locale.
`mblen' requires no supporting OS subroutines.

File: libc.info, Node: mbsrtowcs, Next: mbstowcs, Prev: __malloc_lock, Up: Stdlib
2.27 `mbsrtowcs', `mbsnrtowcs'--convert a character string to a wide-character string
=====================================================================================
*Synopsis*
#include <wchar.h>
size_t mbsrtowcs(wchar_t *DST, const char **SRC, size_t LEN,
mbstate_t *PS);
#include <wchar.h>
size_t _mbsrtowcs_r(struct _reent *PTR, wchar_t *DST,
const char **SRC, size_t LEN,
mbstate_t *PS);
#include <wchar.h>
size_t mbsnrtowcs(wchar_t *DST, const char **SRC,
size_t NMS, size_t LEN, mbstate_t *PS);
#include <wchar.h>
size_t _mbsnrtowcs_r(struct _reent *PTR, wchar_t *DST,
const char **SRC, size_t NMS,
size_t LEN, mbstate_t *PS);
*Description*
The `mbsrtowcs' function converts a sequence of multibyte characters
pointed to indirectly by SRC into a sequence of corresponding wide
characters and stores at most LEN of them in the wchar_t array pointed
to by DST, until it encounters a terminating null character ('\0').
If DST is NULL, no characters are stored.
If DST is not NULL, the pointer pointed to by SRC is updated to point
to the character after the one that conversion stopped at. If
conversion stops because a null character is encountered, *SRC is set
to NULL.
The mbstate_t argument, PS, is used to keep track of the shift
state. If it is NULL, `mbsrtowcs' uses an internal, static mbstate_t
object, which is initialized to the initial conversion state at program
startup.
The `mbsnrtowcs' function behaves identically to `mbsrtowcs', except
that conversion stops after reading at most NMS bytes from the buffer
pointed to by SRC.
*Returns*
The `mbsrtowcs' and `mbsnrtowcs' functions return the number of wide
characters stored in the array pointed to by DST if successful,
otherwise it returns (size_t)-1.
*Portability*
`mbsrtowcs' is defined by the C99 standard. `mbsnrtowcs' is defined by
the POSIX.1-2008 standard.

File: libc.info, Node: mbstowcs, Next: mblen, Prev: mbsrtowcs, Up: Stdlib
2.28 `mbstowcs'--minimal multibyte string to wide char converter
================================================================
*Synopsis*
#include <stdlib.h>
int mbstowcs(wchar_t *PWC, const char *S, size_t N);
*Description*
When _MB_CAPABLE is not defined, this is a minimal ANSI-conforming
implementation of `mbstowcs'. In this case, the only "multi-byte
character sequences" recognized are single bytes, and they are
"converted" to wide-char versions simply by byte extension.
When _MB_CAPABLE is defined, this routine calls `_mbstowcs_r' to
perform the conversion, passing a state variable to allow state
dependent decoding. The result is based on the locale setting which may
be restricted to a defined set of locales.
*Returns*
This implementation of `mbstowcs' returns `0' if S is `NULL' or is the
empty string; it returns `-1' if _MB_CAPABLE and one of the multi-byte
characters is invalid or incomplete; otherwise it returns the minimum
of: `n' or the number of multi-byte characters in `s' plus 1 (to
compensate for the nul character). If the return value is -1, the
state of the `pwc' string is indeterminate. If the input has a length
of 0, the output string will be modified to contain a wchar_t nul
terminator.
*Portability*
`mbstowcs' is required in the ANSI C standard. However, the precise
effects vary with the locale.
`mbstowcs' requires no supporting OS subroutines.

File: libc.info, Node: mbtowc, Next: on_exit, Prev: mblen, Up: Stdlib
2.29 `mbtowc'--minimal multibyte to wide char converter
=======================================================
*Synopsis*
#include <stdlib.h>
int mbtowc(wchar_t *PWC, const char *S, size_t N);
*Description*
When _MB_CAPABLE is not defined, this is a minimal ANSI-conforming
implementation of `mbtowc'. In this case, only "multi-byte character
sequences" recognized are single bytes, and they are "converted" to
themselves. Each call to `mbtowc' copies one character from `*S' to
`*PWC', unless S is a null pointer. The argument n is ignored.
When _MB_CAPABLE is defined, this routine calls `_mbtowc_r' to
perform the conversion, passing a state variable to allow state
dependent decoding. The result is based on the locale setting which may
be restricted to a defined set of locales.
*Returns*
This implementation of `mbtowc' returns `0' if S is `NULL' or is the
empty string; it returns `1' if not _MB_CAPABLE or the character is a
single-byte character; it returns `-1' if n is `0' or the multi-byte
character is invalid; otherwise it returns the number of bytes in the
multibyte character. If the return value is -1, no changes are made to
the `pwc' output string. If the input is the empty string, a wchar_t
nul is placed in the output string and 0 is returned. If the input has
a length of 0, no changes are made to the `pwc' output string.
*Portability*
`mbtowc' is required in the ANSI C standard. However, the precise
effects vary with the locale.
`mbtowc' requires no supporting OS subroutines.

File: libc.info, Node: on_exit, Next: qsort, Prev: mbtowc, Up: Stdlib
2.30 `on_exit'--request execution of function with argument at program exit
===========================================================================
*Synopsis*
#include <stdlib.h>
int on_exit (void (*FUNCTION)(int, void *), void *ARG);
*Description*
You can use `on_exit' to enroll functions in a list of functions that
will be called when your program terminates normally. The argument is
a pointer to a user-defined function which takes two arguments. The
first is the status code passed to exit and the second argument is of
type pointer to void. The function must not return a result. The value
of ARG is registered and passed as the argument to FUNCTION.
The functions are kept in a LIFO stack; that is, the last function
enrolled by `atexit' or `on_exit' will be the first to execute when
your program exits. You can intermix functions using `atexit' and
`on_exit'.
There is no built-in limit to the number of functions you can enroll
in this list; however, after every group of 32 functions is enrolled,
`atexit'/`on_exit' will call `malloc' to get space for the next part of
the list. The initial list of 32 functions is statically allocated, so
you can always count on at least that many slots available.
*Returns*
`on_exit' returns `0' if it succeeds in enrolling your function, `-1'
if it fails (possible only if no space was available for `malloc' to
extend the list of functions).
*Portability*
`on_exit' is a non-standard glibc extension
Supporting OS subroutines required: None

File: libc.info, Node: qsort, Next: rand, Prev: on_exit, Up: Stdlib
2.31 `qsort'--sort an array
===========================
*Synopsis*
#include <stdlib.h>
void qsort(void *BASE, size_t NMEMB, size_t SIZE,
int (*COMPAR)(const void *, const void *) );
*Description*
`qsort' sorts an array (beginning at BASE) of NMEMB objects. SIZE
describes the size of each element of the array.
You must supply a pointer to a comparison function, using the
argument shown as COMPAR. (This permits sorting objects of unknown
properties.) Define the comparison function to accept two arguments,
each a pointer to an element of the array starting at BASE. The result
of `(*COMPAR)' must be negative if the first argument is less than the
second, zero if the two arguments match, and positive if the first
argument is greater than the second (where "less than" and "greater
than" refer to whatever arbitrary ordering is appropriate).
The array is sorted in place; that is, when `qsort' returns, the
array elements beginning at BASE have been reordered.
*Returns*
`qsort' does not return a result.
*Portability*
`qsort' is required by ANSI (without specifying the sorting algorithm).

File: libc.info, Node: rand, Next: rand48, Prev: qsort, Up: Stdlib
2.32 `rand', `srand'--pseudo-random numbers
===========================================
*Synopsis*
#include <stdlib.h>
int rand(void);
void srand(unsigned int SEED);
int rand_r(unsigned int *SEED);
*Description*
`rand' returns a different integer each time it is called; each integer
is chosen by an algorithm designed to be unpredictable, so that you can
use `rand' when you require a random number. The algorithm depends on
a static variable called the "random seed"; starting with a given value
of the random seed always produces the same sequence of numbers in
successive calls to `rand'.
You can set the random seed using `srand'; it does nothing beyond
storing its argument in the static variable used by `rand'. You can
exploit this to make the pseudo-random sequence less predictable, if
you wish, by using some other unpredictable value (often the least
significant parts of a time-varying value) as the random seed before
beginning a sequence of calls to `rand'; or, if you wish to ensure (for
example, while debugging) that successive runs of your program use the
same "random" numbers, you can use `srand' to set the same random seed
at the outset.
*Returns*
`rand' returns the next pseudo-random integer in sequence; it is a
number between `0' and `RAND_MAX' (inclusive).
`srand' does not return a result.
*Portability*
`rand' is required by ANSI, but the algorithm for pseudo-random number
generation is not specified; therefore, even if you use the same random
seed, you cannot expect the same sequence of results on two different
systems.
`rand' requires no supporting OS subroutines.

File: libc.info, Node: rand48, Next: strtod, Prev: rand, Up: Stdlib
2.33 `rand48', `drand48', `erand48', `lrand48', `nrand48', `mrand48', `jrand48', `srand48', `seed48', `lcong48'--pseudo-random number generators and initialization routines
============================================================================================================================================================================
*Synopsis*
#include <stdlib.h>
double drand48(void);
double erand48(unsigned short XSEED[3]);
long lrand48(void);
long nrand48(unsigned short XSEED[3]);
long mrand48(void);
long jrand48(unsigned short XSEED[3]);
void srand48(long SEED);
unsigned short *seed48(unsigned short XSEED[3]);
void lcong48(unsigned short P[7]);
*Description*
The `rand48' family of functions generates pseudo-random numbers using
a linear congruential algorithm working on integers 48 bits in size.
The particular formula employed is r(n+1) = (a * r(n) + c) mod m where
the default values are for the multiplicand a = 0xfdeece66d =
25214903917 and the addend c = 0xb = 11. The modulo is always fixed at
m = 2 ** 48. r(n) is called the seed of the random number generator.
For all the six generator routines described next, the first
computational step is to perform a single iteration of the algorithm.
`drand48' and `erand48' return values of type double. The full 48
bits of r(n+1) are loaded into the mantissa of the returned value, with
the exponent set such that the values produced lie in the interval
[0.0, 1.0].
`lrand48' and `nrand48' return values of type long in the range [0,
2**31-1]. The high-order (31) bits of r(n+1) are loaded into the lower
bits of the returned value, with the topmost (sign) bit set to zero.
`mrand48' and `jrand48' return values of type long in the range
[-2**31, 2**31-1]. The high-order (32) bits of r(n+1) are loaded into
the returned value.
`drand48', `lrand48', and `mrand48' use an internal buffer to store
r(n). For these functions the initial value of r(0) = 0x1234abcd330e =
20017429951246.
On the other hand, `erand48', `nrand48', and `jrand48' use a
user-supplied buffer to store the seed r(n), which consists of an array
of 3 shorts, where the zeroth member holds the least significant bits.
All functions share the same multiplicand and addend.
`srand48' is used to initialize the internal buffer r(n) of
`drand48', `lrand48', and `mrand48' such that the 32 bits of the seed
value are copied into the upper 32 bits of r(n), with the lower 16 bits
of r(n) arbitrarily being set to 0x330e. Additionally, the constant
multiplicand and addend of the algorithm are reset to the default
values given above.
`seed48' also initializes the internal buffer r(n) of `drand48',
`lrand48', and `mrand48', but here all 48 bits of the seed can be
specified in an array of 3 shorts, where the zeroth member specifies
the lowest bits. Again, the constant multiplicand and addend of the
algorithm are reset to the default values given above. `seed48'
returns a pointer to an array of 3 shorts which contains the old seed.
This array is statically allocated, thus its contents are lost after
each new call to `seed48'.
Finally, `lcong48' allows full control over the multiplicand and
addend used in `drand48', `erand48', `lrand48', `nrand48', `mrand48',
and `jrand48', and the seed used in `drand48', `lrand48', and `mrand48'.
An array of 7 shorts is passed as parameter; the first three shorts are
used to initialize the seed; the second three are used to initialize the
multiplicand; and the last short is used to initialize the addend. It
is thus not possible to use values greater than 0xffff as the addend.
Note that all three methods of seeding the random number generator
always also set the multiplicand and addend for any of the six
generator calls.
For a more powerful random number generator, see `random'.
*Portability*
SUS requires these functions.
No supporting OS subroutines are required.

File: libc.info, Node: strtod, Next: strtol, Prev: rand48, Up: Stdlib
2.34 `strtod', `strtof'--string to double or float
==================================================
*Synopsis*
#include <stdlib.h>
double strtod(const char *STR, char **TAIL);
float strtof(const char *STR, char **TAIL);
double _strtod_r(void *REENT,
const char *STR, char **TAIL);
*Description*
The function `strtod' parses the character string STR, producing a
substring which can be converted to a double value. The substring
converted is the longest initial subsequence of STR, beginning with the
first non-whitespace character, that has one of these formats:
[+|-]DIGITS[.[DIGITS]][(e|E)[+|-]DIGITS]
[+|-].DIGITS[(e|E)[+|-]DIGITS]
[+|-](i|I)(n|N)(f|F)[(i|I)(n|N)(i|I)(t|T)(y|Y)]
[+|-](n|N)(a|A)(n|N)[<(>[HEXDIGITS]<)>]
[+|-]0(x|X)HEXDIGITS[.[HEXDIGITS]][(p|P)[+|-]DIGITS]
[+|-]0(x|X).HEXDIGITS[(p|P)[+|-]DIGITS]
The substring contains no characters if STR is empty, consists
entirely of whitespace, or if the first non-whitespace character is
something other than `+', `-', `.', or a digit, and cannot be parsed as
infinity or NaN. If the platform does not support NaN, then NaN is
treated as an empty substring. If the substring is empty, no
conversion is done, and the value of STR is stored in `*TAIL'.
Otherwise, the substring is converted, and a pointer to the final string
(which will contain at least the terminating null character of STR) is
stored in `*TAIL'. If you want no assignment to `*TAIL', pass a null
pointer as TAIL. `strtof' is identical to `strtod' except for its
return type.
This implementation returns the nearest machine number to the input
decimal string. Ties are broken by using the IEEE round-even rule.
However, `strtof' is currently subject to double rounding errors.
The alternate function `_strtod_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`strtod' returns the converted substring value, if any. If no
conversion could be performed, 0 is returned. If the correct value is
out of the range of representable values, plus or minus `HUGE_VAL' is
returned, and `ERANGE' is stored in errno. If the correct value would
cause underflow, 0 is returned and `ERANGE' is stored in errno.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: strtol, Next: strtoll, Prev: strtod, Up: Stdlib
2.35 `strtol'--string to long
=============================
*Synopsis*
#include <stdlib.h>
long strtol(const char *S, char **PTR,int BASE);
long _strtol_r(void *REENT,
const char *S, char **PTR,int BASE);
*Description*
The function `strtol' converts the string `*S' to a `long'. First, it
breaks down the string into three parts: leading whitespace, which is
ignored; a subject string consisting of characters resembling an
integer in the radix specified by BASE; and a trailing portion
consisting of zero or more unparseable characters, and always including
the terminating null character. Then, it attempts to convert the
subject string into a `long' and returns the result.
If the value of BASE is 0, the subject string is expected to look
like a normal C integer constant: an optional sign, a possible ``0x''
indicating a hexadecimal base, and a number. If BASE is between 2 and
36, the expected form of the subject is a sequence of letters and
digits representing an integer in the radix specified by BASE, with an
optional plus or minus sign. The letters `a'-`z' (or, equivalently,
`A'-`Z') are used to signify values from 10 to 35; only letters whose
ascribed values are less than BASE are permitted. If BASE is 16, a
leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible letter or digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`strtol' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading 0 and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. If the subject string begins with
a minus sign, the value is negated. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (or not in acceptable form), no
conversion is performed and the value of S is stored in PTR (if PTR is
not `NULL').
The alternate function `_strtol_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`strtol' returns the converted value, if any. If no conversion was
made, 0 is returned.
`strtol' returns `LONG_MAX' or `LONG_MIN' if the magnitude of the
converted value is too large, and sets `errno' to `ERANGE'.
*Portability*
`strtol' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: strtoll, Next: strtoul, Prev: strtol, Up: Stdlib
2.36 `strtoll'--string to long long
===================================
*Synopsis*
#include <stdlib.h>
long long strtoll(const char *S, char **PTR,int BASE);
long long _strtoll_r(void *REENT,
const char *S, char **PTR,int BASE);
*Description*
The function `strtoll' converts the string `*S' to a `long long'.
First, it breaks down the string into three parts: leading whitespace,
which is ignored; a subject string consisting of characters resembling
an integer in the radix specified by BASE; and a trailing portion
consisting of zero or more unparseable characters, and always including
the terminating null character. Then, it attempts to convert the
subject string into a `long long' and returns the result.
If the value of BASE is 0, the subject string is expected to look
like a normal C integer constant: an optional sign, a possible ``0x''
indicating a hexadecimal base, and a number. If BASE is between 2 and
36, the expected form of the subject is a sequence of letters and
digits representing an integer in the radix specified by BASE, with an
optional plus or minus sign. The letters `a'-`z' (or, equivalently,
`A'-`Z') are used to signify values from 10 to 35; only letters whose
ascribed values are less than BASE are permitted. If BASE is 16, a
leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible letter or digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`strtoll' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading 0 and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. If the subject string begins with
a minus sign, the value is negated. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (or not in acceptable form), no
conversion is performed and the value of S is stored in PTR (if PTR is
not `NULL').
The alternate function `_strtoll_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`strtoll' returns the converted value, if any. If no conversion was
made, 0 is returned.
`strtoll' returns `LONG_LONG_MAX' or `LONG_LONG_MIN' if the
magnitude of the converted value is too large, and sets `errno' to
`ERANGE'.
*Portability*
`strtoll' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: strtoul, Next: strtoull, Prev: strtoll, Up: Stdlib
2.37 `strtoul'--string to unsigned long
=======================================
*Synopsis*
#include <stdlib.h>
unsigned long strtoul(const char *S, char **PTR,
int BASE);
unsigned long _strtoul_r(void *REENT, const char *S,
char **PTR, int BASE);
*Description*
The function `strtoul' converts the string `*S' to an `unsigned long'.
First, it breaks down the string into three parts: leading whitespace,
which is ignored; a subject string consisting of the digits meaningful
in the radix specified by BASE (for example, `0' through `7' if the
value of BASE is 8); and a trailing portion consisting of one or more
unparseable characters, which always includes the terminating null
character. Then, it attempts to convert the subject string into an
unsigned long integer, and returns the result.
If the value of BASE is zero, the subject string is expected to look
like a normal C integer constant (save that no optional sign is
permitted): a possible `0x' indicating hexadecimal radix, and a number.
If BASE is between 2 and 36, the expected form of the subject is a
sequence of digits (which may include letters, depending on the base)
representing an integer in the radix specified by BASE. The letters
`a'-`z' (or `A'-`Z') are used as digits valued from 10 to 35. If BASE
is 16, a leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`strtoul' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading `0' and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (that is, if `*'S does not start with
a substring in acceptable form), no conversion is performed and the
value of S is stored in PTR (if PTR is not `NULL').
The alternate function `_strtoul_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`strtoul' returns the converted value, if any. If no conversion was
made, `0' is returned.
`strtoul' returns `ULONG_MAX' if the magnitude of the converted
value is too large, and sets `errno' to `ERANGE'.
*Portability*
`strtoul' is ANSI.
`strtoul' requires no supporting OS subroutines.

File: libc.info, Node: strtoull, Next: wcsrtombs, Prev: strtoul, Up: Stdlib
2.38 `strtoull'--string to unsigned long long
=============================================
*Synopsis*
#include <stdlib.h>
unsigned long long strtoull(const char *S, char **PTR,
int BASE);
unsigned long long _strtoull_r(void *REENT, const char *S,
char **PTR, int BASE);
*Description*
The function `strtoull' converts the string `*S' to an `unsigned long
long'. First, it breaks down the string into three parts: leading
whitespace, which is ignored; a subject string consisting of the digits
meaningful in the radix specified by BASE (for example, `0' through `7'
if the value of BASE is 8); and a trailing portion consisting of one or
more unparseable characters, which always includes the terminating null
character. Then, it attempts to convert the subject string into an
unsigned long long integer, and returns the result.
If the value of BASE is zero, the subject string is expected to look
like a normal C integer constant (save that no optional sign is
permitted): a possible `0x' indicating hexadecimal radix, and a number.
If BASE is between 2 and 36, the expected form of the subject is a
sequence of digits (which may include letters, depending on the base)
representing an integer in the radix specified by BASE. The letters
`a'-`z' (or `A'-`Z') are used as digits valued from 10 to 35. If BASE
is 16, a leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`strtoull' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading `0' and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (that is, if `*'S does not start with
a substring in acceptable form), no conversion is performed and the
value of S is stored in PTR (if PTR is not `NULL').
The alternate function `_strtoull_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`strtoull' returns the converted value, if any. If no conversion was
made, `0' is returned.
`strtoull' returns `ULONG_LONG_MAX' if the magnitude of the converted
value is too large, and sets `errno' to `ERANGE'.
*Portability*
`strtoull' is ANSI.
`strtoull' requires no supporting OS subroutines.

File: libc.info, Node: wcsrtombs, Next: wcstod, Prev: strtoull, Up: Stdlib
2.39 `wcsrtombs', `wcsnrtombs'--convert a wide-character string to a character string
=====================================================================================
*Synopsis*
#include <wchar.h>
size_t wcsrtombs(char *DST, const wchar_t **SRC, size_t LEN,
mbstate_t *PS);
#include <wchar.h>
size_t _wcsrtombs_r(struct _reent *PTR, char *DST,
const wchar_t **SRC, size_t LEN,
mbstate_t *PS);
#include <wchar.h>
size_t wcsnrtombs(char *DST, const wchar_t **SRC,
size_t NWC, size_t LEN, mbstate_t *PS);
#include <wchar.h>
size_t _wcsnrtombs_r(struct _reent *PTR, char *DST,
const wchar_t **SRC, size_t NWC,
size_t LEN, mbstate_t *PS);
*Description*
The `wcsrtombs' function converts a string of wide characters indirectly
pointed to by SRC to a corresponding multibyte character string stored
in the array pointed to by DST>. No more than LEN bytes are written to
DST>.
If DST> is NULL, no characters are stored.
If DST> is not NULL, the pointer pointed to by SRC is updated to
point to the character after the one that conversion stopped at. If
conversion stops because a null character is encountered, *SRC is set
to NULL.
The mbstate_t argument, PS, is used to keep track of the shift
state. If it is NULL, `wcsrtombs' uses an internal, static mbstate_t
object, which is initialized to the initial conversion state at program
startup.
The `wcsnrtombs' function behaves identically to `wcsrtombs', except
that conversion stops after reading at most NWC characters from the
buffer pointed to by SRC.
*Returns*
The `wcsrtombs' and `wcsnrtombs' functions return the number of bytes
stored in the array pointed to by DST (not including any terminating
null), if successful, otherwise it returns (size_t)-1.
*Portability*
`wcsrtombs' is defined by C99 standard. `wcsnrtombs' is defined by the
POSIX.1-2008 standard.

File: libc.info, Node: wcstod, Next: wcstol, Prev: wcsrtombs, Up: Stdlib
2.40 `wcstod', `wcstof'--wide char string to double or float
============================================================
*Synopsis*
#include <stdlib.h>
double wcstod(const wchar_t *STR, wchar_t **TAIL);
float wcstof(const wchar_t *STR, wchar_t **TAIL);
double _wcstod_r(void *REENT,
const wchar_t *STR, wchar_t **TAIL);
float _wcstof_r(void *REENT,
const wchar_t *STR, wchar_t **TAIL);
*Description*
The function `wcstod' parses the wide character string STR, producing a
substring which can be converted to a double value. The substring
converted is the longest initial subsequence of STR, beginning with the
first non-whitespace character, that has one of these formats:
[+|-]DIGITS[.[DIGITS]][(e|E)[+|-]DIGITS]
[+|-].DIGITS[(e|E)[+|-]DIGITS]
[+|-](i|I)(n|N)(f|F)[(i|I)(n|N)(i|I)(t|T)(y|Y)]
[+|-](n|N)(a|A)(n|N)[<(>[HEXDIGITS]<)>]
[+|-]0(x|X)HEXDIGITS[.[HEXDIGITS]][(p|P)[+|-]DIGITS]
[+|-]0(x|X).HEXDIGITS[(p|P)[+|-]DIGITS]
The substring contains no characters if STR is empty, consists
entirely of whitespace, or if the first non-whitespace character is
something other than `+', `-', `.', or a digit, and cannot be parsed as
infinity or NaN. If the platform does not support NaN, then NaN is
treated as an empty substring. If the substring is empty, no
conversion is done, and the value of STR is stored in `*TAIL'.
Otherwise, the substring is converted, and a pointer to the final string
(which will contain at least the terminating null character of STR) is
stored in `*TAIL'. If you want no assignment to `*TAIL', pass a null
pointer as TAIL. `wcstof' is identical to `wcstod' except for its
return type.
This implementation returns the nearest machine number to the input
decimal string. Ties are broken by using the IEEE round-even rule.
However, `wcstof' is currently subject to double rounding errors.
The alternate functions `_wcstod_r' and `_wcstof_r' are reentrant
versions of `wcstod' and `wcstof', respectively. The extra argument
REENT is a pointer to a reentrancy structure.
*Returns*
Return the converted substring value, if any. If no conversion could
be performed, 0 is returned. If the correct value is out of the range
of representable values, plus or minus `HUGE_VAL' is returned, and
`ERANGE' is stored in errno. If the correct value would cause
underflow, 0 is returned and `ERANGE' is stored in errno.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: wcstol, Next: wcstoll, Prev: wcstod, Up: Stdlib
2.41 `wcstol'--wide string to long
==================================
*Synopsis*
#include <wchar.h>
long wcstol(const wchar_t *S, wchar_t **PTR,int BASE);
long _wcstol_r(void *REENT,
const wchar_t *S, wchar_t **PTR,int BASE);
*Description*
The function `wcstol' converts the wide string `*S' to a `long'. First,
it breaks down the string into three parts: leading whitespace, which
is ignored; a subject string consisting of characters resembling an
integer in the radix specified by BASE; and a trailing portion
consisting of zero or more unparseable characters, and always including
the terminating null character. Then, it attempts to convert the
subject string into a `long' and returns the result.
If the value of BASE is 0, the subject string is expected to look
like a normal C integer constant: an optional sign, a possible ``0x''
indicating a hexadecimal base, and a number. If BASE is between 2 and
36, the expected form of the subject is a sequence of letters and
digits representing an integer in the radix specified by BASE, with an
optional plus or minus sign. The letters `a'-`z' (or, equivalently,
`A'-`Z') are used to signify values from 10 to 35; only letters whose
ascribed values are less than BASE are permitted. If BASE is 16, a
leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible letter or digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`wcstol' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading 0 and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. If the subject string begins with
a minus sign, the value is negated. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (or not in acceptable form), no
conversion is performed and the value of S is stored in PTR (if PTR is
not `NULL').
The alternate function `_wcstol_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`wcstol' returns the converted value, if any. If no conversion was
made, 0 is returned.
`wcstol' returns `LONG_MAX' or `LONG_MIN' if the magnitude of the
converted value is too large, and sets `errno' to `ERANGE'.
*Portability*
`wcstol' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: wcstoll, Next: wcstoul, Prev: wcstol, Up: Stdlib
2.42 `wcstoll'--wide string to long long
========================================
*Synopsis*
#include <wchar.h>
long long wcstoll(const wchar_t *S, wchar_t **PTR,int BASE);
long long _wcstoll_r(void *REENT,
const wchar_t *S, wchar_t **PTR,int BASE);
*Description*
The function `wcstoll' converts the wide string `*S' to a `long long'.
First, it breaks down the string into three parts: leading whitespace,
which is ignored; a subject string consisting of characters resembling
an integer in the radix specified by BASE; and a trailing portion
consisting of zero or more unparseable characters, and always including
the terminating null character. Then, it attempts to convert the
subject string into a `long long' and returns the result.
If the value of BASE is 0, the subject string is expected to look
like a normal C integer constant: an optional sign, a possible ``0x''
indicating a hexadecimal base, and a number. If BASE is between 2 and
36, the expected form of the subject is a sequence of letters and
digits representing an integer in the radix specified by BASE, with an
optional plus or minus sign. The letters `a'-`z' (or, equivalently,
`A'-`Z') are used to signify values from 10 to 35; only letters whose
ascribed values are less than BASE are permitted. If BASE is 16, a
leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible letter or digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`wcstoll' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading 0 and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. If the subject string begins with
a minus sign, the value is negated. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (or not in acceptable form), no
conversion is performed and the value of S is stored in PTR (if PTR is
not `NULL').
The alternate function `_wcstoll_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`wcstoll' returns the converted value, if any. If no conversion was
made, 0 is returned.
`wcstoll' returns `LONG_LONG_MAX' or `LONG_LONG_MIN' if the
magnitude of the converted value is too large, and sets `errno' to
`ERANGE'.
*Portability*
`wcstoll' is ANSI.
No supporting OS subroutines are required.

File: libc.info, Node: wcstoul, Next: wcstoull, Prev: wcstoll, Up: Stdlib
2.43 `wcstoul'--wide string to unsigned long
============================================
*Synopsis*
#include <wchar.h>
unsigned long wcstoul(const wchar_t *S, wchar_t **PTR,
int BASE);
unsigned long _wcstoul_r(void *REENT, const wchar_t *S,
wchar_t **PTR, int BASE);
*Description*
The function `wcstoul' converts the wide string `*S' to an `unsigned
long'. First, it breaks down the string into three parts: leading
whitespace, which is ignored; a subject string consisting of the digits
meaningful in the radix specified by BASE (for example, `0' through `7'
if the value of BASE is 8); and a trailing portion consisting of one or
more unparseable characters, which always includes the terminating null
character. Then, it attempts to convert the subject string into an
unsigned long integer, and returns the result.
If the value of BASE is zero, the subject string is expected to look
like a normal C integer constant (save that no optional sign is
permitted): a possible `0x' indicating hexadecimal radix, and a number.
If BASE is between 2 and 36, the expected form of the subject is a
sequence of digits (which may include letters, depending on the base)
representing an integer in the radix specified by BASE. The letters
`a'-`z' (or `A'-`Z') are used as digits valued from 10 to 35. If BASE
is 16, a leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`wcstoul' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading `0' and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (that is, if `*'S does not start with
a substring in acceptable form), no conversion is performed and the
value of S is stored in PTR (if PTR is not `NULL').
The alternate function `_wcstoul_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`wcstoul' returns the converted value, if any. If no conversion was
made, `0' is returned.
`wcstoul' returns `ULONG_MAX' if the magnitude of the converted
value is too large, and sets `errno' to `ERANGE'.
*Portability*
`wcstoul' is ANSI.
`wcstoul' requires no supporting OS subroutines.

File: libc.info, Node: wcstoull, Next: system, Prev: wcstoul, Up: Stdlib
2.44 `wcstoull'--wide string to unsigned long long
==================================================
*Synopsis*
#include <wchar.h>
unsigned long long wcstoull(const wchar_t *S, wchar_t **PTR,
int BASE);
unsigned long long _wcstoull_r(void *REENT, const wchar_t *S,
wchar_t **PTR, int BASE);
*Description*
The function `wcstoull' converts the wide string `*S' to an `unsigned
long long'. First, it breaks down the string into three parts: leading
whitespace, which is ignored; a subject string consisting of the digits
meaningful in the radix specified by BASE (for example, `0' through `7'
if the value of BASE is 8); and a trailing portion consisting of one or
more unparseable characters, which always includes the terminating null
character. Then, it attempts to convert the subject string into an
unsigned long long integer, and returns the result.
If the value of BASE is zero, the subject string is expected to look
like a normal C integer constant: an optional sign (`+' or `-'), a
possible `0x' indicating hexadecimal radix or a possible <0> indicating
octal radix, and a number. If BASE is between 2 and 36, the expected
form of the subject is a sequence of digits (which may include letters,
depending on the base) representing an integer in the radix specified
by BASE. The letters `a'-`z' (or `A'-`Z') are used as digits valued
from 10 to 35. If BASE is 16, a leading `0x' is permitted.
The subject sequence is the longest initial sequence of the input
string that has the expected form, starting with the first
non-whitespace character. If the string is empty or consists entirely
of whitespace, or if the first non-whitespace character is not a
permissible digit, the subject string is empty.
If the subject string is acceptable, and the value of BASE is zero,
`wcstoull' attempts to determine the radix from the input string. A
string with a leading `0x' is treated as a hexadecimal value; a string
with a leading `0' and no `x' is treated as octal; all other strings are
treated as decimal. If BASE is between 2 and 36, it is used as the
conversion radix, as described above. Finally, a pointer to the first
character past the converted subject string is stored in PTR, if PTR is
not `NULL'.
If the subject string is empty (that is, if `*'S does not start with
a substring in acceptable form), no conversion is performed and the
value of S is stored in PTR (if PTR is not `NULL').
The alternate function `_wcstoull_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`wcstoull' returns `0' and sets `errno' to `EINVAL' if the value of
BASE is not supported.
`wcstoull' returns the converted value, if any. If no conversion was
made, `0' is returned.
`wcstoull' returns `ULLONG_MAX' if the magnitude of the converted
value is too large, and sets `errno' to `ERANGE'.
*Portability*
`wcstoull' is ANSI.
`wcstoull' requires no supporting OS subroutines.

File: libc.info, Node: system, Next: wcstombs, Prev: wcstoull, Up: Stdlib
2.45 `system'--execute command string
=====================================
*Synopsis*
#include <stdlib.h>
int system(char *S);
int _system_r(void *REENT, char *S);
*Description*
Use `system' to pass a command string `*S' to `/bin/sh' on your system,
and wait for it to finish executing.
Use "`system(NULL)'" to test whether your system has `/bin/sh'
available.
The alternate function `_system_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`system(NULL)' returns a non-zero value if `/bin/sh' is available, and
`0' if it is not.
With a command argument, the result of `system' is the exit status
returned by `/bin/sh'.
*Portability*
ANSI C requires `system', but leaves the nature and effects of a
command processor undefined. ANSI C does, however, specify that
`system(NULL)' return zero or nonzero to report on the existence of a
command processor.
POSIX.2 requires `system', and requires that it invoke a `sh'.
Where `sh' is found is left unspecified.
Supporting OS subroutines required: `_exit', `_execve', `_fork_r',
`_wait_r'.

File: libc.info, Node: wcstombs, Next: wctomb, Prev: system, Up: Stdlib
2.46 `wcstombs'--minimal wide char string to multibyte string converter
=======================================================================
*Synopsis*
#include <stdlib.h>
size_t wcstombs(char *S, const wchar_t *PWC, size_t N);
*Description*
When _MB_CAPABLE is not defined, this is a minimal ANSI-conforming
implementation of `wcstombs'. In this case, all wide-characters are
expected to represent single bytes and so are converted simply by
casting to char.
When _MB_CAPABLE is defined, this routine calls `_wcstombs_r' to
perform the conversion, passing a state variable to allow state
dependent decoding. The result is based on the locale setting which may
be restricted to a defined set of locales.
*Returns*
This implementation of `wcstombs' returns `0' if S is `NULL' or is the
empty string; it returns `-1' if _MB_CAPABLE and one of the wide-char
characters does not represent a valid multi-byte character; otherwise
it returns the minimum of: `n' or the number of bytes that are
transferred to `s', not including the nul terminator.
If the return value is -1, the state of the `pwc' string is
indeterminate. If the input has a length of 0, the output string will
be modified to contain a wchar_t nul terminator if `n' > 0.
*Portability*
`wcstombs' is required in the ANSI C standard. However, the precise
effects vary with the locale.
`wcstombs' requires no supporting OS subroutines.

File: libc.info, Node: wctomb, Prev: wcstombs, Up: Stdlib
2.47 `wctomb'--minimal wide char to multibyte converter
=======================================================
*Synopsis*
#include <stdlib.h>
int wctomb(char *S, wchar_t WCHAR);
*Description*
When _MB_CAPABLE is not defined, this is a minimal ANSI-conforming
implementation of `wctomb'. The only "wide characters" recognized are
single bytes, and they are "converted" to themselves.
When _MB_CAPABLE is defined, this routine calls `_wctomb_r' to
perform the conversion, passing a state variable to allow state
dependent decoding. The result is based on the locale setting which may
be restricted to a defined set of locales.
Each call to `wctomb' modifies `*S' unless S is a null pointer or
_MB_CAPABLE is defined and WCHAR is invalid.
*Returns*
This implementation of `wctomb' returns `0' if S is `NULL'; it returns
`-1' if _MB_CAPABLE is enabled and the wchar is not a valid multi-byte
character, it returns `1' if _MB_CAPABLE is not defined or the wchar is
in reality a single byte character, otherwise it returns the number of
bytes in the multi-byte character.
*Portability*
`wctomb' is required in the ANSI C standard. However, the precise
effects vary with the locale.
`wctomb' requires no supporting OS subroutines.

File: libc.info, Node: Ctype, Next: Stdio, Prev: Stdlib, Up: Top
3 Character Type Macros and Functions (`ctype.h')
*************************************************
This chapter groups macros (which are also available as subroutines) to
classify characters into several categories (alphabetic, numeric,
control characters, whitespace, and so on), or to perform simple
character mappings.
The header file `ctype.h' defines the macros.
* Menu:
* isalnum:: Alphanumeric character predicate
* isalpha:: Alphabetic character predicate
* isascii:: ASCII character predicate
* iscntrl:: Control character predicate
* isdigit:: Decimal digit predicate
* islower:: Lowercase character predicate
* isprint:: Printable character predicates (isprint, isgraph)
* ispunct:: Punctuation character predicate
* isspace:: Whitespace character predicate
* isupper:: Uppercase character predicate
* isxdigit:: Hexadecimal digit predicate
* toascii:: Force integers to ASCII range
* tolower:: Translate characters to lowercase
* toupper:: Translate characters to uppercase
* iswalnum:: Alphanumeric wide character predicate
* iswalpha:: Alphabetic wide character predicate
* iswblank:: Blank wide character predicate
* iswcntrl:: Control wide character predicate
* iswdigit:: Decimal digit wide character predicate
* iswgraph:: Graphic wide character predicate
* iswlower:: Lowercase wide character predicate
* iswprint:: Printable wide character predicate
* iswpunct:: Punctuation wide character predicate
* iswspace:: Whitespace wide character predicate
* iswupper:: Uppercase wide character predicate
* iswxdigit:: Hexadecimal digit wide character predicate
* iswctype:: Extensible wide-character test
* wctype:: Compute wide-character test type
* towlower:: Translate wide characters to lowercase
* towupper:: Translate wide characters to uppercase
* towctrans:: Extensible wide-character translation
* wctrans:: Compute wide-character translation type

File: libc.info, Node: isalnum, Next: isalpha, Up: Ctype
3.1 `isalnum'--alphanumeric character predicate
===============================================
*Synopsis*
#include <ctype.h>
int isalnum(int C);
*Description*
`isalnum' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for alphabetic or numeric
ASCII characters, and `0' for other arguments. It is defined for all
integer values.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isalnum''.
*Returns*
`isalnum' returns non-zero if C is a letter (`a'-`z' or `A'-`Z') or a
digit (`0'-`9').
*Portability*
`isalnum' is ANSI C.
No OS subroutines are required.

File: libc.info, Node: isalpha, Next: isascii, Prev: isalnum, Up: Ctype
3.2 `isalpha'--alphabetic character predicate
=============================================
*Synopsis*
#include <ctype.h>
int isalpha(int C);
*Description*
`isalpha' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero when C represents an
alphabetic ASCII character, and 0 otherwise. It is defined only when
`isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isalpha''.
*Returns*
`isalpha' returns non-zero if C is a letter (`A'-`Z' or `a'-`z').
*Portability*
`isalpha' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isascii, Next: iscntrl, Prev: isalpha, Up: Ctype
3.3 `isascii'--ASCII character predicate
========================================
*Synopsis*
#include <ctype.h>
int isascii(int C);
*Description*
`isascii' is a macro which returns non-zero when C is an ASCII
character, and 0 otherwise. It is defined for all integer values.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isascii''.
*Returns*
`isascii' returns non-zero if the low order byte of C is in the range 0
to 127 (`0x00'-`0x7F').
*Portability*
`isascii' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: iscntrl, Next: isdigit, Prev: isascii, Up: Ctype
3.4 `iscntrl'--control character predicate
==========================================
*Synopsis*
#include <ctype.h>
int iscntrl(int C);
*Description*
`iscntrl' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for control characters,
and 0 for other characters. It is defined only when `isascii'(C) is
true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef iscntrl''.
*Returns*
`iscntrl' returns non-zero if C is a delete character or ordinary
control character (`0x7F' or `0x00'-`0x1F').
*Portability*
`iscntrl' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isdigit, Next: islower, Prev: iscntrl, Up: Ctype
3.5 `isdigit'--decimal digit predicate
======================================
*Synopsis*
#include <ctype.h>
int isdigit(int C);
*Description*
`isdigit' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for decimal digits, and 0
for other characters. It is defined only when `isascii'(C) is true or
C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isdigit''.
*Returns*
`isdigit' returns non-zero if C is a decimal digit (`0'-`9').
*Portability*
`isdigit' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: islower, Next: isprint, Prev: isdigit, Up: Ctype
3.6 `islower'--lowercase character predicate
============================================
*Synopsis*
#include <ctype.h>
int islower(int C);
*Description*
`islower' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for minuscules (lowercase
alphabetic characters), and 0 for other characters. It is defined only
when `isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef islower''.
*Returns*
`islower' returns non-zero if C is a lowercase letter (`a'-`z').
*Portability*
`islower' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isprint, Next: ispunct, Prev: islower, Up: Ctype
3.7 `isprint', `isgraph'--printable character predicates
========================================================
*Synopsis*
#include <ctype.h>
int isprint(int C);
int isgraph(int C);
*Description*
`isprint' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for printable characters,
and 0 for other character arguments. It is defined only when
`isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining either macro using ``#undef isprint'' or ``#undef isgraph''.
*Returns*
`isprint' returns non-zero if C is a printing character,
(`0x20'-`0x7E'). `isgraph' behaves identically to `isprint', except
that the space character (`0x20') is excluded.
*Portability*
`isprint' and `isgraph' are ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: ispunct, Next: isspace, Prev: isprint, Up: Ctype
3.8 `ispunct'--punctuation character predicate
==============================================
*Synopsis*
#include <ctype.h>
int ispunct(int C);
*Description*
`ispunct' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for printable punctuation
characters, and 0 for other characters. It is defined only when
`isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef ispunct''.
*Returns*
`ispunct' returns non-zero if C is a printable punctuation character
(`isgraph(C) && !isalnum(C)').
*Portability*
`ispunct' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isspace, Next: isupper, Prev: ispunct, Up: Ctype
3.9 `isspace'--whitespace character predicate
=============================================
*Synopsis*
#include <ctype.h>
int isspace(int C);
*Description*
`isspace' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for whitespace
characters, and 0 for other characters. It is defined only when
`isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isspace''.
*Returns*
`isspace' returns non-zero if C is a space, tab, carriage return, new
line, vertical tab, or formfeed (`0x09'-`0x0D', `0x20').
*Portability*
`isspace' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isupper, Next: isxdigit, Prev: isspace, Up: Ctype
3.10 `isupper'--uppercase character predicate
=============================================
*Synopsis*
#include <ctype.h>
int isupper(int C);
*Description*
`isupper' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for uppercase letters
(`A'-`Z'), and 0 for other characters. It is defined only when
`isascii'(C) is true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isupper''.
*Returns*
`isupper' returns non-zero if C is a uppercase letter (A-Z).
*Portability*
`isupper' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: isxdigit, Next: toascii, Prev: isupper, Up: Ctype
3.11 `isxdigit'--hexadecimal digit predicate
============================================
*Synopsis*
#include <ctype.h>
int isxdigit(int C);
*Description*
`isxdigit' is a macro which classifies ASCII integer values by table
lookup. It is a predicate returning non-zero for hexadecimal digits,
and `0' for other characters. It is defined only when `isascii'(C) is
true or C is EOF.
You can use a compiled subroutine instead of the macro definition by
undefining the macro using ``#undef isxdigit''.
*Returns*
`isxdigit' returns non-zero if C is a hexadecimal digit (`0'-`9',
`a'-`f', or `A'-`F').
*Portability*
`isxdigit' is ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: toascii, Next: tolower, Prev: isxdigit, Up: Ctype
3.12 `toascii'--force integers to ASCII range
=============================================
*Synopsis*
#include <ctype.h>
int toascii(int C);
*Description*
`toascii' is a macro which coerces integers to the ASCII range (0-127)
by zeroing any higher-order bits.
You can use a compiled subroutine instead of the macro definition by
undefining this macro using ``#undef toascii''.
*Returns*
`toascii' returns integers between 0 and 127.
*Portability*
`toascii' is not ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: tolower, Next: toupper, Prev: toascii, Up: Ctype
3.13 `tolower'--translate characters to lowercase
=================================================
*Synopsis*
#include <ctype.h>
int tolower(int C);
int _tolower(int C);
*Description*
`tolower' is a macro which converts uppercase characters to lowercase,
leaving all other characters unchanged. It is only defined when C is
an integer in the range `EOF' to `255'.
You can use a compiled subroutine instead of the macro definition by
undefining this macro using ``#undef tolower''.
`_tolower' performs the same conversion as `tolower', but should
only be used when C is known to be an uppercase character (`A'-`Z').
*Returns*
`tolower' returns the lowercase equivalent of C when it is a character
between `A' and `Z', and C otherwise.
`_tolower' returns the lowercase equivalent of C when it is a
character between `A' and `Z'. If C is not one of these characters,
the behaviour of `_tolower' is undefined.
*Portability*
`tolower' is ANSI C. `_tolower' is not recommended for portable
programs.
No supporting OS subroutines are required.

File: libc.info, Node: toupper, Next: iswalnum, Prev: tolower, Up: Ctype
3.14 `toupper'--translate characters to uppercase
=================================================
*Synopsis*
#include <ctype.h>
int toupper(int C);
int _toupper(int C);
*Description*
`toupper' is a macro which converts lowercase characters to uppercase,
leaving all other characters unchanged. It is only defined when C is
an integer in the range `EOF' to `255'.
You can use a compiled subroutine instead of the macro definition by
undefining this macro using ``#undef toupper''.
`_toupper' performs the same conversion as `toupper', but should
only be used when C is known to be a lowercase character (`a'-`z').
*Returns*
`toupper' returns the uppercase equivalent of C when it is a character
between `a' and `z', and C otherwise.
`_toupper' returns the uppercase equivalent of C when it is a
character between `a' and `z'. If C is not one of these characters,
the behaviour of `_toupper' is undefined.
*Portability*
`toupper' is ANSI C. `_toupper' is not recommended for portable
programs.
No supporting OS subroutines are required.

File: libc.info, Node: iswalnum, Next: iswalpha, Prev: toupper, Up: Ctype
3.15 `iswalnum'--alphanumeric wide character test
=================================================
*Synopsis*
#include <wctype.h>
int iswalnum(wint_t C);
*Description*
`iswalnum' is a function which classifies wide-character values that
are alphanumeric.
*Returns*
`iswalnum' returns non-zero if C is a alphanumeric wide character.
*Portability*
`iswalnum' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswalpha, Next: iswblank, Prev: iswalnum, Up: Ctype
3.16 `iswalpha'--alphabetic wide character test
===============================================
*Synopsis*
#include <wctype.h>
int iswalpha(wint_t C);
*Description*
`iswalpha' is a function which classifies wide-character values that
are alphabetic.
*Returns*
`iswalpha' returns non-zero if C is an alphabetic wide character.
*Portability*
`iswalpha' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswcntrl, Next: iswdigit, Prev: iswblank, Up: Ctype
3.17 `iswcntrl'--control wide character test
============================================
*Synopsis*
#include <wctype.h>
int iswcntrl(wint_t C);
*Description*
`iswcntrl' is a function which classifies wide-character values that
are categorized as control characters.
*Returns*
`iswcntrl' returns non-zero if C is a control wide character.
*Portability*
`iswcntrl' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswblank, Next: iswcntrl, Prev: iswalpha, Up: Ctype
3.18 `iswblank'--blank wide character test
==========================================
*Synopsis*
#include <wctype.h>
int iswblank(wint_t C);
*Description*
`iswblank' is a function which classifies wide-character values that
are categorized as blank.
*Returns*
`iswblank' returns non-zero if C is a blank wide character.
*Portability*
`iswblank' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswdigit, Next: iswgraph, Prev: iswcntrl, Up: Ctype
3.19 `iswdigit'--decimal digit wide character test
==================================================
*Synopsis*
#include <wctype.h>
int iswdigit(wint_t C);
*Description*
`iswdigit' is a function which classifies wide-character values that
are decimal digits.
*Returns*
`iswdigit' returns non-zero if C is a decimal digit wide character.
*Portability*
`iswdigit' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswgraph, Next: iswlower, Prev: iswdigit, Up: Ctype
3.20 `iswgraph'--graphic wide character test
============================================
*Synopsis*
#include <wctype.h>
int iswgraph(wint_t C);
*Description*
`iswgraph' is a function which classifies wide-character values that
are graphic.
*Returns*
`iswgraph' returns non-zero if C is a graphic wide character.
*Portability*
`iswgraph' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswlower, Next: iswprint, Prev: iswgraph, Up: Ctype
3.21 `iswlower'--lowercase wide character test
==============================================
*Synopsis*
#include <wctype.h>
int iswlower(wint_t C);
*Description*
`iswlower' is a function which classifies wide-character values that
have uppercase translations.
*Returns*
`iswlower' returns non-zero if C is a lowercase wide character.
*Portability*
`iswlower' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswprint, Next: iswpunct, Prev: iswlower, Up: Ctype
3.22 `iswprint'--printable wide character test
==============================================
*Synopsis*
#include <wctype.h>
int iswprint(wint_t C);
*Description*
`iswprint' is a function which classifies wide-character values that
are printable.
*Returns*
`iswprint' returns non-zero if C is a printable wide character.
*Portability*
`iswprint' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswpunct, Next: iswspace, Prev: iswprint, Up: Ctype
3.23 `iswpunct'--punctuation wide character test
================================================
*Synopsis*
#include <wctype.h>
int iswpunct(wint_t C);
*Description*
`iswpunct' is a function which classifies wide-character values that
are punctuation.
*Returns*
`iswpunct' returns non-zero if C is a punctuation wide character.
*Portability*
`iswpunct' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswspace, Next: iswupper, Prev: iswpunct, Up: Ctype
3.24 `iswspace'--whitespace wide character test
===============================================
*Synopsis*
#include <wctype.h>
int iswspace(wint_t C);
*Description*
`iswspace' is a function which classifies wide-character values that
are categorized as whitespace.
*Returns*
`iswspace' returns non-zero if C is a whitespace wide character.
*Portability*
`iswspace' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswupper, Next: iswxdigit, Prev: iswspace, Up: Ctype
3.25 `iswupper'--uppercase wide character test
==============================================
*Synopsis*
#include <wctype.h>
int iswupper(wint_t C);
*Description*
`iswupper' is a function which classifies wide-character values that
have uppercase translations.
*Returns*
`iswupper' returns non-zero if C is a uppercase wide character.
*Portability*
`iswupper' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswxdigit, Next: iswctype, Prev: iswupper, Up: Ctype
3.26 `iswxdigit'--hexadecimal digit wide character test
=======================================================
*Synopsis*
#include <wctype.h>
int iswxdigit(wint_t C);
*Description*
`iswxdigit' is a function which classifies wide character values that
are hexadecimal digits.
*Returns*
`iswxdigit' returns non-zero if C is a hexadecimal digit wide character.
*Portability*
`iswxdigit' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: iswctype, Next: wctype, Prev: iswxdigit, Up: Ctype
3.27 `iswctype'--extensible wide-character test
===============================================
*Synopsis*
#include <wctype.h>
int iswctype(wint_t C, wctype_t DESC);
*Description*
`iswctype' is a function which classifies wide-character values using
the wide-character test specified by DESC.
*Returns*
`iswctype' returns non-zero if and only if C matches the test specified
by DESC. If DESC is unknown, zero is returned.
*Portability*
`iswctype' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: wctype, Next: towlower, Prev: iswctype, Up: Ctype
3.28 `wctype'--get wide-character classification type
=====================================================
*Synopsis*
#include <wctype.h>
wctype_t wctype(const char *C);
*Description*
`wctype' is a function which takes a string C and gives back the
appropriate wctype_t type value associated with the string, if one
exists. The following values are guaranteed to be recognized: "alnum",
"alpha", "blank", "cntrl", "digit", "graph", "lower", "print", "punct",
"space", "upper", and "xdigit".
*Returns*
`wctype' returns 0 and sets `errno' to `EINVAL' if the given name is
invalid. Otherwise, it returns a valid non-zero wctype_t value.
*Portability*
`wctype' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: towlower, Next: towupper, Prev: wctype, Up: Ctype
3.29 `towlower'--translate wide characters to lowercase
=======================================================
*Synopsis*
#include <wctype.h>
wint_t towlower(wint_t C);
*Description*
`towlower' is a function which converts uppercase wide characters to
lowercase, leaving all other characters unchanged.
*Returns*
`towlower' returns the lowercase equivalent of C when it is a uppercase
wide character; otherwise, it returns the input character.
*Portability*
`towlower' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: towupper, Next: towctrans, Prev: towlower, Up: Ctype
3.30 `towupper'--translate wide characters to uppercase
=======================================================
*Synopsis*
#include <wctype.h>
wint_t towupper(wint_t C);
*Description*
`towupper' is a function which converts lowercase wide characters to
uppercase, leaving all other characters unchanged.
*Returns*
`towupper' returns the uppercase equivalent of C when it is a lowercase
wide character, otherwise, it returns the input character.
*Portability*
`towupper' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: towctrans, Next: wctrans, Prev: towupper, Up: Ctype
3.31 `towctrans'--extensible wide-character translation
=======================================================
*Synopsis*
#include <wctype.h>
wint_t towctrans(wint_t C, wctrans_t W);
*Description*
`towctrans' is a function which converts wide characters based on a
specified translation type W. If the translation type is invalid or
cannot be applied to the current character, no change to the character
is made.
*Returns*
`towctrans' returns the translated equivalent of C when it is a valid
for the given translation, otherwise, it returns the input character.
When the translation type is invalid, `errno' is set `EINVAL'.
*Portability*
`towctrans' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: wctrans, Prev: towctrans, Up: Ctype
3.32 `wctrans'--get wide-character translation type
===================================================
*Synopsis*
#include <wctype.h>
wctrans_t wctrans(const char *C);
*Description*
`wctrans' is a function which takes a string C and gives back the
appropriate wctrans_t type value associated with the string, if one
exists. The following values are guaranteed to be recognized:
"tolower" and "toupper".
*Returns*
`wctrans' returns 0 and sets `errno' to `EINVAL' if the given name is
invalid. Otherwise, it returns a valid non-zero wctrans_t value.
*Portability*
`wctrans' is C99.
No supporting OS subroutines are required.

File: libc.info, Node: Stdio, Next: Strings, Prev: Ctype, Up: Top
4 Input and Output (`stdio.h')
******************************
This chapter comprises functions to manage files or other input/output
streams. Among these functions are subroutines to generate or scan
strings according to specifications from a format string.
The underlying facilities for input and output depend on the host
system, but these functions provide a uniform interface.
The corresponding declarations are in `stdio.h'.
The reentrant versions of these functions use macros
_stdin_r(REENT)
_stdout_r(REENT)
_stderr_r(REENT)
instead of the globals `stdin', `stdout', and `stderr'. The argument
<[reent]> is a pointer to a reentrancy structure.
* Menu:
* clearerr:: Clear file or stream error indicator
* diprintf:: Print to a file descriptor (integer only)
* dprintf:: Print to a file descriptor
* fclose:: Close a file
* fcloseall:: Close all files
* fdopen:: Turn an open file into a stream
* feof:: Test for end of file
* ferror:: Test whether read/write error has occurred
* fflush:: Flush buffered file output
* fgetc:: Get a character from a file or stream
* fgetpos:: Record position in a stream or file
* fgets:: Get character string from a file or stream
* fgetwc:: Get a wide character from a file or stream
* fgetws:: Get a wide character string from a file or stream
* fileno:: Get file descriptor associated with stream
* fmemopen:: Open a stream around a fixed-length buffer
* fopen:: Open a file
* fopencookie:: Open a stream with custom callbacks
* fpurge:: Discard all pending I/O on a stream
* fputc:: Write a character on a stream or file
* fputs:: Write a character string in a file or stream
* fputwc:: Write a wide character to a file or stream
* fputws:: Write a wide character string to a file or stream
* fread:: Read array elements from a file
* freopen:: Open a file using an existing file descriptor
* fseek:: Set file position
* fsetpos:: Restore position of a stream or file
* ftell:: Return position in a stream or file
* funopen:: Open a stream with custom callbacks
* fwide:: Set and determine the orientation of a FILE stream
* fwrite:: Write array elements from memory to a file or stream
* getc:: Get a character from a file or stream (macro)
* getc_unlocked:: Get a character from a file or stream (macro)
* getchar:: Get a character from standard input (macro)
* getchar_unlocked:: Get a character from standard input (macro)
* getdelim:: Get character string from a file or stream
* getline:: Get character string from a file or stream
* gets:: Get character string from standard input (obsolete)
* getw:: Get a word (int) from a file or stream
* getwchar:: Get a wide character from standard input
* mktemp:: Generate unused file name
* open_memstream:: Open a write stream around an arbitrary-length buffer
* perror:: Print an error message on standard error
* putc:: Write a character on a stream or file (macro)
* putc_unlocked:: Write a character on a stream or file (macro)
* putchar:: Write a character on standard output (macro)
* putchar_unlocked:: Write a character on standard output (macro)
* puts:: Write a character string on standard output
* putw:: Write a word (int) to a file or stream
* putwchar:: Write a wide character to standard output
* remove:: Delete a file's name
* rename:: Rename a file
* rewind:: Reinitialize a file or stream
* setbuf:: Specify full buffering for a file or stream
* setbuffer:: Specify full buffering for a file or stream with size
* setlinebuf:: Specify line buffering for a file or stream
* setvbuf:: Specify buffering for a file or stream
* siprintf:: Write formatted output (integer only)
* siscanf:: Scan and format input (integer only)
* sprintf:: Write formatted output
* sscanf:: Scan and format input
* swprintf:: Write formatted wide character output
* swscanf:: Scan and format wide character input
* tmpfile:: Create a temporary file
* tmpnam:: Generate name for a temporary file
* ungetc:: Push data back into a stream
* ungetwc:: Push wide character data back into a stream
* vfprintf:: Format variable argument list
* vfscanf:: Scan variable argument list
* vfwprintf:: Format variable wide character argument list
* vfwscanf:: Scan and format argument list from wide character input
* viprintf:: Format variable argument list (integer only)
* viscanf:: Scan variable format list (integer only)

File: libc.info, Node: clearerr, Next: diprintf, Up: Stdio
4.1 `clearerr'--clear file or stream error indicator
====================================================
*Synopsis*
#include <stdio.h>
void clearerr(FILE *FP);
*Description*
The `stdio' functions maintain an error indicator with each file
pointer FP, to record whether any read or write errors have occurred on
the associated file or stream. Similarly, it maintains an end-of-file
indicator to record whether there is no more data in the file.
Use `clearerr' to reset both of these indicators.
See `ferror' and `feof' to query the two indicators.
*Returns*
`clearerr' does not return a result.
*Portability*
ANSI C requires `clearerr'.
No supporting OS subroutines are required.

File: libc.info, Node: diprintf, Next: dprintf, Prev: clearerr, Up: Stdio
4.2 `diprintf', `vdiprintf'--print to a file descriptor (integer only)
======================================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int diprintf(int FD, const char *FORMAT, ...);
int vdiprintf(int FD, const char *FORMAT, va_list AP);
int _diprintf_r(struct _reent *PTR, int FD,
const char *FORMAT, ...);
int _vdiprintf_r(struct _reent *PTR, int FD,
const char *FORMAT, va_list AP);
*Description*
`diprintf' and `vdiprintf' are similar to `dprintf' and `vdprintf',
except that only integer format specifiers are processed.
The functions `_diprintf_r' and `_vdiprintf_r' are simply reentrant
versions of the functions above.
*Returns*
Similar to `dprintf' and `vdprintf'.
*Portability*
This set of functions is an integer-only extension, and is not portable.
Supporting OS subroutines required: `sbrk', `write'.

File: libc.info, Node: dprintf, Next: fclose, Prev: diprintf, Up: Stdio
4.3 `dprintf', `vdprintf'--print to a file descriptor
=====================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int dprintf(int FD, const char *FORMAT, ...);
int vdprintf(int FD, const char *FORMAT, va_list AP);
int _dprintf_r(struct _reent *PTR, int FD,
const char *FORMAT, ...);
int _vdprintf_r(struct _reent *PTR, int FD,
const char *FORMAT, va_list AP);
*Description*
`dprintf' and `vdprintf' allow printing a format, similarly to
`printf', but write to a file descriptor instead of to a `FILE' stream.
The functions `_dprintf_r' and `_vdprintf_r' are simply reentrant
versions of the functions above.
*Returns*
The return value and errors are exactly as for `write', except that
`errno' may also be set to `ENOMEM' if the heap is exhausted.
*Portability*
This function is originally a GNU extension in glibc and is not
portable.
Supporting OS subroutines required: `sbrk', `write'.

File: libc.info, Node: fclose, Next: fcloseall, Prev: dprintf, Up: Stdio
4.4 `fclose'--close a file
==========================
*Synopsis*
#include <stdio.h>
int fclose(FILE *FP);
int _fclose_r(struct _reent *REENT, FILE *FP);
*Description*
If the file or stream identified by FP is open, `fclose' closes it,
after first ensuring that any pending data is written (by calling
`fflush(FP)').
The alternate function `_fclose_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`fclose' returns `0' if successful (including when FP is `NULL' or not
an open file); otherwise, it returns `EOF'.
*Portability*
`fclose' is required by ANSI C.
Required OS subroutines: `close', `fstat', `isatty', `lseek',
`read', `sbrk', `write'.

File: libc.info, Node: fcloseall, Next: fdopen, Prev: fclose, Up: Stdio
4.5 `fcloseall'--close all files
================================
*Synopsis*
#include <stdio.h>
int fcloseall(void);
int _fcloseall_r (struct _reent *PTR);
*Description*
`fcloseall' closes all files in the current reentrancy struct's domain.
The function `_fcloseall_r' is the same function, except the reentrancy
struct is passed in as the PTR argument.
This function is not recommended as it closes all streams, including
the std streams.
*Returns*
`fclose' returns `0' if all closes are successful. Otherwise, EOF is
returned.
*Portability*
`fcloseall' is a glibc extension.
Required OS subroutines: `close', `fstat', `isatty', `lseek',
`read', `sbrk', `write'.

File: libc.info, Node: fdopen, Next: feof, Prev: fcloseall, Up: Stdio
4.6 `fdopen'--turn open file into a stream
==========================================
*Synopsis*
#include <stdio.h>
FILE *fdopen(int FD, const char *MODE);
FILE *_fdopen_r(struct _reent *REENT,
int FD, const char *MODE);
*Description*
`fdopen' produces a file descriptor of type `FILE *', from a descriptor
for an already-open file (returned, for example, by the system
subroutine `open' rather than by `fopen'). The MODE argument has the
same meanings as in `fopen'.
*Returns*
File pointer or `NULL', as for `fopen'.
*Portability*
`fdopen' is ANSI.

File: libc.info, Node: feof, Next: ferror, Prev: fdopen, Up: Stdio
4.7 `feof'--test for end of file
================================
*Synopsis*
#include <stdio.h>
int feof(FILE *FP);
*Description*
`feof' tests whether or not the end of the file identified by FP has
been reached.
*Returns*
`feof' returns `0' if the end of file has not yet been reached; if at
end of file, the result is nonzero.
*Portability*
`feof' is required by ANSI C.
No supporting OS subroutines are required.

File: libc.info, Node: ferror, Next: fflush, Prev: feof, Up: Stdio
4.8 `ferror'--test whether read/write error has occurred
========================================================
*Synopsis*
#include <stdio.h>
int ferror(FILE *FP);
*Description*
The `stdio' functions maintain an error indicator with each file
pointer FP, to record whether any read or write errors have occurred on
the associated file or stream. Use `ferror' to query this indicator.
See `clearerr' to reset the error indicator.
*Returns*
`ferror' returns `0' if no errors have occurred; it returns a nonzero
value otherwise.
*Portability*
ANSI C requires `ferror'.
No supporting OS subroutines are required.

File: libc.info, Node: fflush, Next: fgetc, Prev: ferror, Up: Stdio
4.9 `fflush'--flush buffered file output
========================================
*Synopsis*
#include <stdio.h>
int fflush(FILE *FP);
int _fflush_r(struct _reent *REENT, FILE *FP);
*Description*
The `stdio' output functions can buffer output before delivering it to
the host system, in order to minimize the overhead of system calls.
Use `fflush' to deliver any such pending output (for the file or
stream identified by FP) to the host system.
If FP is `NULL', `fflush' delivers pending output from all open
files.
Additionally, if FP is a seekable input stream visiting a file
descriptor, set the position of the file descriptor to match next
unread byte, useful for obeying POSIX semantics when ending a process
without consuming all input from the stream.
The alternate function `_fflush_r' is a reentrant version, where the
extra argument REENT is a pointer to a reentrancy structure, and FP
must not be NULL.
*Returns*
`fflush' returns `0' unless it encounters a write error; in that
situation, it returns `EOF'.
*Portability*
ANSI C requires `fflush'. The behavior on input streams is only
specified by POSIX, and not all implementations follow POSIX rules.
No supporting OS subroutines are required.

File: libc.info, Node: fgetc, Next: fgetpos, Prev: fflush, Up: Stdio
4.10 `fgetc'--get a character from a file or stream
===================================================
*Synopsis*
#include <stdio.h>
int fgetc(FILE *FP);
#include <stdio.h>
int _fgetc_r(struct _reent *PTR, FILE *FP);
*Description*
Use `fgetc' to get the next single character from the file or stream
identified by FP. As a side effect, `fgetc' advances the file's
current position indicator.
For a macro version of this function, see `getc'.
The function `_fgetc_r' is simply a reentrant version of `fgetc'
that is passed the additional reentrant structure pointer argument: PTR.
*Returns*
The next character (read as an `unsigned char', and cast to `int'),
unless there is no more data, or the host system reports a read error;
in either of these situations, `fgetc' returns `EOF'.
You can distinguish the two situations that cause an `EOF' result by
using the `ferror' and `feof' functions.
*Portability*
ANSI C requires `fgetc'.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: fgetpos, Next: fgets, Prev: fgetc, Up: Stdio
4.11 `fgetpos'--record position in a stream or file
===================================================
*Synopsis*
#include <stdio.h>
int fgetpos(FILE *FP, fpos_t *POS);
int _fgetpos_r(struct _reent *PTR, FILE *FP, fpos_t *POS);
*Description*
Objects of type `FILE' can have a "position" that records how much of
the file your program has already read. Many of the `stdio' functions
depend on this position, and many change it as a side effect.
You can use `fgetpos' to report on the current position for a file
identified by FP; `fgetpos' will write a value representing that
position at `*POS'. Later, you can use this value with `fsetpos' to
return the file to this position.
In the current implementation, `fgetpos' simply uses a character
count to represent the file position; this is the same number that
would be returned by `ftell'.
*Returns*
`fgetpos' returns `0' when successful. If `fgetpos' fails, the result
is `1'. Failure occurs on streams that do not support positioning; the
global `errno' indicates this condition with the value `ESPIPE'.
*Portability*
`fgetpos' is required by the ANSI C standard, but the meaning of the
value it records is not specified beyond requiring that it be
acceptable as an argument to `fsetpos'. In particular, other
conforming C implementations may return a different result from `ftell'
than what `fgetpos' writes at `*POS'.
No supporting OS subroutines are required.

File: libc.info, Node: fgets, Next: fgetwc, Prev: fgetpos, Up: Stdio
4.12 `fgets'--get character string from a file or stream
========================================================
*Synopsis*
#include <stdio.h>
char *fgets(char *BUF, int N, FILE *FP);
#include <stdio.h>
char *_fgets_r(struct _reent *PTR, char *BUF, int N, FILE *FP);
*Description*
Reads at most N-1 characters from FP until a newline is found. The
characters including to the newline are stored in BUF. The buffer is
terminated with a 0.
The `_fgets_r' function is simply the reentrant version of `fgets'
and is passed an additional reentrancy structure pointer: PTR.
*Returns*
`fgets' returns the buffer passed to it, with the data filled in. If
end of file occurs with some data already accumulated, the data is
returned with no other indication. If no data are read, NULL is
returned instead.
*Portability*
`fgets' should replace all uses of `gets'. Note however that `fgets'
returns all of the data, while `gets' removes the trailing newline
(with no indication that it has done so.)
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: fgetwc, Next: fgetws, Prev: fgets, Up: Stdio
4.13 `fgetwc', `getwc'--get a wide character from a file or stream
==================================================================
*Synopsis*
#include <stdio.h>
#include <wchar.h>
wint_t fgetwc(FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t _fgetwc_r(struct _reent *PTR, FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t getwc(FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t _getwc_r(struct _reent *PTR, FILE *FP);
*Description*
Use `fgetwc' to get the next wide character from the file or stream
identified by FP. As a side effect, `fgetwc' advances the file's
current position indicator.
The `getwc' function or macro functions identically to `fgetwc'.
It may be implemented as a macro, and may evaluate its argument more
than once. There is no reason ever to use it.
`_fgetwc_r' and `_getwc_r' are simply reentrant versions of `fgetwc'
and `getwc' that are passed the additional reentrant structure pointer
argument: PTR.
*Returns*
The next wide character cast to `wint_t'), unless there is no more data,
or the host system reports a read error; in either of these situations,
`fgetwc' and `getwc' return `WEOF'.
You can distinguish the two situations that cause an `EOF' result by
using the `ferror' and `feof' functions.
*Portability*
C99, POSIX.1-2001

File: libc.info, Node: fgetws, Next: fileno, Prev: fgetwc, Up: Stdio
4.14 `fgetws'--get wide character string from a file or stream
==============================================================
*Synopsis*
#include <wchar.h>
wchar_t *fgetws(wchar_t *WS, int N, FILE *FP);
#include <wchar.h>
wchar_t *_fgetws_r(struct _reent *PTR, wchar_t *WS, int N, FILE *FP);
*Description*
Reads at most N-1 wide characters from FP until a newline is found. The
wide characters including to the newline are stored in WS. The buffer
is terminated with a 0.
The `_fgetws_r' function is simply the reentrant version of `fgetws'
and is passed an additional reentrancy structure pointer: PTR.
*Returns*
`fgetws' returns the buffer passed to it, with the data filled in. If
end of file occurs with some data already accumulated, the data is
returned with no other indication. If no data are read, NULL is
returned instead.
*Portability*
C99, POSIX.1-2001

File: libc.info, Node: fileno, Next: fmemopen, Prev: fgetws, Up: Stdio
4.15 `fileno'--return file descriptor associated with stream
============================================================
*Synopsis*
#include <stdio.h>
int fileno(FILE *FP);
*Description*
You can use `fileno' to return the file descriptor identified by FP.
*Returns*
`fileno' returns a non-negative integer when successful. If FP is not
an open stream, `fileno' returns -1.
*Portability*
`fileno' is not part of ANSI C. POSIX requires `fileno'.
Supporting OS subroutines required: none.

File: libc.info, Node: fmemopen, Next: fopen, Prev: fileno, Up: Stdio
4.16 `fmemopen'--open a stream around a fixed-length string
===========================================================
*Synopsis*
#include <stdio.h>
FILE *fmemopen(void *restrict BUF, size_t SIZE,
const char *restrict MODE);
*Description*
`fmemopen' creates a seekable `FILE' stream that wraps a fixed-length
buffer of SIZE bytes starting at BUF. The stream is opened with MODE
treated as in `fopen', where append mode starts writing at the first
NUL byte. If BUF is NULL, then SIZE bytes are automatically provided
as if by `malloc', with the initial size of 0, and MODE must contain
`+' so that data can be read after it is written.
The stream maintains a current position, which moves according to
bytes read or written, and which can be one past the end of the array.
The stream also maintains a current file size, which is never greater
than SIZE. If MODE starts with `r', the position starts at `0', and
file size starts at SIZE if BUF was provided. If MODE starts with `w',
the position and file size start at `0', and if BUF was provided, the
first byte is set to NUL. If MODE starts with `a', the position and
file size start at the location of the first NUL byte, or else SIZE if
BUF was provided.
When reading, NUL bytes have no significance, and reads cannot exceed
the current file size. When writing, the file size can increase up to
SIZE as needed, and NUL bytes may be embedded in the stream (see
`open_memstream' for an alternative that automatically enlarges the
buffer). When the stream is flushed or closed after a write that
changed the file size, a NUL byte is written at the current position if
there is still room; if the stream is not also open for reading, a NUL
byte is additionally written at the last byte of BUF when the stream
has exceeded SIZE, so that a write-only BUF is always NUL-terminated
when the stream is flushed or closed (and the initial SIZE should take
this into account). It is not possible to seek outside the bounds of
SIZE. A NUL byte written during a flush is restored to its previous
value when seeking elsewhere in the string.
*Returns*
The return value is an open FILE pointer on success. On error, `NULL'
is returned, and `errno' will be set to EINVAL if SIZE is zero or MODE
is invalid, ENOMEM if BUF was NULL and memory could not be allocated,
or EMFILE if too many streams are already open.
*Portability*
This function is being added to POSIX 200x, but is not in POSIX 2001.
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: fopen, Next: fopencookie, Prev: fmemopen, Up: Stdio
4.17 `fopen'--open a file
=========================
*Synopsis*
#include <stdio.h>
FILE *fopen(const char *FILE, const char *MODE);
FILE *_fopen_r(struct _reent *REENT,
const char *FILE, const char *MODE);
*Description*
`fopen' initializes the data structures needed to read or write a file.
Specify the file's name as the string at FILE, and the kind of access
you need to the file with the string at MODE.
The alternate function `_fopen_r' is a reentrant version. The extra
argument REENT is a pointer to a reentrancy structure.
Three fundamental kinds of access are available: read, write, and
append. `*MODE' must begin with one of the three characters ``r'',
``w'', or ``a'', to select one of these:
`r'
Open the file for reading; the operation will fail if the file does
not exist, or if the host system does not permit you to read it.
`w'
Open the file for writing _from the beginning_ of the file:
effectively, this always creates a new file. If the file whose
name you specified already existed, its old contents are discarded.
`a'
Open the file for appending data, that is writing from the end of
file. When you open a file this way, all data always goes to the
current end of file; you cannot change this using `fseek'.
Some host systems distinguish between "binary" and "text" files.
Such systems may perform data transformations on data written to, or
read from, files opened as "text". If your system is one of these,
then you can append a ``b'' to any of the three modes above, to specify
that you are opening the file as a binary file (the default is to open
the file as a text file).
``rb'', then, means "read binary"; ``wb'', "write binary"; and
``ab'', "append binary".
To make C programs more portable, the ``b'' is accepted on all
systems, whether or not it makes a difference.
Finally, you might need to both read and write from the same file.
You can also append a ``+'' to any of the three modes, to permit this.
(If you want to append both ``b'' and ``+'', you can do it in either
order: for example, `"rb+"' means the same thing as `"r+b"' when used
as a mode string.)
Use `"r+"' (or `"rb+"') to permit reading and writing anywhere in an
existing file, without discarding any data; `"w+"' (or `"wb+"') to
create a new file (or begin by discarding all data from an old one)
that permits reading and writing anywhere in it; and `"a+"' (or
`"ab+"') to permit reading anywhere in an existing file, but writing
only at the end.
*Returns*
`fopen' returns a file pointer which you can use for other file
operations, unless the file you requested could not be opened; in that
situation, the result is `NULL'. If the reason for failure was an
invalid string at MODE, `errno' is set to `EINVAL'.
*Portability*
`fopen' is required by ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `open', `read', `sbrk', `write'.

File: libc.info, Node: fopencookie, Next: fpurge, Prev: fopen, Up: Stdio
4.18 `fopencookie'--open a stream with custom callbacks
=======================================================
*Synopsis*
#include <stdio.h>
FILE *fopencookie(const void *COOKIE, const char *MODE,
cookie_io_functions_t FUNCTIONS);
*Description*
`fopencookie' creates a `FILE' stream where I/O is performed using
custom callbacks. The callbacks are registered via the structure:
typedef ssize_t (*cookie_read_function_t)(void *_cookie, char *_buf,
size_t _n); typedef ssize_t (*cookie_write_function_t)(void *_cookie,
const char *_buf, size_t _n); typedef int
(*cookie_seek_function_t)(void *_cookie, off_t *_off, int _whence);
typedef int (*cookie_close_function_t)(void *_cookie);
typedef struct
{
cookie_read_function_t *read;
cookie_write_function_t *write;
cookie_seek_function_t *seek;
cookie_close_function_t *close;
} cookie_io_functions_t;
The stream is opened with MODE treated as in `fopen'. The callbacks
FUNCTIONS.READ and FUNCTIONS.WRITE may only be NULL when MODE does not
require them.
FUNCTIONS.READ should return -1 on failure, or else the number of
bytes read (0 on EOF). It is similar to `read', except that COOKIE
will be passed as the first argument.
FUNCTIONS.WRITE should return -1 on failure, or else the number of
bytes written. It is similar to `write', except that COOKIE will be
passed as the first argument.
FUNCTIONS.SEEK should return -1 on failure, and 0 on success, with
_OFF set to the current file position. It is a cross between `lseek'
and `fseek', with the _WHENCE argument interpreted in the same manner.
A NULL FUNCTIONS.SEEK makes the stream behave similarly to a pipe in
relation to stdio functions that require positioning.
FUNCTIONS.CLOSE should return -1 on failure, or 0 on success. It is
similar to `close', except that COOKIE will be passed as the first
argument. A NULL FUNCTIONS.CLOSE merely flushes all data then lets
`fclose' succeed. A failed close will still invalidate the stream.
Read and write I/O functions are allowed to change the underlying
buffer on fully buffered or line buffered streams by calling `setvbuf'.
They are also not required to completely fill or empty the buffer.
They are not, however, allowed to change streams from unbuffered to
buffered or to change the state of the line buffering flag. They must
also be prepared to have read or write calls occur on buffers other
than the one most recently specified.
*Returns*
The return value is an open FILE pointer on success. On error, `NULL'
is returned, and `errno' will be set to EINVAL if a function pointer is
missing or MODE is invalid, ENOMEM if the stream cannot be created, or
EMFILE if too many streams are already open.
*Portability*
This function is a newlib extension, copying the prototype from Linux.
It is not portable. See also the `funopen' interface from BSD.
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: fpurge, Next: fputc, Prev: fopencookie, Up: Stdio
4.19 `fpurge'--discard pending file I/O
=======================================
*Synopsis*
#include <stdio.h>
int fpurge(FILE *FP);
int _fpurge_r(struct _reent *REENT, FILE *FP);
#include <stdio.h>
#include <stdio_ext.h>
void __fpurge(FILE *FP);
*Description*
Use `fpurge' to clear all buffers of the given stream. For output
streams, this discards data not yet written to disk. For input streams,
this discards any data from `ungetc' and any data retrieved from disk
but not yet read via `getc'. This is more severe than `fflush', and
generally is only needed when manually altering the underlying file
descriptor of a stream.
`__fpurge' behaves exactly like `fpurge' but does not return a value.
The alternate function `_fpurge_r' is a reentrant version, where the
extra argument REENT is a pointer to a reentrancy structure, and FP
must not be NULL.
*Returns*
`fpurge' returns `0' unless FP is not valid, in which case it returns
`EOF' and sets `errno'.
*Portability*
These functions are not portable to any standard.
No supporting OS subroutines are required.

File: libc.info, Node: fputc, Next: fputs, Prev: fpurge, Up: Stdio
4.20 `fputc'--write a character on a stream or file
===================================================
*Synopsis*
#include <stdio.h>
int fputc(int CH, FILE *FP);
#include <stdio.h>
int _fputc_r(struct _rent *PTR, int CH, FILE *FP);
*Description*
`fputc' converts the argument CH from an `int' to an `unsigned char',
then writes it to the file or stream identified by FP.
If the file was opened with append mode (or if the stream cannot
support positioning), then the new character goes at the end of the
file or stream. Otherwise, the new character is written at the current
value of the position indicator, and the position indicator oadvances
by one.
For a macro version of this function, see `putc'.
The `_fputc_r' function is simply a reentrant version of `fputc'
that takes an additional reentrant structure argument: PTR.
*Returns*
If successful, `fputc' returns its argument CH. If an error
intervenes, the result is `EOF'. You can use ``ferror(FP)'' to query
for errors.
*Portability*
`fputc' is required by ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: fputs, Next: fputwc, Prev: fputc, Up: Stdio
4.21 `fputs'--write a character string in a file or stream
==========================================================
*Synopsis*
#include <stdio.h>
int fputs(const char *S, FILE *FP);
#include <stdio.h>
int _fputs_r(struct _reent *PTR, const char *S, FILE *FP);
*Description*
`fputs' writes the string at S (but without the trailing null) to the
file or stream identified by FP.
`_fputs_r' is simply the reentrant version of `fputs' that takes an
additional reentrant struct pointer argument: PTR.
*Returns*
If successful, the result is `0'; otherwise, the result is `EOF'.
*Portability*
ANSI C requires `fputs', but does not specify that the result on
success must be `0'; any non-negative value is permitted.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: fputwc, Next: fputws, Prev: fputs, Up: Stdio
4.22 `fputwc', `putwc'--write a wide character on a stream or file
==================================================================
*Synopsis*
#include <stdio.h>
#include <wchar.h>
wint_t fputwc(wchar_t WC, FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t _fputwc_r(struct _reent *PTR, wchar_t WC, FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t putwc(wchar_t WC, FILE *FP);
#include <stdio.h>
#include <wchar.h>
wint_t _putwc_r(struct _reent *PTR, wchar_t WC, FILE *FP);
*Description*
`fputwc' writes the wide character argument WC to the file or stream
identified by FP.
If the file was opened with append mode (or if the stream cannot
support positioning), then the new wide character goes at the end of the
file or stream. Otherwise, the new wide character is written at the
current value of the position indicator, and the position indicator
oadvances by one.
The `putwc' function or macro functions identically to `fputwc'. It
may be implemented as a macro, and may evaluate its argument more than
once. There is no reason ever to use it.
The `_fputwc_r' and `_putwc_r' functions are simply reentrant
versions of `fputwc' and `putwc' that take an additional reentrant
structure argument: PTR.
*Returns*
If successful, `fputwc' and `putwc' return their argument WC. If an
error intervenes, the result is `EOF'. You can use ``ferror(FP)'' to
query for errors.
*Portability*
C99, POSIX.1-2001

File: libc.info, Node: fputws, Next: fread, Prev: fputwc, Up: Stdio
4.23 `fputws'--write a wide character string in a file or stream
================================================================
*Synopsis*
#include <wchar.h>
int fputws(const wchar_t *WS, FILE *FP);
#include <wchar.h>
int _fputws_r(struct _reent *PTR, const wchar_t *WS, FILE *FP);
*Description*
`fputws' writes the wide character string at WS (but without the
trailing null) to the file or stream identified by FP.
`_fputws_r' is simply the reentrant version of `fputws' that takes
an additional reentrant struct pointer argument: PTR.
*Returns*
If successful, the result is a non-negative integer; otherwise, the
result is `-1' to indicate an error.
*Portability*
C99, POSIX.1-2001

File: libc.info, Node: fread, Next: freopen, Prev: fputws, Up: Stdio
4.24 `fread'--read array elements from a file
=============================================
*Synopsis*
#include <stdio.h>
size_t fread(void *BUF, size_t SIZE, size_t COUNT,
FILE *FP);
#include <stdio.h>
size_t _fread_r(struct _reent *PTR, void *BUF,
size_t SIZE, size_t COUNT, FILE *FP);
*Description*
`fread' attempts to copy, from the file or stream identified by FP,
COUNT elements (each of size SIZE) into memory, starting at BUF.
`fread' may copy fewer elements than COUNT if an error, or end of file,
intervenes.
`fread' also advances the file position indicator (if any) for FP by
the number of _characters_ actually read.
`_fread_r' is simply the reentrant version of `fread' that takes an
additional reentrant structure pointer argument: PTR.
*Returns*
The result of `fread' is the number of elements it succeeded in reading.
*Portability*
ANSI C requires `fread'.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: freopen, Next: fseek, Prev: fread, Up: Stdio
4.25 `freopen'--open a file using an existing file descriptor
=============================================================
*Synopsis*
#include <stdio.h>
FILE *freopen(const char *FILE, const char *MODE,
FILE *FP);
FILE *_freopen_r(struct _reent *PTR, const char *FILE,
const char *MODE, FILE *FP);
*Description*
Use this variant of `fopen' if you wish to specify a particular file
descriptor FP (notably `stdin', `stdout', or `stderr') for the file.
If FP was associated with another file or stream, `freopen' closes
that other file or stream (but ignores any errors while closing it).
FILE and MODE are used just as in `fopen'.
If FILE is `NULL', the underlying stream is modified rather than
closed. The file cannot be given a more permissive access mode (for
example, a MODE of "w" will fail on a read-only file descriptor), but
can change status such as append or binary mode. If modification is
not possible, failure occurs.
*Returns*
If successful, the result is the same as the argument FP. If the file
cannot be opened as specified, the result is `NULL'.
*Portability*
ANSI C requires `freopen'.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `open', `read', `sbrk', `write'.

File: libc.info, Node: fseek, Next: fsetpos, Prev: freopen, Up: Stdio
4.26 `fseek', `fseeko'--set file position
=========================================
*Synopsis*
#include <stdio.h>
int fseek(FILE *FP, long OFFSET, int WHENCE)
int fseeko(FILE *FP, off_t OFFSET, int WHENCE)
int _fseek_r(struct _reent *PTR, FILE *FP,
long OFFSET, int WHENCE)
int _fseeko_r(struct _reent *PTR, FILE *FP,
off_t OFFSET, int WHENCE)
*Description*
Objects of type `FILE' can have a "position" that records how much of
the file your program has already read. Many of the `stdio' functions
depend on this position, and many change it as a side effect.
You can use `fseek'/`fseeko' to set the position for the file
identified by FP. The value of OFFSET determines the new position, in
one of three ways selected by the value of WHENCE (defined as macros in
``stdio.h''):
`SEEK_SET'--OFFSET is the absolute file position (an offset from the
beginning of the file) desired. OFFSET must be positive.
`SEEK_CUR'--OFFSET is relative to the current file position. OFFSET
can meaningfully be either positive or negative.
`SEEK_END'--OFFSET is relative to the current end of file. OFFSET
can meaningfully be either positive (to increase the size of the file)
or negative.
See `ftell'/`ftello' to determine the current file position.
*Returns*
`fseek'/`fseeko' return `0' when successful. On failure, the result is
`EOF'. The reason for failure is indicated in `errno': either `ESPIPE'
(the stream identified by FP doesn't support repositioning) or `EINVAL'
(invalid file position).
*Portability*
ANSI C requires `fseek'.
`fseeko' is defined by the Single Unix specification.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: fsetpos, Next: ftell, Prev: fseek, Up: Stdio
4.27 `fsetpos'--restore position of a stream or file
====================================================
*Synopsis*
#include <stdio.h>
int fsetpos(FILE *FP, const fpos_t *POS);
int _fsetpos_r(struct _reent *PTR, FILE *FP,
const fpos_t *POS);
*Description*
Objects of type `FILE' can have a "position" that records how much of
the file your program has already read. Many of the `stdio' functions
depend on this position, and many change it as a side effect.
You can use `fsetpos' to return the file identified by FP to a
previous position `*POS' (after first recording it with `fgetpos').
See `fseek' for a similar facility.
*Returns*
`fgetpos' returns `0' when successful. If `fgetpos' fails, the result
is `1'. The reason for failure is indicated in `errno': either
`ESPIPE' (the stream identified by FP doesn't support repositioning) or
`EINVAL' (invalid file position).
*Portability*
ANSI C requires `fsetpos', but does not specify the nature of `*POS'
beyond identifying it as written by `fgetpos'.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: ftell, Next: funopen, Prev: fsetpos, Up: Stdio
4.28 `ftell', `ftello'--return position in a stream or file
===========================================================
*Synopsis*
#include <stdio.h>
long ftell(FILE *FP);
off_t ftello(FILE *FP);
long _ftell_r(struct _reent *PTR, FILE *FP);
off_t _ftello_r(struct _reent *PTR, FILE *FP);
*Description*
Objects of type `FILE' can have a "position" that records how much of
the file your program has already read. Many of the `stdio' functions
depend on this position, and many change it as a side effect.
The result of `ftell'/`ftello' is the current position for a file
identified by FP. If you record this result, you can later use it with
`fseek'/`fseeko' to return the file to this position. The difference
between `ftell' and `ftello' is that `ftell' returns `long' and
`ftello' returns `off_t'.
In the current implementation, `ftell'/`ftello' simply uses a
character count to represent the file position; this is the same number
that would be recorded by `fgetpos'.
*Returns*
`ftell'/`ftello' return the file position, if possible. If they cannot
do this, they return `-1L'. Failure occurs on streams that do not
support positioning; the global `errno' indicates this condition with
the value `ESPIPE'.
*Portability*
`ftell' is required by the ANSI C standard, but the meaning of its
result (when successful) is not specified beyond requiring that it be
acceptable as an argument to `fseek'. In particular, other conforming
C implementations may return a different result from `ftell' than what
`fgetpos' records.
`ftello' is defined by the Single Unix specification.
No supporting OS subroutines are required.

File: libc.info, Node: funopen, Next: fwide, Prev: ftell, Up: Stdio
4.29 `funopen', `fropen', `fwopen'--open a stream with custom callbacks
=======================================================================
*Synopsis*
#include <stdio.h>
FILE *funopen(const void *COOKIE,
int (*READFN) (void *cookie, char *buf, int n),
int (*WRITEFN) (void *cookie, const char *buf, int n),
fpos_t (*SEEKFN) (void *cookie, fpos_t off, int whence),
int (*CLOSEFN) (void *cookie));
FILE *fropen(const void *COOKIE,
int (*READFN) (void *cookie, char *buf, int n));
FILE *fwopen(const void *COOKIE,
int (*WRITEFN) (void *cookie, const char *buf, int n));
*Description*
`funopen' creates a `FILE' stream where I/O is performed using custom
callbacks. At least one of READFN and WRITEFN must be provided, which
determines whether the stream behaves with mode <"r">, <"w">, or <"r+">.
READFN should return -1 on failure, or else the number of bytes read
(0 on EOF). It is similar to `read', except that <int> rather than
<size_t> bounds a transaction size, and COOKIE will be passed as the
first argument. A NULL READFN makes attempts to read the stream fail.
WRITEFN should return -1 on failure, or else the number of bytes
written. It is similar to `write', except that <int> rather than
<size_t> bounds a transaction size, and COOKIE will be passed as the
first argument. A NULL WRITEFN makes attempts to write the stream fail.
SEEKFN should return (fpos_t)-1 on failure, or else the current file
position. It is similar to `lseek', except that COOKIE will be passed
as the first argument. A NULL SEEKFN makes the stream behave similarly
to a pipe in relation to stdio functions that require positioning.
This implementation assumes fpos_t and off_t are the same type.
CLOSEFN should return -1 on failure, or 0 on success. It is similar
to `close', except that COOKIE will be passed as the first argument. A
NULL CLOSEFN merely flushes all data then lets `fclose' succeed. A
failed close will still invalidate the stream.
Read and write I/O functions are allowed to change the underlying
buffer on fully buffered or line buffered streams by calling `setvbuf'.
They are also not required to completely fill or empty the buffer.
They are not, however, allowed to change streams from unbuffered to
buffered or to change the state of the line buffering flag. They must
also be prepared to have read or write calls occur on buffers other
than the one most recently specified.
The functions `fropen' and `fwopen' are convenience macros around
`funopen' that only use the specified callback.
*Returns*
The return value is an open FILE pointer on success. On error, `NULL'
is returned, and `errno' will be set to EINVAL if a function pointer is
missing, ENOMEM if the stream cannot be created, or EMFILE if too many
streams are already open.
*Portability*
This function is a newlib extension, copying the prototype from BSD.
It is not portable. See also the `fopencookie' interface from Linux.
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: fwide, Next: fwrite, Prev: funopen, Up: Stdio
4.30 `fwide'--set and determine the orientation of a FILE stream
================================================================
*Synopsis*
#include <wchar.h>
int fwide(FILE *FP, int MODE)
int _fwide_r(struct _reent *PTR, FILE *FP, int MODE)
*Description*
When MODE is zero, the `fwide' function determines the current
orientation of FP. It returns a value > 0 if FP is wide-character
oriented, i.e. if wide character I/O is permitted but char I/O is
disallowed. It returns a value < 0 if FP is byte oriented, i.e. if char
I/O is permitted but wide character I/O is disallowed. It returns zero
if FP has no orientation yet; in this case the next I/O operation might
change the orientation (to byte oriented if it is a char I/O operation,
or to wide-character oriented if it is a wide character I/O operation).
Once a stream has an orientation, it cannot be changed and persists
until the stream is closed, unless the stream is re-opened with freopen,
which removes the orientation of the stream.
When MODE is non-zero, the `fwide' function first attempts to set
FP's orientation (to wide-character oriented if MODE > 0, or to byte
oriented if MODE < 0). It then returns a value denoting the current
orientation, as above.
*Returns*
The `fwide' function returns FP's orientation, after possibly changing
it. A return value > 0 means wide-character oriented. A return value <
0 means byte oriented. A return value of zero means undecided.
*Portability*
C99, POSIX.1-2001.

File: libc.info, Node: fwrite, Next: getc, Prev: fwide, Up: Stdio
4.31 `fwrite'--write array elements
===================================
*Synopsis*
#include <stdio.h>
size_t fwrite(const void *BUF, size_t SIZE,
size_t COUNT, FILE *FP);
#include <stdio.h>
size_t _fwrite_r(struct _reent *PTR, const void *BUF, size_t SIZE,
size_t COUNT, FILE *FP);
*Description*
`fwrite' attempts to copy, starting from the memory location BUF, COUNT
elements (each of size SIZE) into the file or stream identified by FP.
`fwrite' may copy fewer elements than COUNT if an error intervenes.
`fwrite' also advances the file position indicator (if any) for FP
by the number of _characters_ actually written.
`_fwrite_r' is simply the reentrant version of `fwrite' that takes
an additional reentrant structure argument: PTR.
*Returns*
If `fwrite' succeeds in writing all the elements you specify, the
result is the same as the argument COUNT. In any event, the result is
the number of complete elements that `fwrite' copied to the file.
*Portability*
ANSI C requires `fwrite'.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getc, Next: getc_unlocked, Prev: fwrite, Up: Stdio
4.32 `getc'--read a character (macro)
=====================================
*Synopsis*
#include <stdio.h>
int getc(FILE *FP);
#include <stdio.h>
int _getc_r(struct _reent *PTR, FILE *FP);
*Description*
`getc' is a macro, defined in `stdio.h'. You can use `getc' to get the
next single character from the file or stream identified by FP. As a
side effect, `getc' advances the file's current position indicator.
For a subroutine version of this macro, see `fgetc'.
The `_getc_r' function is simply the reentrant version of `getc'
which passes an additional reentrancy structure pointer argument: PTR.
*Returns*
The next character (read as an `unsigned char', and cast to `int'),
unless there is no more data, or the host system reports a read error;
in either of these situations, `getc' returns `EOF'.
You can distinguish the two situations that cause an `EOF' result by
using the `ferror' and `feof' functions.
*Portability*
ANSI C requires `getc'; it suggests, but does not require, that `getc'
be implemented as a macro. The standard explicitly permits macro
implementations of `getc' to use the argument more than once;
therefore, in a portable program, you should not use an expression with
side effects as the `getc' argument.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getc_unlocked, Next: getchar, Prev: getc, Up: Stdio
4.33 `getc_unlocked'--non-thread-safe version of getc (macro)
=============================================================
*Synopsis*
#include <stdio.h>
int getc_unlocked(FILE *FP);
#include <stdio.h>
int _getc_unlocked_r(FILE *FP);
*Description*
`getc_unlocked' is a non-thread-safe version of `getc' declared in
`stdio.h'. `getc_unlocked' may only safely be used within a scope
protected by flockfile() (or ftrylockfile()) and funlockfile(). These
functions may safely be used in a multi-threaded program if and only if
they are called while the invoking thread owns the ( FILE *) object, as
is the case after a successful call to the flockfile() or
ftrylockfile() functions. If threads are disabled, then
`getc_unlocked' is equivalent to `getc'.
The `_getc_unlocked_r' function is simply the reentrant version of
`get_unlocked' which passes an additional reentrancy structure pointer
argument: PTR.
*Returns*
See `getc'.
*Portability*
POSIX 1003.1 requires `getc_unlocked'. `getc_unlocked' may be
implemented as a macro, so arguments should not have side-effects.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getchar, Next: getchar_unlocked, Prev: getc_unlocked, Up: Stdio
4.34 `getchar'--read a character (macro)
========================================
*Synopsis*
#include <stdio.h>
int getchar(void);
int _getchar_r(struct _reent *REENT);
*Description*
`getchar' is a macro, defined in `stdio.h'. You can use `getchar' to
get the next single character from the standard input stream. As a
side effect, `getchar' advances the standard input's current position
indicator.
The alternate function `_getchar_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
The next character (read as an `unsigned char', and cast to `int'),
unless there is no more data, or the host system reports a read error;
in either of these situations, `getchar' returns `EOF'.
You can distinguish the two situations that cause an `EOF' result by
using ``ferror(stdin)'' and ``feof(stdin)''.
*Portability*
ANSI C requires `getchar'; it suggests, but does not require, that
`getchar' be implemented as a macro.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getchar_unlocked, Next: getdelim, Prev: getchar, Up: Stdio
4.35 `getchar_unlocked'--non-thread-safe version of getchar (macro)
===================================================================
*Synopsis*
#include <stdio.h>
int getchar_unlocked();
#include <stdio.h>
int _getchar_unlocked_r(struct _reent *PTR);
*Description*
`getchar_unlocked' is a non-thread-safe version of `getchar' declared
in `stdio.h'. `getchar_unlocked' may only safely be used within a
scope protected by flockfile() (or ftrylockfile()) and funlockfile().
These functions may safely be used in a multi-threaded program if and
only if they are called while the invoking thread owns the ( FILE *)
object, as is the case after a successful call to the flockfile() or
ftrylockfile() functions. If threads are disabled, then
`getchar_unlocked' is equivalent to `getchar'.
The `_getchar_unlocked_r' function is simply the reentrant version of
`getchar_unlocked' which passes an addtional reentrancy structure
pointer argument: PTR.
*Returns*
See `getchar'.
*Portability*
POSIX 1003.1 requires `getchar_unlocked'. `getchar_unlocked' may be
implemented as a macro.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getdelim, Next: getline, Prev: getchar_unlocked, Up: Stdio
4.36 `getdelim'--read a line up to a specified line delimiter
=============================================================
*Synopsis*
#include <stdio.h>
int getdelim(char **BUFPTR, size_t *N,
int DELIM, FILE *FP);
*Description*
`getdelim' reads a file FP up to and possibly including a specified
delimiter DELIM. The line is read into a buffer pointed to by BUFPTR
and designated with size *N. If the buffer is not large enough, it
will be dynamically grown by `getdelim'. As the buffer is grown, the
pointer to the size N will be updated.
*Returns*
`getdelim' returns `-1' if no characters were successfully read;
otherwise, it returns the number of bytes successfully read. At end of
file, the result is nonzero.
*Portability*
`getdelim' is a glibc extension.
No supporting OS subroutines are directly required.

File: libc.info, Node: getline, Next: gets, Prev: getdelim, Up: Stdio
4.37 `getline'--read a line from a file
=======================================
*Synopsis*
#include <stdio.h>
ssize_t getline(char **BUFPTR, size_t *N, FILE *FP);
*Description*
`getline' reads a file FP up to and possibly including the newline
character. The line is read into a buffer pointed to by BUFPTR and
designated with size *N. If the buffer is not large enough, it will be
dynamically grown by `getdelim'. As the buffer is grown, the pointer
to the size N will be updated.
`getline' is equivalent to getdelim(bufptr, n, '\n', fp);
*Returns*
`getline' returns `-1' if no characters were successfully read,
otherwise, it returns the number of bytes successfully read. at end of
file, the result is nonzero.
*Portability*
`getline' is a glibc extension.
No supporting OS subroutines are directly required.

File: libc.info, Node: gets, Next: getw, Prev: getline, Up: Stdio
4.38 `gets'--get character string (obsolete, use `fgets' instead)
=================================================================
*Synopsis*
#include <stdio.h>
char *gets(char *BUF);
char *_gets_r(struct _reent *REENT, char *BUF);
*Description*
Reads characters from standard input until a newline is found. The
characters up to the newline are stored in BUF. The newline is
discarded, and the buffer is terminated with a 0.
This is a _dangerous_ function, as it has no way of checking the
amount of space available in BUF. One of the attacks used by the
Internet Worm of 1988 used this to overrun a buffer allocated on the
stack of the finger daemon and overwrite the return address, causing
the daemon to execute code downloaded into it over the connection.
The alternate function `_gets_r' is a reentrant version. The extra
argument REENT is a pointer to a reentrancy structure.
*Returns*
`gets' returns the buffer passed to it, with the data filled in. If end
of file occurs with some data already accumulated, the data is returned
with no other indication. If end of file occurs with no data in the
buffer, NULL is returned.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: getw, Next: getwchar, Prev: gets, Up: Stdio
4.39 `getw'--read a word (int)
==============================
*Synopsis*
#include <stdio.h>
int getw(FILE *FP);
*Description*
`getw' is a function, defined in `stdio.h'. You can use `getw' to get
the next word from the file or stream identified by FP. As a side
effect, `getw' advances the file's current position indicator.
*Returns*
The next word (read as an `int'), unless there is no more data or the
host system reports a read error; in either of these situations, `getw'
returns `EOF'. Since `EOF' is a valid `int', you must use `ferror' or
`feof' to distinguish these situations.
*Portability*
`getw' is a remnant of K&R C; it is not part of any ISO C Standard.
`fread' should be used instead. In fact, this implementation of `getw'
is based upon `fread'.
Supporting OS subroutines required: `fread'.

File: libc.info, Node: getwchar, Next: mktemp, Prev: getw, Up: Stdio
4.40 `getwchar'--read a wide character from standard input
==========================================================
*Synopsis*
#include <wchar.h>
wint_t getwchar(void);
wint_t _getwchar_r(struct _reent *REENT);
*Description*
`getwchar' function or macro is the wide character equivalent of the
`getchar' function. You can use `getwchar' to get the next wide
character from the standard input stream. As a side effect, `getwchar'
advances the standard input's current position indicator.
The alternate function `_getwchar_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
The next wide character cast to `wint_t', unless there is no more data,
or the host system reports a read error; in either of these situations,
`getwchar' returns `WEOF'.
You can distinguish the two situations that cause an `WEOF' result by
using ``ferror(stdin)'' and ``feof(stdin)''.
*Portability*
C99

File: libc.info, Node: mktemp, Next: open_memstream, Prev: getwchar, Up: Stdio
4.41 `mktemp', `mkstemp', `mkostemp', `mkstemps',
=================================================
*Synopsis*
#include <stdlib.h>
char *mktemp(char *PATH);
char *mkdtemp(char *PATH);
int mkstemp(char *PATH);
int mkstemps(char *PATH, int SUFFIXLEN);
int mkostemp(char *PATH, int FLAGS);
int mkostemps(char *PATH, int SUFFIXLEN, int FLAGS);
char *_mktemp_r(struct _reent *REENT, char *PATH);
char *_mkdtemp_r(struct _reent *REENT, char *PATH);
int *_mkstemp_r(struct _reent *REENT, char *PATH);
int *_mkstemps_r(struct _reent *REENT, char *PATH, int LEN);
int *_mkostemp_r(struct _reent *REENT, char *PATH,
int FLAGS);
int *_mkostemps_r(struct _reent *REENT, char *PATH, int LEN,
int FLAGS);
*Description*
`mktemp', `mkstemp', and `mkstemps' attempt to generate a file name
that is not yet in use for any existing file. `mkstemp' and `mkstemps'
create the file and open it for reading and writing; `mktemp' simply
generates the file name (making `mktemp' a security risk). `mkostemp'
and `mkostemps' allow the addition of other `open' flags, such as
`O_CLOEXEC', `O_APPEND', or `O_SYNC'. On platforms with a separate
text mode, `mkstemp' forces `O_BINARY', while `mkostemp' allows the
choice between `O_BINARY', `O_TEXT', or 0 for default. `mkdtemp'
attempts to create a directory instead of a file, with a permissions
mask of 0700.
You supply a simple pattern for the generated file name, as the
string at PATH. The pattern should be a valid filename (including path
information if you wish) ending with at least six ``X'' characters.
The generated filename will match the leading part of the name you
supply, with the trailing ``X'' characters replaced by some combination
of digits and letters. With `mkstemps', the ``X'' characters end
SUFFIXLEN bytes before the end of the string.
The alternate functions `_mktemp_r', `_mkdtemp_r', `_mkstemp_r',
`_mkostemp_r', `_mkostemps_r', and `_mkstemps_r' are reentrant
versions. The extra argument REENT is a pointer to a reentrancy
structure.
*Returns*
`mktemp' returns the pointer PATH to the modified string representing
an unused filename, unless it could not generate one, or the pattern
you provided is not suitable for a filename; in that case, it returns
`NULL'. Be aware that there is an inherent race between generating the
name and attempting to create a file by that name; you are advised to
use `O_EXCL|O_CREAT'.
`mkdtemp' returns the pointer PATH to the modified string if the
directory was created, otherwise it returns `NULL'.
`mkstemp', `mkstemps', `mkostemp', and `mkostemps' return a file
descriptor to the newly created file, unless it could not generate an
unused filename, or the pattern you provided is not suitable for a
filename; in that case, it returns `-1'.
*Portability*
ANSI C does not require either `mktemp' or `mkstemp'; the System V
Interface Definition requires `mktemp' as of Issue 2. POSIX 2001
requires `mkstemp', and POSIX 2008 requires `mkdtemp' while deprecating
`mktemp'. `mkstemps', `mkostemp', and `mkostemps' are not standardized.
Supporting OS subroutines required: `getpid', `mkdir', `open',
`stat'.

File: libc.info, Node: open_memstream, Next: perror, Prev: mktemp, Up: Stdio
4.42 `open_memstream', `open_wmemstream'--open a write stream around an arbitrary-length string
===============================================================================================
*Synopsis*
#include <stdio.h>
FILE *open_memstream(char **restrict BUF,
size_t *restrict SIZE);
#include <wchar.h>
FILE *open_wmemstream(wchar_t **restrict BUF,
size_t *restrict SIZE);
*Description*
`open_memstream' creates a seekable, byte-oriented `FILE' stream that
wraps an arbitrary-length buffer, created as if by `malloc'. The
current contents of *BUF are ignored; this implementation uses *SIZE as
a hint of the maximum size expected, but does not fail if the hint was
wrong. The parameters BUF and SIZE are later stored through following
any call to `fflush' or `fclose', set to the current address and usable
size of the allocated string; although after fflush, the pointer is
only valid until another stream operation that results in a write.
Behavior is undefined if the user alters either *BUF or *SIZE prior to
`fclose'.
`open_wmemstream' is like `open_memstream' just with the associated
stream being wide-oriented. The size set in SIZE in subsequent
operations is the number of wide characters.
The stream is write-only, since the user can directly read *BUF
after a flush; see `fmemopen' for a way to wrap a string with a
readable stream. The user is responsible for calling `free' on the
final *BUF after `fclose'.
Any time the stream is flushed, a NUL byte is written at the current
position (but is not counted in the buffer length), so that the string
is always NUL-terminated after at most *SIZE bytes (or wide characters
in case of `open_wmemstream'). However, data previously written beyond
the current stream offset is not lost, and the NUL value written during
a flush is restored to its previous value when seeking elsewhere in the
string.
*Returns*
The return value is an open FILE pointer on success. On error, `NULL'
is returned, and `errno' will be set to EINVAL if BUF or SIZE is NULL,
ENOMEM if memory could not be allocated, or EMFILE if too many streams
are already open.
*Portability*
POSIX.1-2008
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: perror, Next: putc, Prev: open_memstream, Up: Stdio
4.43 `perror'--print an error message on standard error
=======================================================
*Synopsis*
#include <stdio.h>
void perror(char *PREFIX);
void _perror_r(struct _reent *REENT, char *PREFIX);
*Description*
Use `perror' to print (on standard error) an error message
corresponding to the current value of the global variable `errno'.
Unless you use `NULL' as the value of the argument PREFIX, the error
message will begin with the string at PREFIX, followed by a colon and a
space (`: '). The remainder of the error message is one of the strings
described for `strerror'.
The alternate function `_perror_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`perror' returns no result.
*Portability*
ANSI C requires `perror', but the strings issued vary from one
implementation to another.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: putc, Next: putc_unlocked, Prev: perror, Up: Stdio
4.44 `putc'--write a character (macro)
======================================
*Synopsis*
#include <stdio.h>
int putc(int CH, FILE *FP);
#include <stdio.h>
int _putc_r(struct _reent *PTR, int CH, FILE *FP);
*Description*
`putc' is a macro, defined in `stdio.h'. `putc' writes the argument CH
to the file or stream identified by FP, after converting it from an
`int' to an `unsigned char'.
If the file was opened with append mode (or if the stream cannot
support positioning), then the new character goes at the end of the
file or stream. Otherwise, the new character is written at the current
value of the position indicator, and the position indicator advances by
one.
For a subroutine version of this macro, see `fputc'.
The `_putc_r' function is simply the reentrant version of `putc'
that takes an additional reentrant structure argument: PTR.
*Returns*
If successful, `putc' returns its argument CH. If an error intervenes,
the result is `EOF'. You can use ``ferror(FP)'' to query for errors.
*Portability*
ANSI C requires `putc'; it suggests, but does not require, that `putc'
be implemented as a macro. The standard explicitly permits macro
implementations of `putc' to use the FP argument more than once;
therefore, in a portable program, you should not use an expression with
side effects as this argument.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: putc_unlocked, Next: putchar, Prev: putc, Up: Stdio
4.45 `putc_unlocked'--non-thread-safe version of putc (macro)
=============================================================
*Synopsis*
#include <stdio.h>
int putc_unlocked(int CH, FILE *FP);
#include <stdio.h>
int _putc_unlocked_r(struct _reent *PTR, int CH, FILE *FP);
*Description*
`putc_unlocked' is a non-thread-safe version of `putc' declared in
`stdio.h'. `putc_unlocked' may only safely be used within a scope
protected by flockfile() (or ftrylockfile()) and funlockfile(). These
functions may safely be used in a multi-threaded program if and only if
they are called while the invoking thread owns the ( FILE *) object, as
is the case after a successful call to the flockfile() or
ftrylockfile() functions. If threads are disabled, then
`putc_unlocked' is equivalent to `putc'.
The function `_putc_unlocked_r' is simply the reentrant version of
`putc_unlocked' that takes an additional reentrant structure pointer
argument: PTR.
*Returns*
See `putc'.
*Portability*
POSIX 1003.1 requires `putc_unlocked'. `putc_unlocked' may be
implemented as a macro, so arguments should not have side-effects.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: putchar, Next: putchar_unlocked, Prev: putc_unlocked, Up: Stdio
4.46 `putchar'--write a character (macro)
=========================================
*Synopsis*
#include <stdio.h>
int putchar(int CH);
int _putchar_r(struct _reent *REENT, int CH);
*Description*
`putchar' is a macro, defined in `stdio.h'. `putchar' writes its
argument to the standard output stream, after converting it from an
`int' to an `unsigned char'.
The alternate function `_putchar_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
If successful, `putchar' returns its argument CH. If an error
intervenes, the result is `EOF'. You can use ``ferror(stdin)'' to
query for errors.
*Portability*
ANSI C requires `putchar'; it suggests, but does not require, that
`putchar' be implemented as a macro.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: putchar_unlocked, Next: puts, Prev: putchar, Up: Stdio
4.47 `putchar_unlocked'--non-thread-safe version of putchar (macro)
===================================================================
*Synopsis*
#include <stdio.h>
int putchar_unlocked(int CH);
*Description*
`putchar_unlocked' is a non-thread-safe version of `putchar' declared
in `stdio.h'. `putchar_unlocked' may only safely be used within a
scope protected by flockfile() (or ftrylockfile()) and funlockfile().
These functions may safely be used in a multi-threaded program if and
only if they are called while the invoking thread owns the ( FILE *)
object, as is the case after a successful call to the flockfile() or
ftrylockfile() functions. If threads are disabled, then
`putchar_unlocked' is equivalent to `putchar'.
*Returns*
See `putchar'.
*Portability*
POSIX 1003.1 requires `putchar_unlocked'. `putchar_unlocked' may be
implemented as a macro.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: puts, Next: putw, Prev: putchar_unlocked, Up: Stdio
4.48 `puts'--write a character string
=====================================
*Synopsis*
#include <stdio.h>
int puts(const char *S);
int _puts_r(struct _reent *REENT, const char *S);
*Description*
`puts' writes the string at S (followed by a newline, instead of the
trailing null) to the standard output stream.
The alternate function `_puts_r' is a reentrant version. The extra
argument REENT is a pointer to a reentrancy structure.
*Returns*
If successful, the result is a nonnegative integer; otherwise, the
result is `EOF'.
*Portability*
ANSI C requires `puts', but does not specify that the result on success
must be `0'; any non-negative value is permitted.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: putw, Next: putwchar, Prev: puts, Up: Stdio
4.49 `putw'--write a word (int)
===============================
*Synopsis*
#include <stdio.h>
int putw(int W, FILE *FP);
*Description*
`putw' is a function, defined in `stdio.h'. You can use `putw' to
write a word to the file or stream identified by FP. As a side effect,
`putw' advances the file's current position indicator.
*Returns*
Zero on success, `EOF' on failure.
*Portability*
`putw' is a remnant of K&R C; it is not part of any ISO C Standard.
`fwrite' should be used instead. In fact, this implementation of
`putw' is based upon `fwrite'.
Supporting OS subroutines required: `fwrite'.

File: libc.info, Node: putwchar, Next: remove, Prev: putw, Up: Stdio
4.50 `putwchar'--write a wide character to standard output
==========================================================
*Synopsis*
#include <wchar.h>
wint_t putwchar(wchar_t WC);
wint_t _putwchar_r(struct _reent *REENT, wchar_t WC);
*Description*
The `putwchar' function or macro is the wide-character equivalent of
the `putchar' function. It writes the wide character wc to stdout.
The alternate function `_putwchar_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
If successful, `putwchar' returns its argument WC. If an error
intervenes, the result is `EOF'. You can use ``ferror(stdin)'' to
query for errors.
*Portability*
C99

File: libc.info, Node: remove, Next: rename, Prev: putwchar, Up: Stdio
4.51 `remove'--delete a file's name
===================================
*Synopsis*
#include <stdio.h>
int remove(char *FILENAME);
int _remove_r(struct _reent *REENT, char *FILENAME);
*Description*
Use `remove' to dissolve the association between a particular filename
(the string at FILENAME) and the file it represents. After calling
`remove' with a particular filename, you will no longer be able to open
the file by that name.
In this implementation, you may use `remove' on an open file without
error; existing file descriptors for the file will continue to access
the file's data until the program using them closes the file.
The alternate function `_remove_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
`remove' returns `0' if it succeeds, `-1' if it fails.
*Portability*
ANSI C requires `remove', but only specifies that the result on failure
be nonzero. The behavior of `remove' when you call it on an open file
may vary among implementations.
Supporting OS subroutine required: `unlink'.

File: libc.info, Node: rename, Next: rewind, Prev: remove, Up: Stdio
4.52 `rename'--rename a file
============================
*Synopsis*
#include <stdio.h>
int rename(const char *OLD, const char *NEW);
*Description*
Use `rename' to establish a new name (the string at NEW) for a file now
known by the string at OLD. After a successful `rename', the file is
no longer accessible by the string at OLD.
If `rename' fails, the file named `*OLD' is unaffected. The
conditions for failure depend on the host operating system.
*Returns*
The result is either `0' (when successful) or `-1' (when the file could
not be renamed).
*Portability*
ANSI C requires `rename', but only specifies that the result on failure
be nonzero. The effects of using the name of an existing file as
`*NEW' may vary from one implementation to another.
Supporting OS subroutines required: `link', `unlink', or `rename'.

File: libc.info, Node: rewind, Next: setbuf, Prev: rename, Up: Stdio
4.53 `rewind'--reinitialize a file or stream
============================================
*Synopsis*
#include <stdio.h>
void rewind(FILE *FP);
void _rewind_r(struct _reent *PTR, FILE *FP);
*Description*
`rewind' returns the file position indicator (if any) for the file or
stream identified by FP to the beginning of the file. It also clears
any error indicator and flushes any pending output.
*Returns*
`rewind' does not return a result.
*Portability*
ANSI C requires `rewind'.
No supporting OS subroutines are required.

File: libc.info, Node: setbuf, Next: setbuffer, Prev: rewind, Up: Stdio
4.54 `setbuf'--specify full buffering for a file or stream
==========================================================
*Synopsis*
#include <stdio.h>
void setbuf(FILE *FP, char *BUF);
*Description*
`setbuf' specifies that output to the file or stream identified by FP
should be fully buffered. All output for this file will go to a buffer
(of size `BUFSIZ', specified in ``stdio.h''). Output will be passed on
to the host system only when the buffer is full, or when an input
operation intervenes.
You may, if you wish, supply your own buffer by passing a pointer to
it as the argument BUF. It must have size `BUFSIZ'. You can also use
`NULL' as the value of BUF, to signal that the `setbuf' function is to
allocate the buffer.
*Warnings*
You may only use `setbuf' before performing any file operation other
than opening the file.
If you supply a non-null BUF, you must ensure that the associated
storage continues to be available until you close the stream identified
by FP.
*Returns*
`setbuf' does not return a result.
*Portability*
Both ANSI C and the System V Interface Definition (Issue 2) require
`setbuf'. However, they differ on the meaning of a `NULL' buffer
pointer: the SVID issue 2 specification says that a `NULL' buffer
pointer requests unbuffered output. For maximum portability, avoid
`NULL' buffer pointers.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: setbuffer, Next: setlinebuf, Prev: setbuf, Up: Stdio
4.55 `setbuffer'--specify full buffering for a file or stream with size
=======================================================================
*Synopsis*
#include <stdio.h>
void setbuffer(FILE *FP, char *BUF, int SIZE);
*Description*
`setbuffer' specifies that output to the file or stream identified by
FP should be fully buffered. All output for this file will go to a
buffer (of size SIZE). Output will be passed on to the host system
only when the buffer is full, or when an input operation intervenes.
You may, if you wish, supply your own buffer by passing a pointer to
it as the argument BUF. It must have size SIZE. You can also use
`NULL' as the value of BUF, to signal that the `setbuffer' function is
to allocate the buffer.
*Warnings*
You may only use `setbuffer' before performing any file operation other
than opening the file.
If you supply a non-null BUF, you must ensure that the associated
storage continues to be available until you close the stream identified
by FP.
*Returns*
`setbuffer' does not return a result.
*Portability*
This function comes from BSD not ANSI or POSIX.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: setlinebuf, Next: setvbuf, Prev: setbuffer, Up: Stdio
4.56 `setlinebuf'--specify line buffering for a file or stream
==============================================================
*Synopsis*
#include <stdio.h>
void setlinebuf(FILE *FP);
*Description*
`setlinebuf' specifies that output to the file or stream identified by
FP should be line buffered. This causes the file or stream to pass on
output to the host system at every newline, as well as when the buffer
is full, or when an input operation intervenes.
*Warnings*
You may only use `setlinebuf' before performing any file operation
other than opening the file.
*Returns*
`setlinebuf' returns as per setvbuf.
*Portability*
This function comes from BSD not ANSI or POSIX.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: setvbuf, Next: siprintf, Prev: setlinebuf, Up: Stdio
4.57 `setvbuf'--specify file or stream buffering
================================================
*Synopsis*
#include <stdio.h>
int setvbuf(FILE *FP, char *BUF,
int MODE, size_t SIZE);
*Description*
Use `setvbuf' to specify what kind of buffering you want for the file
or stream identified by FP, by using one of the following values (from
`stdio.h') as the MODE argument:
`_IONBF'
Do not use a buffer: send output directly to the host system for
the file or stream identified by FP.
`_IOFBF'
Use full output buffering: output will be passed on to the host
system only when the buffer is full, or when an input operation
intervenes.
`_IOLBF'
Use line buffering: pass on output to the host system at every
newline, as well as when the buffer is full, or when an input
operation intervenes.
Use the SIZE argument to specify how large a buffer you wish. You
can supply the buffer itself, if you wish, by passing a pointer to a
suitable area of memory as BUF. Otherwise, you may pass `NULL' as the
BUF argument, and `setvbuf' will allocate the buffer.
*Warnings*
You may only use `setvbuf' before performing any file operation other
than opening the file.
If you supply a non-null BUF, you must ensure that the associated
storage continues to be available until you close the stream identified
by FP.
*Returns*
A `0' result indicates success, `EOF' failure (invalid MODE or SIZE can
cause failure).
*Portability*
Both ANSI C and the System V Interface Definition (Issue 2) require
`setvbuf'. However, they differ on the meaning of a `NULL' buffer
pointer: the SVID issue 2 specification says that a `NULL' buffer
pointer requests unbuffered output. For maximum portability, avoid
`NULL' buffer pointers.
Both specifications describe the result on failure only as a nonzero
value.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: siprintf, Next: siscanf, Prev: setvbuf, Up: Stdio
4.58 `siprintf', `fiprintf', `iprintf', `sniprintf', `asiprintf', `asniprintf'--format output (integer only)
============================================================================================================
*Synopsis*
#include <stdio.h>
int iprintf(const char *FORMAT, ...);
int fiprintf(FILE *FD, const char *FORMAT , ...);
int siprintf(char *STR, const char *FORMAT, ...);
int sniprintf(char *STR, size_t SIZE, const char *FORMAT,
...);
int asiprintf(char **STRP, const char *FORMAT, ...);
char *asniprintf(char *STR, size_t *SIZE,
const char *FORMAT, ...);
int _iprintf_r(struct _reent *PTR, const char *FORMAT, ...);
int _fiprintf_r(struct _reent *PTR, FILE *FD,
const char *FORMAT, ...);
int _siprintf_r(struct _reent *PTR, char *STR,
const char *FORMAT, ...);
int _sniprintf_r(struct _reent *PTR, char *STR, size_t SIZE,
const char *FORMAT, ...);
int _asiprintf_r(struct _reent *PTR, char **STRP,
const char *FORMAT, ...);
char *_asniprintf_r(struct _reent *PTR, char *STR,
size_t *SIZE, const char *FORMAT, ...);
*Description*
`iprintf', `fiprintf', `siprintf', `sniprintf', `asiprintf', and
`asniprintf' are the same as `printf', `fprintf', `sprintf',
`snprintf', `asprintf', and `asnprintf', respectively, except that they
restrict usage to non-floating-point format specifiers.
`_iprintf_r', `_fiprintf_r', `_asiprintf_r', `_siprintf_r',
`_sniprintf_r', `_asniprintf_r' are simply reentrant versions of the
functions above.
*Returns*
Similar to `printf', `fprintf', `sprintf', `snprintf', `asprintf', and
`asnprintf'.
*Portability*
`iprintf', `fiprintf', `siprintf', `sniprintf', `asiprintf', and
`asniprintf' are newlib extensions.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: siscanf, Next: sprintf, Prev: siprintf, Up: Stdio
4.59 `siscanf', `fiscanf', `iscanf'--scan and format non-floating input
=======================================================================
*Synopsis*
#include <stdio.h>
int iscanf(const char *FORMAT, ...);
int fiscanf(FILE *FD, const char *FORMAT, ...);
int siscanf(const char *STR, const char *FORMAT, ...);
int _iscanf_r(struct _reent *PTR, const char *FORMAT, ...);
int _fiscanf_r(struct _reent *PTR, FILE *FD,
const char *FORMAT, ...);
int _siscanf_r(struct _reent *PTR, const char *STR,
const char *FORMAT, ...);
*Description*
`iscanf', `fiscanf', and `siscanf' are the same as `scanf', `fscanf',
and `sscanf' respectively, only that they restrict the available
formats to non-floating-point format specifiers.
The routines `_iscanf_r', `_fiscanf_r', and `_siscanf_r' are
reentrant versions of `iscanf', `fiscanf', and `siscanf' that take an
additional first argument pointing to a reentrancy structure.
*Returns*
`iscanf' returns the number of input fields successfully scanned,
converted and stored; the return value does not include scanned fields
which were not stored.
If `iscanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
*Portability*
`iscanf', `fiscanf', and `siscanf' are newlib extensions.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: sprintf, Next: sscanf, Prev: siscanf, Up: Stdio
4.60 `sprintf', `fprintf', `printf', `snprintf', `asprintf', `asnprintf'--format output
=======================================================================================
*Synopsis*
#include <stdio.h>
int printf(const char *FORMAT, ...);
int fprintf(FILE *FD, const char *FORMAT, ...);
int sprintf(char *STR, const char *FORMAT, ...);
int snprintf(char *STR, size_t SIZE, const char *FORMAT,
...);
int asprintf(char **STRP, const char *FORMAT, ...);
char *asnprintf(char *STR, size_t *SIZE, const char *FORMAT,
...);
int _printf_r(struct _reent *PTR, const char *FORMAT, ...);
int _fprintf_r(struct _reent *PTR, FILE *FD,
const char *FORMAT, ...);
int _sprintf_r(struct _reent *PTR, char *STR,
const char *FORMAT, ...);
int _snprintf_r(struct _reent *PTR, char *STR, size_t SIZE,
const char *FORMAT, ...);
int _asprintf_r(struct _reent *PTR, char **STRP,
const char *FORMAT, ...);
char *_asnprintf_r(struct _reent *PTR, char *STR,
size_t *SIZE, const char *FORMAT, ...);
*Description*
`printf' accepts a series of arguments, applies to each a format
specifier from `*FORMAT', and writes the formatted data to `stdout',
without a terminating NUL character. The behavior of `printf' is
undefined if there are not enough arguments for the format. `printf'
returns when it reaches the end of the format string. If there are
more arguments than the format requires, excess arguments are ignored.
`fprintf' is like `printf', except that output is directed to the
stream FD rather than `stdout'.
`sprintf' is like `printf', except that output is directed to the
buffer STR, and a terminating NUL is output. Behavior is undefined if
more output is generated than the buffer can hold.
`snprintf' is like `sprintf', except that output is limited to at
most SIZE bytes, including the terminating `NUL'. As a special case,
if SIZE is 0, STR can be NULL, and `snprintf' merely calculates how
many bytes would be printed.
`asprintf' is like `sprintf', except that the output is stored in a
dynamically allocated buffer, PSTR, which should be freed later with
`free'.
`asnprintf' is like `sprintf', except that the return type is either
the original STR if it was large enough, or a dynamically allocated
string if the output exceeds *SIZE; the length of the result is
returned in *SIZE. When dynamic allocation occurs, the contents of the
original STR may have been modified.
For `sprintf', `snprintf', and `asnprintf', the behavior is
undefined if the output `*STR' overlaps with one of the arguments.
Behavior is also undefined if the argument for `%n' within `*FORMAT'
overlaps another argument.
FORMAT is a pointer to a character string containing two types of
objects: ordinary characters (other than `%'), which are copied
unchanged to the output, and conversion specifications, each of which
is introduced by `%'. (To include `%' in the output, use `%%' in the
format string.) A conversion specification has the following form:
%[POS][FLAGS][WIDTH][.PREC][SIZE]TYPE
The fields of the conversion specification have the following
meanings:
* POS
Conversions normally consume arguments in the order that they are
presented. However, it is possible to consume arguments out of
order, and reuse an argument for more than one conversion
specification (although the behavior is undefined if the same
argument is requested with different types), by specifying POS,
which is a decimal integer followed by '$'. The integer must be
between 1 and <NL_ARGMAX> from limits.h, and if argument `%n$' is
requested, all earlier arguments must be requested somewhere
within FORMAT. If positional parameters are used, then all
conversion specifications except for `%%' must specify a position.
This positional parameters method is a POSIX extension to the C
standard definition for the functions.
* FLAGS
FLAGS is an optional sequence of characters which control output
justification, numeric signs, decimal points, trailing zeros, and
octal and hex prefixes. The flag characters are minus (`-'), plus
(`+'), space ( ), zero (`0'), sharp (`#'), and quote (`''). They
can appear in any combination, although not all flags can be used
for all conversion specification types.
`''
A POSIX extension to the C standard. However, this
implementation presently treats it as a no-op, which is the
default behavior for the C locale, anyway. (If it did what
it is supposed to, when TYPE were `i', `d', `u', `f', `F',
`g', or `G', the integer portion of the conversion would be
formatted with thousands' grouping wide characters.)
`-'
The result of the conversion is left justified, and the right
is padded with blanks. If you do not use this flag, the
result is right justified, and padded on the left.
`+'
The result of a signed conversion (as determined by TYPE of
`d', `i', `a', `A', `e', `E', `f', `F', `g', or `G') will
always begin with a plus or minus sign. (If you do not use
this flag, positive values do not begin with a plus sign.)
`" " (space)'
If the first character of a signed conversion specification
is not a sign, or if a signed conversion results in no
characters, the result will begin with a space. If the space
( ) flag and the plus (`+') flag both appear, the space flag
is ignored.
`0'
If the TYPE character is `d', `i', `o', `u', `x', `X', `a',
`A', `e', `E', `f', `F', `g', or `G': leading zeros are used
to pad the field width (following any indication of sign or
base); no spaces are used for padding. If the zero (`0') and
minus (`-') flags both appear, the zero (`0') flag will be
ignored. For `d', `i', `o', `u', `x', and `X' conversions,
if a precision PREC is specified, the zero (`0') flag is
ignored.
Note that `0' is interpreted as a flag, not as the beginning
of a field width.
`#'
The result is to be converted to an alternative form,
according to the TYPE character:
`o'
Increases precision to force the first digit of the
result to be a zero.
`x'
A non-zero result will have a `0x' prefix.
`X'
A non-zero result will have a `0X' prefix.
`a, A, e, E, f, or F'
The result will always contain a decimal point even if
no digits follow the point. (Normally, a decimal point
appears only if a digit follows it.) Trailing zeros are
removed.
`g or G'
The result will always contain a decimal point even if
no digits follow the point. Trailing zeros are not
removed.
`all others'
Undefined.
* WIDTH
WIDTH is an optional minimum field width. You can either specify
it directly as a decimal integer, or indirectly by using instead
an asterisk (`*'), in which case an `int' argument is used as the
field width. If positional arguments are used, then the width
must also be specified positionally as `*m$', with m as a decimal
integer. Negative field widths are treated as specifying the
minus (`-') flag for left justfication, along with a positive
field width. The resulting format may be wider than the specified
width.
* PREC
PREC is an optional field; if present, it is introduced with ``.''
(a period). You can specify the precision either directly as a
decimal integer or indirectly by using an asterisk (`*'), in which
case an `int' argument is used as the precision. If positional
arguments are used, then the precision must also be specified
positionally as `*m$', with m as a decimal integer. Supplying a
negative precision is equivalent to omitting the precision. If
only a period is specified the precision is zero. The effect
depends on the conversion TYPE.
`d, i, o, u, x, or X'
Minimum number of digits to appear. If no precision is
given, defaults to 1.
`a or A'
Number of digits to appear after the decimal point. If no
precision is given, the precision defaults to the minimum
needed for an exact representation.
`e, E, f or F'
Number of digits to appear after the decimal point. If no
precision is given, the precision defaults to 6.
`g or G'
Maximum number of significant digits. A precision of 0 is
treated the same as a precision of 1. If no precision is
given, the precision defaults to 6.
`s or S'
Maximum number of characters to print from the string. If no
precision is given, the entire string is printed.
`all others'
undefined.
* SIZE
SIZE is an optional modifier that changes the data type that the
corresponding argument has. Behavior is unspecified if a size is
given that does not match the TYPE.
`hh'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument should be converted to a `signed char' or `unsigned
char' before printing.
With `n', specifies that the argument is a pointer to a
`signed char'.
`h'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument should be converted to a `short' or `unsigned short'
before printing.
With `n', specifies that the argument is a pointer to a
`short'.
`l'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `long' or `unsigned long'.
With `c', specifies that the argument has type `wint_t'.
With `s', specifies that the argument is a pointer to
`wchar_t'.
With `n', specifies that the argument is a pointer to a
`long'.
With `a', `A', `e', `E', `f', `F', `g', or `G', has no effect
(because of vararg promotion rules, there is no need to
distinguish between `float' and `double').
`ll'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `long long' or `unsigned long long'.
With `n', specifies that the argument is a pointer to a `long
long'.
`j'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is an `intmax_t' or `uintmax_t'.
With `n', specifies that the argument is a pointer to an
`intmax_t'.
`z'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `size_t'.
With `n', specifies that the argument is a pointer to a
`size_t'.
`t'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `ptrdiff_t'.
With `n', specifies that the argument is a pointer to a
`ptrdiff_t'.
`L'
With `a', `A', `e', `E', `f', `F', `g', or `G', specifies
that the argument is a `long double'.
* TYPE
TYPE specifies what kind of conversion `printf' performs. Here is
a table of these:
`%'
Prints the percent character (`%').
`c'
Prints ARG as single character. If the `l' size specifier is
in effect, a multibyte character is printed.
`C'
Short for `%lc'. A POSIX extension to the C standard.
`s'
Prints the elements of a pointer to `char' until the
precision or a null character is reached. If the `l' size
specifier is in effect, the pointer is to an array of
`wchar_t', and the string is converted to multibyte
characters before printing.
`S'
Short for `%ls'. A POSIX extension to the C standard.
`d or i'
Prints a signed decimal integer; takes an `int'. Leading
zeros are inserted as necessary to reach the precision. A
value of 0 with a precision of 0 produces an empty string.
`D'
Newlib extension, short for `%ld'.
`o'
Prints an unsigned octal integer; takes an `unsigned'.
Leading zeros are inserted as necessary to reach the
precision. A value of 0 with a precision of 0 produces an
empty string.
`O'
Newlib extension, short for `%lo'.
`u'
Prints an unsigned decimal integer; takes an `unsigned'.
Leading zeros are inserted as necessary to reach the
precision. A value of 0 with a precision of 0 produces an
empty string.
`U'
Newlib extension, short for `%lu'.
`x'
Prints an unsigned hexadecimal integer (using `abcdef' as
digits beyond `9'); takes an `unsigned'. Leading zeros are
inserted as necessary to reach the precision. A value of 0
with a precision of 0 produces an empty string.
`X'
Like `x', but uses `ABCDEF' as digits beyond `9'.
`f'
Prints a signed value of the form `[-]9999.9999', with the
precision determining how many digits follow the decimal
point; takes a `double' (remember that `float' promotes to
`double' as a vararg). The low order digit is rounded to
even. If the precision results in at most DECIMAL_DIG
digits, the result is rounded correctly; if more than
DECIMAL_DIG digits are printed, the result is only guaranteed
to round back to the original value.
If the value is infinite, the result is `inf', and no zero
padding is performed. If the value is not a number, the
result is `nan', and no zero padding is performed.
`F'
Like `f', but uses `INF' and `NAN' for non-finite numbers.
`e'
Prints a signed value of the form `[-]9.9999e[+|-]999'; takes
a `double'. The digit before the decimal point is non-zero
if the value is non-zero. The precision determines how many
digits appear between `.' and `e', and the exponent always
contains at least two digits. The value zero has an exponent
of zero. If the value is not finite, it is printed like `f'.
`E'
Like `e', but using `E' to introduce the exponent, and like
`F' for non-finite values.
`g'
Prints a signed value in either `f' or `e' form, based on the
given value and precision--an exponent less than -4 or
greater than the precision selects the `e' form. Trailing
zeros and the decimal point are printed only if necessary;
takes a `double'.
`G'
Like `g', except use `F' or `E' form.
`a'
Prints a signed value of the form `[-]0x1.ffffp[+|-]9'; takes
a `double'. The letters `abcdef' are used for digits beyond
`9'. The precision determines how many digits appear after
the decimal point. The exponent contains at least one digit,
and is a decimal value representing the power of 2; a value
of 0 has an exponent of 0. Non-finite values are printed
like `f'.
`A'
Like `a', except uses `X', `P', and `ABCDEF' instead of lower
case.
`n'
Takes a pointer to `int', and stores a count of the number of
bytes written so far. No output is created.
`p'
Takes a pointer to `void', and prints it in an
implementation-defined format. This implementation is
similar to `%#tx'), except that `0x' appears even for the
NULL pointer.
`m'
Prints the output of `strerror(errno)'; no argument is
required. A GNU extension.
`_printf_r', `_fprintf_r', `_asprintf_r', `_sprintf_r',
`_snprintf_r', `_asnprintf_r' are simply reentrant versions of the
functions above.
*Returns*
On success, `sprintf' and `asprintf' return the number of bytes in the
output string, except the concluding `NUL' is not counted. `snprintf'
returns the number of bytes that would be in the output string, except
the concluding `NUL' is not counted. `printf' and `fprintf' return the
number of characters transmitted. `asnprintf' returns the original STR
if there was enough room, otherwise it returns an allocated string.
If an error occurs, the result of `printf', `fprintf', `snprintf',
and `asprintf' is a negative value, and the result of `asnprintf' is
NULL. No error returns occur for `sprintf'. For `printf' and
`fprintf', `errno' may be set according to `fputc'. For `asprintf' and
`asnprintf', `errno' may be set to ENOMEM if allocation fails, and for
`snprintf', `errno' may be set to EOVERFLOW if SIZE or the output
length exceeds INT_MAX.
*Bugs*
The ""' (quote) flag does not work when locale's thousands_sep is not
empty.
*Portability*
ANSI C requires `printf', `fprintf', `sprintf', and `snprintf'.
`asprintf' and `asnprintf' are newlib extensions.
The ANSI C standard specifies that implementations must support at
least formatted output of up to 509 characters. This implementation
has no inherent limit.
Depending on how newlib was configured, not all format specifiers are
supported.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: sscanf, Next: swprintf, Prev: sprintf, Up: Stdio
4.61 `sscanf', `fscanf', `scanf'--scan and format input
=======================================================
*Synopsis*
#include <stdio.h>
int scanf(const char *FORMAT, ...);
int fscanf(FILE *FD, const char *FORMAT, ...);
int sscanf(const char *STR, const char *FORMAT, ...);
int _scanf_r(struct _reent *PTR, const char *FORMAT, ...);
int _fscanf_r(struct _reent *PTR, FILE *FD,
const char *FORMAT, ...);
int _sscanf_r(struct _reent *PTR, const char *STR,
const char *FORMAT, ...);
*Description*
`scanf' scans a series of input fields from standard input, one
character at a time. Each field is interpreted according to a format
specifier passed to `scanf' in the format string at `*FORMAT'. `scanf'
stores the interpreted input from each field at the address passed to
it as the corresponding argument following FORMAT. You must supply the
same number of format specifiers and address arguments as there are
input fields.
There must be sufficient address arguments for the given format
specifiers; if not the results are unpredictable and likely
disasterous. Excess address arguments are merely ignored.
`scanf' often produces unexpected results if the input diverges from
an expected pattern. Since the combination of `gets' or `fgets'
followed by `sscanf' is safe and easy, that is the preferred way to be
certain that a program is synchronized with input at the end of a line.
`fscanf' and `sscanf' are identical to `scanf', other than the
source of input: `fscanf' reads from a file, and `sscanf' from a string.
The routines `_scanf_r', `_fscanf_r', and `_sscanf_r' are reentrant
versions of `scanf', `fscanf', and `sscanf' that take an additional
first argument pointing to a reentrancy structure.
The string at `*FORMAT' is a character sequence composed of zero or
more directives. Directives are composed of one or more whitespace
characters, non-whitespace characters, and format specifications.
Whitespace characters are blank (` '), tab (`\t'), or newline (`\n').
When `scanf' encounters a whitespace character in the format string it
will read (but not store) all consecutive whitespace characters up to
the next non-whitespace character in the input.
Non-whitespace characters are all other ASCII characters except the
percent sign (`%'). When `scanf' encounters a non-whitespace character
in the format string it will read, but not store a matching
non-whitespace character.
Format specifications tell `scanf' to read and convert characters
from the input field into specific types of values, and store then in
the locations specified by the address arguments.
Trailing whitespace is left unread unless explicitly matched in the
format string.
The format specifiers must begin with a percent sign (`%') and have
the following form:
%[*][WIDTH][SIZE]TYPE
Each format specification begins with the percent character (`%').
The other fields are:
`*'
an optional marker; if present, it suppresses interpretation and
assignment of this input field.
`WIDTH'
an optional maximum field width: a decimal integer, which controls
the maximum number of characters that will be read before
converting the current input field. If the input field has fewer
than WIDTH characters, `scanf' reads all the characters in the
field, and then proceeds with the next field and its format
specification.
If a whitespace or a non-convertable character occurs before WIDTH
character are read, the characters up to that character are read,
converted, and stored. Then `scanf' proceeds to the next format
specification.
`size'
`h', `j', `l', `L', `t', and `z' are optional size characters
which override the default way that `scanf' interprets the data
type of the corresponding argument.
Modifier Type(s)
hh d, i, o, u, x, n convert input to char,
store in char object
h d, i, o, u, x, n convert input to short,
store in short object
h D, I, O, U, X no effect
e, f, c, s, p
j d, i, o, u, x, n convert input to intmax_t,
store in intmax_t object
j all others no effect
l d, i, o, u, x, n convert input to long,
store in long object
l e, f, g convert input to double
store in a double object
l D, I, O, U, X no effect
c, s, p
ll d, i, o, u, x, n convert to long long,
store in long long
L d, i, o, u, x, n convert to long long,
store in long long
L e, f, g, E, G convert to long double,
store in long double
L all others no effect
t d, i, o, u, x, n convert input to ptrdiff_t,
store in ptrdiff_t object
t all others no effect
z d, i, o, u, x, n convert input to size_t,
store in size_t object
z all others no effect
`TYPE'
A character to specify what kind of conversion `scanf' performs.
Here is a table of the conversion characters:
`%'
No conversion is done; the percent character (`%') is stored.
`c'
Scans one character. Corresponding ARG: `(char *arg)'.
`s'
Reads a character string into the array supplied.
Corresponding ARG: `(char arg[])'.
`[PATTERN]'
Reads a non-empty character string into memory starting at
ARG. This area must be large enough to accept the sequence
and a terminating null character which will be added
automatically. (PATTERN is discussed in the paragraph
following this table). Corresponding ARG: `(char *arg)'.
`d'
Reads a decimal integer into the corresponding ARG: `(int
*arg)'.
`D'
Reads a decimal integer into the corresponding ARG: `(long
*arg)'.
`o'
Reads an octal integer into the corresponding ARG: `(int
*arg)'.
`O'
Reads an octal integer into the corresponding ARG: `(long
*arg)'.
`u'
Reads an unsigned decimal integer into the corresponding ARG:
`(unsigned int *arg)'.
`U'
Reads an unsigned decimal integer into the corresponding ARG:
`(unsigned long *arg)'.
`x,X'
Read a hexadecimal integer into the corresponding ARG: `(int
*arg)'.
`e, f, g'
Read a floating-point number into the corresponding ARG:
`(float *arg)'.
`E, F, G'
Read a floating-point number into the corresponding ARG:
`(double *arg)'.
`i'
Reads a decimal, octal or hexadecimal integer into the
corresponding ARG: `(int *arg)'.
`I'
Reads a decimal, octal or hexadecimal integer into the
corresponding ARG: `(long *arg)'.
`n'
Stores the number of characters read in the corresponding
ARG: `(int *arg)'.
`p'
Stores a scanned pointer. ANSI C leaves the details to each
implementation; this implementation treats `%p' exactly the
same as `%U'. Corresponding ARG: `(void **arg)'.
A PATTERN of characters surrounded by square brackets can be used
instead of the `s' type character. PATTERN is a set of characters
which define a search set of possible characters making up the
`scanf' input field. If the first character in the brackets is a
caret (`^'), the search set is inverted to include all ASCII
characters except those between the brackets. There is also a
range facility which you can use as a shortcut. `%[0-9] ' matches
all decimal digits. The hyphen must not be the first or last
character in the set. The character prior to the hyphen must be
lexically less than the character after it.
Here are some PATTERN examples:
`%[abcd]'
matches strings containing only `a', `b', `c', and `d'.
`%[^abcd]'
matches strings containing any characters except `a', `b',
`c', or `d'
`%[A-DW-Z]'
matches strings containing `A', `B', `C', `D', `W', `X', `Y',
`Z'
`%[z-a]'
matches the characters `z', `-', and `a'
Floating point numbers (for field types `e', `f', `g', `E', `F',
`G') must correspond to the following general form:
[+/-] ddddd[.]ddd [E|e[+|-]ddd]
where objects inclosed in square brackets are optional, and `ddd'
represents decimal, octal, or hexadecimal digits.
*Returns*
`scanf' returns the number of input fields successfully scanned,
converted and stored; the return value does not include scanned fields
which were not stored.
If `scanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
`scanf' might stop scanning a particular field before reaching the
normal field end character, or may terminate entirely.
`scanf' stops scanning and storing the current field and moves to
the next input field (if any) in any of the following situations:
* The assignment suppressing character (`*') appears after the `%'
in the format specification; the current input field is scanned
but not stored.
* WIDTH characters have been read (WIDTH is a width specification, a
positive decimal integer).
* The next character read cannot be converted under the the current
format (for example, if a `Z' is read when the format is decimal).
* The next character in the input field does not appear in the
search set (or does appear in the inverted search set).
When `scanf' stops scanning the current input field for one of these
reasons, the next character is considered unread and used as the first
character of the following input field, or the first character in a
subsequent read operation on the input.
`scanf' will terminate under the following circumstances:
* The next character in the input field conflicts with a
corresponding non-whitespace character in the format string.
* The next character in the input field is `EOF'.
* The format string has been exhausted.
When the format string contains a character sequence that is not
part of a format specification, the same character sequence must appear
in the input; `scanf' will scan but not store the matched characters.
If a conflict occurs, the first conflicting character remains in the
input as if it had never been read.
*Portability*
`scanf' is ANSI C.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: swprintf, Next: swscanf, Prev: sscanf, Up: Stdio
4.62 `swprintf', `fwprintf', `wprintf'--wide character format output
====================================================================
*Synopsis*
#include <wchar.h>
int wprintf(const wchar_t *FORMAT, ...);
int fwprintf(FILE *FD, const wchar_t *FORMAT, ...);
int swprintf(wchar_t *STR, size_t SIZE,
const wchar_t *FORMAT, ...);
int _wprintf_r(struct _reent *PTR, const wchar_t *FORMAT, ...);
int _fwprintf_r(struct _reent *PTR, FILE *FD,
const wchar_t *FORMAT, ...);
int _swprintf_r(struct _reent *PTR, wchar_t *STR,
size_t SIZE, const wchar_t *FORMAT, ...);
*Description*
`wprintf' accepts a series of arguments, applies to each a format
specifier from `*FORMAT', and writes the formatted data to `stdout',
without a terminating NUL wide character. The behavior of `wprintf' is
undefined if there are not enough arguments for the format or if any
argument is not the right type for the corresponding conversion
specifier. `wprintf' returns when it reaches the end of the format
string. If there are more arguments than the format requires, excess
arguments are ignored.
`fwprintf' is like `wprintf', except that output is directed to the
stream FD rather than `stdout'.
`swprintf' is like `wprintf', except that output is directed to the
buffer STR with a terminating wide `NUL', and the resulting string
length is limited to at most SIZE wide characters, including the
terminating `NUL'. It is considered an error if the output (including
the terminating wide-`NULL') does not fit into SIZE wide characters.
(This error behavior is not the same as for `snprintf', which
`swprintf' is otherwise completely analogous to. While `snprintf'
allows the needed size to be known simply by giving SIZE=0, `swprintf'
does not, giving an error instead.)
For `swprintf' the behavior is undefined if the output `*STR'
overlaps with one of the arguments. Behavior is also undefined if the
argument for `%n' within `*FORMAT' overlaps another argument.
FORMAT is a pointer to a wide character string containing two types
of objects: ordinary characters (other than `%'), which are copied
unchanged to the output, and conversion specifications, each of which
is introduced by `%'. (To include `%' in the output, use `%%' in the
format string.) A conversion specification has the following form:
%[POS][FLAGS][WIDTH][.PREC][SIZE]TYPE
The fields of the conversion specification have the following
meanings:
* POS
Conversions normally consume arguments in the order that they are
presented. However, it is possible to consume arguments out of
order, and reuse an argument for more than one conversion
specification (although the behavior is undefined if the same
argument is requested with different types), by specifying POS,
which is a decimal integer followed by '$'. The integer must be
between 1 and <NL_ARGMAX> from limits.h, and if argument `%n$' is
requested, all earlier arguments must be requested somewhere
within FORMAT. If positional parameters are used, then all
conversion specifications except for `%%' must specify a position.
This positional parameters method is a POSIX extension to the C
standard definition for the functions.
* FLAGS
FLAGS is an optional sequence of characters which control output
justification, numeric signs, decimal points, trailing zeros, and
octal and hex prefixes. The flag characters are minus (`-'), plus
(`+'), space ( ), zero (`0'), sharp (`#'), and quote (`''). They
can appear in any combination, although not all flags can be used
for all conversion specification types.
`''
A POSIX extension to the C standard. However, this
implementation presently treats it as a no-op, which is the
default behavior for the C locale, anyway. (If it did what
it is supposed to, when TYPE were `i', `d', `u', `f', `F',
`g', or `G', the integer portion of the conversion would be
formatted with thousands' grouping wide characters.)
`-'
The result of the conversion is left justified, and the right
is padded with blanks. If you do not use this flag, the
result is right justified, and padded on the left.
`+'
The result of a signed conversion (as determined by TYPE of
`d', `i', `a', `A', `e', `E', `f', `F', `g', or `G') will
always begin with a plus or minus sign. (If you do not use
this flag, positive values do not begin with a plus sign.)
`" " (space)'
If the first character of a signed conversion specification
is not a sign, or if a signed conversion results in no
characters, the result will begin with a space. If the space
( ) flag and the plus (`+') flag both appear, the space flag
is ignored.
`0'
If the TYPE character is `d', `i', `o', `u', `x', `X', `a',
`A', `e', `E', `f', `F', `g', or `G': leading zeros are used
to pad the field width (following any indication of sign or
base); no spaces are used for padding. If the zero (`0') and
minus (`-') flags both appear, the zero (`0') flag will be
ignored. For `d', `i', `o', `u', `x', and `X' conversions,
if a precision PREC is specified, the zero (`0') flag is
ignored.
Note that `0' is interpreted as a flag, not as the beginning
of a field width.
`#'
The result is to be converted to an alternative form,
according to the TYPE character:
`o'
Increases precision to force the first digit of the
result to be a zero.
`x'
A non-zero result will have a `0x' prefix.
`X'
A non-zero result will have a `0X' prefix.
`a, A, e, E, f, or F'
The result will always contain a decimal point even if
no digits follow the point. (Normally, a decimal point
appears only if a digit follows it.) Trailing zeros are
removed.
`g or G'
The result will always contain a decimal point even if
no digits follow the point. Trailing zeros are not
removed.
`all others'
Undefined.
* WIDTH
WIDTH is an optional minimum field width. You can either specify
it directly as a decimal integer, or indirectly by using instead
an asterisk (`*'), in which case an `int' argument is used as the
field width. If positional arguments are used, then the width
must also be specified positionally as `*m$', with m as a decimal
integer. Negative field widths are treated as specifying the
minus (`-') flag for left justfication, along with a positive
field width. The resulting format may be wider than the specified
width.
* PREC
PREC is an optional field; if present, it is introduced with ``.''
(a period). You can specify the precision either directly as a
decimal integer or indirectly by using an asterisk (`*'), in which
case an `int' argument is used as the precision. If positional
arguments are used, then the precision must also be specified
positionally as `*m$', with m as a decimal integer. Supplying a
negative precision is equivalent to omitting the precision. If
only a period is specified the precision is zero. The effect
depends on the conversion TYPE.
`d, i, o, u, x, or X'
Minimum number of digits to appear. If no precision is
given, defaults to 1.
`a or A'
Number of digits to appear after the decimal point. If no
precision is given, the precision defaults to the minimum
needed for an exact representation.
`e, E, f or F'
Number of digits to appear after the decimal point. If no
precision is given, the precision defaults to 6.
`g or G'
Maximum number of significant digits. A precision of 0 is
treated the same as a precision of 1. If no precision is
given, the precision defaults to 6.
`s or S'
Maximum number of characters to print from the string. If no
precision is given, the entire string is printed.
`all others'
undefined.
* SIZE
SIZE is an optional modifier that changes the data type that the
corresponding argument has. Behavior is unspecified if a size is
given that does not match the TYPE.
`hh'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument should be converted to a `signed char' or `unsigned
char' before printing.
With `n', specifies that the argument is a pointer to a
`signed char'.
`h'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument should be converted to a `short' or `unsigned short'
before printing.
With `n', specifies that the argument is a pointer to a
`short'.
`l'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `long' or `unsigned long'.
With `c', specifies that the argument has type `wint_t'.
With `s', specifies that the argument is a pointer to
`wchar_t'.
With `n', specifies that the argument is a pointer to a
`long'.
With `a', `A', `e', `E', `f', `F', `g', or `G', has no effect
(because of vararg promotion rules, there is no need to
distinguish between `float' and `double').
`ll'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `long long' or `unsigned long long'.
With `n', specifies that the argument is a pointer to a `long
long'.
`j'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is an `intmax_t' or `uintmax_t'.
With `n', specifies that the argument is a pointer to an
`intmax_t'.
`z'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `size_t'.
With `n', specifies that the argument is a pointer to a
`size_t'.
`t'
With `d', `i', `o', `u', `x', or `X', specifies that the
argument is a `ptrdiff_t'.
With `n', specifies that the argument is a pointer to a
`ptrdiff_t'.
`L'
With `a', `A', `e', `E', `f', `F', `g', or `G', specifies
that the argument is a `long double'.
* TYPE
TYPE specifies what kind of conversion `wprintf' performs. Here
is a table of these:
`%'
Prints the percent character (`%').
`c'
If no `l' qualifier is present, the int argument shall be
converted to a wide character as if by calling the btowc()
function and the resulting wide character shall be written.
Otherwise, the wint_t argument shall be converted to wchar_t,
and written.
`C'
Short for `%lc'. A POSIX extension to the C standard.
`s'
If no `l' qualifier is present, the application shall ensure
that the argument is a pointer to a character array
containing a character sequence beginning in the initial
shift state. Characters from the array shall be converted as
if by repeated calls to the mbrtowc() function, with the
conversion state described by an mbstate_t object initialized
to zero before the first character is converted, and written
up to (but not including) the terminating null wide
character. If the precision is specified, no more than that
many wide characters shall be written. If the precision is
not specified, or is greater than the size of the array, the
application shall ensure that the array contains a null wide
character.
If an `l' qualifier is present, the application shall ensure
that the argument is a pointer to an array of type wchar_t.
Wide characters from the array shall be written up to (but
not including) a terminating null wide character. If no
precision is specified, or is greater than the size of the
array, the application shall ensure that the array contains a
null wide character. If a precision is specified, no more
than that many wide characters shall be written.
`S'
Short for `%ls'. A POSIX extension to the C standard.
`d or i'
Prints a signed decimal integer; takes an `int'. Leading
zeros are inserted as necessary to reach the precision. A
value of 0 with a precision of 0 produces an empty string.
`o'
Prints an unsigned octal integer; takes an `unsigned'.
Leading zeros are inserted as necessary to reach the
precision. A value of 0 with a precision of 0 produces an
empty string.
`u'
Prints an unsigned decimal integer; takes an `unsigned'.
Leading zeros are inserted as necessary to reach the
precision. A value of 0 with a precision of 0 produces an
empty string.
`x'
Prints an unsigned hexadecimal integer (using `abcdef' as
digits beyond `9'); takes an `unsigned'. Leading zeros are
inserted as necessary to reach the precision. A value of 0
with a precision of 0 produces an empty string.
`X'
Like `x', but uses `ABCDEF' as digits beyond `9'.
`f'
Prints a signed value of the form `[-]9999.9999', with the
precision determining how many digits follow the decimal
point; takes a `double' (remember that `float' promotes to
`double' as a vararg). The low order digit is rounded to
even. If the precision results in at most DECIMAL_DIG
digits, the result is rounded correctly; if more than
DECIMAL_DIG digits are printed, the result is only guaranteed
to round back to the original value.
If the value is infinite, the result is `inf', and no zero
padding is performed. If the value is not a number, the
result is `nan', and no zero padding is performed.
`F'
Like `f', but uses `INF' and `NAN' for non-finite numbers.
`e'
Prints a signed value of the form `[-]9.9999e[+|-]999'; takes
a `double'. The digit before the decimal point is non-zero
if the value is non-zero. The precision determines how many
digits appear between `.' and `e', and the exponent always
contains at least two digits. The value zero has an exponent
of zero. If the value is not finite, it is printed like `f'.
`E'
Like `e', but using `E' to introduce the exponent, and like
`F' for non-finite values.
`g'
Prints a signed value in either `f' or `e' form, based on the
given value and precision--an exponent less than -4 or
greater than the precision selects the `e' form. Trailing
zeros and the decimal point are printed only if necessary;
takes a `double'.
`G'
Like `g', except use `F' or `E' form.
`a'
Prints a signed value of the form `[-]0x1.ffffp[+|-]9'; takes
a `double'. The letters `abcdef' are used for digits beyond
`9'. The precision determines how many digits appear after
the decimal point. The exponent contains at least one digit,
and is a decimal value representing the power of 2; a value
of 0 has an exponent of 0. Non-finite values are printed
like `f'.
`A'
Like `a', except uses `X', `P', and `ABCDEF' instead of lower
case.
`n'
Takes a pointer to `int', and stores a count of the number of
bytes written so far. No output is created.
`p'
Takes a pointer to `void', and prints it in an
implementation-defined format. This implementation is
similar to `%#tx'), except that `0x' appears even for the
NULL pointer.
`m'
Prints the output of `strerror(errno)'; no argument is
required. A GNU extension.
`_wprintf_r', `_fwprintf_r', `_swprintf_r', are simply reentrant
versions of the functions above.
*Returns*
On success, `swprintf' return the number of wide characters in the
output string, except the concluding `NUL' is not counted. `wprintf'
and `fwprintf' return the number of characters transmitted.
If an error occurs, the result of `wprintf', `fwprintf', and
`swprintf' is a negative value. For `wprintf' and `fwprintf', `errno'
may be set according to `fputwc'. For `swprintf', `errno' may be set
to EOVERFLOW if SIZE is greater than INT_MAX / sizeof (wchar_t), or
when the output does not fit into SIZE wide characters (including the
terminating wide `NULL').
*Bugs*
The ""' (quote) flag does not work when locale's thousands_sep is not
empty.
*Portability*
POSIX-1.2008 with extensions; C99 (compliant except for POSIX
extensions).
Depending on how newlib was configured, not all format specifiers are
supported.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: swscanf, Next: tmpfile, Prev: swprintf, Up: Stdio
4.63 `swscanf', `fwscanf', `wscanf'--scan and format wide character input
=========================================================================
*Synopsis*
#include <stdio.h>
int wscanf(const wchar_t *FORMAT, ...);
int fwscanf(FILE *FD, const wchar_t *FORMAT, ...);
int swscanf(const wchar_t *STR, const wchar_t *FORMAT, ...);
int _wscanf_r(struct _reent *PTR, const wchar_t *FORMAT, ...);
int _fwscanf_r(struct _reent *PTR, FILE *FD,
const wchar_t *FORMAT, ...);
int _swscanf_r(struct _reent *PTR, const wchar_t *STR,
const wchar_t *FORMAT, ...);
*Description*
`wscanf' scans a series of input fields from standard input, one wide
character at a time. Each field is interpreted according to a format
specifier passed to `wscanf' in the format string at `*FORMAT'.
`wscanf' stores the interpreted input from each field at the address
passed to it as the corresponding argument following FORMAT. You must
supply the same number of format specifiers and address arguments as
there are input fields.
There must be sufficient address arguments for the given format
specifiers; if not the results are unpredictable and likely
disasterous. Excess address arguments are merely ignored.
`wscanf' often produces unexpected results if the input diverges from
an expected pattern. Since the combination of `gets' or `fgets'
followed by `swscanf' is safe and easy, that is the preferred way to be
certain that a program is synchronized with input at the end of a line.
`fwscanf' and `swscanf' are identical to `wscanf', other than the
source of input: `fwscanf' reads from a file, and `swscanf' from a
string.
The routines `_wscanf_r', `_fwscanf_r', and `_swscanf_r' are
reentrant versions of `wscanf', `fwscanf', and `swscanf' that take an
additional first argument pointing to a reentrancy structure.
The string at `*FORMAT' is a wide character sequence composed of
zero or more directives. Directives are composed of one or more
whitespace characters, non-whitespace characters, and format
specifications.
Whitespace characters are blank (` '), tab (`\t'), or newline (`\n').
When `wscanf' encounters a whitespace character in the format string it
will read (but not store) all consecutive whitespace characters up to
the next non-whitespace character in the input.
Non-whitespace characters are all other ASCII characters except the
percent sign (`%'). When `wscanf' encounters a non-whitespace
character in the format string it will read, but not store a matching
non-whitespace character.
Format specifications tell `wscanf' to read and convert characters
from the input field into specific types of values, and store then in
the locations specified by the address arguments.
Trailing whitespace is left unread unless explicitly matched in the
format string.
The format specifiers must begin with a percent sign (`%') and have
the following form:
%[*][WIDTH][SIZE]TYPE
Each format specification begins with the percent character (`%').
The other fields are:
`*'
an optional marker; if present, it suppresses interpretation and
assignment of this input field.
`WIDTH'
an optional maximum field width: a decimal integer, which controls
the maximum number of characters that will be read before
converting the current input field. If the input field has fewer
than WIDTH characters, `wscanf' reads all the characters in the
field, and then proceeds with the next field and its format
specification.
If a whitespace or a non-convertable wide character occurs before
WIDTH character are read, the characters up to that character are
read, converted, and stored. Then `wscanf' proceeds to the next
format specification.
`size'
`h', `j', `l', `L', `t', and `z' are optional size characters
which override the default way that `wscanf' interprets the data
type of the corresponding argument.
Modifier Type(s)
hh d, i, o, u, x, n convert input to char,
store in char object
h d, i, o, u, x, n convert input to short,
store in short object
h e, f, c, s, p no effect
j d, i, o, u, x, n convert input to intmax_t,
store in intmax_t object
j all others no effect
l d, i, o, u, x, n convert input to long,
store in long object
l e, f, g convert input to double
store in a double object
l c, s, [ the input is stored in a wchar_t object
l p no effect
ll d, i, o, u, x, n convert to long long,
store in long long
L d, i, o, u, x, n convert to long long,
store in long long
L e, f, g, E, G convert to long double,
store in long double
L all others no effect
t d, i, o, u, x, n convert input to ptrdiff_t,
store in ptrdiff_t object
t all others no effect
z d, i, o, u, x, n convert input to size_t,
store in size_t object
z all others no effect
`TYPE'
A character to specify what kind of conversion `wscanf' performs.
Here is a table of the conversion characters:
`%'
No conversion is done; the percent character (`%') is stored.
`c'
Scans one wide character. Corresponding ARG: `(char *arg)'.
Otherwise, if an `l' specifier is present, the corresponding
ARG is a `(wchar_t *arg)'.
`s'
Reads a character string into the array supplied.
Corresponding ARG: `(char arg[])'. If an `l' specifier is
present, the corresponding ARG is a `(wchar_t *arg)'.
`[PATTERN]'
Reads a non-empty character string into memory starting at
ARG. This area must be large enough to accept the sequence
and a terminating null character which will be added
automatically. (PATTERN is discussed in the paragraph
following this table). Corresponding ARG: `(char *arg)'. If
an `l' specifier is present, the corresponding ARG is a
`(wchar_t *arg)'.
`d'
Reads a decimal integer into the corresponding ARG: `(int
*arg)'.
`o'
Reads an octal integer into the corresponding ARG: `(int
*arg)'.
`u'
Reads an unsigned decimal integer into the corresponding ARG:
`(unsigned int *arg)'.
`x,X'
Read a hexadecimal integer into the corresponding ARG: `(int
*arg)'.
`e, f, g'
Read a floating-point number into the corresponding ARG:
`(float *arg)'.
`E, F, G'
Read a floating-point number into the corresponding ARG:
`(double *arg)'.
`i'
Reads a decimal, octal or hexadecimal integer into the
corresponding ARG: `(int *arg)'.
`n'
Stores the number of characters read in the corresponding
ARG: `(int *arg)'.
`p'
Stores a scanned pointer. ANSI C leaves the details to each
implementation; this implementation treats `%p' exactly the
same as `%U'. Corresponding ARG: `(void **arg)'.
A PATTERN of characters surrounded by square brackets can be used
instead of the `s' type character. PATTERN is a set of characters
which define a search set of possible characters making up the
`wscanf' input field. If the first character in the brackets is a
caret (`^'), the search set is inverted to include all ASCII
characters except those between the brackets. There is no range
facility as is defined in the corresponding non-wide character
scanf functions. Ranges are not part of the POSIX standard.
Here are some PATTERN examples:
`%[abcd]'
matches wide character strings containing only `a', `b', `c',
and `d'.
`%[^abcd]'
matches wide character strings containing any characters
except `a', `b', `c', or `d'.
`%[A-DW-Z]'
Note: No wide character ranges, so this expression matches
wide character strings containing `A', `-', `D', `W', `Z'.
Floating point numbers (for field types `e', `f', `g', `E', `F',
`G') must correspond to the following general form:
[+/-] ddddd[.]ddd [E|e[+|-]ddd]
where objects inclosed in square brackets are optional, and `ddd'
represents decimal, octal, or hexadecimal digits.
*Returns*
`wscanf' returns the number of input fields successfully scanned,
converted and stored; the return value does not include scanned fields
which were not stored.
If `wscanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
`wscanf' might stop scanning a particular field before reaching the
normal field end character, or may terminate entirely.
`wscanf' stops scanning and storing the current field and moves to
the next input field (if any) in any of the following situations:
* The assignment suppressing character (`*') appears after the `%'
in the format specification; the current input field is scanned
but not stored.
* WIDTH characters have been read (WIDTH is a width specification, a
positive decimal integer).
* The next wide character read cannot be converted under the the
current format (for example, if a `Z' is read when the format is
decimal).
* The next wide character in the input field does not appear in the
search set (or does appear in the inverted search set).
When `wscanf' stops scanning the current input field for one of
these reasons, the next character is considered unread and used as the
first character of the following input field, or the first character in
a subsequent read operation on the input.
`wscanf' will terminate under the following circumstances:
* The next wide character in the input field conflicts with a
corresponding non-whitespace character in the format string.
* The next wide character in the input field is `WEOF'.
* The format string has been exhausted.
When the format string contains a wide character sequence that is
not part of a format specification, the same wide character sequence
must appear in the input; `wscanf' will scan but not store the matched
characters. If a conflict occurs, the first conflicting wide character
remains in the input as if it had never been read.
*Portability*
`wscanf' is C99, POSIX-1.2008.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: tmpfile, Next: tmpnam, Prev: swscanf, Up: Stdio
4.64 `tmpfile'--create a temporary file
=======================================
*Synopsis*
#include <stdio.h>
FILE *tmpfile(void);
FILE *_tmpfile_r(struct _reent *REENT);
*Description*
Create a temporary file (a file which will be deleted automatically),
using a name generated by `tmpnam'. The temporary file is opened with
the mode `"wb+"', permitting you to read and write anywhere in it as a
binary file (without any data transformations the host system may
perform for text files).
The alternate function `_tmpfile_r' is a reentrant version. The
argument REENT is a pointer to a reentrancy structure.
*Returns*
`tmpfile' normally returns a pointer to the temporary file. If no
temporary file could be created, the result is NULL, and `errno'
records the reason for failure.
*Portability*
Both ANSI C and the System V Interface Definition (Issue 2) require
`tmpfile'.
Supporting OS subroutines required: `close', `fstat', `getpid',
`isatty', `lseek', `open', `read', `sbrk', `write'.
`tmpfile' also requires the global pointer `environ'.

File: libc.info, Node: tmpnam, Next: ungetc, Prev: tmpfile, Up: Stdio
4.65 `tmpnam', `tempnam'--name for a temporary file
===================================================
*Synopsis*
#include <stdio.h>
char *tmpnam(char *S);
char *tempnam(char *DIR, char *PFX);
char *_tmpnam_r(struct _reent *REENT, char *S);
char *_tempnam_r(struct _reent *REENT, char *DIR, char *PFX);
*Description*
Use either of these functions to generate a name for a temporary file.
The generated name is guaranteed to avoid collision with other files
(for up to `TMP_MAX' calls of either function).
`tmpnam' generates file names with the value of `P_tmpdir' (defined
in ``stdio.h'') as the leading directory component of the path.
You can use the `tmpnam' argument S to specify a suitable area of
memory for the generated filename; otherwise, you can call
`tmpnam(NULL)' to use an internal static buffer.
`tempnam' allows you more control over the generated filename: you
can use the argument DIR to specify the path to a directory for
temporary files, and you can use the argument PFX to specify a prefix
for the base filename.
If DIR is `NULL', `tempnam' will attempt to use the value of
environment variable `TMPDIR' instead; if there is no such value,
`tempnam' uses the value of `P_tmpdir' (defined in ``stdio.h'').
If you don't need any particular prefix to the basename of temporary
files, you can pass `NULL' as the PFX argument to `tempnam'.
`_tmpnam_r' and `_tempnam_r' are reentrant versions of `tmpnam' and
`tempnam' respectively. The extra argument REENT is a pointer to a
reentrancy structure.
*Warnings*
The generated filenames are suitable for temporary files, but do not in
themselves make files temporary. Files with these names must still be
explicitly removed when you no longer want them.
If you supply your own data area S for `tmpnam', you must ensure
that it has room for at least `L_tmpnam' elements of type `char'.
*Returns*
Both `tmpnam' and `tempnam' return a pointer to the newly generated
filename.
*Portability*
ANSI C requires `tmpnam', but does not specify the use of `P_tmpdir'.
The System V Interface Definition (Issue 2) requires both `tmpnam' and
`tempnam'.
Supporting OS subroutines required: `close', `fstat', `getpid',
`isatty', `lseek', `open', `read', `sbrk', `write'.
The global pointer `environ' is also required.

File: libc.info, Node: ungetc, Next: ungetwc, Prev: tmpnam, Up: Stdio
4.66 `ungetc'--push data back into a stream
===========================================
*Synopsis*
#include <stdio.h>
int ungetc(int C, FILE *STREAM);
int _ungetc_r(struct _reent *REENT, int C, FILE *STREAM);
*Description*
`ungetc' is used to return bytes back to STREAM to be read again. If C
is EOF, the stream is unchanged. Otherwise, the unsigned char C is put
back on the stream, and subsequent reads will see the bytes pushed back
in reverse order. Pushed byes are lost if the stream is repositioned,
such as by `fseek', `fsetpos', or `rewind'.
The underlying file is not changed, but it is possible to push back
something different than what was originally read. Ungetting a
character will clear the end-of-stream marker, and decrement the file
position indicator. Pushing back beyond the beginning of a file gives
unspecified behavior.
The alternate function `_ungetc_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
The character pushed back, or `EOF' on error.
*Portability*
ANSI C requires `ungetc', but only requires a pushback buffer of one
byte; although this implementation can handle multiple bytes, not all
can. Pushing back a signed char is a common application bug.
Supporting OS subroutines required: `sbrk'.

File: libc.info, Node: ungetwc, Next: vfprintf, Prev: ungetc, Up: Stdio
4.67 `ungetwc'--push wide character data back into a stream
===========================================================
*Synopsis*
#include <stdio.h>
#include <wchar.h>
wint_t ungetwc(wint_t WC, FILE *STREAM);
wint_t _ungetwc_r(struct _reent *REENT, wint_t WC, FILE *STREAM);
*Description*
`ungetwc' is used to return wide characters back to STREAM to be read
again. If WC is WEOF, the stream is unchanged. Otherwise, the wide
character WC is put back on the stream, and subsequent reads will see
the wide chars pushed back in reverse order. Pushed wide chars are
lost if the stream is repositioned, such as by `fseek', `fsetpos', or
`rewind'.
The underlying file is not changed, but it is possible to push back
something different than what was originally read. Ungetting a
character will clear the end-of-stream marker, and decrement the file
position indicator. Pushing back beyond the beginning of a file gives
unspecified behavior.
The alternate function `_ungetwc_r' is a reentrant version. The
extra argument REENT is a pointer to a reentrancy structure.
*Returns*
The wide character pushed back, or `WEOF' on error.
*Portability*
C99

File: libc.info, Node: vfprintf, Next: vfscanf, Prev: ungetwc, Up: Stdio
4.68 `vfprintf', `vprintf', `vsprintf', `vsnprintf', `vasprintf', `vasnprintf'--format argument list
====================================================================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int vprintf(const char *FMT, va_list LIST);
int vfprintf(FILE *FP, const char *FMT, va_list LIST);
int vsprintf(char *STR, const char *FMT, va_list LIST);
int vsnprintf(char *STR, size_t SIZE, const char *FMT,
va_list LIST);
int vasprintf(char **STRP, const char *FMT, va_list LIST);
char *vasnprintf(char *STR, size_t *SIZE, const char *FMT,
va_list LIST);
int _vprintf_r(struct _reent *REENT, const char *FMT,
va_list LIST);
int _vfprintf_r(struct _reent *REENT, FILE *FP,
const char *FMT, va_list LIST);
int _vsprintf_r(struct _reent *REENT, char *STR,
const char *FMT, va_list LIST);
int _vasprintf_r(struct _reent *REENT, char **STR,
const char *FMT, va_list LIST);
int _vsnprintf_r(struct _reent *REENT, char *STR,
size_t SIZE, const char *FMT, va_list LIST);
char *_vasnprintf_r(struct _reent *REENT, char *STR,
size_t *SIZE, const char *FMT, va_list LIST);
*Description*
`vprintf', `vfprintf', `vasprintf', `vsprintf', `vsnprintf', and
`vasnprintf' are (respectively) variants of `printf', `fprintf',
`asprintf', `sprintf', `snprintf', and `asnprintf'. They differ only
in allowing their caller to pass the variable argument list as a
`va_list' object (initialized by `va_start') rather than directly
accepting a variable number of arguments. The caller is responsible
for calling `va_end'.
`_vprintf_r', `_vfprintf_r', `_vasprintf_r', `_vsprintf_r',
`_vsnprintf_r', and `_vasnprintf_r' are reentrant versions of the above.
*Returns*
The return values are consistent with the corresponding functions.
*Portability*
ANSI C requires `vprintf', `vfprintf', `vsprintf', and `vsnprintf'.
The remaining functions are newlib extensions.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: vfscanf, Next: vfwprintf, Prev: vfprintf, Up: Stdio
4.69 `vfscanf', `vscanf', `vsscanf'--format argument list
=========================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int vscanf(const char *FMT, va_list LIST);
int vfscanf(FILE *FP, const char *FMT, va_list LIST);
int vsscanf(const char *STR, const char *FMT, va_list LIST);
int _vscanf_r(struct _reent *REENT, const char *FMT,
va_list LIST);
int _vfscanf_r(struct _reent *REENT, FILE *FP, const char *FMT,
va_list LIST);
int _vsscanf_r(struct _reent *REENT, const char *STR,
const char *FMT, va_list LIST);
*Description*
`vscanf', `vfscanf', and `vsscanf' are (respectively) variants of
`scanf', `fscanf', and `sscanf'. They differ only in allowing their
caller to pass the variable argument list as a `va_list' object
(initialized by `va_start') rather than directly accepting a variable
number of arguments.
*Returns*
The return values are consistent with the corresponding functions:
`vscanf' returns the number of input fields successfully scanned,
converted, and stored; the return value does not include scanned fields
which were not stored.
If `vscanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
The routines `_vscanf_r', `_vfscanf_f', and `_vsscanf_r' are
reentrant versions which take an additional first parameter which
points to the reentrancy structure.
*Portability*
These are GNU extensions.
Supporting OS subroutines required:

File: libc.info, Node: vfwprintf, Next: vfwscanf, Prev: vfscanf, Up: Stdio
4.70 `vfwprintf', `vwprintf', `vswprintf'--wide character format argument list
==============================================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
#include <wchar.h>
int vwprintf(const wchar_t *FMT, va_list LIST);
int vfwprintf(FILE *FP, const wchar_t *FMT, va_list LIST);
int vswprintf(wchar_t *STR, size_t SIZE, const wchar_t *FMT,
va_list LIST);
int _vwprintf_r(struct _reent *REENT, const wchar_t *FMT,
va_list LIST);
int _vfwprintf_r(struct _reent *REENT, FILE *FP,
const wchar_t *FMT, va_list LIST);
int _vswprintf_r(struct _reent *REENT, wchar_t *STR,
size_t SIZE, const wchar_t *FMT, va_list LIST);
*Description*
`vwprintf', `vfwprintf' and `vswprintf' are (respectively) variants of
`wprintf', `fwprintf' and `swprintf'. They differ only in allowing
their caller to pass the variable argument list as a `va_list' object
(initialized by `va_start') rather than directly accepting a variable
number of arguments. The caller is responsible for calling `va_end'.
`_vwprintf_r', `_vfwprintf_r' and `_vswprintf_r' are reentrant
versions of the above.
*Returns*
The return values are consistent with the corresponding functions.
*Portability*
POSIX-1.2008 with extensions; C99 (compliant except for POSIX
extensions).
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.
*See Also*
`wprintf', `fwprintf' and `swprintf'.

File: libc.info, Node: vfwscanf, Next: viprintf, Prev: vfwprintf, Up: Stdio
4.71 `vfwscanf', `vwscanf', `vswscanf'--scan and format argument list from wide character input
===============================================================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int vwscanf(const wchar_t *FMT, va_list LIST);
int vfwscanf(FILE *FP, const wchar_t *FMT, va_list LIST);
int vswscanf(const wchar_t *STR, const wchar_t *FMT, va_list LIST);
int _vwscanf(struct _reent *REENT, const wchar_t *FMT,
va_list LIST);
int _vfwscanf(struct _reent *REENT, FILE *FP, const wchar_t *FMT,
va_list LIST);
int _vswscanf(struct _reent *REENT, const wchar_t *STR,
const wchar_t *FMT, va_list LIST);
*Description*
`vwscanf', `vfwscanf', and `vswscanf' are (respectively) variants of
`wscanf', `fwscanf', and `swscanf'. They differ only in allowing their
caller to pass the variable argument list as a `va_list' object
(initialized by `va_start') rather than directly accepting a variable
number of arguments.
*Returns*
The return values are consistent with the corresponding functions:
`vwscanf' returns the number of input fields successfully scanned,
converted, and stored; the return value does not include scanned fields
which were not stored.
If `vwscanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
The routines `_vwscanf', `_vfwscanf', and `_vswscanf' are reentrant
versions which take an additional first parameter which points to the
reentrancy structure.
*Portability*
C99, POSIX-1.2008

File: libc.info, Node: viprintf, Next: viscanf, Prev: vfwscanf, Up: Stdio
4.72 `viprintf', `vfiprintf', `vsiprintf', `vsniprintf', `vasiprintf', `vasniprintf'--format argument list (integer only)
=========================================================================================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int viprintf(const char *FMT, va_list LIST);
int vfiprintf(FILE *FP, const char *FMT, va_list LIST);
int vsiprintf(char *STR, const char *FMT, va_list LIST);
int vsniprintf(char *STR, size_t SIZE, const char *FMT,
va_list LIST);
int vasiprintf(char **STRP, const char *FMT, va_list LIST);
char *vasniprintf(char *STR, size_t *SIZE, const char *FMT,
va_list LIST);
int _viprintf_r(struct _reent *REENT, const char *FMT,
va_list LIST);
int _vfiprintf_r(struct _reent *REENT, FILE *FP,
const char *FMT, va_list LIST);
int _vsiprintf_r(struct _reent *REENT, char *STR,
const char *FMT, va_list LIST);
int _vsniprintf_r(struct _reent *REENT, char *STR,
size_t SIZE, const char *FMT, va_list LIST);
int _vasiprintf_r(struct _reent *REENT, char **STR,
const char *FMT, va_list LIST);
char *_vasniprintf_r(struct _reent *REENT, char *STR,
size_t *SIZE, const char *FMT, va_list LIST);
*Description*
`viprintf', `vfiprintf', `vasiprintf', `vsiprintf', `vsniprintf', and
`vasniprintf' are (respectively) variants of `iprintf', `fiprintf',
`asiprintf', `siprintf', `sniprintf', and `asniprintf'. They differ
only in allowing their caller to pass the variable argument list as a
`va_list' object (initialized by `va_start') rather than directly
accepting a variable number of arguments. The caller is responsible
for calling `va_end'.
`_viprintf_r', `_vfiprintf_r', `_vasiprintf_r', `_vsiprintf_r',
`_vsniprintf_r', and `_vasniprintf_r' are reentrant versions of the
above.
*Returns*
The return values are consistent with the corresponding functions:
*Portability*
All of these functions are newlib extensions.
Supporting OS subroutines required: `close', `fstat', `isatty',
`lseek', `read', `sbrk', `write'.

File: libc.info, Node: viscanf, Prev: viprintf, Up: Stdio
4.73 `viscanf', `vfiscanf', `vsiscanf'--format argument list
============================================================
*Synopsis*
#include <stdio.h>
#include <stdarg.h>
int viscanf(const char *FMT, va_list LIST);
int vfiscanf(FILE *FP, const char *FMT, va_list LIST);
int vsiscanf(const char *STR, const char *FMT, va_list LIST);
int _viscanf_r(struct _reent *REENT, const char *FMT,
va_list LIST);
int _vfiscanf_r(struct _reent *REENT, FILE *FP, const char *FMT,
va_list LIST);
int _vsiscanf_r(struct _reent *REENT, const char *STR,
const char *FMT, va_list LIST);
*Description*
`viscanf', `vfiscanf', and `vsiscanf' are (respectively) variants of
`iscanf', `fiscanf', and `siscanf'. They differ only in allowing their
caller to pass the variable argument list as a `va_list' object
(initialized by `va_start') rather than directly accepting a variable
number of arguments.
*Returns*
The return values are consistent with the corresponding functions:
`viscanf' returns the number of input fields successfully scanned,
converted, and stored; the return value does not include scanned fields
which were not stored.
If `viscanf' attempts to read at end-of-file, the return value is
`EOF'.
If no fields were stored, the return value is `0'.
The routines `_viscanf_r', `_vfiscanf_f', and `_vsiscanf_r' are
reentrant versions which take an additional first parameter which
points to the reentrancy structure.
*Portability*
These are newlib extensions.
Supporting OS subroutines required:

File: libc.info, Node: Strings, Next: Wchar strings, Prev: Stdio, Up: Top
5 Strings and Memory (`string.h')
*********************************
This chapter describes string-handling functions and functions for
managing areas of memory. The corresponding declarations are in
`string.h'.
* Menu:
* bcmp:: Compare two memory areas
* bcopy:: Copy memory regions
* bzero:: Initialize memory to zero
* index:: Search for character in string
* memccpy:: Copy memory regions up to end-token
* memchr:: Find character in memory
* memcmp:: Compare two memory areas
* memcpy:: Copy memory regions
* memmem:: Find memory segment
* memmove:: Move possibly overlapping memory
* mempcpy:: Copy memory regions and locate end
* memset:: Set an area of memory
* rindex:: Reverse search for character in string
* stpcpy:: Copy string returning a pointer to its end
* stpncpy:: Counted copy string returning a pointer to its end
* strcasecmp:: Compare strings ignoring case
* strcasestr:: Find string segment ignoring case
* strcat:: Concatenate strings
* strchr:: Search for character in string
* strcmp:: Character string compare
* strcoll:: Locale-specific character string compare
* strcpy:: Copy string
* strcspn:: Count chars not in string
* strerror:: Convert error number to string
* strerror_r:: Convert error number to string
* strlen:: Character string length
* strlwr:: Convert string to lowercase
* strncasecmp:: Compare strings ignoring case
* strncat:: Concatenate strings
* strncmp:: Character string compare
* strncpy:: Counted copy string
* strnlen:: Character string length
* strpbrk:: Find chars in string
* strrchr:: Reverse search for character in string
* strsignal:: Return signal message string
* strspn:: Find initial match
* strstr:: Find string segment
* strtok:: Get next token from a string
* strupr:: Convert string to upper case
* strxfrm:: Transform string
* swab:: Swap adjacent bytes
* wcscasecmp:: Compare wide character strings ignoring case
* wcsdup:: Wide character string duplicate
* wcsncasecmp:: Compare wide character strings ignoring case

File: libc.info, Node: bcmp, Next: bcopy, Up: Strings
5.1 `bcmp'--compare two memory areas
====================================
*Synopsis*
#include <strings.h>
int bcmp(const void *S1, const void *S2, size_t N);
*Description*
This function compares not more than N bytes of the object pointed to
by S1 with the object pointed to by S2.
This function is identical to `memcmp'.
*Returns*
The function returns an integer greater than, equal to or less than
zero according to whether the object pointed to by S1 is greater than,
equal to or less than the object pointed to by S2.
*Portability*
`bcmp' requires no supporting OS subroutines.

File: libc.info, Node: bcopy, Next: bzero, Prev: bcmp, Up: Strings
5.2 `bcopy'--copy memory regions
================================
*Synopsis*
#include <strings.h>
void bcopy(const void *IN, void *OUT, size_t N);
*Description*
This function copies N bytes from the memory region pointed to by IN to
the memory region pointed to by OUT.
This function is implemented in term of `memmove'.
*Portability*
`bcopy' requires no supporting OS subroutines.

File: libc.info, Node: bzero, Next: index, Prev: bcopy, Up: Strings
5.3 `bzero'--initialize memory to zero
======================================
*Synopsis*
#include <strings.h>
void bzero(void *B, size_t LENGTH);
*Description*
`bzero' initializes LENGTH bytes of memory, starting at address B, to
zero.
*Returns*
`bzero' does not return a result.
*Portability*
`bzero' is in the Berkeley Software Distribution. Neither ANSI C nor
the System V Interface Definition (Issue 2) require `bzero'.
`bzero' requires no supporting OS subroutines.

File: libc.info, Node: index, Next: memccpy, Prev: bzero, Up: Strings
5.4 `index'--search for character in string
===========================================
*Synopsis*
#include <strings.h>
char * index(const char *STRING, int C);
*Description*
This function finds the first occurence of C (converted to a char) in
the string pointed to by STRING (including the terminating null
character).
This function is identical to `strchr'.
*Returns*
Returns a pointer to the located character, or a null pointer if C does
not occur in STRING.
*Portability*
`index' requires no supporting OS subroutines.

File: libc.info, Node: memccpy, Next: memchr, Prev: index, Up: Strings
5.5 `memccpy'--copy memory regions with end-token check
=======================================================
*Synopsis*
#include <string.h>
void* memccpy(void *OUT, const void *IN,
int ENDCHAR, size_t N);
*Description*
This function copies up to N bytes from the memory region pointed to by
IN to the memory region pointed to by OUT. If a byte matching the
ENDCHAR is encountered, the byte is copied and copying stops.
If the regions overlap, the behavior is undefined.
*Returns*
`memccpy' returns a pointer to the first byte following the ENDCHAR in
the OUT region. If no byte matching ENDCHAR was copied, then `NULL' is
returned.
*Portability*
`memccpy' is a GNU extension.
`memccpy' requires no supporting OS subroutines.

File: libc.info, Node: memchr, Next: memcmp, Prev: memccpy, Up: Strings
5.6 `memchr'--find character in memory
======================================
*Synopsis*
#include <string.h>
void *memchr(const void *SRC, int C, size_t LENGTH);
*Description*
This function searches memory starting at `*SRC' for the character C.
The search only ends with the first occurrence of C, or after LENGTH
characters; in particular, `NUL' does not terminate the search.
*Returns*
If the character C is found within LENGTH characters of `*SRC', a
pointer to the character is returned. If C is not found, then `NULL' is
returned.
*Portability*
`memchr' is ANSI C.
`memchr' requires no supporting OS subroutines.

File: libc.info, Node: memcmp, Next: memcpy, Prev: memchr, Up: Strings
5.7 `memcmp'--compare two memory areas
======================================
*Synopsis*
#include <string.h>
int memcmp(const void *S1, const void *S2, size_t N);
*Description*
This function compares not more than N characters of the object pointed
to by S1 with the object pointed to by S2.
*Returns*
The function returns an integer greater than, equal to or less than
zero according to whether the object pointed to by S1 is greater than,
equal to or less than the object pointed to by S2.
*Portability*
`memcmp' is ANSI C.
`memcmp' requires no supporting OS subroutines.

File: libc.info, Node: memcpy, Next: memmem, Prev: memcmp, Up: Strings
5.8 `memcpy'--copy memory regions
=================================
*Synopsis*
#include <string.h>
void* memcpy(void *OUT, const void *IN, size_t N);
*Description*
This function copies N bytes from the memory region pointed to by IN to
the memory region pointed to by OUT.
If the regions overlap, the behavior is undefined.
*Returns*
`memcpy' returns a pointer to the first byte of the OUT region.
*Portability*
`memcpy' is ANSI C.
`memcpy' requires no supporting OS subroutines.

File: libc.info, Node: memmem, Next: memmove, Prev: memcpy, Up: Strings
5.9 `memmem'--find memory segment
=================================
*Synopsis*
#include <string.h>
char *memmem(const void *S1, size_t L1, const void *S2,
size_t L2);
*Description*
Locates the first occurrence in the memory region pointed to by S1 with
length L1 of the sequence of bytes pointed to by S2 of length L2. If
you already know the lengths of your haystack and needle, `memmem' can
be much faster than `strstr'.
*Returns*
Returns a pointer to the located segment, or a null pointer if S2 is
not found. If L2 is 0, S1 is returned.
*Portability*
`memmem' is a newlib extension.
`memmem' requires no supporting OS subroutines.

File: libc.info, Node: memmove, Next: mempcpy, Prev: memmem, Up: Strings
5.10 `memmove'--move possibly overlapping memory
================================================
*Synopsis*
#include <string.h>
void *memmove(void *DST, const void *SRC, size_t LENGTH);
*Description*
This function moves LENGTH characters from the block of memory starting
at `*SRC' to the memory starting at `*DST'. `memmove' reproduces the
characters correctly at `*DST' even if the two areas overlap.
*Returns*
The function returns DST as passed.
*Portability*
`memmove' is ANSI C.
`memmove' requires no supporting OS subroutines.

File: libc.info, Node: mempcpy, Next: memset, Prev: memmove, Up: Strings
5.11 `mempcpy'--copy memory regions and return end pointer
==========================================================
*Synopsis*
#include <string.h>
void* mempcpy(void *OUT, const void *IN, size_t N);
*Description*
This function copies N bytes from the memory region pointed to by IN to
the memory region pointed to by OUT.
If the regions overlap, the behavior is undefined.
*Returns*
`mempcpy' returns a pointer to the byte following the last byte copied
to the OUT region.
*Portability*
`mempcpy' is a GNU extension.
`mempcpy' requires no supporting OS subroutines.

File: libc.info, Node: memset, Next: rindex, Prev: mempcpy, Up: Strings
5.12 `memset'--set an area of memory
====================================
*Synopsis*
#include <string.h>
void *memset(void *DST, int C, size_t LENGTH);
*Description*
This function converts the argument C into an unsigned char and fills
the first LENGTH characters of the array pointed to by DST to the value.
*Returns*
`memset' returns the value of DST.
*Portability*
`memset' is ANSI C.
`memset' requires no supporting OS subroutines.

File: libc.info, Node: rindex, Next: stpcpy, Prev: memset, Up: Strings
5.13 `rindex'--reverse search for character in string
=====================================================
*Synopsis*
#include <string.h>
char * rindex(const char *STRING, int C);
*Description*
This function finds the last occurence of C (converted to a char) in
the string pointed to by STRING (including the terminating null
character).
This function is identical to `strrchr'.
*Returns*
Returns a pointer to the located character, or a null pointer if C does
not occur in STRING.
*Portability*
`rindex' requires no supporting OS subroutines.

File: libc.info, Node: stpcpy, Next: stpncpy, Prev: rindex, Up: Strings
5.14 `stpcpy'--copy string returning a pointer to its end
=========================================================
*Synopsis*
#include <string.h>
char *stpcpy(char *DST, const char *SRC);
*Description*
`stpcpy' copies the string pointed to by SRC (including the terminating
null character) to the array pointed to by DST.
*Returns*
This function returns a pointer to the end of the destination string,
thus pointing to the trailing '\0'.
*Portability*
`stpcpy' is a GNU extension, candidate for inclusion into POSIX/SUSv4.
`stpcpy' requires no supporting OS subroutines.

File: libc.info, Node: stpncpy, Next: strcasecmp, Prev: stpcpy, Up: Strings
5.15 `stpncpy'--counted copy string returning a pointer to its end
==================================================================
*Synopsis*
#include <string.h>
char *stpncpy(char *DST, const char *SRC, size_t LENGTH);
*Description*
`stpncpy' copies not more than LENGTH characters from the the string
pointed to by SRC (including the terminating null character) to the
array pointed to by DST. If the string pointed to by SRC is shorter
than LENGTH characters, null characters are appended to the destination
array until a total of LENGTH characters have been written.
*Returns*
This function returns a pointer to the end of the destination string,
thus pointing to the trailing '\0', or, if the destination string is
not null-terminated, pointing to dst + n.
*Portability*
`stpncpy' is a GNU extension, candidate for inclusion into POSIX/SUSv4.
`stpncpy' requires no supporting OS subroutines.

File: libc.info, Node: strcasecmp, Next: strcasestr, Prev: stpncpy, Up: Strings
5.16 `strcasecmp'--case-insensitive character string compare
============================================================
*Synopsis*
#include <strings.h>
int strcasecmp(const char *A, const char *B);
*Description*
`strcasecmp' compares the string at A to the string at B in a
case-insensitive manner.
*Returns*
If `*A' sorts lexicographically after `*B' (after both are converted to
lowercase), `strcasecmp' returns a number greater than zero. If the
two strings match, `strcasecmp' returns zero. If `*A' sorts
lexicographically before `*B', `strcasecmp' returns a number less than
zero.
*Portability*
`strcasecmp' is in the Berkeley Software Distribution.
`strcasecmp' requires no supporting OS subroutines. It uses
tolower() from elsewhere in this library.

File: libc.info, Node: strcasestr, Next: strcat, Prev: strcasecmp, Up: Strings
5.17 `strcasestr'--case-insensitive character string search
===========================================================
*Synopsis*
#include <string.h>
char *strcasestr(const char *S, const char *FIND);
*Description*
`strcasestr' searchs the string S for the first occurrence of the
sequence FIND. `strcasestr' is identical to `strstr' except the search
is case-insensitive.
*Returns*
A pointer to the first case-insensitive occurrence of the sequence FIND
or `NULL' if no match was found.
*Portability*
`strcasestr' is in the Berkeley Software Distribution.
`strcasestr' requires no supporting OS subroutines. It uses
tolower() from elsewhere in this library.

File: libc.info, Node: strcat, Next: strchr, Prev: strcasestr, Up: Strings
5.18 `strcat'--concatenate strings
==================================
*Synopsis*
#include <string.h>
char *strcat(char *DST, const char *SRC);
*Description*
`strcat' appends a copy of the string pointed to by SRC (including the
terminating null character) to the end of the string pointed to by DST.
The initial character of SRC overwrites the null character at the end
of DST.
*Returns*
This function returns the initial value of DST
*Portability*
`strcat' is ANSI C.
`strcat' requires no supporting OS subroutines.

File: libc.info, Node: strchr, Next: strcmp, Prev: strcat, Up: Strings
5.19 `strchr'--search for character in string
=============================================
*Synopsis*
#include <string.h>
char * strchr(const char *STRING, int C);
*Description*
This function finds the first occurence of C (converted to a char) in
the string pointed to by STRING (including the terminating null
character).
*Returns*
Returns a pointer to the located character, or a null pointer if C does
not occur in STRING.
*Portability*
`strchr' is ANSI C.
`strchr' requires no supporting OS subroutines.

File: libc.info, Node: strcmp, Next: strcoll, Prev: strchr, Up: Strings
5.20 `strcmp'--character string compare
=======================================
*Synopsis*
#include <string.h>
int strcmp(const char *A, const char *B);
*Description*
`strcmp' compares the string at A to the string at B.
*Returns*
If `*A' sorts lexicographically after `*B', `strcmp' returns a number
greater than zero. If the two strings match, `strcmp' returns zero.
If `*A' sorts lexicographically before `*B', `strcmp' returns a number
less than zero.
*Portability*
`strcmp' is ANSI C.
`strcmp' requires no supporting OS subroutines.

File: libc.info, Node: strcoll, Next: strcpy, Prev: strcmp, Up: Strings
5.21 `strcoll'--locale-specific character string compare
========================================================
*Synopsis*
#include <string.h>
int strcoll(const char *STRA, const char * STRB);
*Description*
`strcoll' compares the string pointed to by STRA to the string pointed
to by STRB, using an interpretation appropriate to the current
`LC_COLLATE' state.
*Returns*
If the first string is greater than the second string, `strcoll'
returns a number greater than zero. If the two strings are equivalent,
`strcoll' returns zero. If the first string is less than the second
string, `strcoll' returns a number less than zero.
*Portability*
`strcoll' is ANSI C.
`strcoll' requires no supporting OS subroutines.

File: libc.info, Node: strcpy, Next: strcspn, Prev: strcoll, Up: Strings
5.22 `strcpy'--copy string
==========================
*Synopsis*
#include <string.h>
char *strcpy(char *DST, const char *SRC);
*Description*
`strcpy' copies the string pointed to by SRC (including the terminating
null character) to the array pointed to by DST.
*Returns*
This function returns the initial value of DST.
*Portability*
`strcpy' is ANSI C.
`strcpy' requires no supporting OS subroutines.

File: libc.info, Node: strcspn, Next: strerror, Prev: strcpy, Up: Strings
5.23 `strcspn'--count characters not in string
==============================================
*Synopsis*
size_t strcspn(const char *S1, const char *S2);
*Description*
This function computes the length of the initial part of the string
pointed to by S1 which consists entirely of characters NOT from the
string pointed to by S2 (excluding the terminating null character).
*Returns*
`strcspn' returns the length of the substring found.
*Portability*
`strcspn' is ANSI C.
`strcspn' requires no supporting OS subroutines.

File: libc.info, Node: strerror, Next: strerror_r, Prev: strcspn, Up: Strings
5.24 `strerror'--convert error number to string
===============================================
*Synopsis*
#include <string.h>
char *strerror(int ERRNUM);
char *_strerror_r(struct _reent PTR, int ERRNUM,
int INTERNAL, int *ERROR);
*Description*
`strerror' converts the error number ERRNUM into a string. The value
of ERRNUM is usually a copy of `errno'. If `errnum' is not a known
error number, the result points to an empty string.
This implementation of `strerror' prints out the following strings
for each of the values defined in ``errno.h'':
`0'
Success
`E2BIG'
Arg list too long
`EACCES'
Permission denied
`EADDRINUSE'
Address already in use
`EADV'
Advertise error
`EAFNOSUPPORT'
Address family not supported by protocol family
`EAGAIN'
No more processes
`EALREADY'
Socket already connected
`EBADF'
Bad file number
`EBADMSG'
Bad message
`EBUSY'
Device or resource busy
`ECHILD'
No children
`ECOMM'
Communication error
`ECONNABORTED'
Software caused connection abort
`ECONNREFUSED'
Connection refused
`EDEADLK'
Deadlock
`EDESTADDRREQ'
Destination address required
`EEXIST'
File exists
`EDOM'
Math argument
`EFAULT'
Bad address
`EFBIG'
File too large
`EHOSTDOWN'
Host is down
`EHOSTUNREACH'
Host is unreachable
`EIDRM'
Identifier removed
`EINPROGRESS'
Connection already in progress
`EINTR'
Interrupted system call
`EINVAL'
Invalid argument
`EIO'
I/O error
`EISCONN'
Socket is already connected
`EISDIR'
Is a directory
`ELIBACC'
Cannot access a needed shared library
`ELIBBAD'
Accessing a corrupted shared library
`ELIBEXEC'
Cannot exec a shared library directly
`ELIBMAX'
Attempting to link in more shared libraries than system limit
`ELIBSCN'
`.lib' section in a.out corrupted
`EMFILE'
Too many open files
`EMLINK'
Too many links
`EMSGSIZE'
Message too long
`EMULTIHOP'
Multihop attempted
`ENAMETOOLONG'
File or path name too long
`ENETDOWN'
Network interface not configured
`ENETUNREACH'
Network is unreachable
`ENFILE'
Too many open files in system
`ENODEV'
No such device
`ENOENT'
No such file or directory
`ENOEXEC'
Exec format error
`ENOLCK'
No lock
`ENOLINK'
Virtual circuit is gone
`ENOMEM'
Not enough space
`ENOMSG'
No message of desired type
`ENONET'
Machine is not on the network
`ENOPKG'
No package
`ENOPROTOOPT'
Protocol not available
`ENOSPC'
No space left on device
`ENOSR'
No stream resources
`ENOSTR'
Not a stream
`ENOSYS'
Function not implemented
`ENOTBLK'
Block device required
`ENOTCONN'
Socket is not connected
`ENOTDIR'
Not a directory
`ENOTEMPTY'
Directory not empty
`ENOTSOCK'
Socket operation on non-socket
`ENOTSUP'
Not supported
`ENOTTY'
Not a character device
`ENXIO'
No such device or address
`EPERM'
Not owner
`EPIPE'
Broken pipe
`EPROTO'
Protocol error
`EPROTOTYPE'
Protocol wrong type for socket
`EPROTONOSUPPORT'
Unknown protocol
`ERANGE'
Result too large
`EREMOTE'
Resource is remote
`EROFS'
Read-only file system
`ESHUTDOWN'
Can't send after socket shutdown
`ESOCKTNOSUPPORT'
Socket type not supported
`ESPIPE'
Illegal seek
`ESRCH'
No such process
`ESRMNT'
Srmount error
`ETIME'
Stream ioctl timeout
`ETIMEDOUT'
Connection timed out
`ETXTBSY'
Text file busy
`EXDEV'
Cross-device link
`ECANCELED'
Operation canceled
`ENOTRECOVERABLE'
State not recoverable
`EOWNERDEAD'
Previous owner died
`ESTRPIPE'
Strings pipe error
`_strerror_r' is a reentrant version of the above.
*Returns*
This function returns a pointer to a string. Your application must not
modify that string.
*Portability*
ANSI C requires `strerror', but does not specify the strings used for
each error number.
Although this implementation of `strerror' is reentrant (depending
on `_user_strerror'), ANSI C declares that subsequent calls to
`strerror' may overwrite the result string; therefore portable code
cannot depend on the reentrancy of this subroutine.
Although this implementation of `strerror' guarantees a non-null
result with a NUL-terminator, some implementations return `NULL' on
failure. Although POSIX allows `strerror' to set `errno' to EINVAL on
failure, this implementation does not do so (unless you provide
`_user_strerror').
POSIX recommends that unknown ERRNUM result in a message including
that value, however it is not a requirement and this implementation
does not provide that information (unless you provide `_user_strerror').
This implementation of `strerror' provides for user-defined
extensibility. `errno.h' defines __ELASTERROR, which can be used as a
base for user-defined error values. If the user supplies a routine
named `_user_strerror', and ERRNUM passed to `strerror' does not match
any of the supported values, `_user_strerror' is called with three
arguments. The first is of type INT, and is the ERRNUM value unknown
to `strerror'. The second is of type INT, and matches the INTERNAL
argument of `_strerror_r'; this should be zero if called from `strerror'
and non-zero if called from any other function; `_user_strerror' can
use this information to satisfy the POSIX rule that no other
standardized function can overwrite a static buffer reused by
`strerror'. The third is of type INT *, and matches the ERROR argument
of `_strerror_r'; if a non-zero value is stored into that location
(usually EINVAL), then `strerror' will set `errno' to that value, and
the XPG variant of `strerror_r' will return that value instead of zero
or ERANGE. `_user_strerror' returns a CHAR * value; returning NULL
implies that the user function did not choose to handle ERRNUM. The
default `_user_strerror' returns NULL for all input values. Note that
`_user_sterror' must be thread-safe, and only denote errors via the
third argument rather than modifying `errno', if `strerror' and
`strerror_r' are are to comply with POSIX.
`strerror' requires no supporting OS subroutines.

File: libc.info, Node: strerror_r, Next: strlen, Prev: strerror, Up: Strings
5.25 `strerror_r'--convert error number to string and copy to buffer
====================================================================
*Synopsis*
#include <string.h>
#ifdef _GNU_SOURCE
char *strerror_r(int ERRNUM, char *BUFFER, size_t N);
#else
int strerror_r(int ERRNUM, char *BUFFER, size_t N);
#endif
*Description*
`strerror_r' converts the error number ERRNUM into a string and copies
the result into the supplied BUFFER for a length up to N, including the
NUL terminator. The value of ERRNUM is usually a copy of `errno'. If
`errnum' is not a known error number, the result is the empty string.
See `strerror' for how strings are mapped to `errnum'.
*Returns*
There are two variants: the GNU version always returns a NUL-terminated
string, which is BUFFER if all went well, but which is another pointer
if N was too small (leaving BUFFER untouched). If the return is not
BUFFER, your application must not modify that string. The POSIX
version returns 0 on success, EINVAL if `errnum' was not recognized,
and ERANGE if N was too small. The variant chosen depends on macros
that you define before inclusion of `string.h'.
*Portability*
`strerror_r' with a CHAR * result is a GNU extension. `strerror_r'
with an INT result is required by POSIX 2001. This function is
compliant only if `_user_strerror' is not provided, or if it is
thread-safe and uses separate storage according to whether the second
argument of that function is non-zero. For more details on
`_user_strerror', see the `strerror' documentation.
POSIX states that the contents of BUF are unspecified on error,
although this implementation guarantees a NUL-terminated string for all
except N of 0.
POSIX recommends that unknown ERRNUM result in a message including
that value, however it is not a requirement and this implementation
provides only an empty string (unless you provide `_user_strerror').
POSIX also recommends that unknown ERRNUM fail with EINVAL even when
providing such a message, however it is not a requirement and this
implementation will return success if `_user_strerror' provided a
non-empty alternate string without assigning into its third argument.
`strerror_r' requires no supporting OS subroutines.

File: libc.info, Node: strlen, Next: strlwr, Prev: strerror_r, Up: Strings
5.26 `strlen'--character string length
======================================
*Synopsis*
#include <string.h>
size_t strlen(const char *STR);
*Description*
The `strlen' function works out the length of the string starting at
`*STR' by counting chararacters until it reaches a `NULL' character.
*Returns*
`strlen' returns the character count.
*Portability*
`strlen' is ANSI C.
`strlen' requires no supporting OS subroutines.

File: libc.info, Node: strlwr, Next: strncasecmp, Prev: strlen, Up: Strings
5.27 `strlwr'--force string to lowercase
========================================
*Synopsis*
#include <string.h>
char *strlwr(char *A);
*Description*
`strlwr' converts each character in the string at A to lowercase.
*Returns*
`strlwr' returns its argument, A.
*Portability*
`strlwr' is not widely portable.
`strlwr' requires no supporting OS subroutines.

File: libc.info, Node: strncasecmp, Next: strncat, Prev: strlwr, Up: Strings
5.28 `strncasecmp'--case-insensitive character string compare
=============================================================
*Synopsis*
#include <strings.h>
int strncasecmp(const char *A, const char * B, size_t LENGTH);
*Description*
`strncasecmp' compares up to LENGTH characters from the string at A to
the string at B in a case-insensitive manner.
*Returns*
If `*A' sorts lexicographically after `*B' (after both are converted to
lowercase), `strncasecmp' returns a number greater than zero. If the
two strings are equivalent, `strncasecmp' returns zero. If `*A' sorts
lexicographically before `*B', `strncasecmp' returns a number less than
zero.
*Portability*
`strncasecmp' is in the Berkeley Software Distribution.
`strncasecmp' requires no supporting OS subroutines. It uses
tolower() from elsewhere in this library.

File: libc.info, Node: strncat, Next: strncmp, Prev: strncasecmp, Up: Strings
5.29 `strncat'--concatenate strings
===================================
*Synopsis*
#include <string.h>
char *strncat(char *DST, const char *SRC, size_t LENGTH);
*Description*
`strncat' appends not more than LENGTH characters from the string
pointed to by SRC (including the terminating null character) to the end
of the string pointed to by DST. The initial character of SRC
overwrites the null character at the end of DST. A terminating null
character is always appended to the result
*Warnings*
Note that a null is always appended, so that if the copy is limited by
the LENGTH argument, the number of characters appended to DST is `n +
1'.
*Returns*
This function returns the initial value of DST
*Portability*
`strncat' is ANSI C.
`strncat' requires no supporting OS subroutines.

File: libc.info, Node: strncmp, Next: strncpy, Prev: strncat, Up: Strings
5.30 `strncmp'--character string compare
========================================
*Synopsis*
#include <string.h>
int strncmp(const char *A, const char * B, size_t LENGTH);
*Description*
`strncmp' compares up to LENGTH characters from the string at A to the
string at B.
*Returns*
If `*A' sorts lexicographically after `*B', `strncmp' returns a number
greater than zero. If the two strings are equivalent, `strncmp'
returns zero. If `*A' sorts lexicographically before `*B', `strncmp'
returns a number less than zero.
*Portability*
`strncmp' is ANSI C.
`strncmp' requires no supporting OS subroutines.

File: libc.info, Node: strncpy, Next: strnlen, Prev: strncmp, Up: Strings
5.31 `strncpy'--counted copy string
===================================
*Synopsis*
#include <string.h>
char *strncpy(char *DST, const char *SRC, size_t LENGTH);
*Description*
`strncpy' copies not more than LENGTH characters from the the string
pointed to by SRC (including the terminating null character) to the
array pointed to by DST. If the string pointed to by SRC is shorter
than LENGTH characters, null characters are appended to the destination
array until a total of LENGTH characters have been written.
*Returns*
This function returns the initial value of DST.
*Portability*
`strncpy' is ANSI C.
`strncpy' requires no supporting OS subroutines.

File: libc.info, Node: strnlen, Next: strpbrk, Prev: strncpy, Up: Strings
5.32 `strnlen'--character string length
=======================================
*Synopsis*
#include <string.h>
size_t strnlen(const char *STR, size_t N);
*Description*
The `strnlen' function works out the length of the string starting at
`*STR' by counting chararacters until it reaches a NUL character or the
maximum: N number of characters have been inspected.
*Returns*
`strnlen' returns the character count or N.
*Portability*
`strnlen' is a GNU extension.
`strnlen' requires no supporting OS subroutines.

File: libc.info, Node: strpbrk, Next: strrchr, Prev: strnlen, Up: Strings
5.33 `strpbrk'--find characters in string
=========================================
*Synopsis*
#include <string.h>
char *strpbrk(const char *S1, const char *S2);
*Description*
This function locates the first occurence in the string pointed to by
S1 of any character in string pointed to by S2 (excluding the
terminating null character).
*Returns*
`strpbrk' returns a pointer to the character found in S1, or a null
pointer if no character from S2 occurs in S1.
*Portability*
`strpbrk' requires no supporting OS subroutines.