1 @node Arithmetic, Date and Time, Mathematics, Top
2 @c %MENU% Low level arithmetic functions
3 @chapter Arithmetic Functions
5 This chapter contains information about functions for doing basic
6 arithmetic operations, such as splitting a float into its integer and
7 fractional parts or retrieving the imaginary part of a complex value.
8 These functions are declared in the header files @file{math.h} and
12 * Floating Point Numbers:: Basic concepts. IEEE 754.
13 * Floating Point Classes:: The five kinds of floating-point number.
14 * Floating Point Errors:: When something goes wrong in a calculation.
15 * Rounding:: Controlling how results are rounded.
16 * Control Functions:: Saving and restoring the FPU's state.
17 * Arithmetic Functions:: Fundamental operations provided by the library.
18 * Complex Numbers:: The types. Writing complex constants.
19 * Operations on Complex:: Projection, conjugation, decomposition.
20 * Integer Division:: Integer division with guaranteed rounding.
21 * Parsing of Numbers:: Converting strings to numbers.
22 * System V Number Conversion:: An archaic way to convert numbers to strings.
25 @node Floating Point Numbers
26 @section Floating Point Numbers
27 @cindex floating point
29 @cindex IEEE floating point
31 Most computer hardware has support for two different kinds of numbers:
32 integers (@math{@dots{}-3, -2, -1, 0, 1, 2, 3@dots{}}) and
33 floating-point numbers. Floating-point numbers have three parts: the
34 @dfn{mantissa}, the @dfn{exponent}, and the @dfn{sign bit}. The real
35 number represented by a floating-point value is given by
37 $(s \mathrel? -1 \mathrel: 1) \cdot 2^e \cdot M$
40 @math{(s ? -1 : 1) @mul{} 2^e @mul{} M}
42 where @math{s} is the sign bit, @math{e} the exponent, and @math{M}
43 the mantissa. @xref{Floating Point Concepts}, for details. (It is
44 possible to have a different @dfn{base} for the exponent, but all modern
45 hardware uses @math{2}.)
47 Floating-point numbers can represent a finite subset of the real
48 numbers. While this subset is large enough for most purposes, it is
49 important to remember that the only reals that can be represented
50 exactly are rational numbers that have a terminating binary expansion
51 shorter than the width of the mantissa. Even simple fractions such as
52 @math{1/5} can only be approximated by floating point.
54 Mathematical operations and functions frequently need to produce values
55 that are not representable. Often these values can be approximated
56 closely enough for practical purposes, but sometimes they can't.
57 Historically there was no way to tell when the results of a calculation
58 were inaccurate. Modern computers implement the @w{IEEE 754} standard
59 for numerical computations, which defines a framework for indicating to
60 the program when the results of calculation are not trustworthy. This
61 framework consists of a set of @dfn{exceptions} that indicate why a
62 result could not be represented, and the special values @dfn{infinity}
63 and @dfn{not a number} (NaN).
65 @node Floating Point Classes
66 @section Floating-Point Number Classification Functions
67 @cindex floating-point classes
68 @cindex classes, floating-point
71 @w{ISO C99} defines macros that let you determine what sort of
72 floating-point number a variable holds.
76 @deftypefn {Macro} int fpclassify (@emph{float-type} @var{x})
77 This is a generic macro which works on all floating-point types and
78 which returns a value of type @code{int}. The possible values are:
82 The floating-point number @var{x} is ``Not a Number'' (@pxref{Infinity
85 The value of @var{x} is either plus or minus infinity (@pxref{Infinity
88 The value of @var{x} is zero. In floating-point formats like @w{IEEE
89 754}, where zero can be signed, this value is also returned if
90 @var{x} is negative zero.
92 Numbers whose absolute value is too small to be represented in the
93 normal format are represented in an alternate, @dfn{denormalized} format
94 (@pxref{Floating Point Concepts}). This format is less precise but can
95 represent values closer to zero. @code{fpclassify} returns this value
96 for values of @var{x} in this alternate format.
98 This value is returned for all other values of @var{x}. It indicates
99 that there is nothing special about the number.
104 @code{fpclassify} is most useful if more than one property of a number
105 must be tested. There are more specific macros which only test one
106 property at a time. Generally these macros execute faster than
107 @code{fpclassify}, since there is special hardware support for them.
108 You should therefore use the specific macros whenever possible.
112 @deftypefn {Macro} int isfinite (@emph{float-type} @var{x})
113 This macro returns a nonzero value if @var{x} is finite: not plus or
114 minus infinity, and not NaN. It is equivalent to
117 (fpclassify (x) != FP_NAN && fpclassify (x) != FP_INFINITE)
120 @code{isfinite} is implemented as a macro which accepts any
126 @deftypefn {Macro} int isnormal (@emph{float-type} @var{x})
127 This macro returns a nonzero value if @var{x} is finite and normalized.
131 (fpclassify (x) == FP_NORMAL)
137 @deftypefn {Macro} int isnan (@emph{float-type} @var{x})
138 This macro returns a nonzero value if @var{x} is NaN. It is equivalent
142 (fpclassify (x) == FP_NAN)
146 Another set of floating-point classification functions was provided by
147 BSD. The GNU C library also supports these functions; however, we
148 recommend that you use the ISO C99 macros in new code. Those are standard
149 and will be available more widely. Also, since they are macros, you do
150 not have to worry about the type of their argument.
154 @deftypefun int isinf (double @var{x})
157 @deftypefunx int isinff (float @var{x})
160 @deftypefunx int isinfl (long double @var{x})
161 This function returns @code{-1} if @var{x} represents negative infinity,
162 @code{1} if @var{x} represents positive infinity, and @code{0} otherwise.
167 @deftypefun int isnan (double @var{x})
170 @deftypefunx int isnanf (float @var{x})
173 @deftypefunx int isnanl (long double @var{x})
174 This function returns a nonzero value if @var{x} is a ``not a number''
175 value, and zero otherwise.
177 @strong{Note:} The @code{isnan} macro defined by @w{ISO C99} overrides
178 the BSD function. This is normally not a problem, because the two
179 routines behave identically. However, if you really need to get the BSD
180 function for some reason, you can write
189 @deftypefun int finite (double @var{x})
192 @deftypefunx int finitef (float @var{x})
195 @deftypefunx int finitel (long double @var{x})
196 This function returns a nonzero value if @var{x} is finite or a ``not a
197 number'' value, and zero otherwise.
202 @deftypefun double infnan (int @var{error})
203 This function is provided for compatibility with BSD. Its argument is
204 an error code, @code{EDOM} or @code{ERANGE}; @code{infnan} returns the
205 value that a math function would return if it set @code{errno} to that
206 value. @xref{Math Error Reporting}. @code{-ERANGE} is also acceptable
207 as an argument, and corresponds to @code{-HUGE_VAL} as a value.
209 In the BSD library, on certain machines, @code{infnan} raises a fatal
210 signal in all cases. The GNU library does not do likewise, because that
211 does not fit the @w{ISO C} specification.
214 @strong{Portability Note:} The functions listed in this section are BSD
218 @node Floating Point Errors
219 @section Errors in Floating-Point Calculations
222 * FP Exceptions:: IEEE 754 math exceptions and how to detect them.
223 * Infinity and NaN:: Special values returned by calculations.
224 * Status bit operations:: Checking for exceptions after the fact.
225 * Math Error Reporting:: How the math functions report errors.
229 @subsection FP Exceptions
233 @cindex division by zero
234 @cindex inexact exception
235 @cindex invalid exception
236 @cindex overflow exception
237 @cindex underflow exception
239 The @w{IEEE 754} standard defines five @dfn{exceptions} that can occur
240 during a calculation. Each corresponds to a particular sort of error,
243 When exceptions occur (when exceptions are @dfn{raised}, in the language
244 of the standard), one of two things can happen. By default the
245 exception is simply noted in the floating-point @dfn{status word}, and
246 the program continues as if nothing had happened. The operation
247 produces a default value, which depends on the exception (see the table
248 below). Your program can check the status word to find out which
251 Alternatively, you can enable @dfn{traps} for exceptions. In that case,
252 when an exception is raised, your program will receive the @code{SIGFPE}
253 signal. The default action for this signal is to terminate the
254 program. @xref{Signal Handling}, for how you can change the effect of
258 In the System V math library, the user-defined function @code{matherr}
259 is called when certain exceptions occur inside math library functions.
260 However, the Unix98 standard deprecates this interface. We support it
261 for historical compatibility, but recommend that you do not use it in
265 The exceptions defined in @w{IEEE 754} are:
268 @item Invalid Operation
269 This exception is raised if the given operands are invalid for the
270 operation to be performed. Examples are
271 (see @w{IEEE 754}, @w{section 7}):
274 Addition or subtraction: @math{@infinity{} - @infinity{}}. (But
275 @math{@infinity{} + @infinity{} = @infinity{}}).
277 Multiplication: @math{0 @mul{} @infinity{}}.
279 Division: @math{0/0} or @math{@infinity{}/@infinity{}}.
281 Remainder: @math{x} REM @math{y}, where @math{y} is zero or @math{x} is
284 Square root if the operand is less then zero. More generally, any
285 mathematical function evaluated outside its domain produces this
288 Conversion of a floating-point number to an integer or decimal
289 string, when the number cannot be represented in the target format (due
290 to overflow, infinity, or NaN).
292 Conversion of an unrecognizable input string.
294 Comparison via predicates involving @math{<} or @math{>}, when one or
295 other of the operands is NaN. You can prevent this exception by using
296 the unordered comparison functions instead; see @ref{FP Comparison Functions}.
299 If the exception does not trap, the result of the operation is NaN.
301 @item Division by Zero
302 This exception is raised when a finite nonzero number is divided
303 by zero. If no trap occurs the result is either @math{+@infinity{}} or
304 @math{-@infinity{}}, depending on the signs of the operands.
307 This exception is raised whenever the result cannot be represented
308 as a finite value in the precision format of the destination. If no trap
309 occurs the result depends on the sign of the intermediate result and the
310 current rounding mode (@w{IEEE 754}, @w{section 7.3}):
313 Round to nearest carries all overflows to @math{@infinity{}}
314 with the sign of the intermediate result.
316 Round toward @math{0} carries all overflows to the largest representable
317 finite number with the sign of the intermediate result.
319 Round toward @math{-@infinity{}} carries positive overflows to the
320 largest representable finite number and negative overflows to
324 Round toward @math{@infinity{}} carries negative overflows to the
325 most negative representable finite number and positive overflows
326 to @math{@infinity{}}.
329 Whenever the overflow exception is raised, the inexact exception is also
333 The underflow exception is raised when an intermediate result is too
334 small to be calculated accurately, or if the operation's result rounded
335 to the destination precision is too small to be normalized.
337 When no trap is installed for the underflow exception, underflow is
338 signaled (via the underflow flag) only when both tininess and loss of
339 accuracy have been detected. If no trap handler is installed the
340 operation continues with an imprecise small value, or zero if the
341 destination precision cannot hold the small exact result.
344 This exception is signalled if a rounded result is not exact (such as
345 when calculating the square root of two) or a result overflows without
349 @node Infinity and NaN
350 @subsection Infinity and NaN
355 @w{IEEE 754} floating point numbers can represent positive or negative
356 infinity, and @dfn{NaN} (not a number). These three values arise from
357 calculations whose result is undefined or cannot be represented
358 accurately. You can also deliberately set a floating-point variable to
359 any of them, which is sometimes useful. Some examples of calculations
360 that produce infinity or NaN:
364 @math{1/0 = @infinity{}}
365 @math{log (0) = -@infinity{}}
366 @math{sqrt (-1) = NaN}
370 $${1\over0} = \infty$$
372 $$\sqrt{-1} = \hbox{NaN}$$
375 When a calculation produces any of these values, an exception also
376 occurs; see @ref{FP Exceptions}.
378 The basic operations and math functions all accept infinity and NaN and
379 produce sensible output. Infinities propagate through calculations as
380 one would expect: for example, @math{2 + @infinity{} = @infinity{}},
381 @math{4/@infinity{} = 0}, atan @math{(@infinity{}) = @pi{}/2}. NaN, on
382 the other hand, infects any calculation that involves it. Unless the
383 calculation would produce the same result no matter what real value
384 replaced NaN, the result is NaN.
386 In comparison operations, positive infinity is larger than all values
387 except itself and NaN, and negative infinity is smaller than all values
388 except itself and NaN. NaN is @dfn{unordered}: it is not equal to,
389 greater than, or less than anything, @emph{including itself}. @code{x ==
390 x} is false if the value of @code{x} is NaN. You can use this to test
391 whether a value is NaN or not, but the recommended way to test for NaN
392 is with the @code{isnan} function (@pxref{Floating Point Classes}). In
393 addition, @code{<}, @code{>}, @code{<=}, and @code{>=} will raise an
394 exception when applied to NaNs.
396 @file{math.h} defines macros that allow you to explicitly set a variable
401 @deftypevr Macro float INFINITY
402 An expression representing positive infinity. It is equal to the value
403 produced by mathematical operations like @code{1.0 / 0.0}.
404 @code{-INFINITY} represents negative infinity.
406 You can test whether a floating-point value is infinite by comparing it
407 to this macro. However, this is not recommended; you should use the
408 @code{isfinite} macro instead. @xref{Floating Point Classes}.
410 This macro was introduced in the @w{ISO C99} standard.
415 @deftypevr Macro float NAN
416 An expression representing a value which is ``not a number''. This
417 macro is a GNU extension, available only on machines that support the
418 ``not a number'' value---that is to say, on all machines that support
421 You can use @samp{#ifdef NAN} to test whether the machine supports
422 NaN. (Of course, you must arrange for GNU extensions to be visible,
423 such as by defining @code{_GNU_SOURCE}, and then you must include
427 @w{IEEE 754} also allows for another unusual value: negative zero. This
428 value is produced when you divide a positive number by negative
429 infinity, or when a negative result is smaller than the limits of
430 representation. Negative zero behaves identically to zero in all
431 calculations, unless you explicitly test the sign bit with
432 @code{signbit} or @code{copysign}.
434 @node Status bit operations
435 @subsection Examining the FPU status word
437 @w{ISO C99} defines functions to query and manipulate the
438 floating-point status word. You can use these functions to check for
439 untrapped exceptions when it's convenient, rather than worrying about
440 them in the middle of a calculation.
442 These constants represent the various @w{IEEE 754} exceptions. Not all
443 FPUs report all the different exceptions. Each constant is defined if
444 and only if the FPU you are compiling for supports that exception, so
445 you can test for FPU support with @samp{#ifdef}. They are defined in
452 The inexact exception.
456 The divide by zero exception.
460 The underflow exception.
464 The overflow exception.
468 The invalid exception.
471 The macro @code{FE_ALL_EXCEPT} is the bitwise OR of all exception macros
472 which are supported by the FP implementation.
474 These functions allow you to clear exception flags, test for exceptions,
475 and save and restore the set of exceptions flagged.
479 @deftypefun int feclearexcept (int @var{excepts})
480 This function clears all of the supported exception flags indicated by
483 The function returns zero in case the operation was successful, a
484 non-zero value otherwise.
489 @deftypefun int feraiseexcept (int @var{excepts})
490 This function raises the supported exceptions indicated by
491 @var{excepts}. If more than one exception bit in @var{excepts} is set
492 the order in which the exceptions are raised is undefined except that
493 overflow (@code{FE_OVERFLOW}) or underflow (@code{FE_UNDERFLOW}) are
494 raised before inexact (@code{FE_INEXACT}). Whether for overflow or
495 underflow the inexact exception is also raised is also implementation
498 The function returns zero in case the operation was successful, a
499 non-zero value otherwise.
504 @deftypefun int fetestexcept (int @var{excepts})
505 Test whether the exception flags indicated by the parameter @var{except}
506 are currently set. If any of them are, a nonzero value is returned
507 which specifies which exceptions are set. Otherwise the result is zero.
510 To understand these functions, imagine that the status word is an
511 integer variable named @var{status}. @code{feclearexcept} is then
512 equivalent to @samp{status &= ~excepts} and @code{fetestexcept} is
513 equivalent to @samp{(status & excepts)}. The actual implementation may
514 be very different, of course.
516 Exception flags are only cleared when the program explicitly requests it,
517 by calling @code{feclearexcept}. If you want to check for exceptions
518 from a set of calculations, you should clear all the flags first. Here
519 is a simple example of the way to use @code{fetestexcept}:
525 feclearexcept (FE_ALL_EXCEPT);
527 raised = fetestexcept (FE_OVERFLOW | FE_INVALID);
528 if (raised & FE_OVERFLOW) @{ /* ... */ @}
529 if (raised & FE_INVALID) @{ /* ... */ @}
534 You cannot explicitly set bits in the status word. You can, however,
535 save the entire status word and restore it later. This is done with the
540 @deftypefun int fegetexceptflag (fexcept_t *@var{flagp}, int @var{excepts})
541 This function stores in the variable pointed to by @var{flagp} an
542 implementation-defined value representing the current setting of the
543 exception flags indicated by @var{excepts}.
545 The function returns zero in case the operation was successful, a
546 non-zero value otherwise.
551 @deftypefun int fesetexceptflag (const fexcept_t *@var{flagp}, int
553 This function restores the flags for the exceptions indicated by
554 @var{excepts} to the values stored in the variable pointed to by
557 The function returns zero in case the operation was successful, a
558 non-zero value otherwise.
561 Note that the value stored in @code{fexcept_t} bears no resemblance to
562 the bit mask returned by @code{fetestexcept}. The type may not even be
563 an integer. Do not attempt to modify an @code{fexcept_t} variable.
565 @node Math Error Reporting
566 @subsection Error Reporting by Mathematical Functions
567 @cindex errors, mathematical
571 Many of the math functions are defined only over a subset of the real or
572 complex numbers. Even if they are mathematically defined, their result
573 may be larger or smaller than the range representable by their return
574 type. These are known as @dfn{domain errors}, @dfn{overflows}, and
575 @dfn{underflows}, respectively. Math functions do several things when
576 one of these errors occurs. In this manual we will refer to the
577 complete response as @dfn{signalling} a domain error, overflow, or
580 When a math function suffers a domain error, it raises the invalid
581 exception and returns NaN. It also sets @var{errno} to @code{EDOM};
582 this is for compatibility with old systems that do not support @w{IEEE
583 754} exception handling. Likewise, when overflow occurs, math
584 functions raise the overflow exception and return @math{@infinity{}} or
585 @math{-@infinity{}} as appropriate. They also set @var{errno} to
586 @code{ERANGE}. When underflow occurs, the underflow exception is
587 raised, and zero (appropriately signed) is returned. @var{errno} may be
588 set to @code{ERANGE}, but this is not guaranteed.
590 Some of the math functions are defined mathematically to result in a
591 complex value over parts of their domains. The most familiar example of
592 this is taking the square root of a negative number. The complex math
593 functions, such as @code{csqrt}, will return the appropriate complex value
594 in this case. The real-valued functions, such as @code{sqrt}, will
595 signal a domain error.
597 Some older hardware does not support infinities. On that hardware,
598 overflows instead return a particular very large number (usually the
599 largest representable number). @file{math.h} defines macros you can use
600 to test for overflow on both old and new hardware.
604 @deftypevr Macro double HUGE_VAL
607 @deftypevrx Macro float HUGE_VALF
610 @deftypevrx Macro {long double} HUGE_VALL
611 An expression representing a particular very large number. On machines
612 that use @w{IEEE 754} floating point format, @code{HUGE_VAL} is infinity.
613 On other machines, it's typically the largest positive number that can
616 Mathematical functions return the appropriately typed version of
617 @code{HUGE_VAL} or @code{@minus{}HUGE_VAL} when the result is too large
622 @section Rounding Modes
624 Floating-point calculations are carried out internally with extra
625 precision, and then rounded to fit into the destination type. This
626 ensures that results are as precise as the input data. @w{IEEE 754}
627 defines four possible rounding modes:
630 @item Round to nearest.
631 This is the default mode. It should be used unless there is a specific
632 need for one of the others. In this mode results are rounded to the
633 nearest representable value. If the result is midway between two
634 representable values, the even representable is chosen. @dfn{Even} here
635 means the lowest-order bit is zero. This rounding mode prevents
636 statistical bias and guarantees numeric stability: round-off errors in a
637 lengthy calculation will remain smaller than half of @code{FLT_EPSILON}.
639 @c @item Round toward @math{+@infinity{}}
640 @item Round toward plus Infinity.
641 All results are rounded to the smallest representable value
642 which is greater than the result.
644 @c @item Round toward @math{-@infinity{}}
645 @item Round toward minus Infinity.
646 All results are rounded to the largest representable value which is less
649 @item Round toward zero.
650 All results are rounded to the largest representable value whose
651 magnitude is less than that of the result. In other words, if the
652 result is negative it is rounded up; if it is positive, it is rounded
657 @file{fenv.h} defines constants which you can use to refer to the
658 various rounding modes. Each one will be defined if and only if the FPU
659 supports the corresponding rounding mode.
672 Round toward @math{+@infinity{}}.
678 Round toward @math{-@infinity{}}.
682 @vindex FE_TOWARDZERO
687 Underflow is an unusual case. Normally, @w{IEEE 754} floating point
688 numbers are always normalized (@pxref{Floating Point Concepts}).
689 Numbers smaller than @math{2^r} (where @math{r} is the minimum exponent,
690 @code{FLT_MIN_RADIX-1} for @var{float}) cannot be represented as
691 normalized numbers. Rounding all such numbers to zero or @math{2^r}
692 would cause some algorithms to fail at 0. Therefore, they are left in
693 denormalized form. That produces loss of precision, since some bits of
694 the mantissa are stolen to indicate the decimal point.
696 If a result is too small to be represented as a denormalized number, it
697 is rounded to zero. However, the sign of the result is preserved; if
698 the calculation was negative, the result is @dfn{negative zero}.
699 Negative zero can also result from some operations on infinity, such as
700 @math{4/-@infinity{}}. Negative zero behaves identically to zero except
701 when the @code{copysign} or @code{signbit} functions are used to check
702 the sign bit directly.
704 At any time one of the above four rounding modes is selected. You can
705 find out which one with this function:
709 @deftypefun int fegetround (void)
710 Returns the currently selected rounding mode, represented by one of the
711 values of the defined rounding mode macros.
715 To change the rounding mode, use this function:
719 @deftypefun int fesetround (int @var{round})
720 Changes the currently selected rounding mode to @var{round}. If
721 @var{round} does not correspond to one of the supported rounding modes
722 nothing is changed. @code{fesetround} returns a nonzero value if it
723 changed the rounding mode, zero if the mode is not supported.
726 You should avoid changing the rounding mode if possible. It can be an
727 expensive operation; also, some hardware requires you to compile your
728 program differently for it to work. The resulting code may run slower.
729 See your compiler documentation for details.
730 @c This section used to claim that functions existed to round one number
731 @c in a specific fashion. I can't find any functions in the library
734 @node Control Functions
735 @section Floating-Point Control Functions
737 @w{IEEE 754} floating-point implementations allow the programmer to
738 decide whether traps will occur for each of the exceptions, by setting
739 bits in the @dfn{control word}. In C, traps result in the program
740 receiving the @code{SIGFPE} signal; see @ref{Signal Handling}.
742 @strong{Note:} @w{IEEE 754} says that trap handlers are given details of
743 the exceptional situation, and can set the result value. C signals do
744 not provide any mechanism to pass this information back and forth.
745 Trapping exceptions in C is therefore not very useful.
747 It is sometimes necessary to save the state of the floating-point unit
748 while you perform some calculation. The library provides functions
749 which save and restore the exception flags, the set of exceptions that
750 generate traps, and the rounding mode. This information is known as the
751 @dfn{floating-point environment}.
753 The functions to save and restore the floating-point environment all use
754 a variable of type @code{fenv_t} to store information. This type is
755 defined in @file{fenv.h}. Its size and contents are
756 implementation-defined. You should not attempt to manipulate a variable
757 of this type directly.
759 To save the state of the FPU, use one of these functions:
763 @deftypefun int fegetenv (fenv_t *@var{envp})
764 Store the floating-point environment in the variable pointed to by
767 The function returns zero in case the operation was successful, a
768 non-zero value otherwise.
773 @deftypefun int feholdexcept (fenv_t *@var{envp})
774 Store the current floating-point environment in the object pointed to by
775 @var{envp}. Then clear all exception flags, and set the FPU to trap no
776 exceptions. Not all FPUs support trapping no exceptions; if
777 @code{feholdexcept} cannot set this mode, it returns zero. If it
778 succeeds, it returns a nonzero value.
781 The functions which restore the floating-point environment can take two
786 Pointers to @code{fenv_t} objects, which were initialized previously by a
787 call to @code{fegetenv} or @code{feholdexcept}.
790 The special macro @code{FE_DFL_ENV} which represents the floating-point
791 environment as it was available at program start.
793 Implementation defined macros with names starting with @code{FE_}.
795 @vindex FE_NOMASK_ENV
796 If possible, the GNU C Library defines a macro @code{FE_NOMASK_ENV}
797 which represents an environment where every exception raised causes a
798 trap to occur. You can test for this macro using @code{#ifdef}. It is
799 only defined if @code{_GNU_SOURCE} is defined.
801 Some platforms might define other predefined environments.
805 To set the floating-point environment, you can use either of these
810 @deftypefun int fesetenv (const fenv_t *@var{envp})
811 Set the floating-point environment to that described by @var{envp}.
813 The function returns zero in case the operation was successful, a
814 non-zero value otherwise.
819 @deftypefun int feupdateenv (const fenv_t *@var{envp})
820 Like @code{fesetenv}, this function sets the floating-point environment
821 to that described by @var{envp}. However, if any exceptions were
822 flagged in the status word before @code{feupdateenv} was called, they
823 remain flagged after the call. In other words, after @code{feupdateenv}
824 is called, the status word is the bitwise OR of the previous status word
825 and the one saved in @var{envp}.
827 The function returns zero in case the operation was successful, a
828 non-zero value otherwise.
832 To control for individual exceptions if raising them causes a trap to
833 occur, you can use the following two functions.
835 @strong{Portability Note:} These functions are all GNU extensions.
839 @deftypefun int feenableexcept (int @var{excepts})
840 This functions enables traps for each of the exceptions as indicated by
841 the parameter @var{except}. The individual excepetions are described in
842 @ref{Status bit operations}. Only the specified exceptions are
843 enabled, the status of the other exceptions is not changed.
845 The function returns the previous enabled exceptions in case the
846 operation was successful, @code{-1} otherwise.
851 @deftypefun int fedisableexcept (int @var{excepts})
852 This functions disables traps for each of the exceptions as indicated by
853 the parameter @var{except}. The individual excepetions are described in
854 @ref{Status bit operations}. Only the specified exceptions are
855 disabled, the status of the other exceptions is not changed.
857 The function returns the previous enabled exceptions in case the
858 operation was successful, @code{-1} otherwise.
863 @deftypefun int fegetexcept (int @var{excepts})
864 The function returns a bitmask of all currently enabled exceptions. It
865 returns @code{-1} in case of failure.
868 @node Arithmetic Functions
869 @section Arithmetic Functions
871 The C library provides functions to do basic operations on
872 floating-point numbers. These include absolute value, maximum and minimum,
873 normalization, bit twiddling, rounding, and a few others.
876 * Absolute Value:: Absolute values of integers and floats.
877 * Normalization Functions:: Extracting exponents and putting them back.
878 * Rounding Functions:: Rounding floats to integers.
879 * Remainder Functions:: Remainders on division, precisely defined.
880 * FP Bit Twiddling:: Sign bit adjustment. Adding epsilon.
881 * FP Comparison Functions:: Comparisons without risk of exceptions.
882 * Misc FP Arithmetic:: Max, min, positive difference, multiply-add.
886 @subsection Absolute Value
887 @cindex absolute value functions
889 These functions are provided for obtaining the @dfn{absolute value} (or
890 @dfn{magnitude}) of a number. The absolute value of a real number
891 @var{x} is @var{x} if @var{x} is positive, @minus{}@var{x} if @var{x} is
892 negative. For a complex number @var{z}, whose real part is @var{x} and
893 whose imaginary part is @var{y}, the absolute value is @w{@code{sqrt
894 (@var{x}*@var{x} + @var{y}*@var{y})}}.
898 Prototypes for @code{abs}, @code{labs} and @code{llabs} are in @file{stdlib.h};
899 @code{imaxabs} is declared in @file{inttypes.h};
900 @code{fabs}, @code{fabsf} and @code{fabsl} are declared in @file{math.h}.
901 @code{cabs}, @code{cabsf} and @code{cabsl} are declared in @file{complex.h}.
905 @deftypefun int abs (int @var{number})
908 @deftypefunx {long int} labs (long int @var{number})
911 @deftypefunx {long long int} llabs (long long int @var{number})
914 @deftypefunx intmax_t imaxabs (intmax_t @var{number})
915 These functions return the absolute value of @var{number}.
917 Most computers use a two's complement integer representation, in which
918 the absolute value of @code{INT_MIN} (the smallest possible @code{int})
919 cannot be represented; thus, @w{@code{abs (INT_MIN)}} is not defined.
921 @code{llabs} and @code{imaxdiv} are new to @w{ISO C99}.
926 @deftypefun double fabs (double @var{number})
929 @deftypefunx float fabsf (float @var{number})
932 @deftypefunx {long double} fabsl (long double @var{number})
933 This function returns the absolute value of the floating-point number
939 @deftypefun double cabs (complex double @var{z})
942 @deftypefunx float cabsf (complex float @var{z})
945 @deftypefunx {long double} cabsl (complex long double @var{z})
946 These functions return the absolute value of the complex number @var{z}
947 (@pxref{Complex Numbers}). The absolute value of a complex number is:
950 sqrt (creal (@var{z}) * creal (@var{z}) + cimag (@var{z}) * cimag (@var{z}))
953 This function should always be used instead of the direct formula
954 because it takes special care to avoid losing precision. It may also
955 take advantage of hardware support for this operation. See @code{hypot}
956 in @ref{Exponents and Logarithms}.
959 @node Normalization Functions
960 @subsection Normalization Functions
961 @cindex normalization functions (floating-point)
963 The functions described in this section are primarily provided as a way
964 to efficiently perform certain low-level manipulations on floating point
965 numbers that are represented internally using a binary radix;
966 see @ref{Floating Point Concepts}. These functions are required to
967 have equivalent behavior even if the representation does not use a radix
968 of 2, but of course they are unlikely to be particularly efficient in
972 All these functions are declared in @file{math.h}.
976 @deftypefun double frexp (double @var{value}, int *@var{exponent})
979 @deftypefunx float frexpf (float @var{value}, int *@var{exponent})
982 @deftypefunx {long double} frexpl (long double @var{value}, int *@var{exponent})
983 These functions are used to split the number @var{value}
984 into a normalized fraction and an exponent.
986 If the argument @var{value} is not zero, the return value is @var{value}
987 times a power of two, and is always in the range 1/2 (inclusive) to 1
988 (exclusive). The corresponding exponent is stored in
989 @code{*@var{exponent}}; the return value multiplied by 2 raised to this
990 exponent equals the original number @var{value}.
992 For example, @code{frexp (12.8, &exponent)} returns @code{0.8} and
993 stores @code{4} in @code{exponent}.
995 If @var{value} is zero, then the return value is zero and
996 zero is stored in @code{*@var{exponent}}.
1001 @deftypefun double ldexp (double @var{value}, int @var{exponent})
1004 @deftypefunx float ldexpf (float @var{value}, int @var{exponent})
1007 @deftypefunx {long double} ldexpl (long double @var{value}, int @var{exponent})
1008 These functions return the result of multiplying the floating-point
1009 number @var{value} by 2 raised to the power @var{exponent}. (It can
1010 be used to reassemble floating-point numbers that were taken apart
1013 For example, @code{ldexp (0.8, 4)} returns @code{12.8}.
1016 The following functions, which come from BSD, provide facilities
1017 equivalent to those of @code{ldexp} and @code{frexp}.
1021 @deftypefun double logb (double @var{x})
1024 @deftypefunx float logbf (float @var{x})
1027 @deftypefunx {long double} logbl (long double @var{x})
1028 These functions return the integer part of the base-2 logarithm of
1029 @var{x}, an integer value represented in type @code{double}. This is
1030 the highest integer power of @code{2} contained in @var{x}. The sign of
1031 @var{x} is ignored. For example, @code{logb (3.5)} is @code{1.0} and
1032 @code{logb (4.0)} is @code{2.0}.
1034 When @code{2} raised to this power is divided into @var{x}, it gives a
1035 quotient between @code{1} (inclusive) and @code{2} (exclusive).
1037 If @var{x} is zero, the return value is minus infinity if the machine
1038 supports infinities, and a very small number if it does not. If @var{x}
1039 is infinity, the return value is infinity.
1041 For finite @var{x}, the value returned by @code{logb} is one less than
1042 the value that @code{frexp} would store into @code{*@var{exponent}}.
1047 @deftypefun double scalb (double @var{value}, int @var{exponent})
1050 @deftypefunx float scalbf (float @var{value}, int @var{exponent})
1053 @deftypefunx {long double} scalbl (long double @var{value}, int @var{exponent})
1054 The @code{scalb} function is the BSD name for @code{ldexp}.
1059 @deftypefun {long long int} scalbn (double @var{x}, int n)
1062 @deftypefunx {long long int} scalbnf (float @var{x}, int n)
1065 @deftypefunx {long long int} scalbnl (long double @var{x}, int n)
1066 @code{scalbn} is identical to @code{scalb}, except that the exponent
1067 @var{n} is an @code{int} instead of a floating-point number.
1072 @deftypefun {long long int} scalbln (double @var{x}, long int n)
1075 @deftypefunx {long long int} scalblnf (float @var{x}, long int n)
1078 @deftypefunx {long long int} scalblnl (long double @var{x}, long int n)
1079 @code{scalbln} is identical to @code{scalb}, except that the exponent
1080 @var{n} is a @code{long int} instead of a floating-point number.
1085 @deftypefun {long long int} significand (double @var{x})
1088 @deftypefunx {long long int} significandf (float @var{x})
1091 @deftypefunx {long long int} significandl (long double @var{x})
1092 @code{significand} returns the mantissa of @var{x} scaled to the range
1094 It is equivalent to @w{@code{scalb (@var{x}, (double) -ilogb (@var{x}))}}.
1096 This function exists mainly for use in certain standardized tests
1097 of @w{IEEE 754} conformance.
1100 @node Rounding Functions
1101 @subsection Rounding Functions
1102 @cindex converting floats to integers
1105 The functions listed here perform operations such as rounding and
1106 truncation of floating-point values. Some of these functions convert
1107 floating point numbers to integer values. They are all declared in
1110 You can also convert floating-point numbers to integers simply by
1111 casting them to @code{int}. This discards the fractional part,
1112 effectively rounding towards zero. However, this only works if the
1113 result can actually be represented as an @code{int}---for very large
1114 numbers, this is impossible. The functions listed here return the
1115 result as a @code{double} instead to get around this problem.
1119 @deftypefun double ceil (double @var{x})
1122 @deftypefunx float ceilf (float @var{x})
1125 @deftypefunx {long double} ceill (long double @var{x})
1126 These functions round @var{x} upwards to the nearest integer,
1127 returning that value as a @code{double}. Thus, @code{ceil (1.5)}
1133 @deftypefun double floor (double @var{x})
1136 @deftypefunx float floorf (float @var{x})
1139 @deftypefunx {long double} floorl (long double @var{x})
1140 These functions round @var{x} downwards to the nearest
1141 integer, returning that value as a @code{double}. Thus, @code{floor
1142 (1.5)} is @code{1.0} and @code{floor (-1.5)} is @code{-2.0}.
1147 @deftypefun double trunc (double @var{x})
1150 @deftypefunx float truncf (float @var{x})
1153 @deftypefunx {long double} truncl (long double @var{x})
1154 @code{trunc} is another name for @code{floor}
1159 @deftypefun double rint (double @var{x})
1162 @deftypefunx float rintf (float @var{x})
1165 @deftypefunx {long double} rintl (long double @var{x})
1166 These functions round @var{x} to an integer value according to the
1167 current rounding mode. @xref{Floating Point Parameters}, for
1168 information about the various rounding modes. The default
1169 rounding mode is to round to the nearest integer; some machines
1170 support other modes, but round-to-nearest is always used unless
1171 you explicitly select another.
1173 If @var{x} was not initially an integer, these functions raise the
1179 @deftypefun double nearbyint (double @var{x})
1182 @deftypefunx float nearbyintf (float @var{x})
1185 @deftypefunx {long double} nearbyintl (long double @var{x})
1186 These functions return the same value as the @code{rint} functions, but
1187 do not raise the inexact exception if @var{x} is not an integer.
1192 @deftypefun double round (double @var{x})
1195 @deftypefunx float roundf (float @var{x})
1198 @deftypefunx {long double} roundl (long double @var{x})
1199 These functions are similar to @code{rint}, but they round halfway
1200 cases away from zero instead of to the nearest even integer.
1205 @deftypefun {long int} lrint (double @var{x})
1208 @deftypefunx {long int} lrintf (float @var{x})
1211 @deftypefunx {long int} lrintl (long double @var{x})
1212 These functions are just like @code{rint}, but they return a
1213 @code{long int} instead of a floating-point number.
1218 @deftypefun {long long int} llrint (double @var{x})
1221 @deftypefunx {long long int} llrintf (float @var{x})
1224 @deftypefunx {long long int} llrintl (long double @var{x})
1225 These functions are just like @code{rint}, but they return a
1226 @code{long long int} instead of a floating-point number.
1231 @deftypefun {long int} lround (double @var{x})
1234 @deftypefunx {long int} lroundf (float @var{x})
1237 @deftypefunx {long int} lroundl (long double @var{x})
1238 These functions are just like @code{round}, but they return a
1239 @code{long int} instead of a floating-point number.
1244 @deftypefun {long long int} llround (double @var{x})
1247 @deftypefunx {long long int} llroundf (float @var{x})
1250 @deftypefunx {long long int} llroundl (long double @var{x})
1251 These functions are just like @code{round}, but they return a
1252 @code{long long int} instead of a floating-point number.
1258 @deftypefun double modf (double @var{value}, double *@var{integer-part})
1261 @deftypefunx float modff (float @var{value}, float *@var{integer-part})
1264 @deftypefunx {long double} modfl (long double @var{value}, long double *@var{integer-part})
1265 These functions break the argument @var{value} into an integer part and a
1266 fractional part (between @code{-1} and @code{1}, exclusive). Their sum
1267 equals @var{value}. Each of the parts has the same sign as @var{value},
1268 and the integer part is always rounded toward zero.
1270 @code{modf} stores the integer part in @code{*@var{integer-part}}, and
1271 returns the fractional part. For example, @code{modf (2.5, &intpart)}
1272 returns @code{0.5} and stores @code{2.0} into @code{intpart}.
1275 @node Remainder Functions
1276 @subsection Remainder Functions
1278 The functions in this section compute the remainder on division of two
1279 floating-point numbers. Each is a little different; pick the one that
1284 @deftypefun double fmod (double @var{numerator}, double @var{denominator})
1287 @deftypefunx float fmodf (float @var{numerator}, float @var{denominator})
1290 @deftypefunx {long double} fmodl (long double @var{numerator}, long double @var{denominator})
1291 These functions compute the remainder from the division of
1292 @var{numerator} by @var{denominator}. Specifically, the return value is
1293 @code{@var{numerator} - @w{@var{n} * @var{denominator}}}, where @var{n}
1294 is the quotient of @var{numerator} divided by @var{denominator}, rounded
1295 towards zero to an integer. Thus, @w{@code{fmod (6.5, 2.3)}} returns
1296 @code{1.9}, which is @code{6.5} minus @code{4.6}.
1298 The result has the same sign as the @var{numerator} and has magnitude
1299 less than the magnitude of the @var{denominator}.
1301 If @var{denominator} is zero, @code{fmod} signals a domain error.
1306 @deftypefun double drem (double @var{numerator}, double @var{denominator})
1309 @deftypefunx float dremf (float @var{numerator}, float @var{denominator})
1312 @deftypefunx {long double} dreml (long double @var{numerator}, long double @var{denominator})
1313 These functions are like @code{fmod} except that they rounds the
1314 internal quotient @var{n} to the nearest integer instead of towards zero
1315 to an integer. For example, @code{drem (6.5, 2.3)} returns @code{-0.4},
1316 which is @code{6.5} minus @code{6.9}.
1318 The absolute value of the result is less than or equal to half the
1319 absolute value of the @var{denominator}. The difference between
1320 @code{fmod (@var{numerator}, @var{denominator})} and @code{drem
1321 (@var{numerator}, @var{denominator})} is always either
1322 @var{denominator}, minus @var{denominator}, or zero.
1324 If @var{denominator} is zero, @code{drem} signals a domain error.
1329 @deftypefun double remainder (double @var{numerator}, double @var{denominator})
1332 @deftypefunx float remainderf (float @var{numerator}, float @var{denominator})
1335 @deftypefunx {long double} remainderl (long double @var{numerator}, long double @var{denominator})
1336 This function is another name for @code{drem}.
1339 @node FP Bit Twiddling
1340 @subsection Setting and modifying single bits of FP values
1341 @cindex FP arithmetic
1343 There are some operations that are too complicated or expensive to
1344 perform by hand on floating-point numbers. @w{ISO C99} defines
1345 functions to do these operations, which mostly involve changing single
1350 @deftypefun double copysign (double @var{x}, double @var{y})
1353 @deftypefunx float copysignf (float @var{x}, float @var{y})
1356 @deftypefunx {long double} copysignl (long double @var{x}, long double @var{y})
1357 These functions return @var{x} but with the sign of @var{y}. They work
1358 even if @var{x} or @var{y} are NaN or zero. Both of these can carry a
1359 sign (although not all implementations support it) and this is one of
1360 the few operations that can tell the difference.
1362 @code{copysign} never raises an exception.
1363 @c except signalling NaNs
1365 This function is defined in @w{IEC 559} (and the appendix with
1366 recommended functions in @w{IEEE 754}/@w{IEEE 854}).
1371 @deftypefun int signbit (@emph{float-type} @var{x})
1372 @code{signbit} is a generic macro which can work on all floating-point
1373 types. It returns a nonzero value if the value of @var{x} has its sign
1376 This is not the same as @code{x < 0.0}, because @w{IEEE 754} floating
1377 point allows zero to be signed. The comparison @code{-0.0 < 0.0} is
1378 false, but @code{signbit (-0.0)} will return a nonzero value.
1383 @deftypefun double nextafter (double @var{x}, double @var{y})
1386 @deftypefunx float nextafterf (float @var{x}, float @var{y})
1389 @deftypefunx {long double} nextafterl (long double @var{x}, long double @var{y})
1390 The @code{nextafter} function returns the next representable neighbor of
1391 @var{x} in the direction towards @var{y}. The size of the step between
1392 @var{x} and the result depends on the type of the result. If
1393 @math{@var{x} = @var{y}} the function simply returns @var{x}. If either
1394 value is @code{NaN}, @code{NaN} is returned. Otherwise
1395 a value corresponding to the value of the least significant bit in the
1396 mantissa is added or subtracted, depending on the direction.
1397 @code{nextafter} will signal overflow or underflow if the result goes
1398 outside of the range of normalized numbers.
1400 This function is defined in @w{IEC 559} (and the appendix with
1401 recommended functions in @w{IEEE 754}/@w{IEEE 854}).
1406 @deftypefun double nexttoward (double @var{x}, long double @var{y})
1409 @deftypefunx float nexttowardf (float @var{x}, long double @var{y})
1412 @deftypefunx {long double} nexttowardl (long double @var{x}, long double @var{y})
1413 These functions are identical to the corresponding versions of
1414 @code{nextafter} except that their second argument is a @code{long
1421 @deftypefun double nan (const char *@var{tagp})
1424 @deftypefunx float nanf (const char *@var{tagp})
1427 @deftypefunx {long double} nanl (const char *@var{tagp})
1428 The @code{nan} function returns a representation of NaN, provided that
1429 NaN is supported by the target platform.
1430 @code{nan ("@var{n-char-sequence}")} is equivalent to
1431 @code{strtod ("NAN(@var{n-char-sequence})")}.
1433 The argument @var{tagp} is used in an unspecified manner. On @w{IEEE
1434 754} systems, there are many representations of NaN, and @var{tagp}
1435 selects one. On other systems it may do nothing.
1438 @node FP Comparison Functions
1439 @subsection Floating-Point Comparison Functions
1440 @cindex unordered comparison
1442 The standard C comparison operators provoke exceptions when one or other
1443 of the operands is NaN. For example,
1450 will raise an exception if @var{a} is NaN. (This does @emph{not}
1451 happen with @code{==} and @code{!=}; those merely return false and true,
1452 respectively, when NaN is examined.) Frequently this exception is
1453 undesirable. @w{ISO C99} therefore defines comparison functions that
1454 do not raise exceptions when NaN is examined. All of the functions are
1455 implemented as macros which allow their arguments to be of any
1456 floating-point type. The macros are guaranteed to evaluate their
1457 arguments only once.
1461 @deftypefn Macro int isgreater (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1462 This macro determines whether the argument @var{x} is greater than
1463 @var{y}. It is equivalent to @code{(@var{x}) > (@var{y})}, but no
1464 exception is raised if @var{x} or @var{y} are NaN.
1469 @deftypefn Macro int isgreaterequal (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1470 This macro determines whether the argument @var{x} is greater than or
1471 equal to @var{y}. It is equivalent to @code{(@var{x}) >= (@var{y})}, but no
1472 exception is raised if @var{x} or @var{y} are NaN.
1477 @deftypefn Macro int isless (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1478 This macro determines whether the argument @var{x} is less than @var{y}.
1479 It is equivalent to @code{(@var{x}) < (@var{y})}, but no exception is
1480 raised if @var{x} or @var{y} are NaN.
1485 @deftypefn Macro int islessequal (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1486 This macro determines whether the argument @var{x} is less than or equal
1487 to @var{y}. It is equivalent to @code{(@var{x}) <= (@var{y})}, but no
1488 exception is raised if @var{x} or @var{y} are NaN.
1493 @deftypefn Macro int islessgreater (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1494 This macro determines whether the argument @var{x} is less or greater
1495 than @var{y}. It is equivalent to @code{(@var{x}) < (@var{y}) ||
1496 (@var{x}) > (@var{y})} (although it only evaluates @var{x} and @var{y}
1497 once), but no exception is raised if @var{x} or @var{y} are NaN.
1499 This macro is not equivalent to @code{@var{x} != @var{y}}, because that
1500 expression is true if @var{x} or @var{y} are NaN.
1505 @deftypefn Macro int isunordered (@emph{real-floating} @var{x}, @emph{real-floating} @var{y})
1506 This macro determines whether its arguments are unordered. In other
1507 words, it is true if @var{x} or @var{y} are NaN, and false otherwise.
1510 Not all machines provide hardware support for these operations. On
1511 machines that don't, the macros can be very slow. Therefore, you should
1512 not use these functions when NaN is not a concern.
1514 @strong{Note:} There are no macros @code{isequal} or @code{isunequal}.
1515 They are unnecessary, because the @code{==} and @code{!=} operators do
1516 @emph{not} throw an exception if one or both of the operands are NaN.
1518 @node Misc FP Arithmetic
1519 @subsection Miscellaneous FP arithmetic functions
1522 @cindex positive difference
1523 @cindex multiply-add
1525 The functions in this section perform miscellaneous but common
1526 operations that are awkward to express with C operators. On some
1527 processors these functions can use special machine instructions to
1528 perform these operations faster than the equivalent C code.
1532 @deftypefun double fmin (double @var{x}, double @var{y})
1535 @deftypefunx float fminf (float @var{x}, float @var{y})
1538 @deftypefunx {long double} fminl (long double @var{x}, long double @var{y})
1539 The @code{fmin} function returns the lesser of the two values @var{x}
1540 and @var{y}. It is similar to the expression
1542 ((x) < (y) ? (x) : (y))
1544 except that @var{x} and @var{y} are only evaluated once.
1546 If an argument is NaN, the other argument is returned. If both arguments
1547 are NaN, NaN is returned.
1552 @deftypefun double fmax (double @var{x}, double @var{y})
1555 @deftypefunx float fmaxf (float @var{x}, float @var{y})
1558 @deftypefunx {long double} fmaxl (long double @var{x}, long double @var{y})
1559 The @code{fmax} function returns the greater of the two values @var{x}
1562 If an argument is NaN, the other argument is returned. If both arguments
1563 are NaN, NaN is returned.
1568 @deftypefun double fdim (double @var{x}, double @var{y})
1571 @deftypefunx float fdimf (float @var{x}, float @var{y})
1574 @deftypefunx {long double} fdiml (long double @var{x}, long double @var{y})
1575 The @code{fdim} function returns the positive difference between
1576 @var{x} and @var{y}. The positive difference is @math{@var{x} -
1577 @var{y}} if @var{x} is greater than @var{y}, and @math{0} otherwise.
1579 If @var{x}, @var{y}, or both are NaN, NaN is returned.
1584 @deftypefun double fma (double @var{x}, double @var{y}, double @var{z})
1587 @deftypefunx float fmaf (float @var{x}, float @var{y}, float @var{z})
1590 @deftypefunx {long double} fmal (long double @var{x}, long double @var{y}, long double @var{z})
1592 The @code{fma} function performs floating-point multiply-add. This is
1593 the operation @math{(@var{x} @mul{} @var{y}) + @var{z}}, but the
1594 intermediate result is not rounded to the destination type. This can
1595 sometimes improve the precision of a calculation.
1597 This function was introduced because some processors have a special
1598 instruction to perform multiply-add. The C compiler cannot use it
1599 directly, because the expression @samp{x*y + z} is defined to round the
1600 intermediate result. @code{fma} lets you choose when you want to round
1604 On processors which do not implement multiply-add in hardware,
1605 @code{fma} can be very slow since it must avoid intermediate rounding.
1606 @file{math.h} defines the symbols @code{FP_FAST_FMA},
1607 @code{FP_FAST_FMAF}, and @code{FP_FAST_FMAL} when the corresponding
1608 version of @code{fma} is no slower than the expression @samp{x*y + z}.
1609 In the GNU C library, this always means the operation is implemented in
1613 @node Complex Numbers
1614 @section Complex Numbers
1616 @cindex complex numbers
1618 @w{ISO C99} introduces support for complex numbers in C. This is done
1619 with a new type qualifier, @code{complex}. It is a keyword if and only
1620 if @file{complex.h} has been included. There are three complex types,
1621 corresponding to the three real types: @code{float complex},
1622 @code{double complex}, and @code{long double complex}.
1624 To construct complex numbers you need a way to indicate the imaginary
1625 part of a number. There is no standard notation for an imaginary
1626 floating point constant. Instead, @file{complex.h} defines two macros
1627 that can be used to create complex numbers.
1629 @deftypevr Macro {const float complex} _Complex_I
1630 This macro is a representation of the complex number ``@math{0+1i}''.
1631 Multiplying a real floating-point value by @code{_Complex_I} gives a
1632 complex number whose value is purely imaginary. You can use this to
1633 construct complex constants:
1636 @math{3.0 + 4.0i} = @code{3.0 + 4.0 * _Complex_I}
1639 Note that @code{_Complex_I * _Complex_I} has the value @code{-1}, but
1640 the type of that value is @code{complex}.
1643 @c Put this back in when gcc supports _Imaginary_I. It's too confusing.
1646 Without an optimizing compiler this is more expensive than the use of
1647 @code{_Imaginary_I} but with is better than nothing. You can avoid all
1648 the hassles if you use the @code{I} macro below if the name is not
1651 @deftypevr Macro {const float imaginary} _Imaginary_I
1652 This macro is a representation of the value ``@math{1i}''. I.e., it is
1656 _Imaginary_I * _Imaginary_I = -1
1660 The result is not of type @code{float imaginary} but instead @code{float}.
1661 One can use it to easily construct complex number like in
1664 3.0 - _Imaginary_I * 4.0
1668 which results in the complex number with a real part of 3.0 and a
1669 imaginary part -4.0.
1674 @code{_Complex_I} is a bit of a mouthful. @file{complex.h} also defines
1675 a shorter name for the same constant.
1677 @deftypevr Macro {const float complex} I
1678 This macro has exactly the same value as @code{_Complex_I}. Most of the
1679 time it is preferable. However, it causes problems if you want to use
1680 the identifier @code{I} for something else. You can safely write
1683 #include <complex.h>
1688 if you need @code{I} for your own purposes. (In that case we recommend
1689 you also define some other short name for @code{_Complex_I}, such as
1693 If the implementation does not support the @code{imaginary} types
1694 @code{I} is defined as @code{_Complex_I} which is the second best
1695 solution. It still can be used in the same way but requires a most
1696 clever compiler to get the same results.
1700 @node Operations on Complex
1701 @section Projections, Conjugates, and Decomposing of Complex Numbers
1702 @cindex project complex numbers
1703 @cindex conjugate complex numbers
1704 @cindex decompose complex numbers
1707 @w{ISO C99} also defines functions that perform basic operations on
1708 complex numbers, such as decomposition and conjugation. The prototypes
1709 for all these functions are in @file{complex.h}. All functions are
1710 available in three variants, one for each of the three complex types.
1714 @deftypefun double creal (complex double @var{z})
1717 @deftypefunx float crealf (complex float @var{z})
1720 @deftypefunx {long double} creall (complex long double @var{z})
1721 These functions return the real part of the complex number @var{z}.
1726 @deftypefun double cimag (complex double @var{z})
1729 @deftypefunx float cimagf (complex float @var{z})
1732 @deftypefunx {long double} cimagl (complex long double @var{z})
1733 These functions return the imaginary part of the complex number @var{z}.
1738 @deftypefun {complex double} conj (complex double @var{z})
1741 @deftypefunx {complex float} conjf (complex float @var{z})
1744 @deftypefunx {complex long double} conjl (complex long double @var{z})
1745 These functions return the conjugate value of the complex number
1746 @var{z}. The conjugate of a complex number has the same real part and a
1747 negated imaginary part. In other words, @samp{conj(a + bi) = a + -bi}.
1752 @deftypefun double carg (complex double @var{z})
1755 @deftypefunx float cargf (complex float @var{z})
1758 @deftypefunx {long double} cargl (complex long double @var{z})
1759 These functions return the argument of the complex number @var{z}.
1760 The argument of a complex number is the angle in the complex plane
1761 between the positive real axis and a line passing through zero and the
1762 number. This angle is measured in the usual fashion and ranges from @math{0}
1765 @code{carg} has a branch cut along the positive real axis.
1770 @deftypefun {complex double} cproj (complex double @var{z})
1773 @deftypefunx {complex float} cprojf (complex float @var{z})
1776 @deftypefunx {complex long double} cprojl (complex long double @var{z})
1777 These functions return the projection of the complex value @var{z} onto
1778 the Riemann sphere. Values with a infinite imaginary part are projected
1779 to positive infinity on the real axis, even if the real part is NaN. If
1780 the real part is infinite, the result is equivalent to
1783 INFINITY + I * copysign (0.0, cimag (z))
1787 @node Integer Division
1788 @section Integer Division
1789 @cindex integer division functions
1791 This section describes functions for performing integer division. These
1792 functions are redundant when GNU CC is used, because in GNU C the
1793 @samp{/} operator always rounds towards zero. But in other C
1794 implementations, @samp{/} may round differently with negative arguments.
1795 @code{div} and @code{ldiv} are useful because they specify how to round
1796 the quotient: towards zero. The remainder has the same sign as the
1799 These functions are specified to return a result @var{r} such that the value
1800 @code{@var{r}.quot*@var{denominator} + @var{r}.rem} equals
1804 To use these facilities, you should include the header file
1805 @file{stdlib.h} in your program.
1809 @deftp {Data Type} div_t
1810 This is a structure type used to hold the result returned by the @code{div}
1811 function. It has the following members:
1815 The quotient from the division.
1818 The remainder from the division.
1824 @deftypefun div_t div (int @var{numerator}, int @var{denominator})
1825 This function @code{div} computes the quotient and remainder from
1826 the division of @var{numerator} by @var{denominator}, returning the
1827 result in a structure of type @code{div_t}.
1829 If the result cannot be represented (as in a division by zero), the
1830 behavior is undefined.
1832 Here is an example, albeit not a very useful one.
1836 result = div (20, -6);
1840 Now @code{result.quot} is @code{-3} and @code{result.rem} is @code{2}.
1845 @deftp {Data Type} ldiv_t
1846 This is a structure type used to hold the result returned by the @code{ldiv}
1847 function. It has the following members:
1851 The quotient from the division.
1854 The remainder from the division.
1857 (This is identical to @code{div_t} except that the components are of
1858 type @code{long int} rather than @code{int}.)
1863 @deftypefun ldiv_t ldiv (long int @var{numerator}, long int @var{denominator})
1864 The @code{ldiv} function is similar to @code{div}, except that the
1865 arguments are of type @code{long int} and the result is returned as a
1866 structure of type @code{ldiv_t}.
1871 @deftp {Data Type} lldiv_t
1872 This is a structure type used to hold the result returned by the @code{lldiv}
1873 function. It has the following members:
1876 @item long long int quot
1877 The quotient from the division.
1879 @item long long int rem
1880 The remainder from the division.
1883 (This is identical to @code{div_t} except that the components are of
1884 type @code{long long int} rather than @code{int}.)
1889 @deftypefun lldiv_t lldiv (long long int @var{numerator}, long long int @var{denominator})
1890 The @code{lldiv} function is like the @code{div} function, but the
1891 arguments are of type @code{long long int} and the result is returned as
1892 a structure of type @code{lldiv_t}.
1894 The @code{lldiv} function was added in @w{ISO C99}.
1899 @deftp {Data Type} imaxdiv_t
1900 This is a structure type used to hold the result returned by the @code{imaxdiv}
1901 function. It has the following members:
1905 The quotient from the division.
1908 The remainder from the division.
1911 (This is identical to @code{div_t} except that the components are of
1912 type @code{intmax_t} rather than @code{int}.)
1917 @deftypefun imaxdiv_t imaxdiv (intmax_t @var{numerator}, intmax_t @var{denominator})
1918 The @code{imaxdiv} function is like the @code{div} function, but the
1919 arguments are of type @code{intmax_t} and the result is returned as
1920 a structure of type @code{imaxdiv_t}.
1922 The @code{imaxdiv} function was added in @w{ISO C99}.
1926 @node Parsing of Numbers
1927 @section Parsing of Numbers
1928 @cindex parsing numbers (in formatted input)
1929 @cindex converting strings to numbers
1930 @cindex number syntax, parsing
1931 @cindex syntax, for reading numbers
1933 This section describes functions for ``reading'' integer and
1934 floating-point numbers from a string. It may be more convenient in some
1935 cases to use @code{sscanf} or one of the related functions; see
1936 @ref{Formatted Input}. But often you can make a program more robust by
1937 finding the tokens in the string by hand, then converting the numbers
1941 * Parsing of Integers:: Functions for conversion of integer values.
1942 * Parsing of Floats:: Functions for conversion of floating-point
1946 @node Parsing of Integers
1947 @subsection Parsing of Integers
1950 These functions are declared in @file{stdlib.h}.
1954 @deftypefun {long int} strtol (const char *@var{string}, char **@var{tailptr}, int @var{base})
1955 The @code{strtol} (``string-to-long'') function converts the initial
1956 part of @var{string} to a signed integer, which is returned as a value
1957 of type @code{long int}.
1959 This function attempts to decompose @var{string} as follows:
1963 A (possibly empty) sequence of whitespace characters. Which characters
1964 are whitespace is determined by the @code{isspace} function
1965 (@pxref{Classification of Characters}). These are discarded.
1968 An optional plus or minus sign (@samp{+} or @samp{-}).
1971 A nonempty sequence of digits in the radix specified by @var{base}.
1973 If @var{base} is zero, decimal radix is assumed unless the series of
1974 digits begins with @samp{0} (specifying octal radix), or @samp{0x} or
1975 @samp{0X} (specifying hexadecimal radix); in other words, the same
1976 syntax used for integer constants in C.
1978 Otherwise @var{base} must have a value between @code{2} and @code{35}.
1979 If @var{base} is @code{16}, the digits may optionally be preceded by
1980 @samp{0x} or @samp{0X}. If base has no legal value the value returned
1981 is @code{0l} and the global variable @code{errno} is set to @code{EINVAL}.
1984 Any remaining characters in the string. If @var{tailptr} is not a null
1985 pointer, @code{strtol} stores a pointer to this tail in
1986 @code{*@var{tailptr}}.
1989 If the string is empty, contains only whitespace, or does not contain an
1990 initial substring that has the expected syntax for an integer in the
1991 specified @var{base}, no conversion is performed. In this case,
1992 @code{strtol} returns a value of zero and the value stored in
1993 @code{*@var{tailptr}} is the value of @var{string}.
1995 In a locale other than the standard @code{"C"} locale, this function
1996 may recognize additional implementation-dependent syntax.
1998 If the string has valid syntax for an integer but the value is not
1999 representable because of overflow, @code{strtol} returns either
2000 @code{LONG_MAX} or @code{LONG_MIN} (@pxref{Range of Type}), as
2001 appropriate for the sign of the value. It also sets @code{errno}
2002 to @code{ERANGE} to indicate there was overflow.
2004 You should not check for errors by examining the return value of
2005 @code{strtol}, because the string might be a valid representation of
2006 @code{0l}, @code{LONG_MAX}, or @code{LONG_MIN}. Instead, check whether
2007 @var{tailptr} points to what you expect after the number
2008 (e.g. @code{'\0'} if the string should end after the number). You also
2009 need to clear @var{errno} before the call and check it afterward, in
2010 case there was overflow.
2012 There is an example at the end of this section.
2017 @deftypefun {unsigned long int} strtoul (const char *@var{string}, char **@var{tailptr}, int @var{base})
2018 The @code{strtoul} (``string-to-unsigned-long'') function is like
2019 @code{strtol} except it returns an @code{unsigned long int} value. If
2020 the number has a leading @samp{-} sign, the return value is negated.
2021 The syntax is the same as described above for @code{strtol}. The value
2022 returned on overflow is @code{ULONG_MAX} (@pxref{Range of
2025 @code{strtoul} sets @var{errno} to @code{EINVAL} if @var{base} is out of
2026 range, or @code{ERANGE} on overflow.
2031 @deftypefun {long long int} strtoll (const char *@var{string}, char **@var{tailptr}, int @var{base})
2032 The @code{strtoll} function is like @code{strtol} except that it returns
2033 a @code{long long int} value, and accepts numbers with a correspondingly
2036 If the string has valid syntax for an integer but the value is not
2037 representable because of overflow, @code{strtoll} returns either
2038 @code{LONG_LONG_MAX} or @code{LONG_LONG_MIN} (@pxref{Range of Type}), as
2039 appropriate for the sign of the value. It also sets @code{errno} to
2040 @code{ERANGE} to indicate there was overflow.
2042 The @code{strtoll} function was introduced in @w{ISO C99}.
2047 @deftypefun {long long int} strtoq (const char *@var{string}, char **@var{tailptr}, int @var{base})
2048 @code{strtoq} (``string-to-quad-word'') is the BSD name for @code{strtoll}.
2053 @deftypefun {unsigned long long int} strtoull (const char *@var{string}, char **@var{tailptr}, int @var{base})
2054 The @code{strtoull} function is like @code{strtoul} except that it
2055 returns an @code{unsigned long long int}. The value returned on overflow
2056 is @code{ULONG_LONG_MAX} (@pxref{Range of Type}).
2058 The @code{strtoull} function was introduced in @w{ISO C99}.
2063 @deftypefun {unsigned long long int} strtouq (const char *@var{string}, char **@var{tailptr}, int @var{base})
2064 @code{strtouq} is the BSD name for @code{strtoull}.
2069 @deftypefun {long int} atol (const char *@var{string})
2070 This function is similar to the @code{strtol} function with a @var{base}
2071 argument of @code{10}, except that it need not detect overflow errors.
2072 The @code{atol} function is provided mostly for compatibility with
2073 existing code; using @code{strtol} is more robust.
2078 @deftypefun int atoi (const char *@var{string})
2079 This function is like @code{atol}, except that it returns an @code{int}.
2080 The @code{atoi} function is also considered obsolete; use @code{strtol}
2086 @deftypefun {long long int} atoll (const char *@var{string})
2087 This function is similar to @code{atol}, except it returns a @code{long
2090 The @code{atoll} function was introduced in @w{ISO C99}. It too is
2091 obsolete (despite having just been added); use @code{strtoll} instead.
2094 @c !!! please fact check this paragraph -zw
2099 @cindex parsing numbers and locales
2100 @cindex locales, parsing numbers and
2101 Some locales specify a printed syntax for numbers other than the one
2102 that these functions understand. If you need to read numbers formatted
2103 in some other locale, you can use the @code{strtoX_l} functions. Each
2104 of the @code{strtoX} functions has a counterpart with @samp{_l} added to
2105 its name. The @samp{_l} counterparts take an additional argument: a
2106 pointer to an @code{locale_t} structure, which describes how the numbers
2107 to be read are formatted. @xref{Locales}.
2109 @strong{Portability Note:} These functions are all GNU extensions. You
2110 can also use @code{scanf} or its relatives, which have the @samp{'} flag
2111 for parsing numeric input according to the current locale
2112 (@pxref{Numeric Input Conversions}). This feature is standard.
2114 Here is a function which parses a string as a sequence of integers and
2115 returns the sum of them:
2119 sum_ints_from_string (char *string)
2127 /* @r{Skip whitespace by hand, to detect the end.} */
2128 while (isspace (*string)) string++;
2132 /* @r{There is more nonwhitespace,} */
2133 /* @r{so it ought to be another number.} */
2136 next = strtol (string, &tail, 0);
2137 /* @r{Add it in, if not overflow.} */
2139 printf ("Overflow\n");
2142 /* @r{Advance past it.} */
2150 @node Parsing of Floats
2151 @subsection Parsing of Floats
2154 These functions are declared in @file{stdlib.h}.
2158 @deftypefun double strtod (const char *@var{string}, char **@var{tailptr})
2159 The @code{strtod} (``string-to-double'') function converts the initial
2160 part of @var{string} to a floating-point number, which is returned as a
2161 value of type @code{double}.
2163 This function attempts to decompose @var{string} as follows:
2167 A (possibly empty) sequence of whitespace characters. Which characters
2168 are whitespace is determined by the @code{isspace} function
2169 (@pxref{Classification of Characters}). These are discarded.
2172 An optional plus or minus sign (@samp{+} or @samp{-}).
2174 @item A floating point number in decimal or hexadecimal format. The
2179 A nonempty sequence of digits optionally containing a decimal-point
2180 character---normally @samp{.}, but it depends on the locale
2181 (@pxref{General Numeric}).
2184 An optional exponent part, consisting of a character @samp{e} or
2185 @samp{E}, an optional sign, and a sequence of digits.
2189 The hexadecimal format is as follows:
2193 A 0x or 0X followed by a nonempty sequence of hexadecimal digits
2194 optionally containing a decimal-point character---normally @samp{.}, but
2195 it depends on the locale (@pxref{General Numeric}).
2198 An optional binary-exponent part, consisting of a character @samp{p} or
2199 @samp{P}, an optional sign, and a sequence of digits.
2204 Any remaining characters in the string. If @var{tailptr} is not a null
2205 pointer, a pointer to this tail of the string is stored in
2206 @code{*@var{tailptr}}.
2209 If the string is empty, contains only whitespace, or does not contain an
2210 initial substring that has the expected syntax for a floating-point
2211 number, no conversion is performed. In this case, @code{strtod} returns
2212 a value of zero and the value returned in @code{*@var{tailptr}} is the
2213 value of @var{string}.
2215 In a locale other than the standard @code{"C"} or @code{"POSIX"} locales,
2216 this function may recognize additional locale-dependent syntax.
2218 If the string has valid syntax for a floating-point number but the value
2219 is outside the range of a @code{double}, @code{strtod} will signal
2220 overflow or underflow as described in @ref{Math Error Reporting}.
2222 @code{strtod} recognizes four special input strings. The strings
2223 @code{"inf"} and @code{"infinity"} are converted to @math{@infinity{}},
2224 or to the largest representable value if the floating-point format
2225 doesn't support infinities. You can prepend a @code{"+"} or @code{"-"}
2226 to specify the sign. Case is ignored when scanning these strings.
2228 The strings @code{"nan"} and @code{"nan(@var{chars...})"} are converted
2229 to NaN. Again, case is ignored. If @var{chars...} are provided, they
2230 are used in some unspecified fashion to select a particular
2231 representation of NaN (there can be several).
2233 Since zero is a valid result as well as the value returned on error, you
2234 should check for errors in the same way as for @code{strtol}, by
2235 examining @var{errno} and @var{tailptr}.
2240 @deftypefun float strtof (const char *@var{string}, char **@var{tailptr})
2243 @deftypefunx {long double} strtold (const char *@var{string}, char **@var{tailptr})
2244 These functions are analogous to @code{strtod}, but return @code{float}
2245 and @code{long double} values respectively. They report errors in the
2246 same way as @code{strtod}. @code{strtof} can be substantially faster
2247 than @code{strtod}, but has less precision; conversely, @code{strtold}
2248 can be much slower but has more precision (on systems where @code{long
2249 double} is a separate type).
2251 These functions have been GNU extensions and are new to @w{ISO C99}.
2256 @deftypefun double atof (const char *@var{string})
2257 This function is similar to the @code{strtod} function, except that it
2258 need not detect overflow and underflow errors. The @code{atof} function
2259 is provided mostly for compatibility with existing code; using
2260 @code{strtod} is more robust.
2263 The GNU C library also provides @samp{_l} versions of thse functions,
2264 which take an additional argument, the locale to use in conversion.
2265 @xref{Parsing of Integers}.
2267 @node System V Number Conversion
2268 @section Old-fashioned System V number-to-string functions
2270 The old @w{System V} C library provided three functions to convert
2271 numbers to strings, with unusual and hard-to-use semantics. The GNU C
2272 library also provides these functions and some natural extensions.
2274 These functions are only available in glibc and on systems descended
2275 from AT&T Unix. Therefore, unless these functions do precisely what you
2276 need, it is better to use @code{sprintf}, which is standard.
2278 All these functions are defined in @file{stdlib.h}.
2281 @comment SVID, Unix98
2282 @deftypefun {char *} ecvt (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg})
2283 The function @code{ecvt} converts the floating-point number @var{value}
2284 to a string with at most @var{ndigit} decimal digits. The
2285 returned string contains no decimal point or sign. The first digit of
2286 the string is non-zero (unless @var{value} is actually zero) and the
2287 last digit is rounded to nearest. @code{*@var{decpt}} is set to the
2288 index in the string of the first digit after the decimal point.
2289 @code{*@var{neg}} is set to a nonzero value if @var{value} is negative,
2292 If @var{ndigit} decimal digits would exceed the precision of a
2293 @code{double} it is reduced to a system-specific value.
2295 The returned string is statically allocated and overwritten by each call
2298 If @var{value} is zero, it is implementation defined whether
2299 @code{*@var{decpt}} is @code{0} or @code{1}.
2301 For example: @code{ecvt (12.3, 5, &d, &n)} returns @code{"12300"}
2302 and sets @var{d} to @code{2} and @var{n} to @code{0}.
2306 @comment SVID, Unix98
2307 @deftypefun {char *} fcvt (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg})
2308 The function @code{fcvt} is like @code{ecvt}, but @var{ndigit} specifies
2309 the number of digits after the decimal point. If @var{ndigit} is less
2310 than zero, @var{value} is rounded to the @math{@var{ndigit}+1}'th place to the
2311 left of the decimal point. For example, if @var{ndigit} is @code{-1},
2312 @var{value} will be rounded to the nearest 10. If @var{ndigit} is
2313 negative and larger than the number of digits to the left of the decimal
2314 point in @var{value}, @var{value} will be rounded to one significant digit.
2316 If @var{ndigit} decimal digits would exceed the precision of a
2317 @code{double} it is reduced to a system-specific value.
2319 The returned string is statically allocated and overwritten by each call
2324 @comment SVID, Unix98
2325 @deftypefun {char *} gcvt (double @var{value}, int @var{ndigit}, char *@var{buf})
2326 @code{gcvt} is functionally equivalent to @samp{sprintf(buf, "%*g",
2327 ndigit, value}. It is provided only for compatibility's sake. It
2330 If @var{ndigit} decimal digits would exceed the precision of a
2331 @code{double} it is reduced to a system-specific value.
2334 As extensions, the GNU C library provides versions of these three
2335 functions that take @code{long double} arguments.
2339 @deftypefun {char *} qecvt (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg})
2340 This function is equivalent to @code{ecvt} except that it takes a
2341 @code{long double} for the first parameter and that @var{ndigit} is
2342 restricted by the precision of a @code{long double}.
2347 @deftypefun {char *} qfcvt (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg})
2348 This function is equivalent to @code{fcvt} except that it
2349 takes a @code{long double} for the first parameter and that @var{ndigit} is
2350 restricted by the precision of a @code{long double}.
2355 @deftypefun {char *} qgcvt (long double @var{value}, int @var{ndigit}, char *@var{buf})
2356 This function is equivalent to @code{gcvt} except that it takes a
2357 @code{long double} for the first parameter and that @var{ndigit} is
2358 restricted by the precision of a @code{long double}.
2363 The @code{ecvt} and @code{fcvt} functions, and their @code{long double}
2364 equivalents, all return a string located in a static buffer which is
2365 overwritten by the next call to the function. The GNU C library
2366 provides another set of extended functions which write the converted
2367 string into a user-supplied buffer. These have the conventional
2370 @code{gcvt_r} is not necessary, because @code{gcvt} already uses a
2371 user-supplied buffer.
2375 @deftypefun {char *} ecvt_r (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len})
2376 The @code{ecvt_r} function is the same as @code{ecvt}, except
2377 that it places its result into the user-specified buffer pointed to by
2378 @var{buf}, with length @var{len}.
2380 This function is a GNU extension.
2384 @comment SVID, Unix98
2385 @deftypefun {char *} fcvt_r (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len})
2386 The @code{fcvt_r} function is the same as @code{fcvt}, except
2387 that it places its result into the user-specified buffer pointed to by
2388 @var{buf}, with length @var{len}.
2390 This function is a GNU extension.
2395 @deftypefun {char *} qecvt_r (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len})
2396 The @code{qecvt_r} function is the same as @code{qecvt}, except
2397 that it places its result into the user-specified buffer pointed to by
2398 @var{buf}, with length @var{len}.
2400 This function is a GNU extension.
2405 @deftypefun {char *} qfcvt_r (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len})
2406 The @code{qfcvt_r} function is the same as @code{qfcvt}, except
2407 that it places its result into the user-specified buffer pointed to by
2408 @var{buf}, with length @var{len}.
2410 This function is a GNU extension.