]>
Commit | Line | Data |
---|---|---|
28f540f4 | 1 | @node Arithmetic, Date and Time, Mathematics, Top |
7a68c94a UD |
2 | @c %MENU% Low level arithmetic functions |
3 | @chapter Arithmetic Functions | |
28f540f4 RM |
4 | |
5 | This chapter contains information about functions for doing basic | |
6 | arithmetic operations, such as splitting a float into its integer and | |
b4012b75 UD |
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 | |
9 | @file{complex.h}. | |
28f540f4 RM |
10 | |
11 | @menu | |
0e4ee106 UD |
12 | * Integers:: Basic integer types and concepts |
13 | * Integer Division:: Integer division with guaranteed rounding. | |
7a68c94a UD |
14 | * Floating Point Numbers:: Basic concepts. IEEE 754. |
15 | * Floating Point Classes:: The five kinds of floating-point number. | |
16 | * Floating Point Errors:: When something goes wrong in a calculation. | |
17 | * Rounding:: Controlling how results are rounded. | |
18 | * Control Functions:: Saving and restoring the FPU's state. | |
19 | * Arithmetic Functions:: Fundamental operations provided by the library. | |
20 | * Complex Numbers:: The types. Writing complex constants. | |
21 | * Operations on Complex:: Projection, conjugation, decomposition. | |
7a68c94a UD |
22 | * Parsing of Numbers:: Converting strings to numbers. |
23 | * System V Number Conversion:: An archaic way to convert numbers to strings. | |
28f540f4 RM |
24 | @end menu |
25 | ||
0e4ee106 UD |
26 | @node Integers |
27 | @section Integers | |
28 | @cindex integer | |
29 | ||
30 | The C language defines several integer data types: integer, short integer, | |
31 | long integer, and character, all in both signed and unsigned varieties. | |
e6e81391 UD |
32 | The GNU C compiler extends the language to contain long long integers |
33 | as well. | |
0e4ee106 UD |
34 | @cindex signedness |
35 | ||
36 | The C integer types were intended to allow code to be portable among | |
37 | machines with different inherent data sizes (word sizes), so each type | |
38 | may have different ranges on different machines. The problem with | |
39 | this is that a program often needs to be written for a particular range | |
40 | of integers, and sometimes must be written for a particular size of | |
41 | storage, regardless of what machine the program runs on. | |
42 | ||
43 | To address this problem, the GNU C library contains C type definitions | |
44 | you can use to declare integers that meet your exact needs. Because the | |
45 | GNU C library header files are customized to a specific machine, your | |
46 | program source code doesn't have to be. | |
47 | ||
48 | These @code{typedef}s are in @file{stdint.h}. | |
49 | @pindex stdint.h | |
50 | ||
51 | If you require that an integer be represented in exactly N bits, use one | |
52 | of the following types, with the obvious mapping to bit size and signedness: | |
53 | ||
68979757 | 54 | @itemize @bullet |
0e4ee106 UD |
55 | @item int8_t |
56 | @item int16_t | |
57 | @item int32_t | |
58 | @item int64_t | |
59 | @item uint8_t | |
60 | @item uint16_t | |
61 | @item uint32_t | |
62 | @item uint64_t | |
63 | @end itemize | |
64 | ||
65 | If your C compiler and target machine do not allow integers of a certain | |
66 | size, the corresponding above type does not exist. | |
67 | ||
68 | If you don't need a specific storage size, but want the smallest data | |
69 | structure with @emph{at least} N bits, use one of these: | |
70 | ||
68979757 | 71 | @itemize @bullet |
150f9fb8 AJ |
72 | @item int_least8_t |
73 | @item int_least16_t | |
74 | @item int_least32_t | |
75 | @item int_least64_t | |
76 | @item uint_least8_t | |
77 | @item uint_least16_t | |
78 | @item uint_least32_t | |
79 | @item uint_least64_t | |
0e4ee106 UD |
80 | @end itemize |
81 | ||
e6e81391 | 82 | If you don't need a specific storage size, but want the data structure |
0e4ee106 UD |
83 | that allows the fastest access while having at least N bits (and |
84 | among data structures with the same access speed, the smallest one), use | |
85 | one of these: | |
86 | ||
68979757 | 87 | @itemize @bullet |
150f9fb8 AJ |
88 | @item int_fast8_t |
89 | @item int_fast16_t | |
90 | @item int_fast32_t | |
91 | @item int_fast64_t | |
92 | @item uint_fast8_t | |
93 | @item uint_fast16_t | |
94 | @item uint_fast32_t | |
95 | @item uint_fast64_t | |
0e4ee106 UD |
96 | @end itemize |
97 | ||
e6e81391 | 98 | If you want an integer with the widest range possible on the platform on |
0e4ee106 UD |
99 | which it is being used, use one of the following. If you use these, |
100 | you should write code that takes into account the variable size and range | |
101 | of the integer. | |
102 | ||
68979757 | 103 | @itemize @bullet |
0e4ee106 UD |
104 | @item intmax_t |
105 | @item uintmax_t | |
106 | @end itemize | |
107 | ||
108 | The GNU C library also provides macros that tell you the maximum and | |
109 | minimum possible values for each integer data type. The macro names | |
110 | follow these examples: @code{INT32_MAX}, @code{UINT8_MAX}, | |
111 | @code{INT_FAST32_MIN}, @code{INT_LEAST64_MIN}, @code{UINTMAX_MAX}, | |
112 | @code{INTMAX_MAX}, @code{INTMAX_MIN}. Note that there are no macros for | |
113 | unsigned integer minima. These are always zero. | |
114 | @cindex maximum possible integer | |
0bc93a2f | 115 | @cindex minimum possible integer |
0e4ee106 UD |
116 | |
117 | There are similar macros for use with C's built in integer types which | |
118 | should come with your C compiler. These are described in @ref{Data Type | |
119 | Measurements}. | |
120 | ||
121 | Don't forget you can use the C @code{sizeof} function with any of these | |
122 | data types to get the number of bytes of storage each uses. | |
123 | ||
124 | ||
125 | @node Integer Division | |
126 | @section Integer Division | |
127 | @cindex integer division functions | |
128 | ||
129 | This section describes functions for performing integer division. These | |
130 | functions are redundant when GNU CC is used, because in GNU C the | |
131 | @samp{/} operator always rounds towards zero. But in other C | |
132 | implementations, @samp{/} may round differently with negative arguments. | |
133 | @code{div} and @code{ldiv} are useful because they specify how to round | |
134 | the quotient: towards zero. The remainder has the same sign as the | |
135 | numerator. | |
136 | ||
137 | These functions are specified to return a result @var{r} such that the value | |
138 | @code{@var{r}.quot*@var{denominator} + @var{r}.rem} equals | |
139 | @var{numerator}. | |
140 | ||
141 | @pindex stdlib.h | |
142 | To use these facilities, you should include the header file | |
143 | @file{stdlib.h} in your program. | |
144 | ||
145 | @comment stdlib.h | |
146 | @comment ISO | |
147 | @deftp {Data Type} div_t | |
148 | This is a structure type used to hold the result returned by the @code{div} | |
149 | function. It has the following members: | |
150 | ||
151 | @table @code | |
152 | @item int quot | |
153 | The quotient from the division. | |
154 | ||
155 | @item int rem | |
156 | The remainder from the division. | |
157 | @end table | |
158 | @end deftp | |
159 | ||
160 | @comment stdlib.h | |
161 | @comment ISO | |
162 | @deftypefun div_t div (int @var{numerator}, int @var{denominator}) | |
163 | This function @code{div} computes the quotient and remainder from | |
164 | the division of @var{numerator} by @var{denominator}, returning the | |
165 | result in a structure of type @code{div_t}. | |
166 | ||
167 | If the result cannot be represented (as in a division by zero), the | |
168 | behavior is undefined. | |
169 | ||
170 | Here is an example, albeit not a very useful one. | |
171 | ||
172 | @smallexample | |
173 | div_t result; | |
174 | result = div (20, -6); | |
175 | @end smallexample | |
176 | ||
177 | @noindent | |
178 | Now @code{result.quot} is @code{-3} and @code{result.rem} is @code{2}. | |
179 | @end deftypefun | |
180 | ||
181 | @comment stdlib.h | |
182 | @comment ISO | |
183 | @deftp {Data Type} ldiv_t | |
184 | This is a structure type used to hold the result returned by the @code{ldiv} | |
185 | function. It has the following members: | |
186 | ||
187 | @table @code | |
188 | @item long int quot | |
189 | The quotient from the division. | |
190 | ||
191 | @item long int rem | |
192 | The remainder from the division. | |
193 | @end table | |
194 | ||
195 | (This is identical to @code{div_t} except that the components are of | |
196 | type @code{long int} rather than @code{int}.) | |
197 | @end deftp | |
198 | ||
199 | @comment stdlib.h | |
200 | @comment ISO | |
201 | @deftypefun ldiv_t ldiv (long int @var{numerator}, long int @var{denominator}) | |
202 | The @code{ldiv} function is similar to @code{div}, except that the | |
203 | arguments are of type @code{long int} and the result is returned as a | |
204 | structure of type @code{ldiv_t}. | |
205 | @end deftypefun | |
206 | ||
207 | @comment stdlib.h | |
208 | @comment ISO | |
209 | @deftp {Data Type} lldiv_t | |
210 | This is a structure type used to hold the result returned by the @code{lldiv} | |
211 | function. It has the following members: | |
212 | ||
213 | @table @code | |
214 | @item long long int quot | |
215 | The quotient from the division. | |
216 | ||
217 | @item long long int rem | |
218 | The remainder from the division. | |
219 | @end table | |
220 | ||
221 | (This is identical to @code{div_t} except that the components are of | |
222 | type @code{long long int} rather than @code{int}.) | |
223 | @end deftp | |
224 | ||
225 | @comment stdlib.h | |
226 | @comment ISO | |
227 | @deftypefun lldiv_t lldiv (long long int @var{numerator}, long long int @var{denominator}) | |
228 | The @code{lldiv} function is like the @code{div} function, but the | |
229 | arguments are of type @code{long long int} and the result is returned as | |
230 | a structure of type @code{lldiv_t}. | |
231 | ||
232 | The @code{lldiv} function was added in @w{ISO C99}. | |
233 | @end deftypefun | |
234 | ||
235 | @comment inttypes.h | |
236 | @comment ISO | |
237 | @deftp {Data Type} imaxdiv_t | |
238 | This is a structure type used to hold the result returned by the @code{imaxdiv} | |
239 | function. It has the following members: | |
240 | ||
241 | @table @code | |
242 | @item intmax_t quot | |
243 | The quotient from the division. | |
244 | ||
245 | @item intmax_t rem | |
246 | The remainder from the division. | |
247 | @end table | |
248 | ||
249 | (This is identical to @code{div_t} except that the components are of | |
250 | type @code{intmax_t} rather than @code{int}.) | |
251 | ||
252 | See @ref{Integers} for a description of the @code{intmax_t} type. | |
253 | ||
254 | @end deftp | |
255 | ||
256 | @comment inttypes.h | |
257 | @comment ISO | |
258 | @deftypefun imaxdiv_t imaxdiv (intmax_t @var{numerator}, intmax_t @var{denominator}) | |
259 | The @code{imaxdiv} function is like the @code{div} function, but the | |
260 | arguments are of type @code{intmax_t} and the result is returned as | |
261 | a structure of type @code{imaxdiv_t}. | |
262 | ||
263 | See @ref{Integers} for a description of the @code{intmax_t} type. | |
264 | ||
265 | The @code{imaxdiv} function was added in @w{ISO C99}. | |
266 | @end deftypefun | |
267 | ||
268 | ||
7a68c94a UD |
269 | @node Floating Point Numbers |
270 | @section Floating Point Numbers | |
271 | @cindex floating point | |
272 | @cindex IEEE 754 | |
b4012b75 UD |
273 | @cindex IEEE floating point |
274 | ||
7a68c94a UD |
275 | Most computer hardware has support for two different kinds of numbers: |
276 | integers (@math{@dots{}-3, -2, -1, 0, 1, 2, 3@dots{}}) and | |
277 | floating-point numbers. Floating-point numbers have three parts: the | |
278 | @dfn{mantissa}, the @dfn{exponent}, and the @dfn{sign bit}. The real | |
279 | number represented by a floating-point value is given by | |
280 | @tex | |
281 | $(s \mathrel? -1 \mathrel: 1) \cdot 2^e \cdot M$ | |
282 | @end tex | |
283 | @ifnottex | |
284 | @math{(s ? -1 : 1) @mul{} 2^e @mul{} M} | |
285 | @end ifnottex | |
286 | where @math{s} is the sign bit, @math{e} the exponent, and @math{M} | |
287 | the mantissa. @xref{Floating Point Concepts}, for details. (It is | |
288 | possible to have a different @dfn{base} for the exponent, but all modern | |
289 | hardware uses @math{2}.) | |
290 | ||
291 | Floating-point numbers can represent a finite subset of the real | |
292 | numbers. While this subset is large enough for most purposes, it is | |
293 | important to remember that the only reals that can be represented | |
294 | exactly are rational numbers that have a terminating binary expansion | |
295 | shorter than the width of the mantissa. Even simple fractions such as | |
296 | @math{1/5} can only be approximated by floating point. | |
297 | ||
298 | Mathematical operations and functions frequently need to produce values | |
299 | that are not representable. Often these values can be approximated | |
300 | closely enough for practical purposes, but sometimes they can't. | |
301 | Historically there was no way to tell when the results of a calculation | |
302 | were inaccurate. Modern computers implement the @w{IEEE 754} standard | |
303 | for numerical computations, which defines a framework for indicating to | |
304 | the program when the results of calculation are not trustworthy. This | |
305 | framework consists of a set of @dfn{exceptions} that indicate why a | |
306 | result could not be represented, and the special values @dfn{infinity} | |
307 | and @dfn{not a number} (NaN). | |
308 | ||
309 | @node Floating Point Classes | |
310 | @section Floating-Point Number Classification Functions | |
311 | @cindex floating-point classes | |
312 | @cindex classes, floating-point | |
313 | @pindex math.h | |
b4012b75 | 314 | |
ec751a23 | 315 | @w{ISO C99} defines macros that let you determine what sort of |
7a68c94a | 316 | floating-point number a variable holds. |
b4012b75 UD |
317 | |
318 | @comment math.h | |
319 | @comment ISO | |
7a68c94a UD |
320 | @deftypefn {Macro} int fpclassify (@emph{float-type} @var{x}) |
321 | This is a generic macro which works on all floating-point types and | |
322 | which returns a value of type @code{int}. The possible values are: | |
28f540f4 | 323 | |
7a68c94a UD |
324 | @vtable @code |
325 | @item FP_NAN | |
326 | The floating-point number @var{x} is ``Not a Number'' (@pxref{Infinity | |
327 | and NaN}) | |
328 | @item FP_INFINITE | |
329 | The value of @var{x} is either plus or minus infinity (@pxref{Infinity | |
330 | and NaN}) | |
331 | @item FP_ZERO | |
332 | The value of @var{x} is zero. In floating-point formats like @w{IEEE | |
333 | 754}, where zero can be signed, this value is also returned if | |
334 | @var{x} is negative zero. | |
335 | @item FP_SUBNORMAL | |
336 | Numbers whose absolute value is too small to be represented in the | |
337 | normal format are represented in an alternate, @dfn{denormalized} format | |
338 | (@pxref{Floating Point Concepts}). This format is less precise but can | |
339 | represent values closer to zero. @code{fpclassify} returns this value | |
340 | for values of @var{x} in this alternate format. | |
341 | @item FP_NORMAL | |
342 | This value is returned for all other values of @var{x}. It indicates | |
343 | that there is nothing special about the number. | |
344 | @end vtable | |
28f540f4 | 345 | |
7a68c94a | 346 | @end deftypefn |
28f540f4 | 347 | |
7a68c94a UD |
348 | @code{fpclassify} is most useful if more than one property of a number |
349 | must be tested. There are more specific macros which only test one | |
350 | property at a time. Generally these macros execute faster than | |
351 | @code{fpclassify}, since there is special hardware support for them. | |
352 | You should therefore use the specific macros whenever possible. | |
28f540f4 RM |
353 | |
354 | @comment math.h | |
7a68c94a UD |
355 | @comment ISO |
356 | @deftypefn {Macro} int isfinite (@emph{float-type} @var{x}) | |
357 | This macro returns a nonzero value if @var{x} is finite: not plus or | |
358 | minus infinity, and not NaN. It is equivalent to | |
fe0ec73e UD |
359 | |
360 | @smallexample | |
7a68c94a | 361 | (fpclassify (x) != FP_NAN && fpclassify (x) != FP_INFINITE) |
fe0ec73e UD |
362 | @end smallexample |
363 | ||
7a68c94a UD |
364 | @code{isfinite} is implemented as a macro which accepts any |
365 | floating-point type. | |
366 | @end deftypefn | |
fe0ec73e | 367 | |
7a68c94a UD |
368 | @comment math.h |
369 | @comment ISO | |
370 | @deftypefn {Macro} int isnormal (@emph{float-type} @var{x}) | |
371 | This macro returns a nonzero value if @var{x} is finite and normalized. | |
372 | It is equivalent to | |
b4012b75 UD |
373 | |
374 | @smallexample | |
7a68c94a | 375 | (fpclassify (x) == FP_NORMAL) |
b4012b75 | 376 | @end smallexample |
7a68c94a | 377 | @end deftypefn |
b4012b75 | 378 | |
7a68c94a UD |
379 | @comment math.h |
380 | @comment ISO | |
381 | @deftypefn {Macro} int isnan (@emph{float-type} @var{x}) | |
382 | This macro returns a nonzero value if @var{x} is NaN. It is equivalent | |
383 | to | |
b4012b75 UD |
384 | |
385 | @smallexample | |
7a68c94a | 386 | (fpclassify (x) == FP_NAN) |
b4012b75 | 387 | @end smallexample |
7a68c94a | 388 | @end deftypefn |
b4012b75 | 389 | |
7a68c94a UD |
390 | Another set of floating-point classification functions was provided by |
391 | BSD. The GNU C library also supports these functions; however, we | |
ec751a23 | 392 | recommend that you use the ISO C99 macros in new code. Those are standard |
7a68c94a UD |
393 | and will be available more widely. Also, since they are macros, you do |
394 | not have to worry about the type of their argument. | |
28f540f4 RM |
395 | |
396 | @comment math.h | |
397 | @comment BSD | |
398 | @deftypefun int isinf (double @var{x}) | |
4260bc74 UD |
399 | @comment math.h |
400 | @comment BSD | |
779ae82e | 401 | @deftypefunx int isinff (float @var{x}) |
4260bc74 UD |
402 | @comment math.h |
403 | @comment BSD | |
779ae82e | 404 | @deftypefunx int isinfl (long double @var{x}) |
28f540f4 RM |
405 | This function returns @code{-1} if @var{x} represents negative infinity, |
406 | @code{1} if @var{x} represents positive infinity, and @code{0} otherwise. | |
407 | @end deftypefun | |
408 | ||
409 | @comment math.h | |
410 | @comment BSD | |
411 | @deftypefun int isnan (double @var{x}) | |
4260bc74 UD |
412 | @comment math.h |
413 | @comment BSD | |
779ae82e | 414 | @deftypefunx int isnanf (float @var{x}) |
4260bc74 UD |
415 | @comment math.h |
416 | @comment BSD | |
779ae82e | 417 | @deftypefunx int isnanl (long double @var{x}) |
28f540f4 | 418 | This function returns a nonzero value if @var{x} is a ``not a number'' |
7a68c94a | 419 | value, and zero otherwise. |
b9b49b44 | 420 | |
48b22986 | 421 | @strong{NB:} The @code{isnan} macro defined by @w{ISO C99} overrides |
7a68c94a UD |
422 | the BSD function. This is normally not a problem, because the two |
423 | routines behave identically. However, if you really need to get the BSD | |
424 | function for some reason, you can write | |
b9b49b44 | 425 | |
7a68c94a UD |
426 | @smallexample |
427 | (isnan) (x) | |
428 | @end smallexample | |
28f540f4 RM |
429 | @end deftypefun |
430 | ||
431 | @comment math.h | |
432 | @comment BSD | |
433 | @deftypefun int finite (double @var{x}) | |
4260bc74 UD |
434 | @comment math.h |
435 | @comment BSD | |
779ae82e | 436 | @deftypefunx int finitef (float @var{x}) |
4260bc74 UD |
437 | @comment math.h |
438 | @comment BSD | |
779ae82e | 439 | @deftypefunx int finitel (long double @var{x}) |
28f540f4 RM |
440 | This function returns a nonzero value if @var{x} is finite or a ``not a |
441 | number'' value, and zero otherwise. | |
442 | @end deftypefun | |
443 | ||
28f540f4 RM |
444 | @strong{Portability Note:} The functions listed in this section are BSD |
445 | extensions. | |
446 | ||
b4012b75 | 447 | |
7a68c94a UD |
448 | @node Floating Point Errors |
449 | @section Errors in Floating-Point Calculations | |
450 | ||
451 | @menu | |
452 | * FP Exceptions:: IEEE 754 math exceptions and how to detect them. | |
453 | * Infinity and NaN:: Special values returned by calculations. | |
454 | * Status bit operations:: Checking for exceptions after the fact. | |
455 | * Math Error Reporting:: How the math functions report errors. | |
456 | @end menu | |
457 | ||
458 | @node FP Exceptions | |
459 | @subsection FP Exceptions | |
460 | @cindex exception | |
461 | @cindex signal | |
462 | @cindex zero divide | |
463 | @cindex division by zero | |
464 | @cindex inexact exception | |
465 | @cindex invalid exception | |
466 | @cindex overflow exception | |
467 | @cindex underflow exception | |
468 | ||
469 | The @w{IEEE 754} standard defines five @dfn{exceptions} that can occur | |
470 | during a calculation. Each corresponds to a particular sort of error, | |
471 | such as overflow. | |
472 | ||
473 | When exceptions occur (when exceptions are @dfn{raised}, in the language | |
474 | of the standard), one of two things can happen. By default the | |
475 | exception is simply noted in the floating-point @dfn{status word}, and | |
476 | the program continues as if nothing had happened. The operation | |
477 | produces a default value, which depends on the exception (see the table | |
478 | below). Your program can check the status word to find out which | |
479 | exceptions happened. | |
480 | ||
481 | Alternatively, you can enable @dfn{traps} for exceptions. In that case, | |
482 | when an exception is raised, your program will receive the @code{SIGFPE} | |
483 | signal. The default action for this signal is to terminate the | |
8b7fb588 | 484 | program. @xref{Signal Handling}, for how you can change the effect of |
7a68c94a UD |
485 | the signal. |
486 | ||
487 | @findex matherr | |
488 | In the System V math library, the user-defined function @code{matherr} | |
489 | is called when certain exceptions occur inside math library functions. | |
490 | However, the Unix98 standard deprecates this interface. We support it | |
491 | for historical compatibility, but recommend that you do not use it in | |
492 | new programs. | |
493 | ||
494 | @noindent | |
495 | The exceptions defined in @w{IEEE 754} are: | |
496 | ||
497 | @table @samp | |
498 | @item Invalid Operation | |
499 | This exception is raised if the given operands are invalid for the | |
500 | operation to be performed. Examples are | |
501 | (see @w{IEEE 754}, @w{section 7}): | |
502 | @enumerate | |
503 | @item | |
504 | Addition or subtraction: @math{@infinity{} - @infinity{}}. (But | |
505 | @math{@infinity{} + @infinity{} = @infinity{}}). | |
506 | @item | |
507 | Multiplication: @math{0 @mul{} @infinity{}}. | |
508 | @item | |
509 | Division: @math{0/0} or @math{@infinity{}/@infinity{}}. | |
510 | @item | |
511 | Remainder: @math{x} REM @math{y}, where @math{y} is zero or @math{x} is | |
512 | infinite. | |
513 | @item | |
514 | Square root if the operand is less then zero. More generally, any | |
515 | mathematical function evaluated outside its domain produces this | |
516 | exception. | |
517 | @item | |
518 | Conversion of a floating-point number to an integer or decimal | |
519 | string, when the number cannot be represented in the target format (due | |
520 | to overflow, infinity, or NaN). | |
521 | @item | |
522 | Conversion of an unrecognizable input string. | |
523 | @item | |
524 | Comparison via predicates involving @math{<} or @math{>}, when one or | |
525 | other of the operands is NaN. You can prevent this exception by using | |
526 | the unordered comparison functions instead; see @ref{FP Comparison Functions}. | |
527 | @end enumerate | |
528 | ||
529 | If the exception does not trap, the result of the operation is NaN. | |
530 | ||
531 | @item Division by Zero | |
532 | This exception is raised when a finite nonzero number is divided | |
533 | by zero. If no trap occurs the result is either @math{+@infinity{}} or | |
534 | @math{-@infinity{}}, depending on the signs of the operands. | |
535 | ||
536 | @item Overflow | |
537 | This exception is raised whenever the result cannot be represented | |
538 | as a finite value in the precision format of the destination. If no trap | |
539 | occurs the result depends on the sign of the intermediate result and the | |
540 | current rounding mode (@w{IEEE 754}, @w{section 7.3}): | |
541 | @enumerate | |
542 | @item | |
543 | Round to nearest carries all overflows to @math{@infinity{}} | |
544 | with the sign of the intermediate result. | |
545 | @item | |
546 | Round toward @math{0} carries all overflows to the largest representable | |
547 | finite number with the sign of the intermediate result. | |
548 | @item | |
549 | Round toward @math{-@infinity{}} carries positive overflows to the | |
550 | largest representable finite number and negative overflows to | |
551 | @math{-@infinity{}}. | |
552 | ||
553 | @item | |
554 | Round toward @math{@infinity{}} carries negative overflows to the | |
555 | most negative representable finite number and positive overflows | |
556 | to @math{@infinity{}}. | |
557 | @end enumerate | |
558 | ||
559 | Whenever the overflow exception is raised, the inexact exception is also | |
560 | raised. | |
561 | ||
562 | @item Underflow | |
563 | The underflow exception is raised when an intermediate result is too | |
564 | small to be calculated accurately, or if the operation's result rounded | |
565 | to the destination precision is too small to be normalized. | |
566 | ||
567 | When no trap is installed for the underflow exception, underflow is | |
568 | signaled (via the underflow flag) only when both tininess and loss of | |
569 | accuracy have been detected. If no trap handler is installed the | |
570 | operation continues with an imprecise small value, or zero if the | |
571 | destination precision cannot hold the small exact result. | |
572 | ||
573 | @item Inexact | |
574 | This exception is signalled if a rounded result is not exact (such as | |
575 | when calculating the square root of two) or a result overflows without | |
576 | an overflow trap. | |
577 | @end table | |
578 | ||
579 | @node Infinity and NaN | |
580 | @subsection Infinity and NaN | |
581 | @cindex infinity | |
582 | @cindex not a number | |
583 | @cindex NaN | |
584 | ||
585 | @w{IEEE 754} floating point numbers can represent positive or negative | |
586 | infinity, and @dfn{NaN} (not a number). These three values arise from | |
587 | calculations whose result is undefined or cannot be represented | |
588 | accurately. You can also deliberately set a floating-point variable to | |
589 | any of them, which is sometimes useful. Some examples of calculations | |
590 | that produce infinity or NaN: | |
591 | ||
592 | @ifnottex | |
593 | @smallexample | |
594 | @math{1/0 = @infinity{}} | |
595 | @math{log (0) = -@infinity{}} | |
596 | @math{sqrt (-1) = NaN} | |
597 | @end smallexample | |
598 | @end ifnottex | |
599 | @tex | |
600 | $${1\over0} = \infty$$ | |
601 | $$\log 0 = -\infty$$ | |
602 | $$\sqrt{-1} = \hbox{NaN}$$ | |
603 | @end tex | |
604 | ||
605 | When a calculation produces any of these values, an exception also | |
606 | occurs; see @ref{FP Exceptions}. | |
607 | ||
608 | The basic operations and math functions all accept infinity and NaN and | |
609 | produce sensible output. Infinities propagate through calculations as | |
610 | one would expect: for example, @math{2 + @infinity{} = @infinity{}}, | |
611 | @math{4/@infinity{} = 0}, atan @math{(@infinity{}) = @pi{}/2}. NaN, on | |
612 | the other hand, infects any calculation that involves it. Unless the | |
613 | calculation would produce the same result no matter what real value | |
614 | replaced NaN, the result is NaN. | |
615 | ||
616 | In comparison operations, positive infinity is larger than all values | |
617 | except itself and NaN, and negative infinity is smaller than all values | |
618 | except itself and NaN. NaN is @dfn{unordered}: it is not equal to, | |
619 | greater than, or less than anything, @emph{including itself}. @code{x == | |
620 | x} is false if the value of @code{x} is NaN. You can use this to test | |
621 | whether a value is NaN or not, but the recommended way to test for NaN | |
622 | is with the @code{isnan} function (@pxref{Floating Point Classes}). In | |
623 | addition, @code{<}, @code{>}, @code{<=}, and @code{>=} will raise an | |
624 | exception when applied to NaNs. | |
625 | ||
626 | @file{math.h} defines macros that allow you to explicitly set a variable | |
627 | to infinity or NaN. | |
b4012b75 UD |
628 | |
629 | @comment math.h | |
630 | @comment ISO | |
7a68c94a UD |
631 | @deftypevr Macro float INFINITY |
632 | An expression representing positive infinity. It is equal to the value | |
633 | produced by mathematical operations like @code{1.0 / 0.0}. | |
634 | @code{-INFINITY} represents negative infinity. | |
635 | ||
636 | You can test whether a floating-point value is infinite by comparing it | |
637 | to this macro. However, this is not recommended; you should use the | |
638 | @code{isfinite} macro instead. @xref{Floating Point Classes}. | |
639 | ||
ec751a23 | 640 | This macro was introduced in the @w{ISO C99} standard. |
7a68c94a UD |
641 | @end deftypevr |
642 | ||
643 | @comment math.h | |
644 | @comment GNU | |
645 | @deftypevr Macro float NAN | |
646 | An expression representing a value which is ``not a number''. This | |
647 | macro is a GNU extension, available only on machines that support the | |
648 | ``not a number'' value---that is to say, on all machines that support | |
649 | IEEE floating point. | |
650 | ||
651 | You can use @samp{#ifdef NAN} to test whether the machine supports | |
652 | NaN. (Of course, you must arrange for GNU extensions to be visible, | |
653 | such as by defining @code{_GNU_SOURCE}, and then you must include | |
654 | @file{math.h}.) | |
655 | @end deftypevr | |
656 | ||
657 | @w{IEEE 754} also allows for another unusual value: negative zero. This | |
658 | value is produced when you divide a positive number by negative | |
659 | infinity, or when a negative result is smaller than the limits of | |
660 | representation. Negative zero behaves identically to zero in all | |
661 | calculations, unless you explicitly test the sign bit with | |
662 | @code{signbit} or @code{copysign}. | |
663 | ||
664 | @node Status bit operations | |
665 | @subsection Examining the FPU status word | |
666 | ||
ec751a23 | 667 | @w{ISO C99} defines functions to query and manipulate the |
7a68c94a UD |
668 | floating-point status word. You can use these functions to check for |
669 | untrapped exceptions when it's convenient, rather than worrying about | |
670 | them in the middle of a calculation. | |
671 | ||
672 | These constants represent the various @w{IEEE 754} exceptions. Not all | |
673 | FPUs report all the different exceptions. Each constant is defined if | |
674 | and only if the FPU you are compiling for supports that exception, so | |
675 | you can test for FPU support with @samp{#ifdef}. They are defined in | |
676 | @file{fenv.h}. | |
b4012b75 UD |
677 | |
678 | @vtable @code | |
7a68c94a UD |
679 | @comment fenv.h |
680 | @comment ISO | |
681 | @item FE_INEXACT | |
682 | The inexact exception. | |
683 | @comment fenv.h | |
684 | @comment ISO | |
685 | @item FE_DIVBYZERO | |
686 | The divide by zero exception. | |
687 | @comment fenv.h | |
688 | @comment ISO | |
689 | @item FE_UNDERFLOW | |
690 | The underflow exception. | |
691 | @comment fenv.h | |
692 | @comment ISO | |
693 | @item FE_OVERFLOW | |
694 | The overflow exception. | |
695 | @comment fenv.h | |
696 | @comment ISO | |
697 | @item FE_INVALID | |
698 | The invalid exception. | |
b4012b75 UD |
699 | @end vtable |
700 | ||
7a68c94a UD |
701 | The macro @code{FE_ALL_EXCEPT} is the bitwise OR of all exception macros |
702 | which are supported by the FP implementation. | |
b4012b75 | 703 | |
7a68c94a UD |
704 | These functions allow you to clear exception flags, test for exceptions, |
705 | and save and restore the set of exceptions flagged. | |
b4012b75 | 706 | |
7a68c94a | 707 | @comment fenv.h |
b4012b75 | 708 | @comment ISO |
63ae7b63 | 709 | @deftypefun int feclearexcept (int @var{excepts}) |
7a68c94a UD |
710 | This function clears all of the supported exception flags indicated by |
711 | @var{excepts}. | |
63ae7b63 UD |
712 | |
713 | The function returns zero in case the operation was successful, a | |
714 | non-zero value otherwise. | |
715 | @end deftypefun | |
716 | ||
717 | @comment fenv.h | |
718 | @comment ISO | |
719 | @deftypefun int feraiseexcept (int @var{excepts}) | |
720 | This function raises the supported exceptions indicated by | |
721 | @var{excepts}. If more than one exception bit in @var{excepts} is set | |
722 | the order in which the exceptions are raised is undefined except that | |
723 | overflow (@code{FE_OVERFLOW}) or underflow (@code{FE_UNDERFLOW}) are | |
724 | raised before inexact (@code{FE_INEXACT}). Whether for overflow or | |
725 | underflow the inexact exception is also raised is also implementation | |
726 | dependent. | |
727 | ||
728 | The function returns zero in case the operation was successful, a | |
729 | non-zero value otherwise. | |
7a68c94a UD |
730 | @end deftypefun |
731 | ||
732 | @comment fenv.h | |
733 | @comment ISO | |
734 | @deftypefun int fetestexcept (int @var{excepts}) | |
735 | Test whether the exception flags indicated by the parameter @var{except} | |
736 | are currently set. If any of them are, a nonzero value is returned | |
737 | which specifies which exceptions are set. Otherwise the result is zero. | |
738 | @end deftypefun | |
739 | ||
740 | To understand these functions, imagine that the status word is an | |
741 | integer variable named @var{status}. @code{feclearexcept} is then | |
742 | equivalent to @samp{status &= ~excepts} and @code{fetestexcept} is | |
743 | equivalent to @samp{(status & excepts)}. The actual implementation may | |
744 | be very different, of course. | |
745 | ||
746 | Exception flags are only cleared when the program explicitly requests it, | |
747 | by calling @code{feclearexcept}. If you want to check for exceptions | |
748 | from a set of calculations, you should clear all the flags first. Here | |
749 | is a simple example of the way to use @code{fetestexcept}: | |
b4012b75 UD |
750 | |
751 | @smallexample | |
7a68c94a UD |
752 | @{ |
753 | double f; | |
754 | int raised; | |
755 | feclearexcept (FE_ALL_EXCEPT); | |
756 | f = compute (); | |
757 | raised = fetestexcept (FE_OVERFLOW | FE_INVALID); | |
95fdc6a0 UD |
758 | if (raised & FE_OVERFLOW) @{ /* @dots{} */ @} |
759 | if (raised & FE_INVALID) @{ /* @dots{} */ @} | |
760 | /* @dots{} */ | |
7a68c94a | 761 | @} |
b4012b75 UD |
762 | @end smallexample |
763 | ||
7a68c94a UD |
764 | You cannot explicitly set bits in the status word. You can, however, |
765 | save the entire status word and restore it later. This is done with the | |
766 | following functions: | |
b4012b75 | 767 | |
7a68c94a | 768 | @comment fenv.h |
b4012b75 | 769 | @comment ISO |
63ae7b63 | 770 | @deftypefun int fegetexceptflag (fexcept_t *@var{flagp}, int @var{excepts}) |
7a68c94a UD |
771 | This function stores in the variable pointed to by @var{flagp} an |
772 | implementation-defined value representing the current setting of the | |
773 | exception flags indicated by @var{excepts}. | |
63ae7b63 UD |
774 | |
775 | The function returns zero in case the operation was successful, a | |
776 | non-zero value otherwise. | |
7a68c94a | 777 | @end deftypefun |
b4012b75 | 778 | |
7a68c94a UD |
779 | @comment fenv.h |
780 | @comment ISO | |
9251c568 | 781 | @deftypefun int fesetexceptflag (const fexcept_t *@var{flagp}, int @var{excepts}) |
7a68c94a UD |
782 | This function restores the flags for the exceptions indicated by |
783 | @var{excepts} to the values stored in the variable pointed to by | |
784 | @var{flagp}. | |
63ae7b63 UD |
785 | |
786 | The function returns zero in case the operation was successful, a | |
787 | non-zero value otherwise. | |
7a68c94a UD |
788 | @end deftypefun |
789 | ||
790 | Note that the value stored in @code{fexcept_t} bears no resemblance to | |
791 | the bit mask returned by @code{fetestexcept}. The type may not even be | |
792 | an integer. Do not attempt to modify an @code{fexcept_t} variable. | |
793 | ||
794 | @node Math Error Reporting | |
795 | @subsection Error Reporting by Mathematical Functions | |
796 | @cindex errors, mathematical | |
797 | @cindex domain error | |
798 | @cindex range error | |
799 | ||
800 | Many of the math functions are defined only over a subset of the real or | |
801 | complex numbers. Even if they are mathematically defined, their result | |
802 | may be larger or smaller than the range representable by their return | |
803 | type. These are known as @dfn{domain errors}, @dfn{overflows}, and | |
804 | @dfn{underflows}, respectively. Math functions do several things when | |
805 | one of these errors occurs. In this manual we will refer to the | |
806 | complete response as @dfn{signalling} a domain error, overflow, or | |
807 | underflow. | |
808 | ||
809 | When a math function suffers a domain error, it raises the invalid | |
810 | exception and returns NaN. It also sets @var{errno} to @code{EDOM}; | |
811 | this is for compatibility with old systems that do not support @w{IEEE | |
812 | 754} exception handling. Likewise, when overflow occurs, math | |
813 | functions raise the overflow exception and return @math{@infinity{}} or | |
814 | @math{-@infinity{}} as appropriate. They also set @var{errno} to | |
815 | @code{ERANGE}. When underflow occurs, the underflow exception is | |
816 | raised, and zero (appropriately signed) is returned. @var{errno} may be | |
817 | set to @code{ERANGE}, but this is not guaranteed. | |
818 | ||
819 | Some of the math functions are defined mathematically to result in a | |
820 | complex value over parts of their domains. The most familiar example of | |
821 | this is taking the square root of a negative number. The complex math | |
822 | functions, such as @code{csqrt}, will return the appropriate complex value | |
823 | in this case. The real-valued functions, such as @code{sqrt}, will | |
824 | signal a domain error. | |
825 | ||
826 | Some older hardware does not support infinities. On that hardware, | |
827 | overflows instead return a particular very large number (usually the | |
828 | largest representable number). @file{math.h} defines macros you can use | |
829 | to test for overflow on both old and new hardware. | |
b4012b75 UD |
830 | |
831 | @comment math.h | |
832 | @comment ISO | |
7a68c94a | 833 | @deftypevr Macro double HUGE_VAL |
4260bc74 UD |
834 | @comment math.h |
835 | @comment ISO | |
7a68c94a | 836 | @deftypevrx Macro float HUGE_VALF |
4260bc74 UD |
837 | @comment math.h |
838 | @comment ISO | |
7a68c94a UD |
839 | @deftypevrx Macro {long double} HUGE_VALL |
840 | An expression representing a particular very large number. On machines | |
841 | that use @w{IEEE 754} floating point format, @code{HUGE_VAL} is infinity. | |
842 | On other machines, it's typically the largest positive number that can | |
843 | be represented. | |
844 | ||
845 | Mathematical functions return the appropriately typed version of | |
846 | @code{HUGE_VAL} or @code{@minus{}HUGE_VAL} when the result is too large | |
847 | to be represented. | |
848 | @end deftypevr | |
b4012b75 | 849 | |
7a68c94a UD |
850 | @node Rounding |
851 | @section Rounding Modes | |
852 | ||
853 | Floating-point calculations are carried out internally with extra | |
854 | precision, and then rounded to fit into the destination type. This | |
855 | ensures that results are as precise as the input data. @w{IEEE 754} | |
856 | defines four possible rounding modes: | |
857 | ||
858 | @table @asis | |
859 | @item Round to nearest. | |
860 | This is the default mode. It should be used unless there is a specific | |
861 | need for one of the others. In this mode results are rounded to the | |
862 | nearest representable value. If the result is midway between two | |
863 | representable values, the even representable is chosen. @dfn{Even} here | |
864 | means the lowest-order bit is zero. This rounding mode prevents | |
865 | statistical bias and guarantees numeric stability: round-off errors in a | |
866 | lengthy calculation will remain smaller than half of @code{FLT_EPSILON}. | |
867 | ||
868 | @c @item Round toward @math{+@infinity{}} | |
869 | @item Round toward plus Infinity. | |
870 | All results are rounded to the smallest representable value | |
871 | which is greater than the result. | |
872 | ||
873 | @c @item Round toward @math{-@infinity{}} | |
874 | @item Round toward minus Infinity. | |
875 | All results are rounded to the largest representable value which is less | |
876 | than the result. | |
877 | ||
878 | @item Round toward zero. | |
879 | All results are rounded to the largest representable value whose | |
880 | magnitude is less than that of the result. In other words, if the | |
881 | result is negative it is rounded up; if it is positive, it is rounded | |
882 | down. | |
883 | @end table | |
b4012b75 | 884 | |
7a68c94a UD |
885 | @noindent |
886 | @file{fenv.h} defines constants which you can use to refer to the | |
887 | various rounding modes. Each one will be defined if and only if the FPU | |
888 | supports the corresponding rounding mode. | |
b4012b75 | 889 | |
7a68c94a UD |
890 | @table @code |
891 | @comment fenv.h | |
892 | @comment ISO | |
893 | @vindex FE_TONEAREST | |
894 | @item FE_TONEAREST | |
895 | Round to nearest. | |
b4012b75 | 896 | |
7a68c94a UD |
897 | @comment fenv.h |
898 | @comment ISO | |
899 | @vindex FE_UPWARD | |
900 | @item FE_UPWARD | |
901 | Round toward @math{+@infinity{}}. | |
b4012b75 | 902 | |
7a68c94a UD |
903 | @comment fenv.h |
904 | @comment ISO | |
905 | @vindex FE_DOWNWARD | |
906 | @item FE_DOWNWARD | |
907 | Round toward @math{-@infinity{}}. | |
b4012b75 | 908 | |
7a68c94a UD |
909 | @comment fenv.h |
910 | @comment ISO | |
911 | @vindex FE_TOWARDZERO | |
912 | @item FE_TOWARDZERO | |
913 | Round toward zero. | |
914 | @end table | |
b4012b75 | 915 | |
7a68c94a UD |
916 | Underflow is an unusual case. Normally, @w{IEEE 754} floating point |
917 | numbers are always normalized (@pxref{Floating Point Concepts}). | |
918 | Numbers smaller than @math{2^r} (where @math{r} is the minimum exponent, | |
919 | @code{FLT_MIN_RADIX-1} for @var{float}) cannot be represented as | |
920 | normalized numbers. Rounding all such numbers to zero or @math{2^r} | |
921 | would cause some algorithms to fail at 0. Therefore, they are left in | |
922 | denormalized form. That produces loss of precision, since some bits of | |
923 | the mantissa are stolen to indicate the decimal point. | |
924 | ||
925 | If a result is too small to be represented as a denormalized number, it | |
926 | is rounded to zero. However, the sign of the result is preserved; if | |
927 | the calculation was negative, the result is @dfn{negative zero}. | |
928 | Negative zero can also result from some operations on infinity, such as | |
929 | @math{4/-@infinity{}}. Negative zero behaves identically to zero except | |
930 | when the @code{copysign} or @code{signbit} functions are used to check | |
931 | the sign bit directly. | |
932 | ||
933 | At any time one of the above four rounding modes is selected. You can | |
934 | find out which one with this function: | |
935 | ||
936 | @comment fenv.h | |
937 | @comment ISO | |
938 | @deftypefun int fegetround (void) | |
939 | Returns the currently selected rounding mode, represented by one of the | |
940 | values of the defined rounding mode macros. | |
941 | @end deftypefun | |
b4012b75 | 942 | |
7a68c94a UD |
943 | @noindent |
944 | To change the rounding mode, use this function: | |
b4012b75 | 945 | |
7a68c94a UD |
946 | @comment fenv.h |
947 | @comment ISO | |
948 | @deftypefun int fesetround (int @var{round}) | |
949 | Changes the currently selected rounding mode to @var{round}. If | |
950 | @var{round} does not correspond to one of the supported rounding modes | |
d5655997 UD |
951 | nothing is changed. @code{fesetround} returns zero if it changed the |
952 | rounding mode, a nonzero value if the mode is not supported. | |
7a68c94a | 953 | @end deftypefun |
b4012b75 | 954 | |
7a68c94a UD |
955 | You should avoid changing the rounding mode if possible. It can be an |
956 | expensive operation; also, some hardware requires you to compile your | |
957 | program differently for it to work. The resulting code may run slower. | |
958 | See your compiler documentation for details. | |
959 | @c This section used to claim that functions existed to round one number | |
960 | @c in a specific fashion. I can't find any functions in the library | |
961 | @c that do that. -zw | |
962 | ||
963 | @node Control Functions | |
964 | @section Floating-Point Control Functions | |
965 | ||
966 | @w{IEEE 754} floating-point implementations allow the programmer to | |
967 | decide whether traps will occur for each of the exceptions, by setting | |
968 | bits in the @dfn{control word}. In C, traps result in the program | |
969 | receiving the @code{SIGFPE} signal; see @ref{Signal Handling}. | |
970 | ||
48b22986 | 971 | @strong{NB:} @w{IEEE 754} says that trap handlers are given details of |
7a68c94a UD |
972 | the exceptional situation, and can set the result value. C signals do |
973 | not provide any mechanism to pass this information back and forth. | |
974 | Trapping exceptions in C is therefore not very useful. | |
975 | ||
976 | It is sometimes necessary to save the state of the floating-point unit | |
977 | while you perform some calculation. The library provides functions | |
978 | which save and restore the exception flags, the set of exceptions that | |
979 | generate traps, and the rounding mode. This information is known as the | |
980 | @dfn{floating-point environment}. | |
981 | ||
982 | The functions to save and restore the floating-point environment all use | |
983 | a variable of type @code{fenv_t} to store information. This type is | |
984 | defined in @file{fenv.h}. Its size and contents are | |
985 | implementation-defined. You should not attempt to manipulate a variable | |
986 | of this type directly. | |
987 | ||
988 | To save the state of the FPU, use one of these functions: | |
989 | ||
990 | @comment fenv.h | |
b4012b75 | 991 | @comment ISO |
63ae7b63 | 992 | @deftypefun int fegetenv (fenv_t *@var{envp}) |
7a68c94a UD |
993 | Store the floating-point environment in the variable pointed to by |
994 | @var{envp}. | |
63ae7b63 UD |
995 | |
996 | The function returns zero in case the operation was successful, a | |
997 | non-zero value otherwise. | |
b4012b75 UD |
998 | @end deftypefun |
999 | ||
7a68c94a | 1000 | @comment fenv.h |
b4012b75 | 1001 | @comment ISO |
7a68c94a UD |
1002 | @deftypefun int feholdexcept (fenv_t *@var{envp}) |
1003 | Store the current floating-point environment in the object pointed to by | |
1004 | @var{envp}. Then clear all exception flags, and set the FPU to trap no | |
1005 | exceptions. Not all FPUs support trapping no exceptions; if | |
0f6b172f UD |
1006 | @code{feholdexcept} cannot set this mode, it returns nonzero value. If it |
1007 | succeeds, it returns zero. | |
b4012b75 UD |
1008 | @end deftypefun |
1009 | ||
7a7a7ee5 | 1010 | The functions which restore the floating-point environment can take these |
7a68c94a | 1011 | kinds of arguments: |
b4012b75 | 1012 | |
7a68c94a UD |
1013 | @itemize @bullet |
1014 | @item | |
1015 | Pointers to @code{fenv_t} objects, which were initialized previously by a | |
1016 | call to @code{fegetenv} or @code{feholdexcept}. | |
1017 | @item | |
1018 | @vindex FE_DFL_ENV | |
1019 | The special macro @code{FE_DFL_ENV} which represents the floating-point | |
1020 | environment as it was available at program start. | |
1021 | @item | |
7a7a7ee5 AJ |
1022 | Implementation defined macros with names starting with @code{FE_} and |
1023 | having type @code{fenv_t *}. | |
b4012b75 | 1024 | |
7a68c94a UD |
1025 | @vindex FE_NOMASK_ENV |
1026 | If possible, the GNU C Library defines a macro @code{FE_NOMASK_ENV} | |
1027 | which represents an environment where every exception raised causes a | |
1028 | trap to occur. You can test for this macro using @code{#ifdef}. It is | |
1029 | only defined if @code{_GNU_SOURCE} is defined. | |
1030 | ||
1031 | Some platforms might define other predefined environments. | |
1032 | @end itemize | |
1033 | ||
1034 | @noindent | |
1035 | To set the floating-point environment, you can use either of these | |
1036 | functions: | |
1037 | ||
1038 | @comment fenv.h | |
b4012b75 | 1039 | @comment ISO |
63ae7b63 | 1040 | @deftypefun int fesetenv (const fenv_t *@var{envp}) |
7a68c94a | 1041 | Set the floating-point environment to that described by @var{envp}. |
63ae7b63 UD |
1042 | |
1043 | The function returns zero in case the operation was successful, a | |
1044 | non-zero value otherwise. | |
b4012b75 UD |
1045 | @end deftypefun |
1046 | ||
7a68c94a | 1047 | @comment fenv.h |
b4012b75 | 1048 | @comment ISO |
63ae7b63 | 1049 | @deftypefun int feupdateenv (const fenv_t *@var{envp}) |
7a68c94a UD |
1050 | Like @code{fesetenv}, this function sets the floating-point environment |
1051 | to that described by @var{envp}. However, if any exceptions were | |
1052 | flagged in the status word before @code{feupdateenv} was called, they | |
1053 | remain flagged after the call. In other words, after @code{feupdateenv} | |
1054 | is called, the status word is the bitwise OR of the previous status word | |
1055 | and the one saved in @var{envp}. | |
63ae7b63 UD |
1056 | |
1057 | The function returns zero in case the operation was successful, a | |
1058 | non-zero value otherwise. | |
b4012b75 UD |
1059 | @end deftypefun |
1060 | ||
05ef7ce9 UD |
1061 | @noindent |
1062 | To control for individual exceptions if raising them causes a trap to | |
1063 | occur, you can use the following two functions. | |
1064 | ||
1065 | @strong{Portability Note:} These functions are all GNU extensions. | |
1066 | ||
1067 | @comment fenv.h | |
1068 | @comment GNU | |
1069 | @deftypefun int feenableexcept (int @var{excepts}) | |
1070 | This functions enables traps for each of the exceptions as indicated by | |
1071 | the parameter @var{except}. The individual excepetions are described in | |
6e8afc1c | 1072 | @ref{Status bit operations}. Only the specified exceptions are |
05ef7ce9 UD |
1073 | enabled, the status of the other exceptions is not changed. |
1074 | ||
1075 | The function returns the previous enabled exceptions in case the | |
1076 | operation was successful, @code{-1} otherwise. | |
1077 | @end deftypefun | |
1078 | ||
1079 | @comment fenv.h | |
1080 | @comment GNU | |
1081 | @deftypefun int fedisableexcept (int @var{excepts}) | |
1082 | This functions disables traps for each of the exceptions as indicated by | |
1083 | the parameter @var{except}. The individual excepetions are described in | |
6e8afc1c | 1084 | @ref{Status bit operations}. Only the specified exceptions are |
05ef7ce9 UD |
1085 | disabled, the status of the other exceptions is not changed. |
1086 | ||
1087 | The function returns the previous enabled exceptions in case the | |
1088 | operation was successful, @code{-1} otherwise. | |
1089 | @end deftypefun | |
1090 | ||
1091 | @comment fenv.h | |
1092 | @comment GNU | |
1093 | @deftypefun int fegetexcept (int @var{excepts}) | |
1094 | The function returns a bitmask of all currently enabled exceptions. It | |
1095 | returns @code{-1} in case of failure. | |
6e8afc1c | 1096 | @end deftypefun |
05ef7ce9 | 1097 | |
7a68c94a UD |
1098 | @node Arithmetic Functions |
1099 | @section Arithmetic Functions | |
b4012b75 | 1100 | |
7a68c94a UD |
1101 | The C library provides functions to do basic operations on |
1102 | floating-point numbers. These include absolute value, maximum and minimum, | |
1103 | normalization, bit twiddling, rounding, and a few others. | |
b4012b75 | 1104 | |
7a68c94a UD |
1105 | @menu |
1106 | * Absolute Value:: Absolute values of integers and floats. | |
1107 | * Normalization Functions:: Extracting exponents and putting them back. | |
1108 | * Rounding Functions:: Rounding floats to integers. | |
1109 | * Remainder Functions:: Remainders on division, precisely defined. | |
1110 | * FP Bit Twiddling:: Sign bit adjustment. Adding epsilon. | |
1111 | * FP Comparison Functions:: Comparisons without risk of exceptions. | |
1112 | * Misc FP Arithmetic:: Max, min, positive difference, multiply-add. | |
1113 | @end menu | |
b4012b75 | 1114 | |
28f540f4 | 1115 | @node Absolute Value |
7a68c94a | 1116 | @subsection Absolute Value |
28f540f4 RM |
1117 | @cindex absolute value functions |
1118 | ||
1119 | These functions are provided for obtaining the @dfn{absolute value} (or | |
1120 | @dfn{magnitude}) of a number. The absolute value of a real number | |
2d26e9eb | 1121 | @var{x} is @var{x} if @var{x} is positive, @minus{}@var{x} if @var{x} is |
28f540f4 RM |
1122 | negative. For a complex number @var{z}, whose real part is @var{x} and |
1123 | whose imaginary part is @var{y}, the absolute value is @w{@code{sqrt | |
1124 | (@var{x}*@var{x} + @var{y}*@var{y})}}. | |
1125 | ||
1126 | @pindex math.h | |
1127 | @pindex stdlib.h | |
fe0ec73e | 1128 | Prototypes for @code{abs}, @code{labs} and @code{llabs} are in @file{stdlib.h}; |
e518937a | 1129 | @code{imaxabs} is declared in @file{inttypes.h}; |
7a68c94a | 1130 | @code{fabs}, @code{fabsf} and @code{fabsl} are declared in @file{math.h}. |
b4012b75 | 1131 | @code{cabs}, @code{cabsf} and @code{cabsl} are declared in @file{complex.h}. |
28f540f4 RM |
1132 | |
1133 | @comment stdlib.h | |
f65fd747 | 1134 | @comment ISO |
28f540f4 | 1135 | @deftypefun int abs (int @var{number}) |
4260bc74 UD |
1136 | @comment stdlib.h |
1137 | @comment ISO | |
7a68c94a | 1138 | @deftypefunx {long int} labs (long int @var{number}) |
4260bc74 UD |
1139 | @comment stdlib.h |
1140 | @comment ISO | |
7a68c94a | 1141 | @deftypefunx {long long int} llabs (long long int @var{number}) |
e518937a UD |
1142 | @comment inttypes.h |
1143 | @comment ISO | |
1144 | @deftypefunx intmax_t imaxabs (intmax_t @var{number}) | |
7a68c94a | 1145 | These functions return the absolute value of @var{number}. |
28f540f4 RM |
1146 | |
1147 | Most computers use a two's complement integer representation, in which | |
1148 | the absolute value of @code{INT_MIN} (the smallest possible @code{int}) | |
1149 | cannot be represented; thus, @w{@code{abs (INT_MIN)}} is not defined. | |
28f540f4 | 1150 | |
ec751a23 | 1151 | @code{llabs} and @code{imaxdiv} are new to @w{ISO C99}. |
0e4ee106 UD |
1152 | |
1153 | See @ref{Integers} for a description of the @code{intmax_t} type. | |
1154 | ||
fe0ec73e UD |
1155 | @end deftypefun |
1156 | ||
28f540f4 | 1157 | @comment math.h |
f65fd747 | 1158 | @comment ISO |
28f540f4 | 1159 | @deftypefun double fabs (double @var{number}) |
4260bc74 UD |
1160 | @comment math.h |
1161 | @comment ISO | |
779ae82e | 1162 | @deftypefunx float fabsf (float @var{number}) |
4260bc74 UD |
1163 | @comment math.h |
1164 | @comment ISO | |
779ae82e | 1165 | @deftypefunx {long double} fabsl (long double @var{number}) |
28f540f4 RM |
1166 | This function returns the absolute value of the floating-point number |
1167 | @var{number}. | |
1168 | @end deftypefun | |
1169 | ||
b4012b75 UD |
1170 | @comment complex.h |
1171 | @comment ISO | |
1172 | @deftypefun double cabs (complex double @var{z}) | |
4260bc74 UD |
1173 | @comment complex.h |
1174 | @comment ISO | |
779ae82e | 1175 | @deftypefunx float cabsf (complex float @var{z}) |
4260bc74 UD |
1176 | @comment complex.h |
1177 | @comment ISO | |
779ae82e | 1178 | @deftypefunx {long double} cabsl (complex long double @var{z}) |
7a68c94a UD |
1179 | These functions return the absolute value of the complex number @var{z} |
1180 | (@pxref{Complex Numbers}). The absolute value of a complex number is: | |
28f540f4 RM |
1181 | |
1182 | @smallexample | |
b4012b75 | 1183 | sqrt (creal (@var{z}) * creal (@var{z}) + cimag (@var{z}) * cimag (@var{z})) |
28f540f4 | 1184 | @end smallexample |
dfd2257a | 1185 | |
7a68c94a UD |
1186 | This function should always be used instead of the direct formula |
1187 | because it takes special care to avoid losing precision. It may also | |
1188 | take advantage of hardware support for this operation. See @code{hypot} | |
8b7fb588 | 1189 | in @ref{Exponents and Logarithms}. |
28f540f4 RM |
1190 | @end deftypefun |
1191 | ||
1192 | @node Normalization Functions | |
7a68c94a | 1193 | @subsection Normalization Functions |
28f540f4 RM |
1194 | @cindex normalization functions (floating-point) |
1195 | ||
1196 | The functions described in this section are primarily provided as a way | |
1197 | to efficiently perform certain low-level manipulations on floating point | |
1198 | numbers that are represented internally using a binary radix; | |
1199 | see @ref{Floating Point Concepts}. These functions are required to | |
1200 | have equivalent behavior even if the representation does not use a radix | |
1201 | of 2, but of course they are unlikely to be particularly efficient in | |
1202 | those cases. | |
1203 | ||
1204 | @pindex math.h | |
1205 | All these functions are declared in @file{math.h}. | |
1206 | ||
1207 | @comment math.h | |
f65fd747 | 1208 | @comment ISO |
28f540f4 | 1209 | @deftypefun double frexp (double @var{value}, int *@var{exponent}) |
4260bc74 UD |
1210 | @comment math.h |
1211 | @comment ISO | |
779ae82e | 1212 | @deftypefunx float frexpf (float @var{value}, int *@var{exponent}) |
4260bc74 UD |
1213 | @comment math.h |
1214 | @comment ISO | |
779ae82e | 1215 | @deftypefunx {long double} frexpl (long double @var{value}, int *@var{exponent}) |
b4012b75 | 1216 | These functions are used to split the number @var{value} |
28f540f4 RM |
1217 | into a normalized fraction and an exponent. |
1218 | ||
1219 | If the argument @var{value} is not zero, the return value is @var{value} | |
1220 | times a power of two, and is always in the range 1/2 (inclusive) to 1 | |
1221 | (exclusive). The corresponding exponent is stored in | |
1222 | @code{*@var{exponent}}; the return value multiplied by 2 raised to this | |
1223 | exponent equals the original number @var{value}. | |
1224 | ||
1225 | For example, @code{frexp (12.8, &exponent)} returns @code{0.8} and | |
1226 | stores @code{4} in @code{exponent}. | |
1227 | ||
1228 | If @var{value} is zero, then the return value is zero and | |
1229 | zero is stored in @code{*@var{exponent}}. | |
1230 | @end deftypefun | |
1231 | ||
1232 | @comment math.h | |
f65fd747 | 1233 | @comment ISO |
28f540f4 | 1234 | @deftypefun double ldexp (double @var{value}, int @var{exponent}) |
4260bc74 UD |
1235 | @comment math.h |
1236 | @comment ISO | |
779ae82e | 1237 | @deftypefunx float ldexpf (float @var{value}, int @var{exponent}) |
4260bc74 UD |
1238 | @comment math.h |
1239 | @comment ISO | |
779ae82e | 1240 | @deftypefunx {long double} ldexpl (long double @var{value}, int @var{exponent}) |
b4012b75 | 1241 | These functions return the result of multiplying the floating-point |
28f540f4 RM |
1242 | number @var{value} by 2 raised to the power @var{exponent}. (It can |
1243 | be used to reassemble floating-point numbers that were taken apart | |
1244 | by @code{frexp}.) | |
1245 | ||
1246 | For example, @code{ldexp (0.8, 4)} returns @code{12.8}. | |
1247 | @end deftypefun | |
1248 | ||
7a68c94a | 1249 | The following functions, which come from BSD, provide facilities |
b7d03293 UD |
1250 | equivalent to those of @code{ldexp} and @code{frexp}. See also the |
1251 | @w{ISO C} function @code{logb} which originally also appeared in BSD. | |
7a68c94a UD |
1252 | |
1253 | @comment math.h | |
1254 | @comment BSD | |
1255 | @deftypefun double scalb (double @var{value}, int @var{exponent}) | |
4260bc74 UD |
1256 | @comment math.h |
1257 | @comment BSD | |
7a68c94a | 1258 | @deftypefunx float scalbf (float @var{value}, int @var{exponent}) |
4260bc74 UD |
1259 | @comment math.h |
1260 | @comment BSD | |
7a68c94a UD |
1261 | @deftypefunx {long double} scalbl (long double @var{value}, int @var{exponent}) |
1262 | The @code{scalb} function is the BSD name for @code{ldexp}. | |
1263 | @end deftypefun | |
1264 | ||
1265 | @comment math.h | |
1266 | @comment BSD | |
cc6e48bc | 1267 | @deftypefun {long long int} scalbn (double @var{x}, int @var{n}) |
4260bc74 UD |
1268 | @comment math.h |
1269 | @comment BSD | |
cc6e48bc | 1270 | @deftypefunx {long long int} scalbnf (float @var{x}, int @var{n}) |
4260bc74 UD |
1271 | @comment math.h |
1272 | @comment BSD | |
cc6e48bc | 1273 | @deftypefunx {long long int} scalbnl (long double @var{x}, int @var{n}) |
7a68c94a UD |
1274 | @code{scalbn} is identical to @code{scalb}, except that the exponent |
1275 | @var{n} is an @code{int} instead of a floating-point number. | |
1276 | @end deftypefun | |
1277 | ||
1278 | @comment math.h | |
1279 | @comment BSD | |
cc6e48bc | 1280 | @deftypefun {long long int} scalbln (double @var{x}, long int @var{n}) |
4260bc74 UD |
1281 | @comment math.h |
1282 | @comment BSD | |
cc6e48bc | 1283 | @deftypefunx {long long int} scalblnf (float @var{x}, long int @var{n}) |
4260bc74 UD |
1284 | @comment math.h |
1285 | @comment BSD | |
cc6e48bc | 1286 | @deftypefunx {long long int} scalblnl (long double @var{x}, long int @var{n}) |
7a68c94a UD |
1287 | @code{scalbln} is identical to @code{scalb}, except that the exponent |
1288 | @var{n} is a @code{long int} instead of a floating-point number. | |
1289 | @end deftypefun | |
28f540f4 | 1290 | |
7a68c94a UD |
1291 | @comment math.h |
1292 | @comment BSD | |
1293 | @deftypefun {long long int} significand (double @var{x}) | |
4260bc74 UD |
1294 | @comment math.h |
1295 | @comment BSD | |
7a68c94a | 1296 | @deftypefunx {long long int} significandf (float @var{x}) |
4260bc74 UD |
1297 | @comment math.h |
1298 | @comment BSD | |
7a68c94a UD |
1299 | @deftypefunx {long long int} significandl (long double @var{x}) |
1300 | @code{significand} returns the mantissa of @var{x} scaled to the range | |
1301 | @math{[1, 2)}. | |
1302 | It is equivalent to @w{@code{scalb (@var{x}, (double) -ilogb (@var{x}))}}. | |
1303 | ||
1304 | This function exists mainly for use in certain standardized tests | |
1305 | of @w{IEEE 754} conformance. | |
28f540f4 RM |
1306 | @end deftypefun |
1307 | ||
7a68c94a UD |
1308 | @node Rounding Functions |
1309 | @subsection Rounding Functions | |
28f540f4 RM |
1310 | @cindex converting floats to integers |
1311 | ||
1312 | @pindex math.h | |
7a68c94a UD |
1313 | The functions listed here perform operations such as rounding and |
1314 | truncation of floating-point values. Some of these functions convert | |
1315 | floating point numbers to integer values. They are all declared in | |
1316 | @file{math.h}. | |
28f540f4 RM |
1317 | |
1318 | You can also convert floating-point numbers to integers simply by | |
1319 | casting them to @code{int}. This discards the fractional part, | |
1320 | effectively rounding towards zero. However, this only works if the | |
1321 | result can actually be represented as an @code{int}---for very large | |
1322 | numbers, this is impossible. The functions listed here return the | |
1323 | result as a @code{double} instead to get around this problem. | |
1324 | ||
1325 | @comment math.h | |
f65fd747 | 1326 | @comment ISO |
28f540f4 | 1327 | @deftypefun double ceil (double @var{x}) |
4260bc74 UD |
1328 | @comment math.h |
1329 | @comment ISO | |
779ae82e | 1330 | @deftypefunx float ceilf (float @var{x}) |
4260bc74 UD |
1331 | @comment math.h |
1332 | @comment ISO | |
779ae82e | 1333 | @deftypefunx {long double} ceill (long double @var{x}) |
b4012b75 | 1334 | These functions round @var{x} upwards to the nearest integer, |
28f540f4 RM |
1335 | returning that value as a @code{double}. Thus, @code{ceil (1.5)} |
1336 | is @code{2.0}. | |
1337 | @end deftypefun | |
1338 | ||
1339 | @comment math.h | |
f65fd747 | 1340 | @comment ISO |
28f540f4 | 1341 | @deftypefun double floor (double @var{x}) |
4260bc74 UD |
1342 | @comment math.h |
1343 | @comment ISO | |
779ae82e | 1344 | @deftypefunx float floorf (float @var{x}) |
4260bc74 UD |
1345 | @comment math.h |
1346 | @comment ISO | |
779ae82e | 1347 | @deftypefunx {long double} floorl (long double @var{x}) |
b4012b75 | 1348 | These functions round @var{x} downwards to the nearest |
28f540f4 RM |
1349 | integer, returning that value as a @code{double}. Thus, @code{floor |
1350 | (1.5)} is @code{1.0} and @code{floor (-1.5)} is @code{-2.0}. | |
1351 | @end deftypefun | |
1352 | ||
7a68c94a UD |
1353 | @comment math.h |
1354 | @comment ISO | |
1355 | @deftypefun double trunc (double @var{x}) | |
4260bc74 UD |
1356 | @comment math.h |
1357 | @comment ISO | |
7a68c94a | 1358 | @deftypefunx float truncf (float @var{x}) |
4260bc74 UD |
1359 | @comment math.h |
1360 | @comment ISO | |
7a68c94a | 1361 | @deftypefunx {long double} truncl (long double @var{x}) |
e6e81391 UD |
1362 | The @code{trunc} functions round @var{x} towards zero to the nearest |
1363 | integer (returned in floating-point format). Thus, @code{trunc (1.5)} | |
1364 | is @code{1.0} and @code{trunc (-1.5)} is @code{-1.0}. | |
7a68c94a UD |
1365 | @end deftypefun |
1366 | ||
28f540f4 | 1367 | @comment math.h |
b4012b75 | 1368 | @comment ISO |
28f540f4 | 1369 | @deftypefun double rint (double @var{x}) |
4260bc74 UD |
1370 | @comment math.h |
1371 | @comment ISO | |
779ae82e | 1372 | @deftypefunx float rintf (float @var{x}) |
4260bc74 UD |
1373 | @comment math.h |
1374 | @comment ISO | |
779ae82e | 1375 | @deftypefunx {long double} rintl (long double @var{x}) |
b4012b75 | 1376 | These functions round @var{x} to an integer value according to the |
28f540f4 RM |
1377 | current rounding mode. @xref{Floating Point Parameters}, for |
1378 | information about the various rounding modes. The default | |
1379 | rounding mode is to round to the nearest integer; some machines | |
1380 | support other modes, but round-to-nearest is always used unless | |
7a68c94a UD |
1381 | you explicitly select another. |
1382 | ||
1383 | If @var{x} was not initially an integer, these functions raise the | |
1384 | inexact exception. | |
28f540f4 RM |
1385 | @end deftypefun |
1386 | ||
b4012b75 UD |
1387 | @comment math.h |
1388 | @comment ISO | |
1389 | @deftypefun double nearbyint (double @var{x}) | |
4260bc74 UD |
1390 | @comment math.h |
1391 | @comment ISO | |
779ae82e | 1392 | @deftypefunx float nearbyintf (float @var{x}) |
4260bc74 UD |
1393 | @comment math.h |
1394 | @comment ISO | |
779ae82e | 1395 | @deftypefunx {long double} nearbyintl (long double @var{x}) |
7a68c94a UD |
1396 | These functions return the same value as the @code{rint} functions, but |
1397 | do not raise the inexact exception if @var{x} is not an integer. | |
1398 | @end deftypefun | |
1399 | ||
1400 | @comment math.h | |
1401 | @comment ISO | |
1402 | @deftypefun double round (double @var{x}) | |
4260bc74 UD |
1403 | @comment math.h |
1404 | @comment ISO | |
7a68c94a | 1405 | @deftypefunx float roundf (float @var{x}) |
4260bc74 UD |
1406 | @comment math.h |
1407 | @comment ISO | |
7a68c94a UD |
1408 | @deftypefunx {long double} roundl (long double @var{x}) |
1409 | These functions are similar to @code{rint}, but they round halfway | |
713df3d5 RM |
1410 | cases away from zero instead of to the nearest integer (or other |
1411 | current rounding mode). | |
7a68c94a UD |
1412 | @end deftypefun |
1413 | ||
1414 | @comment math.h | |
1415 | @comment ISO | |
1416 | @deftypefun {long int} lrint (double @var{x}) | |
4260bc74 UD |
1417 | @comment math.h |
1418 | @comment ISO | |
7a68c94a | 1419 | @deftypefunx {long int} lrintf (float @var{x}) |
4260bc74 UD |
1420 | @comment math.h |
1421 | @comment ISO | |
7a68c94a UD |
1422 | @deftypefunx {long int} lrintl (long double @var{x}) |
1423 | These functions are just like @code{rint}, but they return a | |
1424 | @code{long int} instead of a floating-point number. | |
1425 | @end deftypefun | |
1426 | ||
1427 | @comment math.h | |
1428 | @comment ISO | |
1429 | @deftypefun {long long int} llrint (double @var{x}) | |
4260bc74 UD |
1430 | @comment math.h |
1431 | @comment ISO | |
7a68c94a | 1432 | @deftypefunx {long long int} llrintf (float @var{x}) |
4260bc74 UD |
1433 | @comment math.h |
1434 | @comment ISO | |
7a68c94a UD |
1435 | @deftypefunx {long long int} llrintl (long double @var{x}) |
1436 | These functions are just like @code{rint}, but they return a | |
1437 | @code{long long int} instead of a floating-point number. | |
b4012b75 UD |
1438 | @end deftypefun |
1439 | ||
7a68c94a UD |
1440 | @comment math.h |
1441 | @comment ISO | |
1442 | @deftypefun {long int} lround (double @var{x}) | |
4260bc74 UD |
1443 | @comment math.h |
1444 | @comment ISO | |
7a68c94a | 1445 | @deftypefunx {long int} lroundf (float @var{x}) |
4260bc74 UD |
1446 | @comment math.h |
1447 | @comment ISO | |
7a68c94a UD |
1448 | @deftypefunx {long int} lroundl (long double @var{x}) |
1449 | These functions are just like @code{round}, but they return a | |
1450 | @code{long int} instead of a floating-point number. | |
1451 | @end deftypefun | |
1452 | ||
1453 | @comment math.h | |
1454 | @comment ISO | |
1455 | @deftypefun {long long int} llround (double @var{x}) | |
4260bc74 UD |
1456 | @comment math.h |
1457 | @comment ISO | |
7a68c94a | 1458 | @deftypefunx {long long int} llroundf (float @var{x}) |
4260bc74 UD |
1459 | @comment math.h |
1460 | @comment ISO | |
7a68c94a UD |
1461 | @deftypefunx {long long int} llroundl (long double @var{x}) |
1462 | These functions are just like @code{round}, but they return a | |
1463 | @code{long long int} instead of a floating-point number. | |
1464 | @end deftypefun | |
1465 | ||
1466 | ||
28f540f4 | 1467 | @comment math.h |
f65fd747 | 1468 | @comment ISO |
28f540f4 | 1469 | @deftypefun double modf (double @var{value}, double *@var{integer-part}) |
4260bc74 UD |
1470 | @comment math.h |
1471 | @comment ISO | |
f2ea0f5b | 1472 | @deftypefunx float modff (float @var{value}, float *@var{integer-part}) |
4260bc74 UD |
1473 | @comment math.h |
1474 | @comment ISO | |
779ae82e | 1475 | @deftypefunx {long double} modfl (long double @var{value}, long double *@var{integer-part}) |
b4012b75 | 1476 | These functions break the argument @var{value} into an integer part and a |
28f540f4 RM |
1477 | fractional part (between @code{-1} and @code{1}, exclusive). Their sum |
1478 | equals @var{value}. Each of the parts has the same sign as @var{value}, | |
7a68c94a | 1479 | and the integer part is always rounded toward zero. |
28f540f4 RM |
1480 | |
1481 | @code{modf} stores the integer part in @code{*@var{integer-part}}, and | |
1482 | returns the fractional part. For example, @code{modf (2.5, &intpart)} | |
1483 | returns @code{0.5} and stores @code{2.0} into @code{intpart}. | |
1484 | @end deftypefun | |
1485 | ||
7a68c94a UD |
1486 | @node Remainder Functions |
1487 | @subsection Remainder Functions | |
1488 | ||
1489 | The functions in this section compute the remainder on division of two | |
1490 | floating-point numbers. Each is a little different; pick the one that | |
1491 | suits your problem. | |
1492 | ||
28f540f4 | 1493 | @comment math.h |
f65fd747 | 1494 | @comment ISO |
28f540f4 | 1495 | @deftypefun double fmod (double @var{numerator}, double @var{denominator}) |
4260bc74 UD |
1496 | @comment math.h |
1497 | @comment ISO | |
779ae82e | 1498 | @deftypefunx float fmodf (float @var{numerator}, float @var{denominator}) |
4260bc74 UD |
1499 | @comment math.h |
1500 | @comment ISO | |
779ae82e | 1501 | @deftypefunx {long double} fmodl (long double @var{numerator}, long double @var{denominator}) |
b4012b75 | 1502 | These functions compute the remainder from the division of |
28f540f4 RM |
1503 | @var{numerator} by @var{denominator}. Specifically, the return value is |
1504 | @code{@var{numerator} - @w{@var{n} * @var{denominator}}}, where @var{n} | |
1505 | is the quotient of @var{numerator} divided by @var{denominator}, rounded | |
1506 | towards zero to an integer. Thus, @w{@code{fmod (6.5, 2.3)}} returns | |
1507 | @code{1.9}, which is @code{6.5} minus @code{4.6}. | |
1508 | ||
1509 | The result has the same sign as the @var{numerator} and has magnitude | |
1510 | less than the magnitude of the @var{denominator}. | |
1511 | ||
7a68c94a | 1512 | If @var{denominator} is zero, @code{fmod} signals a domain error. |
28f540f4 RM |
1513 | @end deftypefun |
1514 | ||
1515 | @comment math.h | |
1516 | @comment BSD | |
1517 | @deftypefun double drem (double @var{numerator}, double @var{denominator}) | |
4260bc74 UD |
1518 | @comment math.h |
1519 | @comment BSD | |
779ae82e | 1520 | @deftypefunx float dremf (float @var{numerator}, float @var{denominator}) |
4260bc74 UD |
1521 | @comment math.h |
1522 | @comment BSD | |
779ae82e | 1523 | @deftypefunx {long double} dreml (long double @var{numerator}, long double @var{denominator}) |
76cf9889 | 1524 | These functions are like @code{fmod} except that they round the |
28f540f4 RM |
1525 | internal quotient @var{n} to the nearest integer instead of towards zero |
1526 | to an integer. For example, @code{drem (6.5, 2.3)} returns @code{-0.4}, | |
1527 | which is @code{6.5} minus @code{6.9}. | |
1528 | ||
1529 | The absolute value of the result is less than or equal to half the | |
1530 | absolute value of the @var{denominator}. The difference between | |
1531 | @code{fmod (@var{numerator}, @var{denominator})} and @code{drem | |
1532 | (@var{numerator}, @var{denominator})} is always either | |
1533 | @var{denominator}, minus @var{denominator}, or zero. | |
1534 | ||
7a68c94a | 1535 | If @var{denominator} is zero, @code{drem} signals a domain error. |
28f540f4 RM |
1536 | @end deftypefun |
1537 | ||
7a68c94a UD |
1538 | @comment math.h |
1539 | @comment BSD | |
1540 | @deftypefun double remainder (double @var{numerator}, double @var{denominator}) | |
4260bc74 UD |
1541 | @comment math.h |
1542 | @comment BSD | |
7a68c94a | 1543 | @deftypefunx float remainderf (float @var{numerator}, float @var{denominator}) |
4260bc74 UD |
1544 | @comment math.h |
1545 | @comment BSD | |
7a68c94a UD |
1546 | @deftypefunx {long double} remainderl (long double @var{numerator}, long double @var{denominator}) |
1547 | This function is another name for @code{drem}. | |
1548 | @end deftypefun | |
28f540f4 | 1549 | |
7a68c94a UD |
1550 | @node FP Bit Twiddling |
1551 | @subsection Setting and modifying single bits of FP values | |
fe0ec73e UD |
1552 | @cindex FP arithmetic |
1553 | ||
7a68c94a | 1554 | There are some operations that are too complicated or expensive to |
ec751a23 | 1555 | perform by hand on floating-point numbers. @w{ISO C99} defines |
7a68c94a UD |
1556 | functions to do these operations, which mostly involve changing single |
1557 | bits. | |
fe0ec73e UD |
1558 | |
1559 | @comment math.h | |
1560 | @comment ISO | |
1561 | @deftypefun double copysign (double @var{x}, double @var{y}) | |
4260bc74 UD |
1562 | @comment math.h |
1563 | @comment ISO | |
fe0ec73e | 1564 | @deftypefunx float copysignf (float @var{x}, float @var{y}) |
4260bc74 UD |
1565 | @comment math.h |
1566 | @comment ISO | |
fe0ec73e | 1567 | @deftypefunx {long double} copysignl (long double @var{x}, long double @var{y}) |
7a68c94a UD |
1568 | These functions return @var{x} but with the sign of @var{y}. They work |
1569 | even if @var{x} or @var{y} are NaN or zero. Both of these can carry a | |
1570 | sign (although not all implementations support it) and this is one of | |
1571 | the few operations that can tell the difference. | |
fe0ec73e | 1572 | |
7a68c94a UD |
1573 | @code{copysign} never raises an exception. |
1574 | @c except signalling NaNs | |
fe0ec73e UD |
1575 | |
1576 | This function is defined in @w{IEC 559} (and the appendix with | |
1577 | recommended functions in @w{IEEE 754}/@w{IEEE 854}). | |
1578 | @end deftypefun | |
1579 | ||
1580 | @comment math.h | |
1581 | @comment ISO | |
1582 | @deftypefun int signbit (@emph{float-type} @var{x}) | |
1583 | @code{signbit} is a generic macro which can work on all floating-point | |
1584 | types. It returns a nonzero value if the value of @var{x} has its sign | |
1585 | bit set. | |
1586 | ||
7a68c94a UD |
1587 | This is not the same as @code{x < 0.0}, because @w{IEEE 754} floating |
1588 | point allows zero to be signed. The comparison @code{-0.0 < 0.0} is | |
1589 | false, but @code{signbit (-0.0)} will return a nonzero value. | |
fe0ec73e UD |
1590 | @end deftypefun |
1591 | ||
1592 | @comment math.h | |
1593 | @comment ISO | |
1594 | @deftypefun double nextafter (double @var{x}, double @var{y}) | |
4260bc74 UD |
1595 | @comment math.h |
1596 | @comment ISO | |
fe0ec73e | 1597 | @deftypefunx float nextafterf (float @var{x}, float @var{y}) |
4260bc74 UD |
1598 | @comment math.h |
1599 | @comment ISO | |
fe0ec73e UD |
1600 | @deftypefunx {long double} nextafterl (long double @var{x}, long double @var{y}) |
1601 | The @code{nextafter} function returns the next representable neighbor of | |
7a68c94a UD |
1602 | @var{x} in the direction towards @var{y}. The size of the step between |
1603 | @var{x} and the result depends on the type of the result. If | |
0a7fef01 | 1604 | @math{@var{x} = @var{y}} the function simply returns @var{y}. If either |
7a68c94a UD |
1605 | value is @code{NaN}, @code{NaN} is returned. Otherwise |
1606 | a value corresponding to the value of the least significant bit in the | |
1607 | mantissa is added or subtracted, depending on the direction. | |
1608 | @code{nextafter} will signal overflow or underflow if the result goes | |
1609 | outside of the range of normalized numbers. | |
fe0ec73e UD |
1610 | |
1611 | This function is defined in @w{IEC 559} (and the appendix with | |
1612 | recommended functions in @w{IEEE 754}/@w{IEEE 854}). | |
1613 | @end deftypefun | |
1614 | ||
7a68c94a UD |
1615 | @comment math.h |
1616 | @comment ISO | |
36fe9ac9 | 1617 | @deftypefun double nexttoward (double @var{x}, long double @var{y}) |
4260bc74 UD |
1618 | @comment math.h |
1619 | @comment ISO | |
36fe9ac9 | 1620 | @deftypefunx float nexttowardf (float @var{x}, long double @var{y}) |
4260bc74 UD |
1621 | @comment math.h |
1622 | @comment ISO | |
36fe9ac9 | 1623 | @deftypefunx {long double} nexttowardl (long double @var{x}, long double @var{y}) |
7a68c94a UD |
1624 | These functions are identical to the corresponding versions of |
1625 | @code{nextafter} except that their second argument is a @code{long | |
1626 | double}. | |
1627 | @end deftypefun | |
1628 | ||
fe0ec73e UD |
1629 | @cindex NaN |
1630 | @comment math.h | |
1631 | @comment ISO | |
1632 | @deftypefun double nan (const char *@var{tagp}) | |
4260bc74 UD |
1633 | @comment math.h |
1634 | @comment ISO | |
fe0ec73e | 1635 | @deftypefunx float nanf (const char *@var{tagp}) |
4260bc74 UD |
1636 | @comment math.h |
1637 | @comment ISO | |
fe0ec73e | 1638 | @deftypefunx {long double} nanl (const char *@var{tagp}) |
7a68c94a UD |
1639 | The @code{nan} function returns a representation of NaN, provided that |
1640 | NaN is supported by the target platform. | |
1641 | @code{nan ("@var{n-char-sequence}")} is equivalent to | |
1642 | @code{strtod ("NAN(@var{n-char-sequence})")}. | |
1643 | ||
1644 | The argument @var{tagp} is used in an unspecified manner. On @w{IEEE | |
1645 | 754} systems, there are many representations of NaN, and @var{tagp} | |
1646 | selects one. On other systems it may do nothing. | |
fe0ec73e UD |
1647 | @end deftypefun |
1648 | ||
7a68c94a UD |
1649 | @node FP Comparison Functions |
1650 | @subsection Floating-Point Comparison Functions | |
1651 | @cindex unordered comparison | |
fe0ec73e | 1652 | |
7a68c94a UD |
1653 | The standard C comparison operators provoke exceptions when one or other |
1654 | of the operands is NaN. For example, | |
1655 | ||
1656 | @smallexample | |
1657 | int v = a < 1.0; | |
1658 | @end smallexample | |
1659 | ||
1660 | @noindent | |
1661 | will raise an exception if @var{a} is NaN. (This does @emph{not} | |
1662 | happen with @code{==} and @code{!=}; those merely return false and true, | |
1663 | respectively, when NaN is examined.) Frequently this exception is | |
ec751a23 | 1664 | undesirable. @w{ISO C99} therefore defines comparison functions that |
7a68c94a UD |
1665 | do not raise exceptions when NaN is examined. All of the functions are |
1666 | implemented as macros which allow their arguments to be of any | |
1667 | floating-point type. The macros are guaranteed to evaluate their | |
1668 | arguments only once. | |
1669 | ||
1670 | @comment math.h | |
1671 | @comment ISO | |
1672 | @deftypefn Macro int isgreater (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1673 | This macro determines whether the argument @var{x} is greater than | |
1674 | @var{y}. It is equivalent to @code{(@var{x}) > (@var{y})}, but no | |
1675 | exception is raised if @var{x} or @var{y} are NaN. | |
1676 | @end deftypefn | |
1677 | ||
1678 | @comment math.h | |
1679 | @comment ISO | |
1680 | @deftypefn Macro int isgreaterequal (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1681 | This macro determines whether the argument @var{x} is greater than or | |
1682 | equal to @var{y}. It is equivalent to @code{(@var{x}) >= (@var{y})}, but no | |
1683 | exception is raised if @var{x} or @var{y} are NaN. | |
1684 | @end deftypefn | |
1685 | ||
1686 | @comment math.h | |
1687 | @comment ISO | |
1688 | @deftypefn Macro int isless (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1689 | This macro determines whether the argument @var{x} is less than @var{y}. | |
1690 | It is equivalent to @code{(@var{x}) < (@var{y})}, but no exception is | |
1691 | raised if @var{x} or @var{y} are NaN. | |
1692 | @end deftypefn | |
1693 | ||
1694 | @comment math.h | |
1695 | @comment ISO | |
1696 | @deftypefn Macro int islessequal (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1697 | This macro determines whether the argument @var{x} is less than or equal | |
1698 | to @var{y}. It is equivalent to @code{(@var{x}) <= (@var{y})}, but no | |
1699 | exception is raised if @var{x} or @var{y} are NaN. | |
1700 | @end deftypefn | |
1701 | ||
1702 | @comment math.h | |
1703 | @comment ISO | |
1704 | @deftypefn Macro int islessgreater (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1705 | This macro determines whether the argument @var{x} is less or greater | |
1706 | than @var{y}. It is equivalent to @code{(@var{x}) < (@var{y}) || | |
1707 | (@var{x}) > (@var{y})} (although it only evaluates @var{x} and @var{y} | |
1708 | once), but no exception is raised if @var{x} or @var{y} are NaN. | |
1709 | ||
1710 | This macro is not equivalent to @code{@var{x} != @var{y}}, because that | |
1711 | expression is true if @var{x} or @var{y} are NaN. | |
1712 | @end deftypefn | |
1713 | ||
1714 | @comment math.h | |
1715 | @comment ISO | |
1716 | @deftypefn Macro int isunordered (@emph{real-floating} @var{x}, @emph{real-floating} @var{y}) | |
1717 | This macro determines whether its arguments are unordered. In other | |
1718 | words, it is true if @var{x} or @var{y} are NaN, and false otherwise. | |
1719 | @end deftypefn | |
1720 | ||
1721 | Not all machines provide hardware support for these operations. On | |
1722 | machines that don't, the macros can be very slow. Therefore, you should | |
1723 | not use these functions when NaN is not a concern. | |
1724 | ||
48b22986 | 1725 | @strong{NB:} There are no macros @code{isequal} or @code{isunequal}. |
7a68c94a UD |
1726 | They are unnecessary, because the @code{==} and @code{!=} operators do |
1727 | @emph{not} throw an exception if one or both of the operands are NaN. | |
1728 | ||
1729 | @node Misc FP Arithmetic | |
1730 | @subsection Miscellaneous FP arithmetic functions | |
fe0ec73e UD |
1731 | @cindex minimum |
1732 | @cindex maximum | |
7a68c94a UD |
1733 | @cindex positive difference |
1734 | @cindex multiply-add | |
fe0ec73e | 1735 | |
7a68c94a UD |
1736 | The functions in this section perform miscellaneous but common |
1737 | operations that are awkward to express with C operators. On some | |
1738 | processors these functions can use special machine instructions to | |
1739 | perform these operations faster than the equivalent C code. | |
fe0ec73e UD |
1740 | |
1741 | @comment math.h | |
1742 | @comment ISO | |
1743 | @deftypefun double fmin (double @var{x}, double @var{y}) | |
4260bc74 UD |
1744 | @comment math.h |
1745 | @comment ISO | |
fe0ec73e | 1746 | @deftypefunx float fminf (float @var{x}, float @var{y}) |
4260bc74 UD |
1747 | @comment math.h |
1748 | @comment ISO | |
fe0ec73e | 1749 | @deftypefunx {long double} fminl (long double @var{x}, long double @var{y}) |
7a68c94a UD |
1750 | The @code{fmin} function returns the lesser of the two values @var{x} |
1751 | and @var{y}. It is similar to the expression | |
1752 | @smallexample | |
1753 | ((x) < (y) ? (x) : (y)) | |
1754 | @end smallexample | |
1755 | except that @var{x} and @var{y} are only evaluated once. | |
fe0ec73e | 1756 | |
7a68c94a UD |
1757 | If an argument is NaN, the other argument is returned. If both arguments |
1758 | are NaN, NaN is returned. | |
fe0ec73e UD |
1759 | @end deftypefun |
1760 | ||
1761 | @comment math.h | |
1762 | @comment ISO | |
1763 | @deftypefun double fmax (double @var{x}, double @var{y}) | |
4260bc74 UD |
1764 | @comment math.h |
1765 | @comment ISO | |
fe0ec73e | 1766 | @deftypefunx float fmaxf (float @var{x}, float @var{y}) |
4260bc74 UD |
1767 | @comment math.h |
1768 | @comment ISO | |
fe0ec73e | 1769 | @deftypefunx {long double} fmaxl (long double @var{x}, long double @var{y}) |
7a68c94a UD |
1770 | The @code{fmax} function returns the greater of the two values @var{x} |
1771 | and @var{y}. | |
fe0ec73e | 1772 | |
7a68c94a UD |
1773 | If an argument is NaN, the other argument is returned. If both arguments |
1774 | are NaN, NaN is returned. | |
fe0ec73e UD |
1775 | @end deftypefun |
1776 | ||
1777 | @comment math.h | |
1778 | @comment ISO | |
1779 | @deftypefun double fdim (double @var{x}, double @var{y}) | |
4260bc74 UD |
1780 | @comment math.h |
1781 | @comment ISO | |
fe0ec73e | 1782 | @deftypefunx float fdimf (float @var{x}, float @var{y}) |
4260bc74 UD |
1783 | @comment math.h |
1784 | @comment ISO | |
fe0ec73e | 1785 | @deftypefunx {long double} fdiml (long double @var{x}, long double @var{y}) |
7a68c94a UD |
1786 | The @code{fdim} function returns the positive difference between |
1787 | @var{x} and @var{y}. The positive difference is @math{@var{x} - | |
1788 | @var{y}} if @var{x} is greater than @var{y}, and @math{0} otherwise. | |
fe0ec73e | 1789 | |
7a68c94a | 1790 | If @var{x}, @var{y}, or both are NaN, NaN is returned. |
fe0ec73e UD |
1791 | @end deftypefun |
1792 | ||
1793 | @comment math.h | |
1794 | @comment ISO | |
1795 | @deftypefun double fma (double @var{x}, double @var{y}, double @var{z}) | |
4260bc74 UD |
1796 | @comment math.h |
1797 | @comment ISO | |
fe0ec73e | 1798 | @deftypefunx float fmaf (float @var{x}, float @var{y}, float @var{z}) |
4260bc74 UD |
1799 | @comment math.h |
1800 | @comment ISO | |
fe0ec73e UD |
1801 | @deftypefunx {long double} fmal (long double @var{x}, long double @var{y}, long double @var{z}) |
1802 | @cindex butterfly | |
7a68c94a UD |
1803 | The @code{fma} function performs floating-point multiply-add. This is |
1804 | the operation @math{(@var{x} @mul{} @var{y}) + @var{z}}, but the | |
1805 | intermediate result is not rounded to the destination type. This can | |
1806 | sometimes improve the precision of a calculation. | |
1807 | ||
1808 | This function was introduced because some processors have a special | |
1809 | instruction to perform multiply-add. The C compiler cannot use it | |
1810 | directly, because the expression @samp{x*y + z} is defined to round the | |
1811 | intermediate result. @code{fma} lets you choose when you want to round | |
1812 | only once. | |
fe0ec73e UD |
1813 | |
1814 | @vindex FP_FAST_FMA | |
7a68c94a UD |
1815 | On processors which do not implement multiply-add in hardware, |
1816 | @code{fma} can be very slow since it must avoid intermediate rounding. | |
1817 | @file{math.h} defines the symbols @code{FP_FAST_FMA}, | |
1818 | @code{FP_FAST_FMAF}, and @code{FP_FAST_FMAL} when the corresponding | |
1819 | version of @code{fma} is no slower than the expression @samp{x*y + z}. | |
1820 | In the GNU C library, this always means the operation is implemented in | |
1821 | hardware. | |
fe0ec73e UD |
1822 | @end deftypefun |
1823 | ||
7a68c94a UD |
1824 | @node Complex Numbers |
1825 | @section Complex Numbers | |
1826 | @pindex complex.h | |
1827 | @cindex complex numbers | |
1828 | ||
ec751a23 | 1829 | @w{ISO C99} introduces support for complex numbers in C. This is done |
7a68c94a UD |
1830 | with a new type qualifier, @code{complex}. It is a keyword if and only |
1831 | if @file{complex.h} has been included. There are three complex types, | |
1832 | corresponding to the three real types: @code{float complex}, | |
1833 | @code{double complex}, and @code{long double complex}. | |
1834 | ||
1835 | To construct complex numbers you need a way to indicate the imaginary | |
1836 | part of a number. There is no standard notation for an imaginary | |
1837 | floating point constant. Instead, @file{complex.h} defines two macros | |
1838 | that can be used to create complex numbers. | |
1839 | ||
1840 | @deftypevr Macro {const float complex} _Complex_I | |
1841 | This macro is a representation of the complex number ``@math{0+1i}''. | |
1842 | Multiplying a real floating-point value by @code{_Complex_I} gives a | |
1843 | complex number whose value is purely imaginary. You can use this to | |
1844 | construct complex constants: | |
1845 | ||
1846 | @smallexample | |
1847 | @math{3.0 + 4.0i} = @code{3.0 + 4.0 * _Complex_I} | |
1848 | @end smallexample | |
1849 | ||
1850 | Note that @code{_Complex_I * _Complex_I} has the value @code{-1}, but | |
1851 | the type of that value is @code{complex}. | |
1852 | @end deftypevr | |
1853 | ||
1854 | @c Put this back in when gcc supports _Imaginary_I. It's too confusing. | |
1855 | @ignore | |
1856 | @noindent | |
1857 | Without an optimizing compiler this is more expensive than the use of | |
1858 | @code{_Imaginary_I} but with is better than nothing. You can avoid all | |
1859 | the hassles if you use the @code{I} macro below if the name is not | |
1860 | problem. | |
1861 | ||
1862 | @deftypevr Macro {const float imaginary} _Imaginary_I | |
1863 | This macro is a representation of the value ``@math{1i}''. I.e., it is | |
1864 | the value for which | |
1865 | ||
1866 | @smallexample | |
1867 | _Imaginary_I * _Imaginary_I = -1 | |
1868 | @end smallexample | |
1869 | ||
1870 | @noindent | |
1871 | The result is not of type @code{float imaginary} but instead @code{float}. | |
1872 | One can use it to easily construct complex number like in | |
1873 | ||
1874 | @smallexample | |
1875 | 3.0 - _Imaginary_I * 4.0 | |
1876 | @end smallexample | |
1877 | ||
1878 | @noindent | |
1879 | which results in the complex number with a real part of 3.0 and a | |
1880 | imaginary part -4.0. | |
1881 | @end deftypevr | |
1882 | @end ignore | |
1883 | ||
1884 | @noindent | |
1885 | @code{_Complex_I} is a bit of a mouthful. @file{complex.h} also defines | |
1886 | a shorter name for the same constant. | |
1887 | ||
1888 | @deftypevr Macro {const float complex} I | |
1889 | This macro has exactly the same value as @code{_Complex_I}. Most of the | |
1890 | time it is preferable. However, it causes problems if you want to use | |
1891 | the identifier @code{I} for something else. You can safely write | |
1892 | ||
1893 | @smallexample | |
1894 | #include <complex.h> | |
1895 | #undef I | |
1896 | @end smallexample | |
1897 | ||
1898 | @noindent | |
1899 | if you need @code{I} for your own purposes. (In that case we recommend | |
1900 | you also define some other short name for @code{_Complex_I}, such as | |
1901 | @code{J}.) | |
1902 | ||
1903 | @ignore | |
1904 | If the implementation does not support the @code{imaginary} types | |
1905 | @code{I} is defined as @code{_Complex_I} which is the second best | |
1906 | solution. It still can be used in the same way but requires a most | |
1907 | clever compiler to get the same results. | |
1908 | @end ignore | |
1909 | @end deftypevr | |
1910 | ||
1911 | @node Operations on Complex | |
1912 | @section Projections, Conjugates, and Decomposing of Complex Numbers | |
1913 | @cindex project complex numbers | |
1914 | @cindex conjugate complex numbers | |
1915 | @cindex decompose complex numbers | |
1916 | @pindex complex.h | |
1917 | ||
ec751a23 | 1918 | @w{ISO C99} also defines functions that perform basic operations on |
7a68c94a UD |
1919 | complex numbers, such as decomposition and conjugation. The prototypes |
1920 | for all these functions are in @file{complex.h}. All functions are | |
1921 | available in three variants, one for each of the three complex types. | |
1922 | ||
1923 | @comment complex.h | |
1924 | @comment ISO | |
1925 | @deftypefun double creal (complex double @var{z}) | |
4260bc74 UD |
1926 | @comment complex.h |
1927 | @comment ISO | |
7a68c94a | 1928 | @deftypefunx float crealf (complex float @var{z}) |
4260bc74 UD |
1929 | @comment complex.h |
1930 | @comment ISO | |
7a68c94a UD |
1931 | @deftypefunx {long double} creall (complex long double @var{z}) |
1932 | These functions return the real part of the complex number @var{z}. | |
1933 | @end deftypefun | |
1934 | ||
1935 | @comment complex.h | |
1936 | @comment ISO | |
1937 | @deftypefun double cimag (complex double @var{z}) | |
4260bc74 UD |
1938 | @comment complex.h |
1939 | @comment ISO | |
7a68c94a | 1940 | @deftypefunx float cimagf (complex float @var{z}) |
4260bc74 UD |
1941 | @comment complex.h |
1942 | @comment ISO | |
7a68c94a UD |
1943 | @deftypefunx {long double} cimagl (complex long double @var{z}) |
1944 | These functions return the imaginary part of the complex number @var{z}. | |
1945 | @end deftypefun | |
1946 | ||
1947 | @comment complex.h | |
1948 | @comment ISO | |
1949 | @deftypefun {complex double} conj (complex double @var{z}) | |
4260bc74 UD |
1950 | @comment complex.h |
1951 | @comment ISO | |
7a68c94a | 1952 | @deftypefunx {complex float} conjf (complex float @var{z}) |
4260bc74 UD |
1953 | @comment complex.h |
1954 | @comment ISO | |
7a68c94a UD |
1955 | @deftypefunx {complex long double} conjl (complex long double @var{z}) |
1956 | These functions return the conjugate value of the complex number | |
1957 | @var{z}. The conjugate of a complex number has the same real part and a | |
1958 | negated imaginary part. In other words, @samp{conj(a + bi) = a + -bi}. | |
1959 | @end deftypefun | |
1960 | ||
1961 | @comment complex.h | |
1962 | @comment ISO | |
1963 | @deftypefun double carg (complex double @var{z}) | |
4260bc74 UD |
1964 | @comment complex.h |
1965 | @comment ISO | |
7a68c94a | 1966 | @deftypefunx float cargf (complex float @var{z}) |
4260bc74 UD |
1967 | @comment complex.h |
1968 | @comment ISO | |
7a68c94a UD |
1969 | @deftypefunx {long double} cargl (complex long double @var{z}) |
1970 | These functions return the argument of the complex number @var{z}. | |
1971 | The argument of a complex number is the angle in the complex plane | |
1972 | between the positive real axis and a line passing through zero and the | |
1973 | number. This angle is measured in the usual fashion and ranges from @math{0} | |
1974 | to @math{2@pi{}}. | |
1975 | ||
1976 | @code{carg} has a branch cut along the positive real axis. | |
1977 | @end deftypefun | |
1978 | ||
1979 | @comment complex.h | |
1980 | @comment ISO | |
1981 | @deftypefun {complex double} cproj (complex double @var{z}) | |
4260bc74 UD |
1982 | @comment complex.h |
1983 | @comment ISO | |
7a68c94a | 1984 | @deftypefunx {complex float} cprojf (complex float @var{z}) |
4260bc74 UD |
1985 | @comment complex.h |
1986 | @comment ISO | |
7a68c94a UD |
1987 | @deftypefunx {complex long double} cprojl (complex long double @var{z}) |
1988 | These functions return the projection of the complex value @var{z} onto | |
1989 | the Riemann sphere. Values with a infinite imaginary part are projected | |
1990 | to positive infinity on the real axis, even if the real part is NaN. If | |
1991 | the real part is infinite, the result is equivalent to | |
1992 | ||
1993 | @smallexample | |
1994 | INFINITY + I * copysign (0.0, cimag (z)) | |
1995 | @end smallexample | |
1996 | @end deftypefun | |
fe0ec73e | 1997 | |
28f540f4 RM |
1998 | @node Parsing of Numbers |
1999 | @section Parsing of Numbers | |
2000 | @cindex parsing numbers (in formatted input) | |
2001 | @cindex converting strings to numbers | |
2002 | @cindex number syntax, parsing | |
2003 | @cindex syntax, for reading numbers | |
2004 | ||
2005 | This section describes functions for ``reading'' integer and | |
2006 | floating-point numbers from a string. It may be more convenient in some | |
2007 | cases to use @code{sscanf} or one of the related functions; see | |
2008 | @ref{Formatted Input}. But often you can make a program more robust by | |
2009 | finding the tokens in the string by hand, then converting the numbers | |
2010 | one by one. | |
2011 | ||
2012 | @menu | |
2013 | * Parsing of Integers:: Functions for conversion of integer values. | |
2014 | * Parsing of Floats:: Functions for conversion of floating-point | |
2015 | values. | |
2016 | @end menu | |
2017 | ||
2018 | @node Parsing of Integers | |
2019 | @subsection Parsing of Integers | |
2020 | ||
2021 | @pindex stdlib.h | |
b642f101 UD |
2022 | @pindex wchar.h |
2023 | The @samp{str} functions are declared in @file{stdlib.h} and those | |
2024 | beginning with @samp{wcs} are declared in @file{wchar.h}. One might | |
2025 | wonder about the use of @code{restrict} in the prototypes of the | |
2026 | functions in this section. It is seemingly useless but the @w{ISO C} | |
2027 | standard uses it (for the functions defined there) so we have to do it | |
2028 | as well. | |
28f540f4 RM |
2029 | |
2030 | @comment stdlib.h | |
f65fd747 | 2031 | @comment ISO |
b642f101 | 2032 | @deftypefun {long int} strtol (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
28f540f4 RM |
2033 | The @code{strtol} (``string-to-long'') function converts the initial |
2034 | part of @var{string} to a signed integer, which is returned as a value | |
b8fe19fa | 2035 | of type @code{long int}. |
28f540f4 RM |
2036 | |
2037 | This function attempts to decompose @var{string} as follows: | |
2038 | ||
2039 | @itemize @bullet | |
b8fe19fa | 2040 | @item |
28f540f4 RM |
2041 | A (possibly empty) sequence of whitespace characters. Which characters |
2042 | are whitespace is determined by the @code{isspace} function | |
2043 | (@pxref{Classification of Characters}). These are discarded. | |
2044 | ||
b8fe19fa | 2045 | @item |
28f540f4 RM |
2046 | An optional plus or minus sign (@samp{+} or @samp{-}). |
2047 | ||
b8fe19fa | 2048 | @item |
28f540f4 RM |
2049 | A nonempty sequence of digits in the radix specified by @var{base}. |
2050 | ||
2051 | If @var{base} is zero, decimal radix is assumed unless the series of | |
2052 | digits begins with @samp{0} (specifying octal radix), or @samp{0x} or | |
2053 | @samp{0X} (specifying hexadecimal radix); in other words, the same | |
2054 | syntax used for integer constants in C. | |
2055 | ||
600a7457 | 2056 | Otherwise @var{base} must have a value between @code{2} and @code{36}. |
28f540f4 | 2057 | If @var{base} is @code{16}, the digits may optionally be preceded by |
2c6fe0bd UD |
2058 | @samp{0x} or @samp{0X}. If base has no legal value the value returned |
2059 | is @code{0l} and the global variable @code{errno} is set to @code{EINVAL}. | |
28f540f4 | 2060 | |
b8fe19fa | 2061 | @item |
28f540f4 RM |
2062 | Any remaining characters in the string. If @var{tailptr} is not a null |
2063 | pointer, @code{strtol} stores a pointer to this tail in | |
2064 | @code{*@var{tailptr}}. | |
2065 | @end itemize | |
2066 | ||
2067 | If the string is empty, contains only whitespace, or does not contain an | |
2068 | initial substring that has the expected syntax for an integer in the | |
2069 | specified @var{base}, no conversion is performed. In this case, | |
2070 | @code{strtol} returns a value of zero and the value stored in | |
2071 | @code{*@var{tailptr}} is the value of @var{string}. | |
2072 | ||
2073 | In a locale other than the standard @code{"C"} locale, this function | |
2074 | may recognize additional implementation-dependent syntax. | |
2075 | ||
2076 | If the string has valid syntax for an integer but the value is not | |
2077 | representable because of overflow, @code{strtol} returns either | |
2078 | @code{LONG_MAX} or @code{LONG_MIN} (@pxref{Range of Type}), as | |
2079 | appropriate for the sign of the value. It also sets @code{errno} | |
2080 | to @code{ERANGE} to indicate there was overflow. | |
2081 | ||
7a68c94a UD |
2082 | You should not check for errors by examining the return value of |
2083 | @code{strtol}, because the string might be a valid representation of | |
2084 | @code{0l}, @code{LONG_MAX}, or @code{LONG_MIN}. Instead, check whether | |
2085 | @var{tailptr} points to what you expect after the number | |
2086 | (e.g. @code{'\0'} if the string should end after the number). You also | |
2087 | need to clear @var{errno} before the call and check it afterward, in | |
2088 | case there was overflow. | |
2c6fe0bd | 2089 | |
28f540f4 RM |
2090 | There is an example at the end of this section. |
2091 | @end deftypefun | |
2092 | ||
b642f101 UD |
2093 | @comment wchar.h |
2094 | @comment ISO | |
2095 | @deftypefun {long int} wcstol (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2096 | The @code{wcstol} function is equivalent to the @code{strtol} function |
2097 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2098 | |
2099 | The @code{wcstol} function was introduced in @w{Amendment 1} of @w{ISO C90}. | |
2100 | @end deftypefun | |
2101 | ||
28f540f4 | 2102 | @comment stdlib.h |
f65fd747 | 2103 | @comment ISO |
b642f101 | 2104 | @deftypefun {unsigned long int} strtoul (const char *retrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
28f540f4 | 2105 | The @code{strtoul} (``string-to-unsigned-long'') function is like |
0e4ee106 | 2106 | @code{strtol} except it converts to an @code{unsigned long int} value. |
7a68c94a | 2107 | The syntax is the same as described above for @code{strtol}. The value |
0e4ee106 UD |
2108 | returned on overflow is @code{ULONG_MAX} (@pxref{Range of Type}). |
2109 | ||
2110 | If @var{string} depicts a negative number, @code{strtoul} acts the same | |
2111 | as @var{strtol} but casts the result to an unsigned integer. That means | |
2112 | for example that @code{strtoul} on @code{"-1"} returns @code{ULONG_MAX} | |
e6e81391 | 2113 | and an input more negative than @code{LONG_MIN} returns |
0e4ee106 | 2114 | (@code{ULONG_MAX} + 1) / 2. |
7a68c94a UD |
2115 | |
2116 | @code{strtoul} sets @var{errno} to @code{EINVAL} if @var{base} is out of | |
2117 | range, or @code{ERANGE} on overflow. | |
2c6fe0bd UD |
2118 | @end deftypefun |
2119 | ||
b642f101 UD |
2120 | @comment wchar.h |
2121 | @comment ISO | |
2122 | @deftypefun {unsigned long int} wcstoul (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2123 | The @code{wcstoul} function is equivalent to the @code{strtoul} function |
2124 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2125 | |
2126 | The @code{wcstoul} function was introduced in @w{Amendment 1} of @w{ISO C90}. | |
2127 | @end deftypefun | |
2128 | ||
2c6fe0bd | 2129 | @comment stdlib.h |
7a68c94a | 2130 | @comment ISO |
b642f101 | 2131 | @deftypefun {long long int} strtoll (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
7a68c94a UD |
2132 | The @code{strtoll} function is like @code{strtol} except that it returns |
2133 | a @code{long long int} value, and accepts numbers with a correspondingly | |
2134 | larger range. | |
2c6fe0bd UD |
2135 | |
2136 | If the string has valid syntax for an integer but the value is not | |
fe7bdd63 | 2137 | representable because of overflow, @code{strtoll} returns either |
2c6fe0bd UD |
2138 | @code{LONG_LONG_MAX} or @code{LONG_LONG_MIN} (@pxref{Range of Type}), as |
2139 | appropriate for the sign of the value. It also sets @code{errno} to | |
2140 | @code{ERANGE} to indicate there was overflow. | |
2c6fe0bd | 2141 | |
ec751a23 | 2142 | The @code{strtoll} function was introduced in @w{ISO C99}. |
2c6fe0bd UD |
2143 | @end deftypefun |
2144 | ||
b642f101 UD |
2145 | @comment wchar.h |
2146 | @comment ISO | |
2147 | @deftypefun {long long int} wcstoll (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2148 | The @code{wcstoll} function is equivalent to the @code{strtoll} function |
2149 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2150 | |
2151 | The @code{wcstoll} function was introduced in @w{Amendment 1} of @w{ISO C90}. | |
2152 | @end deftypefun | |
2153 | ||
2c6fe0bd UD |
2154 | @comment stdlib.h |
2155 | @comment BSD | |
b642f101 | 2156 | @deftypefun {long long int} strtoq (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
7a68c94a | 2157 | @code{strtoq} (``string-to-quad-word'') is the BSD name for @code{strtoll}. |
2c6fe0bd UD |
2158 | @end deftypefun |
2159 | ||
b642f101 UD |
2160 | @comment wchar.h |
2161 | @comment GNU | |
2162 | @deftypefun {long long int} wcstoq (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2163 | The @code{wcstoq} function is equivalent to the @code{strtoq} function |
2164 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2165 | |
2166 | The @code{wcstoq} function is a GNU extension. | |
2167 | @end deftypefun | |
2168 | ||
2c6fe0bd | 2169 | @comment stdlib.h |
7a68c94a | 2170 | @comment ISO |
b642f101 | 2171 | @deftypefun {unsigned long long int} strtoull (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
0e4ee106 UD |
2172 | The @code{strtoull} function is related to @code{strtoll} the same way |
2173 | @code{strtoul} is related to @code{strtol}. | |
fe7bdd63 | 2174 | |
ec751a23 | 2175 | The @code{strtoull} function was introduced in @w{ISO C99}. |
fe7bdd63 UD |
2176 | @end deftypefun |
2177 | ||
b642f101 UD |
2178 | @comment wchar.h |
2179 | @comment ISO | |
2180 | @deftypefun {unsigned long long int} wcstoull (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2181 | The @code{wcstoull} function is equivalent to the @code{strtoull} function |
2182 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2183 | |
2184 | The @code{wcstoull} function was introduced in @w{Amendment 1} of @w{ISO C90}. | |
2185 | @end deftypefun | |
2186 | ||
fe7bdd63 UD |
2187 | @comment stdlib.h |
2188 | @comment BSD | |
b642f101 | 2189 | @deftypefun {unsigned long long int} strtouq (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) |
7a68c94a | 2190 | @code{strtouq} is the BSD name for @code{strtoull}. |
28f540f4 RM |
2191 | @end deftypefun |
2192 | ||
b642f101 UD |
2193 | @comment wchar.h |
2194 | @comment GNU | |
2195 | @deftypefun {unsigned long long int} wcstouq (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2196 | The @code{wcstouq} function is equivalent to the @code{strtouq} function |
2197 | in nearly all aspects but handles wide character strings. | |
b642f101 | 2198 | |
f5708cb0 | 2199 | The @code{wcstouq} function is a GNU extension. |
b642f101 UD |
2200 | @end deftypefun |
2201 | ||
0e4ee106 | 2202 | @comment inttypes.h |
b642f101 UD |
2203 | @comment ISO |
2204 | @deftypefun intmax_t strtoimax (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) | |
0e4ee106 UD |
2205 | The @code{strtoimax} function is like @code{strtol} except that it returns |
2206 | a @code{intmax_t} value, and accepts numbers of a corresponding range. | |
2207 | ||
2208 | If the string has valid syntax for an integer but the value is not | |
2209 | representable because of overflow, @code{strtoimax} returns either | |
2210 | @code{INTMAX_MAX} or @code{INTMAX_MIN} (@pxref{Integers}), as | |
2211 | appropriate for the sign of the value. It also sets @code{errno} to | |
2212 | @code{ERANGE} to indicate there was overflow. | |
2213 | ||
b642f101 UD |
2214 | See @ref{Integers} for a description of the @code{intmax_t} type. The |
2215 | @code{strtoimax} function was introduced in @w{ISO C99}. | |
2216 | @end deftypefun | |
0e4ee106 | 2217 | |
b642f101 UD |
2218 | @comment wchar.h |
2219 | @comment ISO | |
2220 | @deftypefun intmax_t wcstoimax (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2221 | The @code{wcstoimax} function is equivalent to the @code{strtoimax} function |
2222 | in nearly all aspects but handles wide character strings. | |
0e4ee106 | 2223 | |
b642f101 | 2224 | The @code{wcstoimax} function was introduced in @w{ISO C99}. |
0e4ee106 UD |
2225 | @end deftypefun |
2226 | ||
2227 | @comment inttypes.h | |
b642f101 UD |
2228 | @comment ISO |
2229 | @deftypefun uintmax_t strtoumax (const char *restrict @var{string}, char **restrict @var{tailptr}, int @var{base}) | |
0e4ee106 UD |
2230 | The @code{strtoumax} function is related to @code{strtoimax} |
2231 | the same way that @code{strtoul} is related to @code{strtol}. | |
2232 | ||
b642f101 UD |
2233 | See @ref{Integers} for a description of the @code{intmax_t} type. The |
2234 | @code{strtoumax} function was introduced in @w{ISO C99}. | |
2235 | @end deftypefun | |
0e4ee106 | 2236 | |
b642f101 UD |
2237 | @comment wchar.h |
2238 | @comment ISO | |
2239 | @deftypefun uintmax_t wcstoumax (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}, int @var{base}) | |
3554743a AJ |
2240 | The @code{wcstoumax} function is equivalent to the @code{strtoumax} function |
2241 | in nearly all aspects but handles wide character strings. | |
b642f101 UD |
2242 | |
2243 | The @code{wcstoumax} function was introduced in @w{ISO C99}. | |
0e4ee106 UD |
2244 | @end deftypefun |
2245 | ||
28f540f4 | 2246 | @comment stdlib.h |
f65fd747 | 2247 | @comment ISO |
28f540f4 RM |
2248 | @deftypefun {long int} atol (const char *@var{string}) |
2249 | This function is similar to the @code{strtol} function with a @var{base} | |
2250 | argument of @code{10}, except that it need not detect overflow errors. | |
2251 | The @code{atol} function is provided mostly for compatibility with | |
2252 | existing code; using @code{strtol} is more robust. | |
2253 | @end deftypefun | |
2254 | ||
2255 | @comment stdlib.h | |
f65fd747 | 2256 | @comment ISO |
28f540f4 | 2257 | @deftypefun int atoi (const char *@var{string}) |
7a68c94a UD |
2258 | This function is like @code{atol}, except that it returns an @code{int}. |
2259 | The @code{atoi} function is also considered obsolete; use @code{strtol} | |
2260 | instead. | |
28f540f4 RM |
2261 | @end deftypefun |
2262 | ||
fe7bdd63 | 2263 | @comment stdlib.h |
7a68c94a | 2264 | @comment ISO |
fe7bdd63 UD |
2265 | @deftypefun {long long int} atoll (const char *@var{string}) |
2266 | This function is similar to @code{atol}, except it returns a @code{long | |
7a68c94a | 2267 | long int}. |
fe7bdd63 | 2268 | |
ec751a23 | 2269 | The @code{atoll} function was introduced in @w{ISO C99}. It too is |
7a68c94a | 2270 | obsolete (despite having just been added); use @code{strtoll} instead. |
fe7bdd63 UD |
2271 | @end deftypefun |
2272 | ||
b642f101 UD |
2273 | All the functions mentioned in this section so far do not handle |
2274 | alternative representations of characters as described in the locale | |
2275 | data. Some locales specify thousands separator and the way they have to | |
2276 | be used which can help to make large numbers more readable. To read | |
2277 | such numbers one has to use the @code{scanf} functions with the @samp{'} | |
2278 | flag. | |
2c6fe0bd | 2279 | |
28f540f4 RM |
2280 | Here is a function which parses a string as a sequence of integers and |
2281 | returns the sum of them: | |
2282 | ||
2283 | @smallexample | |
2284 | int | |
2285 | sum_ints_from_string (char *string) | |
2286 | @{ | |
2287 | int sum = 0; | |
2288 | ||
2289 | while (1) @{ | |
2290 | char *tail; | |
2291 | int next; | |
2292 | ||
2293 | /* @r{Skip whitespace by hand, to detect the end.} */ | |
2294 | while (isspace (*string)) string++; | |
2295 | if (*string == 0) | |
2296 | break; | |
2297 | ||
2298 | /* @r{There is more nonwhitespace,} */ | |
2299 | /* @r{so it ought to be another number.} */ | |
2300 | errno = 0; | |
2301 | /* @r{Parse it.} */ | |
2302 | next = strtol (string, &tail, 0); | |
2303 | /* @r{Add it in, if not overflow.} */ | |
2304 | if (errno) | |
2305 | printf ("Overflow\n"); | |
2306 | else | |
2307 | sum += next; | |
2308 | /* @r{Advance past it.} */ | |
2309 | string = tail; | |
2310 | @} | |
2311 | ||
2312 | return sum; | |
2313 | @} | |
2314 | @end smallexample | |
2315 | ||
2316 | @node Parsing of Floats | |
2317 | @subsection Parsing of Floats | |
2318 | ||
2319 | @pindex stdlib.h | |
b642f101 UD |
2320 | The @samp{str} functions are declared in @file{stdlib.h} and those |
2321 | beginning with @samp{wcs} are declared in @file{wchar.h}. One might | |
2322 | wonder about the use of @code{restrict} in the prototypes of the | |
2323 | functions in this section. It is seemingly useless but the @w{ISO C} | |
2324 | standard uses it (for the functions defined there) so we have to do it | |
2325 | as well. | |
28f540f4 RM |
2326 | |
2327 | @comment stdlib.h | |
f65fd747 | 2328 | @comment ISO |
b642f101 | 2329 | @deftypefun double strtod (const char *restrict @var{string}, char **restrict @var{tailptr}) |
28f540f4 RM |
2330 | The @code{strtod} (``string-to-double'') function converts the initial |
2331 | part of @var{string} to a floating-point number, which is returned as a | |
b8fe19fa | 2332 | value of type @code{double}. |
28f540f4 RM |
2333 | |
2334 | This function attempts to decompose @var{string} as follows: | |
2335 | ||
2336 | @itemize @bullet | |
b8fe19fa | 2337 | @item |
28f540f4 RM |
2338 | A (possibly empty) sequence of whitespace characters. Which characters |
2339 | are whitespace is determined by the @code{isspace} function | |
2340 | (@pxref{Classification of Characters}). These are discarded. | |
2341 | ||
2342 | @item | |
2343 | An optional plus or minus sign (@samp{+} or @samp{-}). | |
2344 | ||
0c34b1e9 UD |
2345 | @item A floating point number in decimal or hexadecimal format. The |
2346 | decimal format is: | |
2347 | @itemize @minus | |
2348 | ||
28f540f4 RM |
2349 | @item |
2350 | A nonempty sequence of digits optionally containing a decimal-point | |
2351 | character---normally @samp{.}, but it depends on the locale | |
85c165be | 2352 | (@pxref{General Numeric}). |
28f540f4 RM |
2353 | |
2354 | @item | |
2355 | An optional exponent part, consisting of a character @samp{e} or | |
2356 | @samp{E}, an optional sign, and a sequence of digits. | |
2357 | ||
0c34b1e9 UD |
2358 | @end itemize |
2359 | ||
2360 | The hexadecimal format is as follows: | |
2361 | @itemize @minus | |
2362 | ||
2363 | @item | |
2364 | A 0x or 0X followed by a nonempty sequence of hexadecimal digits | |
2365 | optionally containing a decimal-point character---normally @samp{.}, but | |
2366 | it depends on the locale (@pxref{General Numeric}). | |
2367 | ||
2368 | @item | |
2369 | An optional binary-exponent part, consisting of a character @samp{p} or | |
2370 | @samp{P}, an optional sign, and a sequence of digits. | |
2371 | ||
2372 | @end itemize | |
2373 | ||
28f540f4 RM |
2374 | @item |
2375 | Any remaining characters in the string. If @var{tailptr} is not a null | |
2376 | pointer, a pointer to this tail of the string is stored in | |
2377 | @code{*@var{tailptr}}. | |
2378 | @end itemize | |
2379 | ||
2380 | If the string is empty, contains only whitespace, or does not contain an | |
2381 | initial substring that has the expected syntax for a floating-point | |
2382 | number, no conversion is performed. In this case, @code{strtod} returns | |
2383 | a value of zero and the value returned in @code{*@var{tailptr}} is the | |
2384 | value of @var{string}. | |
2385 | ||
26761c28 | 2386 | In a locale other than the standard @code{"C"} or @code{"POSIX"} locales, |
2c6fe0bd | 2387 | this function may recognize additional locale-dependent syntax. |
28f540f4 RM |
2388 | |
2389 | If the string has valid syntax for a floating-point number but the value | |
7a68c94a UD |
2390 | is outside the range of a @code{double}, @code{strtod} will signal |
2391 | overflow or underflow as described in @ref{Math Error Reporting}. | |
2392 | ||
2393 | @code{strtod} recognizes four special input strings. The strings | |
2394 | @code{"inf"} and @code{"infinity"} are converted to @math{@infinity{}}, | |
2395 | or to the largest representable value if the floating-point format | |
2396 | doesn't support infinities. You can prepend a @code{"+"} or @code{"-"} | |
2397 | to specify the sign. Case is ignored when scanning these strings. | |
2398 | ||
95fdc6a0 UD |
2399 | The strings @code{"nan"} and @code{"nan(@var{chars@dots{}})"} are converted |
2400 | to NaN. Again, case is ignored. If @var{chars@dots{}} are provided, they | |
7a68c94a UD |
2401 | are used in some unspecified fashion to select a particular |
2402 | representation of NaN (there can be several). | |
2403 | ||
2404 | Since zero is a valid result as well as the value returned on error, you | |
2405 | should check for errors in the same way as for @code{strtol}, by | |
2406 | examining @var{errno} and @var{tailptr}. | |
28f540f4 RM |
2407 | @end deftypefun |
2408 | ||
2c6fe0bd | 2409 | @comment stdlib.h |
ec751a23 | 2410 | @comment ISO |
2c6fe0bd | 2411 | @deftypefun float strtof (const char *@var{string}, char **@var{tailptr}) |
4260bc74 | 2412 | @comment stdlib.h |
ec751a23 | 2413 | @comment ISO |
7a68c94a UD |
2414 | @deftypefunx {long double} strtold (const char *@var{string}, char **@var{tailptr}) |
2415 | These functions are analogous to @code{strtod}, but return @code{float} | |
2416 | and @code{long double} values respectively. They report errors in the | |
2417 | same way as @code{strtod}. @code{strtof} can be substantially faster | |
2418 | than @code{strtod}, but has less precision; conversely, @code{strtold} | |
2419 | can be much slower but has more precision (on systems where @code{long | |
2420 | double} is a separate type). | |
2421 | ||
ec751a23 | 2422 | These functions have been GNU extensions and are new to @w{ISO C99}. |
2c6fe0bd UD |
2423 | @end deftypefun |
2424 | ||
b642f101 UD |
2425 | @comment wchar.h |
2426 | @comment ISO | |
2427 | @deftypefun double wcstod (const wchar_t *restrict @var{string}, wchar_t **restrict @var{tailptr}) | |
2428 | @comment stdlib.h | |
2429 | @comment ISO | |
2430 | @deftypefunx float wcstof (const wchar_t *@var{string}, wchar_t **@var{tailptr}) | |
2431 | @comment stdlib.h | |
2432 | @comment ISO | |
2433 | @deftypefunx {long double} wcstold (const wchar_t *@var{string}, wchar_t **@var{tailptr}) | |
2434 | The @code{wcstod}, @code{wcstof}, and @code{wcstol} functions are | |
2435 | equivalent in nearly all aspect to the @code{strtod}, @code{strtof}, and | |
2436 | @code{strtold} functions but it handles wide character string. | |
2437 | ||
2438 | The @code{wcstod} function was introduced in @w{Amendment 1} of @w{ISO | |
2439 | C90}. The @code{wcstof} and @code{wcstold} functions were introduced in | |
2440 | @w{ISO C99}. | |
2441 | @end deftypefun | |
2442 | ||
28f540f4 | 2443 | @comment stdlib.h |
f65fd747 | 2444 | @comment ISO |
28f540f4 RM |
2445 | @deftypefun double atof (const char *@var{string}) |
2446 | This function is similar to the @code{strtod} function, except that it | |
2447 | need not detect overflow and underflow errors. The @code{atof} function | |
2448 | is provided mostly for compatibility with existing code; using | |
2449 | @code{strtod} is more robust. | |
2450 | @end deftypefun | |
880f421f | 2451 | |
49c091e5 | 2452 | The GNU C library also provides @samp{_l} versions of these functions, |
7a68c94a UD |
2453 | which take an additional argument, the locale to use in conversion. |
2454 | @xref{Parsing of Integers}. | |
880f421f | 2455 | |
7a68c94a UD |
2456 | @node System V Number Conversion |
2457 | @section Old-fashioned System V number-to-string functions | |
880f421f | 2458 | |
7a68c94a UD |
2459 | The old @w{System V} C library provided three functions to convert |
2460 | numbers to strings, with unusual and hard-to-use semantics. The GNU C | |
2461 | library also provides these functions and some natural extensions. | |
880f421f | 2462 | |
7a68c94a UD |
2463 | These functions are only available in glibc and on systems descended |
2464 | from AT&T Unix. Therefore, unless these functions do precisely what you | |
2465 | need, it is better to use @code{sprintf}, which is standard. | |
880f421f | 2466 | |
7a68c94a | 2467 | All these functions are defined in @file{stdlib.h}. |
880f421f UD |
2468 | |
2469 | @comment stdlib.h | |
2470 | @comment SVID, Unix98 | |
7a68c94a | 2471 | @deftypefun {char *} ecvt (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}) |
880f421f | 2472 | The function @code{ecvt} converts the floating-point number @var{value} |
0ea5db4f UD |
2473 | to a string with at most @var{ndigit} decimal digits. The |
2474 | returned string contains no decimal point or sign. The first digit of | |
2475 | the string is non-zero (unless @var{value} is actually zero) and the | |
2476 | last digit is rounded to nearest. @code{*@var{decpt}} is set to the | |
7a68c94a | 2477 | index in the string of the first digit after the decimal point. |
0ea5db4f UD |
2478 | @code{*@var{neg}} is set to a nonzero value if @var{value} is negative, |
2479 | zero otherwise. | |
880f421f | 2480 | |
67994d6f UD |
2481 | If @var{ndigit} decimal digits would exceed the precision of a |
2482 | @code{double} it is reduced to a system-specific value. | |
2483 | ||
880f421f UD |
2484 | The returned string is statically allocated and overwritten by each call |
2485 | to @code{ecvt}. | |
2486 | ||
0ea5db4f UD |
2487 | If @var{value} is zero, it is implementation defined whether |
2488 | @code{*@var{decpt}} is @code{0} or @code{1}. | |
880f421f | 2489 | |
0ea5db4f UD |
2490 | For example: @code{ecvt (12.3, 5, &d, &n)} returns @code{"12300"} |
2491 | and sets @var{d} to @code{2} and @var{n} to @code{0}. | |
880f421f UD |
2492 | @end deftypefun |
2493 | ||
880f421f UD |
2494 | @comment stdlib.h |
2495 | @comment SVID, Unix98 | |
0ea5db4f | 2496 | @deftypefun {char *} fcvt (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}) |
7a68c94a UD |
2497 | The function @code{fcvt} is like @code{ecvt}, but @var{ndigit} specifies |
2498 | the number of digits after the decimal point. If @var{ndigit} is less | |
2499 | than zero, @var{value} is rounded to the @math{@var{ndigit}+1}'th place to the | |
2500 | left of the decimal point. For example, if @var{ndigit} is @code{-1}, | |
2501 | @var{value} will be rounded to the nearest 10. If @var{ndigit} is | |
2502 | negative and larger than the number of digits to the left of the decimal | |
2503 | point in @var{value}, @var{value} will be rounded to one significant digit. | |
880f421f | 2504 | |
67994d6f UD |
2505 | If @var{ndigit} decimal digits would exceed the precision of a |
2506 | @code{double} it is reduced to a system-specific value. | |
2507 | ||
880f421f UD |
2508 | The returned string is statically allocated and overwritten by each call |
2509 | to @code{fcvt}. | |
880f421f UD |
2510 | @end deftypefun |
2511 | ||
2512 | @comment stdlib.h | |
2513 | @comment SVID, Unix98 | |
2514 | @deftypefun {char *} gcvt (double @var{value}, int @var{ndigit}, char *@var{buf}) | |
7a68c94a UD |
2515 | @code{gcvt} is functionally equivalent to @samp{sprintf(buf, "%*g", |
2516 | ndigit, value}. It is provided only for compatibility's sake. It | |
2517 | returns @var{buf}. | |
67994d6f UD |
2518 | |
2519 | If @var{ndigit} decimal digits would exceed the precision of a | |
2520 | @code{double} it is reduced to a system-specific value. | |
880f421f UD |
2521 | @end deftypefun |
2522 | ||
7a68c94a UD |
2523 | As extensions, the GNU C library provides versions of these three |
2524 | functions that take @code{long double} arguments. | |
880f421f UD |
2525 | |
2526 | @comment stdlib.h | |
2527 | @comment GNU | |
7a68c94a | 2528 | @deftypefun {char *} qecvt (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}) |
67994d6f UD |
2529 | This function is equivalent to @code{ecvt} except that it takes a |
2530 | @code{long double} for the first parameter and that @var{ndigit} is | |
2531 | restricted by the precision of a @code{long double}. | |
880f421f UD |
2532 | @end deftypefun |
2533 | ||
2534 | @comment stdlib.h | |
2535 | @comment GNU | |
0ea5db4f | 2536 | @deftypefun {char *} qfcvt (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}) |
7a68c94a | 2537 | This function is equivalent to @code{fcvt} except that it |
67994d6f UD |
2538 | takes a @code{long double} for the first parameter and that @var{ndigit} is |
2539 | restricted by the precision of a @code{long double}. | |
880f421f UD |
2540 | @end deftypefun |
2541 | ||
2542 | @comment stdlib.h | |
2543 | @comment GNU | |
2544 | @deftypefun {char *} qgcvt (long double @var{value}, int @var{ndigit}, char *@var{buf}) | |
67994d6f UD |
2545 | This function is equivalent to @code{gcvt} except that it takes a |
2546 | @code{long double} for the first parameter and that @var{ndigit} is | |
2547 | restricted by the precision of a @code{long double}. | |
880f421f UD |
2548 | @end deftypefun |
2549 | ||
2550 | ||
2551 | @cindex gcvt_r | |
7a68c94a UD |
2552 | The @code{ecvt} and @code{fcvt} functions, and their @code{long double} |
2553 | equivalents, all return a string located in a static buffer which is | |
2554 | overwritten by the next call to the function. The GNU C library | |
2555 | provides another set of extended functions which write the converted | |
2556 | string into a user-supplied buffer. These have the conventional | |
2557 | @code{_r} suffix. | |
2558 | ||
2559 | @code{gcvt_r} is not necessary, because @code{gcvt} already uses a | |
2560 | user-supplied buffer. | |
880f421f UD |
2561 | |
2562 | @comment stdlib.h | |
2563 | @comment GNU | |
5c1c368f | 2564 | @deftypefun int ecvt_r (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len}) |
7a68c94a UD |
2565 | The @code{ecvt_r} function is the same as @code{ecvt}, except |
2566 | that it places its result into the user-specified buffer pointed to by | |
5c1c368f UD |
2567 | @var{buf}, with length @var{len}. The return value is @code{-1} in |
2568 | case of an error and zero otherwise. | |
880f421f | 2569 | |
7a68c94a | 2570 | This function is a GNU extension. |
880f421f UD |
2571 | @end deftypefun |
2572 | ||
2573 | @comment stdlib.h | |
2574 | @comment SVID, Unix98 | |
5c1c368f UD |
2575 | @deftypefun int fcvt_r (double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len}) |
2576 | The @code{fcvt_r} function is the same as @code{fcvt}, except that it | |
2577 | places its result into the user-specified buffer pointed to by | |
2578 | @var{buf}, with length @var{len}. The return value is @code{-1} in | |
2579 | case of an error and zero otherwise. | |
880f421f | 2580 | |
7a68c94a | 2581 | This function is a GNU extension. |
880f421f UD |
2582 | @end deftypefun |
2583 | ||
2584 | @comment stdlib.h | |
2585 | @comment GNU | |
5c1c368f | 2586 | @deftypefun int qecvt_r (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len}) |
7a68c94a UD |
2587 | The @code{qecvt_r} function is the same as @code{qecvt}, except |
2588 | that it places its result into the user-specified buffer pointed to by | |
5c1c368f UD |
2589 | @var{buf}, with length @var{len}. The return value is @code{-1} in |
2590 | case of an error and zero otherwise. | |
880f421f | 2591 | |
7a68c94a | 2592 | This function is a GNU extension. |
880f421f UD |
2593 | @end deftypefun |
2594 | ||
2595 | @comment stdlib.h | |
2596 | @comment GNU | |
5c1c368f | 2597 | @deftypefun int qfcvt_r (long double @var{value}, int @var{ndigit}, int *@var{decpt}, int *@var{neg}, char *@var{buf}, size_t @var{len}) |
7a68c94a UD |
2598 | The @code{qfcvt_r} function is the same as @code{qfcvt}, except |
2599 | that it places its result into the user-specified buffer pointed to by | |
5c1c368f UD |
2600 | @var{buf}, with length @var{len}. The return value is @code{-1} in |
2601 | case of an error and zero otherwise. | |
880f421f | 2602 | |
7a68c94a | 2603 | This function is a GNU extension. |
880f421f | 2604 | @end deftypefun |