]>
Commit | Line | Data |
---|---|---|
5ce8f203 | 1 | @node Date and Time, Resource Usage And Limitation, Arithmetic, Top |
7a68c94a | 2 | @c %MENU% Functions for getting the date and time and formatting them nicely |
28f540f4 RM |
3 | @chapter Date and Time |
4 | ||
5 | This chapter describes functions for manipulating dates and times, | |
99a20616 UD |
6 | including functions for determining what time it is and conversion |
7 | between different time representations. | |
28f540f4 RM |
8 | |
9 | @menu | |
99a20616 UD |
10 | * Time Basics:: Concepts and definitions. |
11 | * Elapsed Time:: Data types to represent elapsed times | |
12 | * Processor And CPU Time:: Time a program has spent executing. | |
28f540f4 RM |
13 | * Calendar Time:: Manipulation of ``real'' dates and times. |
14 | * Setting an Alarm:: Sending a signal after a specified time. | |
15 | * Sleeping:: Waiting for a period of time. | |
28f540f4 RM |
16 | @end menu |
17 | ||
5ce8f203 | 18 | |
99a20616 UD |
19 | @node Time Basics |
20 | @section Time Basics | |
21 | @cindex time | |
5ce8f203 | 22 | |
99a20616 UD |
23 | Discussing time in a technical manual can be difficult because the word |
24 | ``time'' in English refers to lots of different things. In this manual, | |
25 | we use a rigorous terminology to avoid confusion, and the only thing we | |
26 | use the simple word ``time'' for is to talk about the abstract concept. | |
27 | ||
28 | A @dfn{calendar time} is a point in the time continuum, for example | |
29 | November 4, 1990 at 18:02.5 UTC. Sometimes this is called ``absolute | |
30 | time''. | |
31 | @cindex calendar time | |
32 | ||
33 | We don't speak of a ``date'', because that is inherent in a calendar | |
34 | time. | |
35 | @cindex date | |
36 | ||
37 | An @dfn{interval} is a contiguous part of the time continuum between two | |
38 | calendar times, for example the hour between 9:00 and 10:00 on July 4, | |
39 | 1980. | |
40 | @cindex interval | |
41 | ||
42 | An @dfn{elapsed time} is the length of an interval, for example, 35 | |
43 | minutes. People sometimes sloppily use the word ``interval'' to refer | |
44 | to the elapsed time of some interval. | |
45 | @cindex elapsed time | |
46 | @cindex time, elapsed | |
47 | ||
48 | An @dfn{amount of time} is a sum of elapsed times, which need not be of | |
49 | any specific intervals. For example, the amount of time it takes to | |
50 | read a book might be 9 hours, independently of when and in how many | |
51 | sittings it is read. | |
52 | ||
53 | A @dfn{period} is the elapsed time of an interval between two events, | |
54 | especially when they are part of a sequence of regularly repeating | |
55 | events. | |
56 | @cindex period of time | |
57 | ||
58 | @dfn{CPU time} is like calendar time, except that it is based on the | |
59 | subset of the time continuum when a particular process is actively | |
60 | using a CPU. CPU time is, therefore, relative to a process. | |
28f540f4 | 61 | @cindex CPU time |
99a20616 UD |
62 | |
63 | @dfn{Processor time} is an amount of time that a CPU is in use. In | |
64 | fact, it's a basic system resource, since there's a limit to how much | |
65 | can exist in any given interval (that limit is the elapsed time of the | |
66 | interval times the number of CPUs in the processor). People often call | |
67 | this CPU time, but we reserve the latter term in this manual for the | |
68 | definition above. | |
28f540f4 | 69 | @cindex processor time |
99a20616 UD |
70 | |
71 | @node Elapsed Time | |
72 | @section Elapsed Time | |
73 | @cindex elapsed time | |
74 | ||
75 | One way to represent an elapsed time is with a simple arithmetic data | |
76 | type, as with the following function to compute the elapsed time between | |
77 | two calendar times. This function is declared in @file{time.h}. | |
78 | ||
79 | @comment time.h | |
80 | @comment ISO | |
81 | @deftypefun double difftime (time_t @var{time1}, time_t @var{time0}) | |
82 | The @code{difftime} function returns the number of seconds of elapsed | |
83 | time between calendar time @var{time1} and calendar time @var{time0}, as | |
84 | a value of type @code{double}. The difference ignores leap seconds | |
85 | unless leap second support is enabled. | |
86 | ||
87 | In the GNU system, you can simply subtract @code{time_t} values. But on | |
88 | other systems, the @code{time_t} data type might use some other encoding | |
89 | where subtraction doesn't work directly. | |
90 | @end deftypefun | |
91 | ||
92 | The GNU C library provides two data types specifically for representing | |
93 | an elapsed time. They are used by various GNU C library functions, and | |
94 | you can use them for your own purposes too. They're exactly the same | |
95 | except that one has a resolution in microseconds, and the other, newer | |
96 | one, is in nanoseconds. | |
97 | ||
98 | @comment sys/time.h | |
99 | @comment BSD | |
100 | @deftp {Data Type} {struct timeval} | |
101 | @cindex timeval | |
102 | The @code{struct timeval} structure represents an elapsed time. It is | |
103 | declared in @file{sys/time.h} and has the following members: | |
104 | ||
105 | @table @code | |
106 | @item long int tv_sec | |
107 | This represents the number of whole seconds of elapsed time. | |
108 | ||
109 | @item long int tv_usec | |
110 | This is the rest of the elapsed time (a fraction of a second), | |
111 | represented as the number of microseconds. It is always less than one | |
112 | million. | |
113 | ||
114 | @end table | |
115 | @end deftp | |
116 | ||
117 | @comment sys/time.h | |
118 | @comment POSIX.1 | |
119 | @deftp {Data Type} {struct timespec} | |
120 | @cindex timespec | |
121 | The @code{struct timespec} structure represents an elapsed time. It is | |
122 | declared in @file{time.h} and has the following members: | |
123 | ||
124 | @table @code | |
125 | @item long int tv_sec | |
126 | This represents the number of whole seconds of elapsed time. | |
127 | ||
128 | @item long int tv_nsec | |
129 | This is the rest of the elapsed time (a fraction of a second), | |
130 | represented as the number of nanoseconds. It is always less than one | |
131 | billion. | |
132 | ||
133 | @end table | |
134 | @end deftp | |
135 | ||
136 | It is often necessary to subtract two values of type @w{@code{struct | |
137 | timeval}} or @w{@code{struct timespec}}. Here is the best way to do | |
138 | this. It works even on some peculiar operating systems where the | |
139 | @code{tv_sec} member has an unsigned type. | |
140 | ||
141 | @smallexample | |
142 | /* @r{Subtract the `struct timeval' values X and Y,} | |
143 | @r{storing the result in RESULT.} | |
144 | @r{Return 1 if the difference is negative, otherwise 0.} */ | |
145 | ||
146 | int | |
147 | timeval_subtract (result, x, y) | |
148 | struct timeval *result, *x, *y; | |
149 | @{ | |
150 | /* @r{Perform the carry for the later subtraction by updating @var{y}.} */ | |
151 | if (x->tv_usec < y->tv_usec) @{ | |
152 | int nsec = (y->tv_usec - x->tv_usec) / 1000000 + 1; | |
153 | y->tv_usec -= 1000000 * nsec; | |
154 | y->tv_sec += nsec; | |
155 | @} | |
156 | if (x->tv_usec - y->tv_usec > 1000000) @{ | |
157 | int nsec = (x->tv_usec - y->tv_usec) / 1000000; | |
158 | y->tv_usec += 1000000 * nsec; | |
159 | y->tv_sec -= nsec; | |
160 | @} | |
161 | ||
162 | /* @r{Compute the time remaining to wait.} | |
163 | @r{@code{tv_usec} is certainly positive.} */ | |
164 | result->tv_sec = x->tv_sec - y->tv_sec; | |
165 | result->tv_usec = x->tv_usec - y->tv_usec; | |
166 | ||
167 | /* @r{Return 1 if result is negative.} */ | |
168 | return x->tv_sec < y->tv_sec; | |
169 | @} | |
170 | @end smallexample | |
171 | ||
b642f101 | 172 | Common functions that use @code{struct timeval} are @code{gettimeofday} |
99a20616 UD |
173 | and @code{settimeofday}. |
174 | ||
175 | ||
176 | There are no GNU C library functions specifically oriented toward | |
177 | dealing with elapsed times, but the calendar time, processor time, and | |
178 | alarm and sleeping functions have a lot to do with them. | |
179 | ||
180 | ||
181 | @node Processor And CPU Time | |
182 | @section Processor And CPU Time | |
183 | ||
184 | If you're trying to optimize your program or measure its efficiency, | |
185 | it's very useful to know how much processor time it uses. For that, | |
186 | calendar time and elapsed times are useless because a process may spend | |
187 | time waiting for I/O or for other processes to use the CPU. However, | |
188 | you can get the information with the functions in this section. | |
189 | ||
190 | CPU time (@pxref{Time Basics}) is represented by the data type | |
191 | @code{clock_t}, which is a number of @dfn{clock ticks}. It gives the | |
192 | total amount of time a process has actively used a CPU since some | |
193 | arbitrary event. On the GNU system, that event is the creation of the | |
194 | process. While arbitrary in general, the event is always the same event | |
195 | for any particular process, so you can always measure how much time on | |
32c075e1 | 196 | the CPU a particular computation takes by examinining the process' CPU |
99a20616 UD |
197 | time before and after the computation. |
198 | @cindex CPU time | |
28f540f4 RM |
199 | @cindex clock ticks |
200 | @cindex ticks, clock | |
99a20616 UD |
201 | |
202 | In the GNU system, @code{clock_t} is equivalent to @code{long int} and | |
203 | @code{CLOCKS_PER_SEC} is an integer value. But in other systems, both | |
204 | @code{clock_t} and the macro @code{CLOCKS_PER_SEC} can be either integer | |
205 | or floating-point types. Casting CPU time values to @code{double}, as | |
206 | in the example above, makes sure that operations such as arithmetic and | |
207 | printing work properly and consistently no matter what the underlying | |
208 | representation is. | |
209 | ||
210 | Note that the clock can wrap around. On a 32bit system with | |
211 | @code{CLOCKS_PER_SEC} set to one million this function will return the | |
212 | same value approximately every 72 minutes. | |
213 | ||
214 | For additional functions to examine a process' use of processor time, | |
215 | and to control it, @xref{Resource Usage And Limitation}. | |
216 | ||
28f540f4 RM |
217 | |
218 | @menu | |
99a20616 UD |
219 | * CPU Time:: The @code{clock} function. |
220 | * Processor Time:: The @code{times} function. | |
28f540f4 RM |
221 | @end menu |
222 | ||
99a20616 UD |
223 | @node CPU Time |
224 | @subsection CPU Time Inquiry | |
28f540f4 | 225 | |
99a20616 UD |
226 | To get a process' CPU time, you can use the @code{clock} function. This |
227 | facility is declared in the header file @file{time.h}. | |
28f540f4 RM |
228 | @pindex time.h |
229 | ||
99a20616 UD |
230 | In typical usage, you call the @code{clock} function at the beginning |
231 | and end of the interval you want to time, subtract the values, and then | |
232 | divide by @code{CLOCKS_PER_SEC} (the number of clock ticks per second) | |
233 | to get processor time, like this: | |
28f540f4 RM |
234 | |
235 | @smallexample | |
236 | @group | |
237 | #include <time.h> | |
238 | ||
239 | clock_t start, end; | |
99a20616 | 240 | double cpu_time_used; |
28f540f4 RM |
241 | |
242 | start = clock(); | |
243 | @dots{} /* @r{Do the work.} */ | |
244 | end = clock(); | |
99a20616 | 245 | cpu_time_used = ((double) (end - start)) / CLOCKS_PER_SEC; |
28f540f4 RM |
246 | @end group |
247 | @end smallexample | |
248 | ||
99a20616 UD |
249 | Do not use a single CPU time as an amount of time; it doesn't work that |
250 | way. Either do a subtraction as shown above or query processor time | |
251 | directly. @xref{Processor Time}. | |
252 | ||
28f540f4 | 253 | Different computers and operating systems vary wildly in how they keep |
99a20616 UD |
254 | track of CPU time. It's common for the internal processor clock |
255 | to have a resolution somewhere between a hundredth and millionth of a | |
28f540f4 RM |
256 | second. |
257 | ||
28f540f4 | 258 | @comment time.h |
f65fd747 | 259 | @comment ISO |
28f540f4 RM |
260 | @deftypevr Macro int CLOCKS_PER_SEC |
261 | The value of this macro is the number of clock ticks per second measured | |
99a20616 | 262 | by the @code{clock} function. POSIX requires that this value be one |
ce33f7be | 263 | million independent of the actual resolution. |
28f540f4 RM |
264 | @end deftypevr |
265 | ||
266 | @comment time.h | |
267 | @comment POSIX.1 | |
268 | @deftypevr Macro int CLK_TCK | |
01cdeca0 | 269 | This is an obsolete name for @code{CLOCKS_PER_SEC}. |
28f540f4 RM |
270 | @end deftypevr |
271 | ||
272 | @comment time.h | |
f65fd747 | 273 | @comment ISO |
28f540f4 RM |
274 | @deftp {Data Type} clock_t |
275 | This is the type of the value returned by the @code{clock} function. | |
99a20616 | 276 | Values of type @code{clock_t} are numbers of clock ticks. |
28f540f4 RM |
277 | @end deftp |
278 | ||
279 | @comment time.h | |
f65fd747 | 280 | @comment ISO |
28f540f4 | 281 | @deftypefun clock_t clock (void) |
99a20616 | 282 | This function returns the calling process' current CPU time. If the CPU |
28f540f4 RM |
283 | time is not available or cannot be represented, @code{clock} returns the |
284 | value @code{(clock_t)(-1)}. | |
285 | @end deftypefun | |
286 | ||
287 | ||
99a20616 UD |
288 | @node Processor Time |
289 | @subsection Processor Time Inquiry | |
28f540f4 | 290 | |
99a20616 UD |
291 | The @code{times} function returns information about a process' |
292 | consumption of processor time in a @w{@code{struct tms}} object, in | |
293 | addition to the process' CPU time. @xref{Time Basics}. You should | |
28f540f4 | 294 | include the header file @file{sys/times.h} to use this facility. |
99a20616 UD |
295 | @cindex processor time |
296 | @cindex CPU time | |
28f540f4 RM |
297 | @pindex sys/times.h |
298 | ||
299 | @comment sys/times.h | |
300 | @comment POSIX.1 | |
301 | @deftp {Data Type} {struct tms} | |
302 | The @code{tms} structure is used to return information about process | |
303 | times. It contains at least the following members: | |
304 | ||
305 | @table @code | |
306 | @item clock_t tms_utime | |
99a20616 UD |
307 | This is the total processor time the calling process has used in |
308 | executing the instructions of its program. | |
28f540f4 RM |
309 | |
310 | @item clock_t tms_stime | |
99a20616 UD |
311 | This is the processor time the system has used on behalf of the calling |
312 | process. | |
28f540f4 RM |
313 | |
314 | @item clock_t tms_cutime | |
315 | This is the sum of the @code{tms_utime} values and the @code{tms_cutime} | |
316 | values of all terminated child processes of the calling process, whose | |
317 | status has been reported to the parent process by @code{wait} or | |
318 | @code{waitpid}; see @ref{Process Completion}. In other words, it | |
99a20616 UD |
319 | represents the total processor time used in executing the instructions |
320 | of all the terminated child processes of the calling process, excluding | |
321 | child processes which have not yet been reported by @code{wait} or | |
28f540f4 | 322 | @code{waitpid}. |
99a20616 | 323 | @cindex child process |
28f540f4 RM |
324 | |
325 | @item clock_t tms_cstime | |
99a20616 UD |
326 | This is similar to @code{tms_cutime}, but represents the total processor |
327 | time system has used on behalf of all the terminated child processes | |
328 | of the calling process. | |
28f540f4 RM |
329 | @end table |
330 | ||
99a20616 UD |
331 | All of the times are given in numbers of clock ticks. Unlike CPU time, |
332 | these are the actual amounts of time; not relative to any event. | |
333 | @xref{Creating a Process}. | |
28f540f4 RM |
334 | @end deftp |
335 | ||
336 | @comment sys/times.h | |
337 | @comment POSIX.1 | |
338 | @deftypefun clock_t times (struct tms *@var{buffer}) | |
339 | The @code{times} function stores the processor time information for | |
340 | the calling process in @var{buffer}. | |
341 | ||
99a20616 UD |
342 | The return value is the calling process' CPU time (the same value you |
343 | get from @code{clock()}. @code{times} returns @code{(clock_t)(-1)} to | |
344 | indicate failure. | |
28f540f4 RM |
345 | @end deftypefun |
346 | ||
347 | @strong{Portability Note:} The @code{clock} function described in | |
99a20616 | 348 | @ref{CPU Time} is specified by the @w{ISO C} standard. The |
28f540f4 | 349 | @code{times} function is a feature of POSIX.1. In the GNU system, the |
99a20616 UD |
350 | CPU time is defined to be equivalent to the sum of the @code{tms_utime} |
351 | and @code{tms_stime} fields returned by @code{times}. | |
28f540f4 RM |
352 | |
353 | @node Calendar Time | |
354 | @section Calendar Time | |
355 | ||
99a20616 UD |
356 | This section describes facilities for keeping track of calendar time. |
357 | @xref{Time Basics}. | |
28f540f4 | 358 | |
3566d33c | 359 | The GNU C library represents calendar time three ways: |
28f540f4 RM |
360 | |
361 | @itemize @bullet | |
01cdeca0 | 362 | @item |
3566d33c | 363 | @dfn{Simple time} (the @code{time_t} data type) is a compact |
99a20616 UD |
364 | representation, typically giving the number of seconds of elapsed time |
365 | since some implementation-specific base time. | |
3566d33c | 366 | @cindex simple time |
28f540f4 RM |
367 | |
368 | @item | |
99a20616 UD |
369 | There is also a "high-resolution time" representation. Like simple |
370 | time, this represents a calendar time as an elapsed time since a base | |
371 | time, but instead of measuring in whole seconds, it uses a @code{struct | |
372 | timeval} data type, which includes fractions of a second. Use this time | |
373 | representation instead of simple time when you need greater precision. | |
28f540f4 RM |
374 | @cindex high-resolution time |
375 | ||
376 | @item | |
3566d33c | 377 | @dfn{Local time} or @dfn{broken-down time} (the @code{struct tm} data |
99a20616 | 378 | type) represents a calendar time as a set of components specifying the |
3566d33c | 379 | year, month, and so on in the Gregorian calendar, for a specific time |
99a20616 UD |
380 | zone. This calendar time representation is usually used only to |
381 | communicate with people. | |
28f540f4 RM |
382 | @cindex local time |
383 | @cindex broken-down time | |
3566d33c UD |
384 | @cindex Gregorian calendar |
385 | @cindex calendar, Gregorian | |
28f540f4 RM |
386 | @end itemize |
387 | ||
388 | @menu | |
389 | * Simple Calendar Time:: Facilities for manipulating calendar time. | |
390 | * High-Resolution Calendar:: A time representation with greater precision. | |
391 | * Broken-down Time:: Facilities for manipulating local time. | |
3566d33c | 392 | * High Accuracy Clock:: Maintaining a high accuracy system clock. |
99a20616 | 393 | * Formatting Calendar Time:: Converting times to strings. |
e9dcb080 UD |
394 | * Parsing Date and Time:: Convert textual time and date information back |
395 | into broken-down time values. | |
28f540f4 | 396 | * TZ Variable:: How users specify the time zone. |
01cdeca0 | 397 | * Time Zone Functions:: Functions to examine or specify the time zone. |
28f540f4 RM |
398 | * Time Functions Example:: An example program showing use of some of |
399 | the time functions. | |
400 | @end menu | |
401 | ||
402 | @node Simple Calendar Time | |
403 | @subsection Simple Calendar Time | |
404 | ||
3566d33c UD |
405 | This section describes the @code{time_t} data type for representing calendar |
406 | time as simple time, and the functions which operate on simple time objects. | |
28f540f4 RM |
407 | These facilities are declared in the header file @file{time.h}. |
408 | @pindex time.h | |
409 | ||
410 | @cindex epoch | |
411 | @comment time.h | |
f65fd747 | 412 | @comment ISO |
28f540f4 | 413 | @deftp {Data Type} time_t |
3566d33c | 414 | This is the data type used to represent simple time. Sometimes, it also |
99a20616 UD |
415 | represents an elapsed time. When interpreted as a calendar time value, |
416 | it represents the number of seconds elapsed since 00:00:00 on January 1, | |
417 | 1970, Coordinated Universal Time. (This calendar time is sometimes | |
418 | referred to as the @dfn{epoch}.) POSIX requires that this count not | |
419 | include leap seconds, but on some systems this count includes leap seconds | |
60092701 | 420 | if you set @code{TZ} to certain values (@pxref{TZ Variable}). |
28f540f4 | 421 | |
99a20616 UD |
422 | Note that a simple time has no concept of local time zone. Calendar |
423 | Time @var{T} is the same instant in time regardless of where on the | |
424 | globe the computer is. | |
3566d33c | 425 | |
60092701 | 426 | In the GNU C library, @code{time_t} is equivalent to @code{long int}. |
28f540f4 RM |
427 | In other systems, @code{time_t} might be either an integer or |
428 | floating-point type. | |
429 | @end deftp | |
430 | ||
99a20616 UD |
431 | The function @code{difftime} tells you the elapsed time between two |
432 | simple calendar times, which is not always as easy to compute as just | |
433 | subtracting. @xref{Elapsed Time}. | |
28f540f4 RM |
434 | |
435 | @comment time.h | |
f65fd747 | 436 | @comment ISO |
28f540f4 | 437 | @deftypefun time_t time (time_t *@var{result}) |
99a20616 UD |
438 | The @code{time} function returns the current calendar time as a value of |
439 | type @code{time_t}. If the argument @var{result} is not a null pointer, | |
440 | the calendar time value is also stored in @code{*@var{result}}. If the | |
441 | current calendar time is not available, the value | |
442 | @w{@code{(time_t)(-1)}} is returned. | |
28f540f4 RM |
443 | @end deftypefun |
444 | ||
3566d33c UD |
445 | @c The GNU C library implements stime() with a call to settimeofday() on |
446 | @c Linux. | |
447 | @comment time.h | |
448 | @comment SVID, XPG | |
449 | @deftypefun int stime (time_t *@var{newtime}) | |
32c075e1 | 450 | @code{stime} sets the system clock, i.e. it tells the system that the |
99a20616 | 451 | current calendar time is @var{newtime}, where @code{newtime} is |
3566d33c UD |
452 | interpreted as described in the above definition of @code{time_t}. |
453 | ||
454 | @code{settimeofday} is a newer function which sets the system clock to | |
455 | better than one second precision. @code{settimeofday} is generally a | |
456 | better choice than @code{stime}. @xref{High-Resolution Calendar}. | |
457 | ||
458 | Only the superuser can set the system clock. | |
459 | ||
460 | If the function succeeds, the return value is zero. Otherwise, it is | |
461 | @code{-1} and @code{errno} is set accordingly: | |
462 | ||
463 | @table @code | |
464 | @item EPERM | |
465 | The process is not superuser. | |
466 | @end table | |
467 | @end deftypefun | |
468 | ||
469 | ||
28f540f4 RM |
470 | |
471 | @node High-Resolution Calendar | |
472 | @subsection High-Resolution Calendar | |
473 | ||
3566d33c | 474 | The @code{time_t} data type used to represent simple times has a |
28f540f4 RM |
475 | resolution of only one second. Some applications need more precision. |
476 | ||
477 | So, the GNU C library also contains functions which are capable of | |
478 | representing calendar times to a higher resolution than one second. The | |
479 | functions and the associated data types described in this section are | |
480 | declared in @file{sys/time.h}. | |
481 | @pindex sys/time.h | |
482 | ||
28f540f4 RM |
483 | @comment sys/time.h |
484 | @comment BSD | |
485 | @deftp {Data Type} {struct timezone} | |
486 | The @code{struct timezone} structure is used to hold minimal information | |
487 | about the local time zone. It has the following members: | |
488 | ||
489 | @table @code | |
490 | @item int tz_minuteswest | |
f0f1bf85 | 491 | This is the number of minutes west of UTC. |
28f540f4 RM |
492 | |
493 | @item int tz_dsttime | |
76c23bac | 494 | If nonzero, Daylight Saving Time applies during some part of the year. |
28f540f4 RM |
495 | @end table |
496 | ||
497 | The @code{struct timezone} type is obsolete and should never be used. | |
498 | Instead, use the facilities described in @ref{Time Zone Functions}. | |
499 | @end deftp | |
500 | ||
28f540f4 RM |
501 | @comment sys/time.h |
502 | @comment BSD | |
503 | @deftypefun int gettimeofday (struct timeval *@var{tp}, struct timezone *@var{tzp}) | |
99a20616 UD |
504 | The @code{gettimeofday} function returns the current calendar time as |
505 | the elapsed time since the epoch in the @code{struct timeval} structure | |
506 | indicated by @var{tp}. (@pxref{Elapsed Time} for a description of | |
0cb71e02 | 507 | @code{struct timeval}). Information about the time zone is returned in |
99a20616 UD |
508 | the structure pointed at @var{tzp}. If the @var{tzp} argument is a null |
509 | pointer, time zone information is ignored. | |
28f540f4 RM |
510 | |
511 | The return value is @code{0} on success and @code{-1} on failure. The | |
512 | following @code{errno} error condition is defined for this function: | |
513 | ||
514 | @table @code | |
515 | @item ENOSYS | |
516 | The operating system does not support getting time zone information, and | |
517 | @var{tzp} is not a null pointer. The GNU operating system does not | |
518 | support using @w{@code{struct timezone}} to represent time zone | |
519 | information; that is an obsolete feature of 4.3 BSD. | |
520 | Instead, use the facilities described in @ref{Time Zone Functions}. | |
521 | @end table | |
522 | @end deftypefun | |
523 | ||
524 | @comment sys/time.h | |
525 | @comment BSD | |
526 | @deftypefun int settimeofday (const struct timeval *@var{tp}, const struct timezone *@var{tzp}) | |
99a20616 UD |
527 | The @code{settimeofday} function sets the current calendar time in the |
528 | system clock according to the arguments. As for @code{gettimeofday}, | |
529 | the calendar time is represented as the elapsed time since the epoch. | |
530 | As for @code{gettimeofday}, time zone information is ignored if | |
531 | @var{tzp} is a null pointer. | |
28f540f4 RM |
532 | |
533 | You must be a privileged user in order to use @code{settimeofday}. | |
534 | ||
3566d33c UD |
535 | Some kernels automatically set the system clock from some source such as |
536 | a hardware clock when they start up. Others, including Linux, place the | |
99a20616 | 537 | system clock in an ``invalid'' state (in which attempts to read the clock |
3566d33c UD |
538 | fail). A call of @code{stime} removes the system clock from an invalid |
539 | state, and system startup scripts typically run a program that calls | |
540 | @code{stime}. | |
541 | ||
542 | @code{settimeofday} causes a sudden jump forwards or backwards, which | |
543 | can cause a variety of problems in a system. Use @code{adjtime} (below) | |
544 | to make a smooth transition from one time to another by temporarily | |
545 | speeding up or slowing down the clock. | |
546 | ||
547 | With a Linux kernel, @code{adjtimex} does the same thing and can also | |
548 | make permanent changes to the speed of the system clock so it doesn't | |
549 | need to be corrected as often. | |
550 | ||
28f540f4 RM |
551 | The return value is @code{0} on success and @code{-1} on failure. The |
552 | following @code{errno} error conditions are defined for this function: | |
553 | ||
554 | @table @code | |
555 | @item EPERM | |
99a20616 | 556 | This process cannot set the clock because it is not privileged. |
28f540f4 RM |
557 | |
558 | @item ENOSYS | |
559 | The operating system does not support setting time zone information, and | |
560 | @var{tzp} is not a null pointer. | |
561 | @end table | |
562 | @end deftypefun | |
563 | ||
3566d33c | 564 | @c On Linux, GNU libc implements adjtime() as a call to adjtimex(). |
28f540f4 RM |
565 | @comment sys/time.h |
566 | @comment BSD | |
567 | @deftypefun int adjtime (const struct timeval *@var{delta}, struct timeval *@var{olddelta}) | |
568 | This function speeds up or slows down the system clock in order to make | |
99a20616 UD |
569 | a gradual adjustment. This ensures that the calendar time reported by |
570 | the system clock is always monotonically increasing, which might not | |
571 | happen if you simply set the clock. | |
28f540f4 RM |
572 | |
573 | The @var{delta} argument specifies a relative adjustment to be made to | |
99a20616 UD |
574 | the clock time. If negative, the system clock is slowed down for a |
575 | while until it has lost this much elapsed time. If positive, the system | |
576 | clock is speeded up for a while. | |
28f540f4 RM |
577 | |
578 | If the @var{olddelta} argument is not a null pointer, the @code{adjtime} | |
579 | function returns information about any previous time adjustment that | |
580 | has not yet completed. | |
581 | ||
582 | This function is typically used to synchronize the clocks of computers | |
583 | in a local network. You must be a privileged user to use it. | |
3566d33c UD |
584 | |
585 | With a Linux kernel, you can use the @code{adjtimex} function to | |
586 | permanently change the clock speed. | |
587 | ||
28f540f4 RM |
588 | The return value is @code{0} on success and @code{-1} on failure. The |
589 | following @code{errno} error condition is defined for this function: | |
590 | ||
591 | @table @code | |
592 | @item EPERM | |
593 | You do not have privilege to set the time. | |
594 | @end table | |
595 | @end deftypefun | |
596 | ||
597 | @strong{Portability Note:} The @code{gettimeofday}, @code{settimeofday}, | |
01cdeca0 | 598 | and @code{adjtime} functions are derived from BSD. |
28f540f4 RM |
599 | |
600 | ||
3566d33c UD |
601 | Symbols for the following function are declared in @file{sys/timex.h}. |
602 | ||
603 | @comment sys/timex.h | |
604 | @comment GNU | |
605 | @deftypefun int adjtimex (struct timex *@var{timex}) | |
606 | ||
607 | @code{adjtimex} is functionally identical to @code{ntp_adjtime}. | |
608 | @xref{High Accuracy Clock}. | |
609 | ||
610 | This function is present only with a Linux kernel. | |
611 | ||
612 | @end deftypefun | |
613 | ||
28f540f4 RM |
614 | @node Broken-down Time |
615 | @subsection Broken-down Time | |
616 | @cindex broken-down time | |
617 | @cindex calendar time and broken-down time | |
618 | ||
99a20616 UD |
619 | Calendar time is represented by the usual GNU C library functions as an |
620 | elapsed time since a fixed base calendar time. This is convenient for | |
621 | computation, but has no relation to the way people normally think of | |
622 | calendar time. By contrast, @dfn{broken-down time} is a binary | |
623 | representation of calendar time separated into year, month, day, and so | |
624 | on. Broken-down time values are not useful for calculations, but they | |
625 | are useful for printing human readable time information. | |
28f540f4 | 626 | |
3566d33c | 627 | A broken-down time value is always relative to a choice of time |
99a20616 | 628 | zone, and it also indicates which time zone that is. |
28f540f4 RM |
629 | |
630 | The symbols in this section are declared in the header file @file{time.h}. | |
631 | ||
632 | @comment time.h | |
f65fd747 | 633 | @comment ISO |
28f540f4 RM |
634 | @deftp {Data Type} {struct tm} |
635 | This is the data type used to represent a broken-down time. The structure | |
99a20616 | 636 | contains at least the following members, which can appear in any order. |
28f540f4 RM |
637 | |
638 | @table @code | |
639 | @item int tm_sec | |
99a20616 UD |
640 | This is the number of full seconds since the top of the minute (normally |
641 | in the range @code{0} through @code{59}, but the actual upper limit is | |
642 | @code{60}, to allow for leap seconds if leap second support is | |
643 | available). | |
28f540f4 RM |
644 | @cindex leap second |
645 | ||
646 | @item int tm_min | |
99a20616 UD |
647 | This is the number of full minutes since the top of the hour (in the |
648 | range @code{0} through @code{59}). | |
28f540f4 RM |
649 | |
650 | @item int tm_hour | |
99a20616 UD |
651 | This is the number of full hours past midnight (in the range @code{0} through |
652 | @code{23}). | |
28f540f4 RM |
653 | |
654 | @item int tm_mday | |
99a20616 UD |
655 | This is the ordinal day of the month (in the range @code{1} through @code{31}). |
656 | Watch out for this one! As the only ordinal number in the structure, it is | |
657 | inconsistent with the rest of the structure. | |
28f540f4 RM |
658 | |
659 | @item int tm_mon | |
99a20616 UD |
660 | This is the number of full calendar months since the beginning of the |
661 | year (in the range @code{0} through @code{11}). Watch out for this one! | |
662 | People usually use ordinal numbers for month-of-year (where January = 1). | |
28f540f4 RM |
663 | |
664 | @item int tm_year | |
99a20616 | 665 | This is the number of full calendar years since 1900. |
28f540f4 RM |
666 | |
667 | @item int tm_wday | |
99a20616 UD |
668 | This is the number of full days since Sunday (in the range @code{0} through |
669 | @code{6}). | |
28f540f4 RM |
670 | |
671 | @item int tm_yday | |
99a20616 UD |
672 | This is the number of full days since the beginning of the year (in the |
673 | range @code{0} through @code{365}). | |
28f540f4 RM |
674 | |
675 | @item int tm_isdst | |
676 | @cindex Daylight Saving Time | |
677 | @cindex summer time | |
678 | This is a flag that indicates whether Daylight Saving Time is (or was, or | |
679 | will be) in effect at the time described. The value is positive if | |
680 | Daylight Saving Time is in effect, zero if it is not, and negative if the | |
681 | information is not available. | |
682 | ||
683 | @item long int tm_gmtoff | |
684 | This field describes the time zone that was used to compute this | |
f0f1bf85 UD |
685 | broken-down time value, including any adjustment for daylight saving; it |
686 | is the number of seconds that you must add to UTC to get local time. | |
687 | You can also think of this as the number of seconds east of UTC. For | |
688 | example, for U.S. Eastern Standard Time, the value is @code{-5*60*60}. | |
689 | The @code{tm_gmtoff} field is derived from BSD and is a GNU library | |
f65fd747 | 690 | extension; it is not visible in a strict @w{ISO C} environment. |
28f540f4 RM |
691 | |
692 | @item const char *tm_zone | |
f0f1bf85 UD |
693 | This field is the name for the time zone that was used to compute this |
694 | broken-down time value. Like @code{tm_gmtoff}, this field is a BSD and | |
f65fd747 | 695 | GNU extension, and is not visible in a strict @w{ISO C} environment. |
28f540f4 RM |
696 | @end table |
697 | @end deftp | |
698 | ||
99a20616 | 699 | |
28f540f4 | 700 | @comment time.h |
f65fd747 | 701 | @comment ISO |
28f540f4 | 702 | @deftypefun {struct tm *} localtime (const time_t *@var{time}) |
3566d33c | 703 | The @code{localtime} function converts the simple time pointed to by |
28f540f4 RM |
704 | @var{time} to broken-down time representation, expressed relative to the |
705 | user's specified time zone. | |
706 | ||
707 | The return value is a pointer to a static broken-down time structure, which | |
60092701 RM |
708 | might be overwritten by subsequent calls to @code{ctime}, @code{gmtime}, |
709 | or @code{localtime}. (But no other library function overwrites the contents | |
710 | of this object.) | |
28f540f4 | 711 | |
fe0ec73e UD |
712 | The return value is the null pointer if @var{time} cannot be represented |
713 | as a broken-down time; typically this is because the year cannot fit into | |
714 | an @code{int}. | |
715 | ||
28f540f4 RM |
716 | Calling @code{localtime} has one other effect: it sets the variable |
717 | @code{tzname} with information about the current time zone. @xref{Time | |
718 | Zone Functions}. | |
719 | @end deftypefun | |
720 | ||
0413b54c UD |
721 | Using the @code{localtime} function is a big problem in multi-threaded |
722 | programs. The result is returned in a static buffer and this is used in | |
76c23bac | 723 | all threads. POSIX.1c introduced a variant of this function. |
0413b54c UD |
724 | |
725 | @comment time.h | |
726 | @comment POSIX.1c | |
727 | @deftypefun {struct tm *} localtime_r (const time_t *@var{time}, struct tm *@var{resultp}) | |
728 | The @code{localtime_r} function works just like the @code{localtime} | |
3566d33c | 729 | function. It takes a pointer to a variable containing a simple time |
0413b54c UD |
730 | and converts it to the broken-down time format. |
731 | ||
732 | But the result is not placed in a static buffer. Instead it is placed | |
733 | in the object of type @code{struct tm} to which the parameter | |
734 | @var{resultp} points. | |
735 | ||
736 | If the conversion is successful the function returns a pointer to the | |
737 | object the result was written into, i.e., it returns @var{resultp}. | |
738 | @end deftypefun | |
739 | ||
740 | ||
28f540f4 | 741 | @comment time.h |
f65fd747 | 742 | @comment ISO |
28f540f4 RM |
743 | @deftypefun {struct tm *} gmtime (const time_t *@var{time}) |
744 | This function is similar to @code{localtime}, except that the broken-down | |
3566d33c UD |
745 | time is expressed as Coordinated Universal Time (UTC) (formerly called |
746 | Greenwich Mean Time (GMT)) rather than relative to a local time zone. | |
28f540f4 | 747 | |
28f540f4 RM |
748 | @end deftypefun |
749 | ||
0413b54c | 750 | As for the @code{localtime} function we have the problem that the result |
f2ea0f5b | 751 | is placed in a static variable. POSIX.1c also provides a replacement for |
0413b54c UD |
752 | @code{gmtime}. |
753 | ||
754 | @comment time.h | |
755 | @comment POSIX.1c | |
756 | @deftypefun {struct tm *} gmtime_r (const time_t *@var{time}, struct tm *@var{resultp}) | |
757 | This function is similar to @code{localtime_r}, except that it converts | |
758 | just like @code{gmtime} the given time as Coordinated Universal Time. | |
759 | ||
760 | If the conversion is successful the function returns a pointer to the | |
761 | object the result was written into, i.e., it returns @var{resultp}. | |
762 | @end deftypefun | |
763 | ||
764 | ||
28f540f4 | 765 | @comment time.h |
f65fd747 | 766 | @comment ISO |
28f540f4 RM |
767 | @deftypefun time_t mktime (struct tm *@var{brokentime}) |
768 | The @code{mktime} function is used to convert a broken-down time structure | |
3566d33c | 769 | to a simple time representation. It also ``normalizes'' the contents of |
28f540f4 RM |
770 | the broken-down time structure, by filling in the day of week and day of |
771 | year based on the other date and time components. | |
772 | ||
773 | The @code{mktime} function ignores the specified contents of the | |
774 | @code{tm_wday} and @code{tm_yday} members of the broken-down time | |
3566d33c | 775 | structure. It uses the values of the other components to determine the |
28f540f4 | 776 | calendar time; it's permissible for these components to have |
76c23bac | 777 | unnormalized values outside their normal ranges. The last thing that |
28f540f4 RM |
778 | @code{mktime} does is adjust the components of the @var{brokentime} |
779 | structure (including the @code{tm_wday} and @code{tm_yday}). | |
780 | ||
3566d33c | 781 | If the specified broken-down time cannot be represented as a simple time, |
28f540f4 RM |
782 | @code{mktime} returns a value of @code{(time_t)(-1)} and does not modify |
783 | the contents of @var{brokentime}. | |
784 | ||
785 | Calling @code{mktime} also sets the variable @code{tzname} with | |
786 | information about the current time zone. @xref{Time Zone Functions}. | |
787 | @end deftypefun | |
788 | ||
3566d33c UD |
789 | @comment time.h |
790 | @comment ??? | |
791 | @deftypefun time_t timelocal (struct tm *@var{brokentime}) | |
792 | ||
793 | @code{timelocal} is functionally identical to @code{mktime}, but more | |
794 | mnemonically named. Note that it is the inverse of the @code{localtime} | |
795 | function. | |
796 | ||
797 | @strong{Portability note:} @code{mktime} is essentially universally | |
798 | available. @code{timelocal} is rather rare. | |
799 | ||
800 | @end deftypefun | |
801 | ||
802 | @comment time.h | |
803 | @comment ??? | |
804 | @deftypefun time_t timegm (struct tm *@var{brokentime}) | |
805 | ||
806 | @code{timegm} is functionally identical to @code{mktime} except it | |
807 | always takes the input values to be Coordinated Universal Time (UTC) | |
808 | regardless of any local time zone setting. | |
809 | ||
810 | Note that @code{timegm} is the inverse of @code{gmtime}. | |
811 | ||
812 | @strong{Portability note:} @code{mktime} is essentially universally | |
813 | available. @code{timegm} is rather rare. For the most portable | |
814 | conversion from a UTC broken-down time to a simple time, set | |
815 | the @code{TZ} environment variable to UTC, call @code{mktime}, then set | |
816 | @code{TZ} back. | |
817 | ||
818 | @end deftypefun | |
819 | ||
820 | ||
821 | ||
822 | @node High Accuracy Clock | |
823 | @subsection High Accuracy Clock | |
824 | ||
825 | @cindex time, high precision | |
826 | @cindex clock, high accuracy | |
827 | @pindex sys/timex.h | |
828 | @c On Linux, GNU libc implements ntp_gettime() and npt_adjtime() as calls | |
829 | @c to adjtimex(). | |
830 | The @code{ntp_gettime} and @code{ntp_adjtime} functions provide an | |
831 | interface to monitor and manipulate the system clock to maintain high | |
832 | accuracy time. For example, you can fine tune the speed of the clock | |
833 | or synchronize it with another time source. | |
834 | ||
835 | A typical use of these functions is by a server implementing the Network | |
836 | Time Protocol to synchronize the clocks of multiple systems and high | |
837 | precision clocks. | |
838 | ||
839 | These functions are declared in @file{sys/timex.h}. | |
840 | ||
841 | @tindex struct ntptimeval | |
842 | @deftp {Data Type} {struct ntptimeval} | |
99a20616 UD |
843 | This structure is used for information about the system clock. It |
844 | contains the following members: | |
3566d33c UD |
845 | @table @code |
846 | @item struct timeval time | |
99a20616 UD |
847 | This is the current calendar time, expressed as the elapsed time since |
848 | the epoch. The @code{struct timeval} data type is described in | |
849 | @ref{Elapsed Time}. | |
3566d33c UD |
850 | |
851 | @item long int maxerror | |
852 | This is the maximum error, measured in microseconds. Unless updated | |
853 | via @code{ntp_adjtime} periodically, this value will reach some | |
854 | platform-specific maximum value. | |
855 | ||
856 | @item long int esterror | |
857 | This is the estimated error, measured in microseconds. This value can | |
858 | be set by @code{ntp_adjtime} to indicate the estimated offset of the | |
99a20616 | 859 | system clock from the true calendar time. |
3566d33c UD |
860 | @end table |
861 | @end deftp | |
862 | ||
863 | @comment sys/timex.h | |
864 | @comment GNU | |
865 | @deftypefun int ntp_gettime (struct ntptimeval *@var{tptr}) | |
866 | The @code{ntp_gettime} function sets the structure pointed to by | |
867 | @var{tptr} to current values. The elements of the structure afterwards | |
868 | contain the values the timer implementation in the kernel assumes. They | |
869 | might or might not be correct. If they are not a @code{ntp_adjtime} | |
870 | call is necessary. | |
871 | ||
872 | The return value is @code{0} on success and other values on failure. The | |
873 | following @code{errno} error conditions are defined for this function: | |
874 | ||
875 | @table @code | |
876 | @item TIME_ERROR | |
877 | The precision clock model is not properly set up at the moment, thus the | |
878 | clock must be considered unsynchronized, and the values should be | |
879 | treated with care. | |
880 | @end table | |
881 | @end deftypefun | |
882 | ||
883 | @tindex struct timex | |
884 | @deftp {Data Type} {struct timex} | |
885 | This structure is used to control and monitor the system clock. It | |
886 | contains the following members: | |
887 | @table @code | |
888 | @item unsigned int modes | |
889 | This variable controls whether and which values are set. Several | |
890 | symbolic constants have to be combined with @emph{binary or} to specify | |
891 | the effective mode. These constants start with @code{MOD_}. | |
892 | ||
893 | @item long int offset | |
99a20616 UD |
894 | This value indicates the current offset of the system clock from the true |
895 | calendar time. The value is given in microseconds. If bit | |
896 | @code{MOD_OFFSET} is set in @code{modes}, the offset (and possibly other | |
897 | dependent values) can be set. The offset's absolute value must not | |
898 | exceed @code{MAXPHASE}. | |
3566d33c UD |
899 | |
900 | ||
901 | @item long int frequency | |
99a20616 UD |
902 | This value indicates the difference in frequency between the true |
903 | calendar time and the system clock. The value is expressed as scaled | |
904 | PPM (parts per million, 0.0001%). The scaling is @code{1 << | |
905 | SHIFT_USEC}. The value can be set with bit @code{MOD_FREQUENCY}, but | |
906 | the absolute value must not exceed @code{MAXFREQ}. | |
3566d33c UD |
907 | |
908 | @item long int maxerror | |
909 | This is the maximum error, measured in microseconds. A new value can be | |
910 | set using bit @code{MOD_MAXERROR}. Unless updated via | |
911 | @code{ntp_adjtime} periodically, this value will increase steadily | |
912 | and reach some platform-specific maximum value. | |
913 | ||
914 | @item long int esterror | |
915 | This is the estimated error, measured in microseconds. This value can | |
916 | be set using bit @code{MOD_ESTERROR}. | |
917 | ||
918 | @item int status | |
919 | This variable reflects the various states of the clock machinery. There | |
920 | are symbolic constants for the significant bits, starting with | |
921 | @code{STA_}. Some of these flags can be updated using the | |
922 | @code{MOD_STATUS} bit. | |
923 | ||
924 | @item long int constant | |
925 | This value represents the bandwidth or stiffness of the PLL (phase | |
926 | locked loop) implemented in the kernel. The value can be changed using | |
927 | bit @code{MOD_TIMECONST}. | |
928 | ||
929 | @item long int precision | |
930 | This value represents the accuracy or the maximum error when reading the | |
931 | system clock. The value is expressed in microseconds. | |
932 | ||
933 | @item long int tolerance | |
934 | This value represents the maximum frequency error of the system clock in | |
935 | scaled PPM. This value is used to increase the @code{maxerror} every | |
936 | second. | |
937 | ||
938 | @item struct timeval time | |
99a20616 | 939 | The current calendar time. |
3566d33c UD |
940 | |
941 | @item long int tick | |
99a20616 | 942 | The elapsed time between clock ticks in microseconds. A clock tick is a |
3566d33c UD |
943 | periodic timer interrupt on which the system clock is based. |
944 | ||
945 | @item long int ppsfreq | |
946 | This is the first of a few optional variables that are present only if | |
947 | the system clock can use a PPS (pulse per second) signal to discipline | |
99a20616 UD |
948 | the system clock. The value is expressed in scaled PPM and it denotes |
949 | the difference in frequency between the system clock and the PPS signal. | |
3566d33c UD |
950 | |
951 | @item long int jitter | |
952 | This value expresses a median filtered average of the PPS signal's | |
953 | dispersion in microseconds. | |
954 | ||
955 | @item int shift | |
956 | This value is a binary exponent for the duration of the PPS calibration | |
957 | interval, ranging from @code{PPS_SHIFT} to @code{PPS_SHIFTMAX}. | |
958 | ||
959 | @item long int stabil | |
960 | This value represents the median filtered dispersion of the PPS | |
961 | frequency in scaled PPM. | |
962 | ||
963 | @item long int jitcnt | |
964 | This counter represents the number of pulses where the jitter exceeded | |
965 | the allowed maximum @code{MAXTIME}. | |
966 | ||
967 | @item long int calcnt | |
968 | This counter reflects the number of successful calibration intervals. | |
969 | ||
970 | @item long int errcnt | |
971 | This counter represents the number of calibration errors (caused by | |
972 | large offsets or jitter). | |
973 | ||
974 | @item long int stbcnt | |
975 | This counter denotes the number of of calibrations where the stability | |
976 | exceeded the threshold. | |
977 | @end table | |
978 | @end deftp | |
979 | ||
980 | @comment sys/timex.h | |
981 | @comment GNU | |
982 | @deftypefun int ntp_adjtime (struct timex *@var{tptr}) | |
983 | The @code{ntp_adjtime} function sets the structure specified by | |
984 | @var{tptr} to current values. | |
985 | ||
986 | In addition, @code{ntp_adjtime} updates some settings to match what you | |
987 | pass to it in *@var{tptr}. Use the @code{modes} element of *@var{tptr} | |
988 | to select what settings to update. You can set @code{offset}, | |
989 | @code{freq}, @code{maxerror}, @code{esterror}, @code{status}, | |
990 | @code{constant}, and @code{tick}. | |
991 | ||
992 | @code{modes} = zero means set nothing. | |
993 | ||
994 | Only the superuser can update settings. | |
995 | ||
996 | @c On Linux, ntp_adjtime() also does the adjtime() function if you set | |
997 | @c modes = ADJ_OFFSET_SINGLESHOT (in fact, that is how GNU libc implements | |
998 | @c adjtime()). But this should be considered an internal function because | |
999 | @c it's so inconsistent with the rest of what ntp_adjtime() does and is | |
1000 | @c forced in an ugly way into the struct timex. So we don't document it | |
1001 | @c and instead document adjtime() as the way to achieve the function. | |
1002 | ||
1003 | The return value is @code{0} on success and other values on failure. The | |
1004 | following @code{errno} error conditions are defined for this function: | |
1005 | ||
1006 | @table @code | |
1007 | @item TIME_ERROR | |
1008 | The high accuracy clock model is not properly set up at the moment, thus the | |
1009 | clock must be considered unsynchronized, and the values should be | |
1010 | treated with care. Another reason could be that the specified new values | |
1011 | are not allowed. | |
1012 | ||
1013 | @item EPERM | |
1014 | The process specified a settings update, but is not superuser. | |
1015 | ||
1016 | @end table | |
1017 | ||
1018 | For more details see RFC1305 (Network Time Protocol, Version 3) and | |
1019 | related documents. | |
1020 | ||
1021 | @strong{Portability note:} Early versions of the GNU C library did not | |
1022 | have this function but did have the synonymous @code{adjtimex}. | |
1023 | ||
1024 | @end deftypefun | |
1025 | ||
1026 | ||
99a20616 UD |
1027 | @node Formatting Calendar Time |
1028 | @subsection Formatting Calendar Time | |
28f540f4 | 1029 | |
99a20616 UD |
1030 | The functions described in this section format calendar time values as |
1031 | strings. These functions are declared in the header file @file{time.h}. | |
28f540f4 RM |
1032 | @pindex time.h |
1033 | ||
1034 | @comment time.h | |
f65fd747 | 1035 | @comment ISO |
28f540f4 RM |
1036 | @deftypefun {char *} asctime (const struct tm *@var{brokentime}) |
1037 | The @code{asctime} function converts the broken-down time value that | |
1038 | @var{brokentime} points to into a string in a standard format: | |
1039 | ||
1040 | @smallexample | |
1041 | "Tue May 21 13:46:22 1991\n" | |
1042 | @end smallexample | |
1043 | ||
1044 | The abbreviations for the days of week are: @samp{Sun}, @samp{Mon}, | |
1045 | @samp{Tue}, @samp{Wed}, @samp{Thu}, @samp{Fri}, and @samp{Sat}. | |
1046 | ||
1047 | The abbreviations for the months are: @samp{Jan}, @samp{Feb}, | |
1048 | @samp{Mar}, @samp{Apr}, @samp{May}, @samp{Jun}, @samp{Jul}, @samp{Aug}, | |
1049 | @samp{Sep}, @samp{Oct}, @samp{Nov}, and @samp{Dec}. | |
1050 | ||
1051 | The return value points to a statically allocated string, which might be | |
60092701 | 1052 | overwritten by subsequent calls to @code{asctime} or @code{ctime}. |
28f540f4 RM |
1053 | (But no other library function overwrites the contents of this |
1054 | string.) | |
1055 | @end deftypefun | |
1056 | ||
0413b54c UD |
1057 | @comment time.h |
1058 | @comment POSIX.1c | |
1059 | @deftypefun {char *} asctime_r (const struct tm *@var{brokentime}, char *@var{buffer}) | |
1060 | This function is similar to @code{asctime} but instead of placing the | |
1061 | result in a static buffer it writes the string in the buffer pointed to | |
76c23bac UD |
1062 | by the parameter @var{buffer}. This buffer should have room |
1063 | for at least 26 bytes, including the terminating null. | |
0413b54c UD |
1064 | |
1065 | If no error occurred the function returns a pointer to the string the | |
1066 | result was written into, i.e., it returns @var{buffer}. Otherwise | |
1067 | return @code{NULL}. | |
1068 | @end deftypefun | |
1069 | ||
1070 | ||
28f540f4 | 1071 | @comment time.h |
f65fd747 | 1072 | @comment ISO |
28f540f4 | 1073 | @deftypefun {char *} ctime (const time_t *@var{time}) |
99a20616 UD |
1074 | The @code{ctime} function is similar to @code{asctime}, except that you |
1075 | specify the calendar time argument as a @code{time_t} simple time value | |
1076 | rather than in broken-down local time format. It is equivalent to | |
28f540f4 RM |
1077 | |
1078 | @smallexample | |
1079 | asctime (localtime (@var{time})) | |
1080 | @end smallexample | |
1081 | ||
1082 | @code{ctime} sets the variable @code{tzname}, because @code{localtime} | |
1083 | does so. @xref{Time Zone Functions}. | |
1084 | @end deftypefun | |
1085 | ||
0413b54c UD |
1086 | @comment time.h |
1087 | @comment POSIX.1c | |
1088 | @deftypefun {char *} ctime_r (const time_t *@var{time}, char *@var{buffer}) | |
99a20616 UD |
1089 | This function is similar to @code{ctime}, but places the result in the |
1090 | string pointed to by @var{buffer}. It is equivalent to (written using | |
1091 | gcc extensions, @pxref{Statement Exprs,,,gcc,Porting and Using gcc}): | |
0413b54c UD |
1092 | |
1093 | @smallexample | |
1094 | (@{ struct tm tm; asctime_r (localtime_r (time, &tm), buf); @}) | |
1095 | @end smallexample | |
1096 | ||
1097 | If no error occurred the function returns a pointer to the string the | |
1098 | result was written into, i.e., it returns @var{buffer}. Otherwise | |
1099 | return @code{NULL}. | |
1100 | @end deftypefun | |
1101 | ||
1102 | ||
28f540f4 | 1103 | @comment time.h |
f65fd747 | 1104 | @comment ISO |
28f540f4 RM |
1105 | @deftypefun size_t strftime (char *@var{s}, size_t @var{size}, const char *@var{template}, const struct tm *@var{brokentime}) |
1106 | This function is similar to the @code{sprintf} function (@pxref{Formatted | |
1107 | Input}), but the conversion specifications that can appear in the format | |
1108 | template @var{template} are specialized for printing components of the date | |
1109 | and time @var{brokentime} according to the locale currently specified for | |
1110 | time conversion (@pxref{Locales}). | |
1111 | ||
1112 | Ordinary characters appearing in the @var{template} are copied to the | |
1113 | output string @var{s}; this can include multibyte character sequences. | |
ec4b0518 | 1114 | Conversion specifiers are introduced by a @samp{%} character, followed |
a2b08ee5 UD |
1115 | by an optional flag which can be one of the following. These flags |
1116 | are all GNU extensions. The first three affect only the output of | |
1117 | numbers: | |
2de99474 UD |
1118 | |
1119 | @table @code | |
1120 | @item _ | |
1121 | The number is padded with spaces. | |
1122 | ||
1123 | @item - | |
1124 | The number is not padded at all. | |
860d3729 UD |
1125 | |
1126 | @item 0 | |
7e3be507 | 1127 | The number is padded with zeros even if the format specifies padding |
860d3729 | 1128 | with spaces. |
7e3be507 UD |
1129 | |
1130 | @item ^ | |
1131 | The output uses uppercase characters, but only if this is possible | |
1132 | (@pxref{Case Conversion}). | |
2de99474 UD |
1133 | @end table |
1134 | ||
ec4b0518 UD |
1135 | The default action is to pad the number with zeros to keep it a constant |
1136 | width. Numbers that do not have a range indicated below are never | |
1137 | padded, since there is no natural width for them. | |
1138 | ||
860d3729 UD |
1139 | Following the flag an optional specification of the width is possible. |
1140 | This is specified in decimal notation. If the natural size of the | |
5b826692 | 1141 | output is of the field has less than the specified number of characters, |
860d3729 UD |
1142 | the result is written right adjusted and space padded to the given |
1143 | size. | |
1144 | ||
1145 | An optional modifier can follow the optional flag and width | |
ba737b94 UD |
1146 | specification. The modifiers, which were first standardized by |
1147 | POSIX.2-1992 and by @w{ISO C99}, are: | |
ec4b0518 UD |
1148 | |
1149 | @table @code | |
1150 | @item E | |
1151 | Use the locale's alternate representation for date and time. This | |
1152 | modifier applies to the @code{%c}, @code{%C}, @code{%x}, @code{%X}, | |
1153 | @code{%y} and @code{%Y} format specifiers. In a Japanese locale, for | |
1154 | example, @code{%Ex} might yield a date format based on the Japanese | |
1155 | Emperors' reigns. | |
1156 | ||
1157 | @item O | |
1158 | Use the locale's alternate numeric symbols for numbers. This modifier | |
1159 | applies only to numeric format specifiers. | |
1160 | @end table | |
1161 | ||
860d3729 UD |
1162 | If the format supports the modifier but no alternate representation |
1163 | is available, it is ignored. | |
ec4b0518 UD |
1164 | |
1165 | The conversion specifier ends with a format specifier taken from the | |
1166 | following list. The whole @samp{%} sequence is replaced in the output | |
1167 | string as follows: | |
28f540f4 RM |
1168 | |
1169 | @table @code | |
1170 | @item %a | |
1171 | The abbreviated weekday name according to the current locale. | |
1172 | ||
1173 | @item %A | |
1174 | The full weekday name according to the current locale. | |
1175 | ||
1176 | @item %b | |
1177 | The abbreviated month name according to the current locale. | |
1178 | ||
1179 | @item %B | |
1180 | The full month name according to the current locale. | |
1181 | ||
c093ea4f UD |
1182 | Using @code{%B} together with @code{%d} produces grammatically |
1183 | incorrect results for some locales. | |
1184 | ||
28f540f4 | 1185 | @item %c |
99a20616 | 1186 | The preferred calendar time representation for the current locale. |
28f540f4 | 1187 | |
2de99474 | 1188 | @item %C |
ec4b0518 UD |
1189 | The century of the year. This is equivalent to the greatest integer not |
1190 | greater than the year divided by 100. | |
1191 | ||
ba737b94 | 1192 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 | 1193 | |
28f540f4 | 1194 | @item %d |
ec4b0518 | 1195 | The day of the month as a decimal number (range @code{01} through @code{31}). |
28f540f4 | 1196 | |
2de99474 UD |
1197 | @item %D |
1198 | The date using the format @code{%m/%d/%y}. | |
1199 | ||
ba737b94 | 1200 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 | 1201 | |
ec4b0518 | 1202 | @item %e |
2de99474 | 1203 | The day of the month like with @code{%d}, but padded with blank (range |
ec4b0518 UD |
1204 | @code{ 1} through @code{31}). |
1205 | ||
ba737b94 | 1206 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
ec4b0518 | 1207 | |
fe0ec73e UD |
1208 | @item %F |
1209 | The date using the format @code{%Y-%m-%d}. This is the form specified | |
1210 | in the @w{ISO 8601} standard and is the preferred form for all uses. | |
1211 | ||
ba737b94 | 1212 | This format was first standardized by @w{ISO C99} and by POSIX.1-2001. |
fe0ec73e | 1213 | |
ec4b0518 UD |
1214 | @item %g |
1215 | The year corresponding to the ISO week number, but without the century | |
1216 | (range @code{00} through @code{99}). This has the same format and value | |
1217 | as @code{%y}, except that if the ISO week number (see @code{%V}) belongs | |
1218 | to the previous or next year, that year is used instead. | |
1219 | ||
ba737b94 | 1220 | This format was first standardized by @w{ISO C99} and by POSIX.1-2001. |
ec4b0518 UD |
1221 | |
1222 | @item %G | |
1223 | The year corresponding to the ISO week number. This has the same format | |
1224 | and value as @code{%Y}, except that if the ISO week number (see | |
1225 | @code{%V}) belongs to the previous or next year, that year is used | |
1226 | instead. | |
2de99474 | 1227 | |
ba737b94 UD |
1228 | This format was first standardized by @w{ISO C99} and by POSIX.1-2001 |
1229 | but was previously available as a GNU extension. | |
2de99474 UD |
1230 | |
1231 | @item %h | |
1232 | The abbreviated month name according to the current locale. The action | |
1233 | is the same as for @code{%b}. | |
1234 | ||
ba737b94 | 1235 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 | 1236 | |
28f540f4 | 1237 | @item %H |
ec4b0518 | 1238 | The hour as a decimal number, using a 24-hour clock (range @code{00} through |
28f540f4 RM |
1239 | @code{23}). |
1240 | ||
1241 | @item %I | |
ec4b0518 | 1242 | The hour as a decimal number, using a 12-hour clock (range @code{01} through |
28f540f4 RM |
1243 | @code{12}). |
1244 | ||
1245 | @item %j | |
ec4b0518 | 1246 | The day of the year as a decimal number (range @code{001} through @code{366}). |
28f540f4 | 1247 | |
2de99474 UD |
1248 | @item %k |
1249 | The hour as a decimal number, using a 24-hour clock like @code{%H}, but | |
ec4b0518 | 1250 | padded with blank (range @code{ 0} through @code{23}). |
2de99474 UD |
1251 | |
1252 | This format is a GNU extension. | |
1253 | ||
1254 | @item %l | |
1255 | The hour as a decimal number, using a 12-hour clock like @code{%I}, but | |
ec4b0518 | 1256 | padded with blank (range @code{ 1} through @code{12}). |
2de99474 UD |
1257 | |
1258 | This format is a GNU extension. | |
1259 | ||
28f540f4 | 1260 | @item %m |
ec4b0518 | 1261 | The month as a decimal number (range @code{01} through @code{12}). |
28f540f4 RM |
1262 | |
1263 | @item %M | |
ec4b0518 | 1264 | The minute as a decimal number (range @code{00} through @code{59}). |
28f540f4 | 1265 | |
2de99474 UD |
1266 | @item %n |
1267 | A single @samp{\n} (newline) character. | |
1268 | ||
ba737b94 | 1269 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 | 1270 | |
28f540f4 | 1271 | @item %p |
ec4b0518 UD |
1272 | Either @samp{AM} or @samp{PM}, according to the given time value; or the |
1273 | corresponding strings for the current locale. Noon is treated as | |
eabea972 UD |
1274 | @samp{PM} and midnight as @samp{AM}. In most locales |
1275 | @samp{AM}/@samp{PM} format is not supported, in such cases @code{"%p"} | |
1276 | yields an empty string. | |
28f540f4 | 1277 | |
7e3be507 UD |
1278 | @ignore |
1279 | We currently have a problem with makeinfo. Write @samp{AM} and @samp{am} | |
1280 | both results in `am'. I.e., the difference in case is not visible anymore. | |
1281 | @end ignore | |
1282 | @item %P | |
1283 | Either @samp{am} or @samp{pm}, according to the given time value; or the | |
1284 | corresponding strings for the current locale, printed in lowercase | |
eabea972 UD |
1285 | characters. Noon is treated as @samp{pm} and midnight as @samp{am}. In |
1286 | most locales @samp{AM}/@samp{PM} format is not supported, in such cases | |
1287 | @code{"%P"} yields an empty string. | |
7e3be507 | 1288 | |
ba737b94 | 1289 | This format is a GNU extension. |
7e3be507 | 1290 | |
2de99474 | 1291 | @item %r |
99a20616 | 1292 | The complete calendar time using the AM/PM format of the current locale. |
2de99474 | 1293 | |
ba737b94 UD |
1294 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
1295 | In the POSIX locale, this format is equivalent to @code{%I:%M:%S %p}. | |
2de99474 UD |
1296 | |
1297 | @item %R | |
1298 | The hour and minute in decimal numbers using the format @code{%H:%M}. | |
1299 | ||
ba737b94 UD |
1300 | This format was first standardized by @w{ISO C99} and by POSIX.1-2001 |
1301 | but was previously available as a GNU extension. | |
2de99474 UD |
1302 | |
1303 | @item %s | |
ec4b0518 UD |
1304 | The number of seconds since the epoch, i.e., since 1970-01-01 00:00:00 UTC. |
1305 | Leap seconds are not counted unless leap second support is available. | |
2de99474 UD |
1306 | |
1307 | This format is a GNU extension. | |
1308 | ||
28f540f4 | 1309 | @item %S |
e9dcb080 | 1310 | The seconds as a decimal number (range @code{00} through @code{60}). |
28f540f4 | 1311 | |
2de99474 UD |
1312 | @item %t |
1313 | A single @samp{\t} (tabulator) character. | |
1314 | ||
ba737b94 | 1315 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 UD |
1316 | |
1317 | @item %T | |
99a20616 | 1318 | The time of day using decimal numbers using the format @code{%H:%M:%S}. |
2de99474 | 1319 | |
ba737b94 | 1320 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
ec4b0518 UD |
1321 | |
1322 | @item %u | |
1323 | The day of the week as a decimal number (range @code{1} through | |
1324 | @code{7}), Monday being @code{1}. | |
1325 | ||
ba737b94 | 1326 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
2de99474 | 1327 | |
28f540f4 | 1328 | @item %U |
ec4b0518 UD |
1329 | The week number of the current year as a decimal number (range @code{00} |
1330 | through @code{53}), starting with the first Sunday as the first day of | |
1331 | the first week. Days preceding the first Sunday in the year are | |
1332 | considered to be in week @code{00}. | |
7c713e28 RM |
1333 | |
1334 | @item %V | |
ec4b0518 UD |
1335 | The @w{ISO 8601:1988} week number as a decimal number (range @code{01} |
1336 | through @code{53}). ISO weeks start with Monday and end with Sunday. | |
1337 | Week @code{01} of a year is the first week which has the majority of its | |
1338 | days in that year; this is equivalent to the week containing the year's | |
1339 | first Thursday, and it is also equivalent to the week containing January | |
1340 | 4. Week @code{01} of a year can contain days from the previous year. | |
1341 | The week before week @code{01} of a year is the last week (@code{52} or | |
1342 | @code{53}) of the previous year even if it contains days from the new | |
1343 | year. | |
1344 | ||
ba737b94 | 1345 | This format was first standardized by POSIX.2-1992 and by @w{ISO C99}. |
28f540f4 | 1346 | |
2de99474 | 1347 | @item %w |
ec4b0518 UD |
1348 | The day of the week as a decimal number (range @code{0} through |
1349 | @code{6}), Sunday being @code{0}. | |
2de99474 | 1350 | |
28f540f4 | 1351 | @item %W |
ec4b0518 UD |
1352 | The week number of the current year as a decimal number (range @code{00} |
1353 | through @code{53}), starting with the first Monday as the first day of | |
1354 | the first week. All days preceding the first Monday in the year are | |
1355 | considered to be in week @code{00}. | |
28f540f4 | 1356 | |
28f540f4 | 1357 | @item %x |
99a20616 | 1358 | The preferred date representation for the current locale. |
28f540f4 RM |
1359 | |
1360 | @item %X | |
99a20616 | 1361 | The preferred time of day representation for the current locale. |
28f540f4 RM |
1362 | |
1363 | @item %y | |
ec4b0518 UD |
1364 | The year without a century as a decimal number (range @code{00} through |
1365 | @code{99}). This is equivalent to the year modulo 100. | |
28f540f4 RM |
1366 | |
1367 | @item %Y | |
ec4b0518 UD |
1368 | The year as a decimal number, using the Gregorian calendar. Years |
1369 | before the year @code{1} are numbered @code{0}, @code{-1}, and so on. | |
28f540f4 | 1370 | |
2de99474 | 1371 | @item %z |
f0f1bf85 UD |
1372 | @w{RFC 822}/@w{ISO 8601:1988} style numeric time zone (e.g., |
1373 | @code{-0600} or @code{+0100}), or nothing if no time zone is | |
1374 | determinable. | |
2de99474 | 1375 | |
ba737b94 UD |
1376 | This format was first standardized by @w{ISO C99} and by POSIX.1-2001 |
1377 | but was previously available as a GNU extension. | |
ec4b0518 | 1378 | |
ba737b94 | 1379 | In the POSIX locale, a full @w{RFC 822} timestamp is generated by the format |
e9dcb080 | 1380 | @w{@samp{"%a, %d %b %Y %H:%M:%S %z"}} (or the equivalent |
eeabe877 | 1381 | @w{@samp{"%a, %d %b %Y %T %z"}}). |
e1350332 | 1382 | |
28f540f4 | 1383 | @item %Z |
ec4b0518 | 1384 | The time zone abbreviation (empty if the time zone can't be determined). |
28f540f4 RM |
1385 | |
1386 | @item %% | |
1387 | A literal @samp{%} character. | |
1388 | @end table | |
1389 | ||
1390 | The @var{size} parameter can be used to specify the maximum number of | |
1391 | characters to be stored in the array @var{s}, including the terminating | |
1392 | null character. If the formatted time requires more than @var{size} | |
76c23bac UD |
1393 | characters, @code{strftime} returns zero and the contents of the array |
1394 | @var{s} are undefined. Otherwise the return value indicates the | |
55c14926 UD |
1395 | number of characters placed in the array @var{s}, not including the |
1396 | terminating null character. | |
1397 | ||
1398 | @emph{Warning:} This convention for the return value which is prescribed | |
1399 | in @w{ISO C} can lead to problems in some situations. For certain | |
1400 | format strings and certain locales the output really can be the empty | |
1401 | string and this cannot be discovered by testing the return value only. | |
1402 | E.g., in most locales the AM/PM time format is not supported (most of | |
1403 | the world uses the 24 hour time representation). In such locales | |
1404 | @code{"%p"} will return the empty string, i.e., the return value is | |
1405 | zero. To detect situations like this something similar to the following | |
1406 | code should be used: | |
1407 | ||
1408 | @smallexample | |
1409 | buf[0] = '\1'; | |
1410 | len = strftime (buf, bufsize, format, tp); | |
1411 | if (len == 0 && buf[0] != '\0') | |
1412 | @{ | |
1413 | /* Something went wrong in the strftime call. */ | |
1414 | @dots{} | |
1415 | @} | |
1416 | @end smallexample | |
28f540f4 RM |
1417 | |
1418 | If @var{s} is a null pointer, @code{strftime} does not actually write | |
1419 | anything, but instead returns the number of characters it would have written. | |
1420 | ||
860d3729 UD |
1421 | According to POSIX.1 every call to @code{strftime} implies a call to |
1422 | @code{tzset}. So the contents of the environment variable @code{TZ} | |
1423 | is examined before any output is produced. | |
1424 | ||
28f540f4 RM |
1425 | For an example of @code{strftime}, see @ref{Time Functions Example}. |
1426 | @end deftypefun | |
1427 | ||
d64b6ad0 UD |
1428 | @comment time.h |
1429 | @comment ISO/Amend1 | |
1430 | @deftypefun size_t wcsftime (wchar_t *@var{s}, size_t @var{size}, const wchar_t *@var{template}, const struct tm *@var{brokentime}) | |
1431 | The @code{wcsftime} function is equivalent to the @code{strftime} | |
76c23bac | 1432 | function with the difference that it operates on wide character |
d64b6ad0 UD |
1433 | strings. The buffer where the result is stored, pointed to by @var{s}, |
1434 | must be an array of wide characters. The parameter @var{size} which | |
1435 | specifies the size of the output buffer gives the number of wide | |
1436 | character, not the number of bytes. | |
1437 | ||
1438 | Also the format string @var{template} is a wide character string. Since | |
1439 | all characters needed to specify the format string are in the basic | |
76c23bac | 1440 | character set it is portably possible to write format strings in the C |
95fdc6a0 | 1441 | source code using the @code{L"@dots{}"} notation. The parameter |
d64b6ad0 UD |
1442 | @var{brokentime} has the same meaning as in the @code{strftime} call. |
1443 | ||
1444 | The @code{wcsftime} function supports the same flags, modifiers, and | |
1445 | format specifiers as the @code{strftime} function. | |
1446 | ||
1447 | The return value of @code{wcsftime} is the number of wide characters | |
1448 | stored in @code{s}. When more characters would have to be written than | |
1449 | can be placed in the buffer @var{s} the return value is zero, with the | |
1450 | same problems indicated in the @code{strftime} documentation. | |
1451 | @end deftypefun | |
1452 | ||
e9dcb080 UD |
1453 | @node Parsing Date and Time |
1454 | @subsection Convert textual time and date information back | |
1455 | ||
1456 | The @w{ISO C} standard does not specify any functions which can convert | |
1457 | the output of the @code{strftime} function back into a binary format. | |
76c23bac UD |
1458 | This led to a variety of more-or-less successful implementations with |
1459 | different interfaces over the years. Then the Unix standard was | |
1460 | extended by the addition of two functions: @code{strptime} and | |
1461 | @code{getdate}. Both have strange interfaces but at least they are | |
1462 | widely available. | |
e9dcb080 UD |
1463 | |
1464 | @menu | |
1465 | * Low-Level Time String Parsing:: Interpret string according to given format. | |
1466 | * General Time String Parsing:: User-friendly function to parse data and | |
1467 | time strings. | |
1468 | @end menu | |
1469 | ||
1470 | @node Low-Level Time String Parsing | |
1471 | @subsubsection Interpret string according to given format | |
1472 | ||
32c075e1 | 1473 | he first function is rather low-level. It is nevertheless frequently |
76c23bac UD |
1474 | used in software since it is better known. Its interface and |
1475 | implementation are heavily influenced by the @code{getdate} function, | |
1476 | which is defined and implemented in terms of calls to @code{strptime}. | |
e9dcb080 UD |
1477 | |
1478 | @comment time.h | |
1479 | @comment XPG4 | |
1480 | @deftypefun {char *} strptime (const char *@var{s}, const char *@var{fmt}, struct tm *@var{tp}) | |
1481 | The @code{strptime} function parses the input string @var{s} according | |
76c23bac | 1482 | to the format string @var{fmt} and stores its results in the |
e9dcb080 UD |
1483 | structure @var{tp}. |
1484 | ||
76c23bac UD |
1485 | The input string could be generated by a @code{strftime} call or |
1486 | obtained any other way. It does not need to be in a human-recognizable | |
1487 | format; e.g. a date passed as @code{"02:1999:9"} is acceptable, even | |
1488 | though it is ambiguous without context. As long as the format string | |
1489 | @var{fmt} matches the input string the function will succeed. | |
e9dcb080 | 1490 | |
45e0579f UD |
1491 | The user has to make sure, though, that the input can be parsed in a |
1492 | unambiguous way. The string @code{"1999112"} can be parsed using the | |
1493 | format @code{"%Y%m%d"} as 1999-1-12, 1999-11-2, or even 19991-1-2. It | |
1494 | is necessary to add appropriate separators to reliably get results. | |
1495 | ||
e9dcb080 | 1496 | The format string consists of the same components as the format string |
76c23bac | 1497 | of the @code{strftime} function. The only difference is that the flags |
e9dcb080 UD |
1498 | @code{_}, @code{-}, @code{0}, and @code{^} are not allowed. |
1499 | @comment Is this really the intention? --drepper | |
76c23bac UD |
1500 | Several of the distinct formats of @code{strftime} do the same work in |
1501 | @code{strptime} since differences like case of the input do not matter. | |
1502 | For reasons of symmetry all formats are supported, though. | |
e9dcb080 UD |
1503 | |
1504 | The modifiers @code{E} and @code{O} are also allowed everywhere the | |
1505 | @code{strftime} function allows them. | |
1506 | ||
1507 | The formats are: | |
1508 | ||
1509 | @table @code | |
1510 | @item %a | |
1511 | @itemx %A | |
1512 | The weekday name according to the current locale, in abbreviated form or | |
1513 | the full name. | |
1514 | ||
1515 | @item %b | |
1516 | @itemx %B | |
1517 | @itemx %h | |
1518 | The month name according to the current locale, in abbreviated form or | |
1519 | the full name. | |
1520 | ||
1521 | @item %c | |
1522 | The date and time representation for the current locale. | |
1523 | ||
1524 | @item %Ec | |
1525 | Like @code{%c} but the locale's alternative date and time format is used. | |
1526 | ||
1527 | @item %C | |
1528 | The century of the year. | |
1529 | ||
1530 | It makes sense to use this format only if the format string also | |
1531 | contains the @code{%y} format. | |
1532 | ||
1533 | @item %EC | |
1534 | The locale's representation of the period. | |
1535 | ||
76c23bac UD |
1536 | Unlike @code{%C} it sometimes makes sense to use this format since some |
1537 | cultures represent years relative to the beginning of eras instead of | |
1538 | using the Gregorian years. | |
e9dcb080 UD |
1539 | |
1540 | @item %d | |
1541 | @item %e | |
1542 | The day of the month as a decimal number (range @code{1} through @code{31}). | |
1543 | Leading zeroes are permitted but not required. | |
1544 | ||
1545 | @item %Od | |
1546 | @itemx %Oe | |
76c23bac | 1547 | Same as @code{%d} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1548 | |
1549 | Leading zeroes are permitted but not required. | |
1550 | ||
1551 | @item %D | |
76c23bac | 1552 | Equivalent to @code{%m/%d/%y}. |
e9dcb080 UD |
1553 | |
1554 | @item %F | |
76c23bac | 1555 | Equivalent to @code{%Y-%m-%d}, which is the @w{ISO 8601} date |
e9dcb080 UD |
1556 | format. |
1557 | ||
ec751a23 | 1558 | This is a GNU extension following an @w{ISO C99} extension to |
e9dcb080 UD |
1559 | @code{strftime}. |
1560 | ||
1561 | @item %g | |
1562 | The year corresponding to the ISO week number, but without the century | |
1563 | (range @code{00} through @code{99}). | |
1564 | ||
76c23bac | 1565 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1566 | recognized, input is consumed but no field in @var{tm} is set. |
1567 | ||
1568 | This format is a GNU extension following a GNU extension of @code{strftime}. | |
1569 | ||
1570 | @item %G | |
1571 | The year corresponding to the ISO week number. | |
1572 | ||
76c23bac | 1573 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1574 | recognized, input is consumed but no field in @var{tm} is set. |
1575 | ||
1576 | This format is a GNU extension following a GNU extension of @code{strftime}. | |
1577 | ||
1578 | @item %H | |
1579 | @itemx %k | |
1580 | The hour as a decimal number, using a 24-hour clock (range @code{00} through | |
1581 | @code{23}). | |
1582 | ||
1583 | @code{%k} is a GNU extension following a GNU extension of @code{strftime}. | |
1584 | ||
1585 | @item %OH | |
76c23bac | 1586 | Same as @code{%H} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1587 | |
1588 | @item %I | |
1589 | @itemx %l | |
1590 | The hour as a decimal number, using a 12-hour clock (range @code{01} through | |
1591 | @code{12}). | |
1592 | ||
1593 | @code{%l} is a GNU extension following a GNU extension of @code{strftime}. | |
1594 | ||
1595 | @item %OI | |
76c23bac | 1596 | Same as @code{%I} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1597 | |
1598 | @item %j | |
1599 | The day of the year as a decimal number (range @code{1} through @code{366}). | |
1600 | ||
1601 | Leading zeroes are permitted but not required. | |
1602 | ||
1603 | @item %m | |
1604 | The month as a decimal number (range @code{1} through @code{12}). | |
1605 | ||
1606 | Leading zeroes are permitted but not required. | |
1607 | ||
1608 | @item %Om | |
76c23bac | 1609 | Same as @code{%m} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1610 | |
1611 | @item %M | |
1612 | The minute as a decimal number (range @code{0} through @code{59}). | |
1613 | ||
1614 | Leading zeroes are permitted but not required. | |
1615 | ||
1616 | @item %OM | |
76c23bac | 1617 | Same as @code{%M} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1618 | |
1619 | @item %n | |
1620 | @itemx %t | |
1621 | Matches any white space. | |
1622 | ||
1623 | @item %p | |
1624 | @item %P | |
1625 | The locale-dependent equivalent to @samp{AM} or @samp{PM}. | |
1626 | ||
1627 | This format is not useful unless @code{%I} or @code{%l} is also used. | |
1628 | Another complication is that the locale might not define these values at | |
1629 | all and therefore the conversion fails. | |
1630 | ||
1631 | @code{%P} is a GNU extension following a GNU extension to @code{strftime}. | |
1632 | ||
1633 | @item %r | |
1634 | The complete time using the AM/PM format of the current locale. | |
1635 | ||
1636 | A complication is that the locale might not define this format at all | |
1637 | and therefore the conversion fails. | |
1638 | ||
1639 | @item %R | |
1640 | The hour and minute in decimal numbers using the format @code{%H:%M}. | |
1641 | ||
1642 | @code{%R} is a GNU extension following a GNU extension to @code{strftime}. | |
1643 | ||
1644 | @item %s | |
1645 | The number of seconds since the epoch, i.e., since 1970-01-01 00:00:00 UTC. | |
1646 | Leap seconds are not counted unless leap second support is available. | |
1647 | ||
1648 | @code{%s} is a GNU extension following a GNU extension to @code{strftime}. | |
1649 | ||
1650 | @item %S | |
99a20616 | 1651 | The seconds as a decimal number (range @code{0} through @code{60}). |
e9dcb080 UD |
1652 | |
1653 | Leading zeroes are permitted but not required. | |
1654 | ||
99a20616 UD |
1655 | @strong{Note:} The Unix specification says the upper bound on this value |
1656 | is @code{61}, a result of a decision to allow double leap seconds. You | |
1657 | will not see the value @code{61} because no minute has more than one | |
1658 | leap second, but the myth persists. | |
e9dcb080 UD |
1659 | |
1660 | @item %OS | |
76c23bac | 1661 | Same as @code{%S} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1662 | |
1663 | @item %T | |
1664 | Equivalent to the use of @code{%H:%M:%S} in this place. | |
1665 | ||
1666 | @item %u | |
1667 | The day of the week as a decimal number (range @code{1} through | |
1668 | @code{7}), Monday being @code{1}. | |
1669 | ||
1670 | Leading zeroes are permitted but not required. | |
1671 | ||
76c23bac | 1672 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1673 | recognized, input is consumed but no field in @var{tm} is set. |
1674 | ||
1675 | @item %U | |
1676 | The week number of the current year as a decimal number (range @code{0} | |
1677 | through @code{53}). | |
1678 | ||
1679 | Leading zeroes are permitted but not required. | |
1680 | ||
1681 | @item %OU | |
76c23bac | 1682 | Same as @code{%U} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1683 | |
1684 | @item %V | |
1685 | The @w{ISO 8601:1988} week number as a decimal number (range @code{1} | |
1686 | through @code{53}). | |
1687 | ||
1688 | Leading zeroes are permitted but not required. | |
1689 | ||
76c23bac | 1690 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1691 | recognized, input is consumed but no field in @var{tm} is set. |
1692 | ||
1693 | @item %w | |
1694 | The day of the week as a decimal number (range @code{0} through | |
1695 | @code{6}), Sunday being @code{0}. | |
1696 | ||
1697 | Leading zeroes are permitted but not required. | |
1698 | ||
76c23bac | 1699 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1700 | recognized, input is consumed but no field in @var{tm} is set. |
1701 | ||
1702 | @item %Ow | |
76c23bac | 1703 | Same as @code{%w} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1704 | |
1705 | @item %W | |
1706 | The week number of the current year as a decimal number (range @code{0} | |
1707 | through @code{53}). | |
1708 | ||
1709 | Leading zeroes are permitted but not required. | |
1710 | ||
76c23bac | 1711 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1712 | recognized, input is consumed but no field in @var{tm} is set. |
1713 | ||
1714 | @item %OW | |
76c23bac | 1715 | Same as @code{%W} but using the locale's alternative numeric symbols. |
e9dcb080 UD |
1716 | |
1717 | @item %x | |
1718 | The date using the locale's date format. | |
1719 | ||
1720 | @item %Ex | |
1721 | Like @code{%x} but the locale's alternative data representation is used. | |
1722 | ||
1723 | @item %X | |
1724 | The time using the locale's time format. | |
1725 | ||
1726 | @item %EX | |
1727 | Like @code{%X} but the locale's alternative time representation is used. | |
1728 | ||
1729 | @item %y | |
1730 | The year without a century as a decimal number (range @code{0} through | |
1731 | @code{99}). | |
1732 | ||
1733 | Leading zeroes are permitted but not required. | |
1734 | ||
76c23bac | 1735 | Note that it is questionable to use this format without |
e9dcb080 UD |
1736 | the @code{%C} format. The @code{strptime} function does regard input |
1737 | values in the range @math{68} to @math{99} as the years @math{1969} to | |
1738 | @math{1999} and the values @math{0} to @math{68} as the years | |
1739 | @math{2000} to @math{2068}. But maybe this heuristic fails for some | |
1740 | input data. | |
1741 | ||
1742 | Therefore it is best to avoid @code{%y} completely and use @code{%Y} | |
1743 | instead. | |
1744 | ||
1745 | @item %Ey | |
1746 | The offset from @code{%EC} in the locale's alternative representation. | |
1747 | ||
1748 | @item %Oy | |
1749 | The offset of the year (from @code{%C}) using the locale's alternative | |
1750 | numeric symbols. | |
1751 | ||
1752 | @item %Y | |
1753 | The year as a decimal number, using the Gregorian calendar. | |
1754 | ||
1755 | @item %EY | |
1756 | The full alternative year representation. | |
1757 | ||
1758 | @item %z | |
107d41a9 | 1759 | The offset from GMT in @w{ISO 8601}/RFC822 format. |
e9dcb080 UD |
1760 | |
1761 | @item %Z | |
1762 | The timezone name. | |
1763 | ||
76c23bac | 1764 | @emph{Note:} Currently, this is not fully implemented. The format is |
e9dcb080 UD |
1765 | recognized, input is consumed but no field in @var{tm} is set. |
1766 | ||
1767 | @item %% | |
1768 | A literal @samp{%} character. | |
1769 | @end table | |
1770 | ||
1771 | All other characters in the format string must have a matching character | |
1772 | in the input string. Exceptions are white spaces in the input string | |
05957bbd UD |
1773 | which can match zero or more whitespace characters in the format string. |
1774 | ||
1775 | @strong{Portability Note:} The XPG standard advises applications to use | |
1776 | at least one whitespace character (as specified by @code{isspace}) or | |
1777 | other non-alphanumeric characters between any two conversion | |
1778 | specifications. The @w{GNU C Library} does not have this limitation but | |
1779 | other libraries might have trouble parsing formats like | |
1780 | @code{"%d%m%Y%H%M%S"}. | |
e9dcb080 UD |
1781 | |
1782 | The @code{strptime} function processes the input string from right to | |
1783 | left. Each of the three possible input elements (white space, literal, | |
1784 | or format) are handled one after the other. If the input cannot be | |
1785 | matched to the format string the function stops. The remainder of the | |
1786 | format and input strings are not processed. | |
1787 | ||
76c23bac UD |
1788 | The function returns a pointer to the first character it was unable to |
1789 | process. If the input string contains more characters than required by | |
1790 | the format string the return value points right after the last consumed | |
1791 | input character. If the whole input string is consumed the return value | |
1792 | points to the @code{NULL} byte at the end of the string. If an error | |
32c075e1 | 1793 | occurs, i.e. @code{strptime} fails to match all of the format string, |
76c23bac | 1794 | the function returns @code{NULL}. |
e9dcb080 UD |
1795 | @end deftypefun |
1796 | ||
76c23bac UD |
1797 | The specification of the function in the XPG standard is rather vague, |
1798 | leaving out a few important pieces of information. Most importantly, it | |
e9dcb080 | 1799 | does not specify what happens to those elements of @var{tm} which are |
76c23bac | 1800 | not directly initialized by the different formats. The |
e9dcb080 UD |
1801 | implementations on different Unix systems vary here. |
1802 | ||
1803 | The GNU libc implementation does not touch those fields which are not | |
1804 | directly initialized. Exceptions are the @code{tm_wday} and | |
76c23bac | 1805 | @code{tm_yday} elements, which are recomputed if any of the year, month, |
e9dcb080 UD |
1806 | or date elements changed. This has two implications: |
1807 | ||
1808 | @itemize @bullet | |
1809 | @item | |
76c23bac UD |
1810 | Before calling the @code{strptime} function for a new input string, you |
1811 | should prepare the @var{tm} structure you pass. Normally this will mean | |
1812 | initializing all values are to zero. Alternatively, you can set all | |
1813 | fields to values like @code{INT_MAX}, allowing you to determine which | |
1814 | elements were set by the function call. Zero does not work here since | |
1815 | it is a valid value for many of the fields. | |
1816 | ||
1817 | Careful initialization is necessary if you want to find out whether a | |
e9dcb080 UD |
1818 | certain field in @var{tm} was initialized by the function call. |
1819 | ||
1820 | @item | |
76c23bac UD |
1821 | You can construct a @code{struct tm} value with several consecutive |
1822 | @code{strptime} calls. A useful application of this is e.g. the parsing | |
1823 | of two separate strings, one containing date information and the other | |
1824 | time information. By parsing one after the other without clearing the | |
1825 | structure in-between, you can construct a complete broken-down time. | |
e9dcb080 UD |
1826 | @end itemize |
1827 | ||
1828 | The following example shows a function which parses a string which is | |
76c23bac | 1829 | contains the date information in either US style or @w{ISO 8601} form: |
e9dcb080 UD |
1830 | |
1831 | @smallexample | |
1832 | const char * | |
1833 | parse_date (const char *input, struct tm *tm) | |
1834 | @{ | |
1835 | const char *cp; | |
1836 | ||
1837 | /* @r{First clear the result structure.} */ | |
1838 | memset (tm, '\0', sizeof (*tm)); | |
1839 | ||
1840 | /* @r{Try the ISO format first.} */ | |
1841 | cp = strptime (input, "%F", tm); | |
1842 | if (cp == NULL) | |
1843 | @{ | |
1844 | /* @r{Does not match. Try the US form.} */ | |
1845 | cp = strptime (input, "%D", tm); | |
1846 | @} | |
1847 | ||
1848 | return cp; | |
1849 | @} | |
1850 | @end smallexample | |
1851 | ||
1852 | @node General Time String Parsing | |
76c23bac | 1853 | @subsubsection A More User-friendly Way to Parse Times and Dates |
e9dcb080 | 1854 | |
76c23bac UD |
1855 | The Unix standard defines another function for parsing date strings. |
1856 | The interface is weird, but if the function happens to suit your | |
1857 | application it is just fine. It is problematic to use this function | |
1858 | in multi-threaded programs or libraries, since it returns a pointer to | |
1859 | a static variable, and uses a global variable and global state (an | |
1860 | environment variable). | |
e9dcb080 UD |
1861 | |
1862 | @comment time.h | |
1863 | @comment Unix98 | |
1864 | @defvar getdate_err | |
76c23bac UD |
1865 | This variable of type @code{int} contains the error code of the last |
1866 | unsuccessful call to @code{getdate}. Defined values are: | |
e9dcb080 UD |
1867 | |
1868 | @table @math | |
1869 | @item 1 | |
1870 | The environment variable @code{DATEMSK} is not defined or null. | |
1871 | @item 2 | |
1872 | The template file denoted by the @code{DATEMSK} environment variable | |
1873 | cannot be opened. | |
1874 | @item 3 | |
1875 | Information about the template file cannot retrieved. | |
1876 | @item 4 | |
76c23bac | 1877 | The template file is not a regular file. |
e9dcb080 UD |
1878 | @item 5 |
1879 | An I/O error occurred while reading the template file. | |
1880 | @item 6 | |
1881 | Not enough memory available to execute the function. | |
1882 | @item 7 | |
1883 | The template file contains no matching template. | |
1884 | @item 8 | |
76c23bac UD |
1885 | The input date is invalid, but would match a template otherwise. This |
1886 | includes dates like February 31st, and dates which cannot be represented | |
1887 | in a @code{time_t} variable. | |
e9dcb080 UD |
1888 | @end table |
1889 | @end defvar | |
1890 | ||
1891 | @comment time.h | |
1892 | @comment Unix98 | |
1893 | @deftypefun {struct tm *} getdate (const char *@var{string}) | |
76c23bac UD |
1894 | The interface to @code{getdate} is the simplest possible for a function |
1895 | to parse a string and return the value. @var{string} is the input | |
1896 | string and the result is returned in a statically-allocated variable. | |
e9dcb080 | 1897 | |
76c23bac UD |
1898 | The details about how the string is processed are hidden from the user. |
1899 | In fact, they can be outside the control of the program. Which formats | |
e9dcb080 | 1900 | are recognized is controlled by the file named by the environment |
76c23bac | 1901 | variable @code{DATEMSK}. This file should contain |
e9dcb080 UD |
1902 | lines of valid format strings which could be passed to @code{strptime}. |
1903 | ||
1904 | The @code{getdate} function reads these format strings one after the | |
1905 | other and tries to match the input string. The first line which | |
1906 | completely matches the input string is used. | |
1907 | ||
76c23bac UD |
1908 | Elements not initialized through the format string retain the values |
1909 | present at the time of the @code{getdate} function call. | |
e9dcb080 | 1910 | |
76c23bac | 1911 | The formats recognized by @code{getdate} are the same as for |
e9dcb080 | 1912 | @code{strptime}. See above for an explanation. There are only a few |
76c23bac | 1913 | extensions to the @code{strptime} behavior: |
e9dcb080 UD |
1914 | |
1915 | @itemize @bullet | |
1916 | @item | |
1917 | If the @code{%Z} format is given the broken-down time is based on the | |
76c23bac | 1918 | current time of the timezone matched, not of the current timezone of the |
e9dcb080 UD |
1919 | runtime environment. |
1920 | ||
1921 | @emph{Note}: This is not implemented (currently). The problem is that | |
1922 | timezone names are not unique. If a fixed timezone is assumed for a | |
76c23bac | 1923 | given string (say @code{EST} meaning US East Coast time), then uses for |
e9dcb080 | 1924 | countries other than the USA will fail. So far we have found no good |
76c23bac | 1925 | solution to this. |
e9dcb080 UD |
1926 | |
1927 | @item | |
1928 | If only the weekday is specified the selected day depends on the current | |
1929 | date. If the current weekday is greater or equal to the @code{tm_wday} | |
76c23bac | 1930 | value the current week's day is chosen, otherwise the day next week is chosen. |
e9dcb080 UD |
1931 | |
1932 | @item | |
76c23bac UD |
1933 | A similar heuristic is used when only the month is given and not the |
1934 | year. If the month is greater than or equal to the current month, then | |
1935 | the current year is used. Otherwise it wraps to next year. The first | |
1936 | day of the month is assumed if one is not explicitly specified. | |
e9dcb080 UD |
1937 | |
1938 | @item | |
76c23bac | 1939 | The current hour, minute, and second are used if the appropriate value is |
e9dcb080 UD |
1940 | not set through the format. |
1941 | ||
1942 | @item | |
76c23bac UD |
1943 | If no date is given tomorrow's date is used if the time is |
1944 | smaller than the current time. Otherwise today's date is taken. | |
e9dcb080 UD |
1945 | @end itemize |
1946 | ||
1947 | It should be noted that the format in the template file need not only | |
1948 | contain format elements. The following is a list of possible format | |
1949 | strings (taken from the Unix standard): | |
1950 | ||
1951 | @smallexample | |
1952 | %m | |
1953 | %A %B %d, %Y %H:%M:%S | |
1954 | %A | |
1955 | %B | |
1956 | %m/%d/%y %I %p | |
1957 | %d,%m,%Y %H:%M | |
1958 | at %A the %dst of %B in %Y | |
1959 | run job at %I %p,%B %dnd | |
1960 | %A den %d. %B %Y %H.%M Uhr | |
1961 | @end smallexample | |
1962 | ||
76c23bac | 1963 | As you can see, the template list can contain very specific strings like |
e9dcb080 | 1964 | @code{run job at %I %p,%B %dnd}. Using the above list of templates and |
76c23bac | 1965 | assuming the current time is Mon Sep 22 12:19:47 EDT 1986 we can obtain the |
16a8520f | 1966 | following results for the given input. |
e9dcb080 | 1967 | |
16a8520f | 1968 | @multitable {xxxxxxxxxxxx} {xxxxxxxxxx} {xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx} |
76c23bac | 1969 | @item Input @tab Match @tab Result |
e9dcb080 UD |
1970 | @item Mon @tab %a @tab Mon Sep 22 12:19:47 EDT 1986 |
1971 | @item Sun @tab %a @tab Sun Sep 28 12:19:47 EDT 1986 | |
1972 | @item Fri @tab %a @tab Fri Sep 26 12:19:47 EDT 1986 | |
1973 | @item September @tab %B @tab Mon Sep 1 12:19:47 EDT 1986 | |
1974 | @item January @tab %B @tab Thu Jan 1 12:19:47 EST 1987 | |
1975 | @item December @tab %B @tab Mon Dec 1 12:19:47 EST 1986 | |
1976 | @item Sep Mon @tab %b %a @tab Mon Sep 1 12:19:47 EDT 1986 | |
1977 | @item Jan Fri @tab %b %a @tab Fri Jan 2 12:19:47 EST 1987 | |
1978 | @item Dec Mon @tab %b %a @tab Mon Dec 1 12:19:47 EST 1986 | |
1979 | @item Jan Wed 1989 @tab %b %a %Y @tab Wed Jan 4 12:19:47 EST 1989 | |
1980 | @item Fri 9 @tab %a %H @tab Fri Sep 26 09:00:00 EDT 1986 | |
1981 | @item Feb 10:30 @tab %b %H:%S @tab Sun Feb 1 10:00:30 EST 1987 | |
1982 | @item 10:30 @tab %H:%M @tab Tue Sep 23 10:30:00 EDT 1986 | |
1983 | @item 13:30 @tab %H:%M @tab Mon Sep 22 13:30:00 EDT 1986 | |
1984 | @end multitable | |
1985 | ||
1986 | The return value of the function is a pointer to a static variable of | |
76c23bac UD |
1987 | type @w{@code{struct tm}}, or a null pointer if an error occurred. The |
1988 | result is only valid until the next @code{getdate} call, making this | |
1989 | function unusable in multi-threaded applications. | |
e9dcb080 UD |
1990 | |
1991 | The @code{errno} variable is @emph{not} changed. Error conditions are | |
76c23bac | 1992 | stored in the global variable @code{getdate_err}. See the |
e9dcb080 | 1993 | description above for a list of the possible error values. |
c730d678 UD |
1994 | |
1995 | @emph{Warning:} The @code{getdate} function should @emph{never} be | |
1996 | used in SUID-programs. The reason is obvious: using the | |
76c23bac | 1997 | @code{DATEMSK} environment variable you can get the function to open |
16a8520f | 1998 | any arbitrary file and chances are high that with some bogus input |
c730d678 UD |
1999 | (such as a binary file) the program will crash. |
2000 | @end deftypefun | |
2001 | ||
2002 | @comment time.h | |
2003 | @comment GNU | |
2004 | @deftypefun int getdate_r (const char *@var{string}, struct tm *@var{tp}) | |
2005 | The @code{getdate_r} function is the reentrant counterpart of | |
2006 | @code{getdate}. It does not use the global variable @code{getdate_err} | |
76c23bac UD |
2007 | to signal an error, but instead returns an error code. The same error |
2008 | codes as described in the @code{getdate_err} documentation above are | |
2009 | used, with 0 meaning success. | |
c730d678 | 2010 | |
76c23bac UD |
2011 | Moreover, @code{getdate_r} stores the broken-down time in the variable |
2012 | of type @code{struct tm} pointed to by the second argument, rather than | |
2013 | in a static variable. | |
c730d678 UD |
2014 | |
2015 | This function is not defined in the Unix standard. Nevertheless it is | |
2016 | available on some other Unix systems as well. | |
2017 | ||
76c23bac UD |
2018 | The warning against using @code{getdate} in SUID-programs applies to |
2019 | @code{getdate_r} as well. | |
e9dcb080 UD |
2020 | @end deftypefun |
2021 | ||
28f540f4 RM |
2022 | @node TZ Variable |
2023 | @subsection Specifying the Time Zone with @code{TZ} | |
2024 | ||
2025 | In POSIX systems, a user can specify the time zone by means of the | |
2026 | @code{TZ} environment variable. For information about how to set | |
2027 | environment variables, see @ref{Environment Variables}. The functions | |
2028 | for accessing the time zone are declared in @file{time.h}. | |
2029 | @pindex time.h | |
2030 | @cindex time zone | |
2031 | ||
2032 | You should not normally need to set @code{TZ}. If the system is | |
f0f1bf85 | 2033 | configured properly, the default time zone will be correct. You might |
76c23bac UD |
2034 | set @code{TZ} if you are using a computer over a network from a |
2035 | different time zone, and would like times reported to you in the time | |
2036 | zone local to you, rather than what is local to the computer. | |
28f540f4 | 2037 | |
76c23bac | 2038 | In POSIX.1 systems the value of the @code{TZ} variable can be in one of |
28f540f4 RM |
2039 | three formats. With the GNU C library, the most common format is the |
2040 | last one, which can specify a selection from a large database of time | |
2041 | zone information for many regions of the world. The first two formats | |
2042 | are used to describe the time zone information directly, which is both | |
2043 | more cumbersome and less precise. But the POSIX.1 standard only | |
2044 | specifies the details of the first two formats, so it is good to be | |
2045 | familiar with them in case you come across a POSIX.1 system that doesn't | |
2046 | support a time zone information database. | |
2047 | ||
2048 | The first format is used when there is no Daylight Saving Time (or | |
2049 | summer time) in the local time zone: | |
2050 | ||
2051 | @smallexample | |
2052 | @r{@var{std} @var{offset}} | |
2053 | @end smallexample | |
2054 | ||
2055 | The @var{std} string specifies the name of the time zone. It must be | |
76c23bac UD |
2056 | three or more characters long and must not contain a leading colon, |
2057 | embedded digits, commas, nor plus and minus signs. There is no space | |
28f540f4 RM |
2058 | character separating the time zone name from the @var{offset}, so these |
2059 | restrictions are necessary to parse the specification correctly. | |
2060 | ||
76c23bac | 2061 | The @var{offset} specifies the time value you must add to the local time |
28f540f4 RM |
2062 | to get a Coordinated Universal Time value. It has syntax like |
2063 | [@code{+}|@code{-}]@var{hh}[@code{:}@var{mm}[@code{:}@var{ss}]]. This | |
2064 | is positive if the local time zone is west of the Prime Meridian and | |
2065 | negative if it is east. The hour must be between @code{0} and | |
01cdeca0 | 2066 | @code{23}, and the minute and seconds between @code{0} and @code{59}. |
28f540f4 RM |
2067 | |
2068 | For example, here is how we would specify Eastern Standard Time, but | |
76c23bac | 2069 | without any Daylight Saving Time alternative: |
28f540f4 RM |
2070 | |
2071 | @smallexample | |
2072 | EST+5 | |
2073 | @end smallexample | |
2074 | ||
2075 | The second format is used when there is Daylight Saving Time: | |
2076 | ||
2077 | @smallexample | |
2078 | @r{@var{std} @var{offset} @var{dst} [@var{offset}]@code{,}@var{start}[@code{/}@var{time}]@code{,}@var{end}[@code{/}@var{time}]} | |
2079 | @end smallexample | |
2080 | ||
2081 | The initial @var{std} and @var{offset} specify the standard time zone, as | |
2082 | described above. The @var{dst} string and @var{offset} specify the name | |
76c23bac | 2083 | and offset for the corresponding Daylight Saving Time zone; if the |
28f540f4 RM |
2084 | @var{offset} is omitted, it defaults to one hour ahead of standard time. |
2085 | ||
76c23bac UD |
2086 | The remainder of the specification describes when Daylight Saving Time is |
2087 | in effect. The @var{start} field is when Daylight Saving Time goes into | |
28f540f4 RM |
2088 | effect and the @var{end} field is when the change is made back to standard |
2089 | time. The following formats are recognized for these fields: | |
2090 | ||
2091 | @table @code | |
2092 | @item J@var{n} | |
2093 | This specifies the Julian day, with @var{n} between @code{1} and @code{365}. | |
2094 | February 29 is never counted, even in leap years. | |
2095 | ||
2096 | @item @var{n} | |
2097 | This specifies the Julian day, with @var{n} between @code{0} and @code{365}. | |
2098 | February 29 is counted in leap years. | |
2099 | ||
2100 | @item M@var{m}.@var{w}.@var{d} | |
2101 | This specifies day @var{d} of week @var{w} of month @var{m}. The day | |
2102 | @var{d} must be between @code{0} (Sunday) and @code{6}. The week | |
2103 | @var{w} must be between @code{1} and @code{5}; week @code{1} is the | |
2104 | first week in which day @var{d} occurs, and week @code{5} specifies the | |
2105 | @emph{last} @var{d} day in the month. The month @var{m} should be | |
2106 | between @code{1} and @code{12}. | |
2107 | @end table | |
2108 | ||
2109 | The @var{time} fields specify when, in the local time currently in | |
2110 | effect, the change to the other time occurs. If omitted, the default is | |
2111 | @code{02:00:00}. | |
2112 | ||
76c23bac UD |
2113 | For example, here is how you would specify the Eastern time zone in the |
2114 | United States, including the appropriate Daylight Saving Time and its dates | |
f0f1bf85 | 2115 | of applicability. The normal offset from UTC is 5 hours; since this is |
28f540f4 RM |
2116 | west of the prime meridian, the sign is positive. Summer time begins on |
2117 | the first Sunday in April at 2:00am, and ends on the last Sunday in October | |
2118 | at 2:00am. | |
2119 | ||
2120 | @smallexample | |
f0f1bf85 | 2121 | EST+5EDT,M4.1.0/2,M10.5.0/2 |
28f540f4 RM |
2122 | @end smallexample |
2123 | ||
76c23bac | 2124 | The schedule of Daylight Saving Time in any particular jurisdiction has |
28f540f4 RM |
2125 | changed over the years. To be strictly correct, the conversion of dates |
2126 | and times in the past should be based on the schedule that was in effect | |
2127 | then. However, this format has no facilities to let you specify how the | |
2128 | schedule has changed from year to year. The most you can do is specify | |
2129 | one particular schedule---usually the present day schedule---and this is | |
2130 | used to convert any date, no matter when. For precise time zone | |
2131 | specifications, it is best to use the time zone information database | |
2132 | (see below). | |
2133 | ||
2134 | The third format looks like this: | |
2135 | ||
2136 | @smallexample | |
2137 | :@var{characters} | |
2138 | @end smallexample | |
2139 | ||
2140 | Each operating system interprets this format differently; in the GNU C | |
2141 | library, @var{characters} is the name of a file which describes the time | |
2142 | zone. | |
2143 | ||
2144 | @pindex /etc/localtime | |
2145 | @pindex localtime | |
2146 | If the @code{TZ} environment variable does not have a value, the | |
2147 | operation chooses a time zone by default. In the GNU C library, the | |
2148 | default time zone is like the specification @samp{TZ=:/etc/localtime} | |
2149 | (or @samp{TZ=:/usr/local/etc/localtime}, depending on how GNU C library | |
2150 | was configured; @pxref{Installation}). Other C libraries use their own | |
2151 | rule for choosing the default time zone, so there is little we can say | |
2152 | about them. | |
2153 | ||
2154 | @cindex time zone database | |
2155 | @pindex /share/lib/zoneinfo | |
2156 | @pindex zoneinfo | |
2157 | If @var{characters} begins with a slash, it is an absolute file name; | |
2158 | otherwise the library looks for the file | |
2159 | @w{@file{/share/lib/zoneinfo/@var{characters}}}. The @file{zoneinfo} | |
2160 | directory contains data files describing local time zones in many | |
2161 | different parts of the world. The names represent major cities, with | |
2162 | subdirectories for geographical areas; for example, | |
2163 | @file{America/New_York}, @file{Europe/London}, @file{Asia/Hong_Kong}. | |
2164 | These data files are installed by the system administrator, who also | |
2165 | sets @file{/etc/localtime} to point to the data file for the local time | |
2166 | zone. The GNU C library comes with a large database of time zone | |
2167 | information for most regions of the world, which is maintained by a | |
2168 | community of volunteers and put in the public domain. | |
2169 | ||
2170 | @node Time Zone Functions | |
2171 | @subsection Functions and Variables for Time Zones | |
2172 | ||
2173 | @comment time.h | |
2174 | @comment POSIX.1 | |
2c6fe0bd | 2175 | @deftypevar {char *} tzname [2] |
28f540f4 | 2176 | The array @code{tzname} contains two strings, which are the standard |
76c23bac UD |
2177 | names of the pair of time zones (standard and Daylight |
2178 | Saving) that the user has selected. @code{tzname[0]} is the name of | |
28f540f4 | 2179 | the standard time zone (for example, @code{"EST"}), and @code{tzname[1]} |
76c23bac | 2180 | is the name for the time zone when Daylight Saving Time is in use (for |
28f540f4 | 2181 | example, @code{"EDT"}). These correspond to the @var{std} and @var{dst} |
f0f1bf85 | 2182 | strings (respectively) from the @code{TZ} environment variable. If |
76c23bac | 2183 | Daylight Saving Time is never used, @code{tzname[1]} is the empty string. |
28f540f4 RM |
2184 | |
2185 | The @code{tzname} array is initialized from the @code{TZ} environment | |
2186 | variable whenever @code{tzset}, @code{ctime}, @code{strftime}, | |
f0f1bf85 UD |
2187 | @code{mktime}, or @code{localtime} is called. If multiple abbreviations |
2188 | have been used (e.g. @code{"EWT"} and @code{"EDT"} for U.S. Eastern War | |
2189 | Time and Eastern Daylight Time), the array contains the most recent | |
2190 | abbreviation. | |
2191 | ||
2192 | The @code{tzname} array is required for POSIX.1 compatibility, but in | |
2193 | GNU programs it is better to use the @code{tm_zone} member of the | |
2194 | broken-down time structure, since @code{tm_zone} reports the correct | |
2195 | abbreviation even when it is not the latest one. | |
2196 | ||
76c23bac | 2197 | Though the strings are declared as @code{char *} the user must refrain |
f2ea0f5b | 2198 | from modifying these strings. Modifying the strings will almost certainly |
0413b54c UD |
2199 | lead to trouble. |
2200 | ||
28f540f4 RM |
2201 | @end deftypevar |
2202 | ||
2203 | @comment time.h | |
2204 | @comment POSIX.1 | |
2205 | @deftypefun void tzset (void) | |
2206 | The @code{tzset} function initializes the @code{tzname} variable from | |
2207 | the value of the @code{TZ} environment variable. It is not usually | |
2208 | necessary for your program to call this function, because it is called | |
2209 | automatically when you use the other time conversion functions that | |
2210 | depend on the time zone. | |
2211 | @end deftypefun | |
2212 | ||
2213 | The following variables are defined for compatibility with System V | |
f0f1bf85 UD |
2214 | Unix. Like @code{tzname}, these variables are set by calling |
2215 | @code{tzset} or the other time conversion functions. | |
28f540f4 RM |
2216 | |
2217 | @comment time.h | |
2218 | @comment SVID | |
2219 | @deftypevar {long int} timezone | |
f0f1bf85 UD |
2220 | This contains the difference between UTC and the latest local standard |
2221 | time, in seconds west of UTC. For example, in the U.S. Eastern time | |
2222 | zone, the value is @code{5*60*60}. Unlike the @code{tm_gmtoff} member | |
2223 | of the broken-down time structure, this value is not adjusted for | |
2224 | daylight saving, and its sign is reversed. In GNU programs it is better | |
2225 | to use @code{tm_gmtoff}, since it contains the correct offset even when | |
2226 | it is not the latest one. | |
28f540f4 RM |
2227 | @end deftypevar |
2228 | ||
2229 | @comment time.h | |
2230 | @comment SVID | |
2231 | @deftypevar int daylight | |
76c23bac UD |
2232 | This variable has a nonzero value if Daylight Saving Time rules apply. |
2233 | A nonzero value does not necessarily mean that Daylight Saving Time is | |
2234 | now in effect; it means only that Daylight Saving Time is sometimes in | |
60092701 | 2235 | effect. |
28f540f4 RM |
2236 | @end deftypevar |
2237 | ||
2238 | @node Time Functions Example | |
2239 | @subsection Time Functions Example | |
2240 | ||
3566d33c UD |
2241 | Here is an example program showing the use of some of the calendar time |
2242 | functions. | |
28f540f4 RM |
2243 | |
2244 | @smallexample | |
2245 | @include strftim.c.texi | |
2246 | @end smallexample | |
2247 | ||
2248 | It produces output like this: | |
2249 | ||
2250 | @smallexample | |
2251 | Wed Jul 31 13:02:36 1991 | |
2252 | Today is Wednesday, July 31. | |
2253 | The time is 01:02 PM. | |
2254 | @end smallexample | |
2255 | ||
2256 | ||
2257 | @node Setting an Alarm | |
2258 | @section Setting an Alarm | |
2259 | ||
2260 | The @code{alarm} and @code{setitimer} functions provide a mechanism for a | |
99a20616 | 2261 | process to interrupt itself in the future. They do this by setting a |
28f540f4 RM |
2262 | timer; when the timer expires, the process receives a signal. |
2263 | ||
2264 | @cindex setting an alarm | |
2265 | @cindex interval timer, setting | |
2266 | @cindex alarms, setting | |
2267 | @cindex timers, setting | |
2268 | Each process has three independent interval timers available: | |
2269 | ||
2270 | @itemize @bullet | |
01cdeca0 | 2271 | @item |
99a20616 | 2272 | A real-time timer that counts elapsed time. This timer sends a |
28f540f4 RM |
2273 | @code{SIGALRM} signal to the process when it expires. |
2274 | @cindex real-time timer | |
2275 | @cindex timer, real-time | |
2276 | ||
01cdeca0 | 2277 | @item |
99a20616 | 2278 | A virtual timer that counts processor time used by the process. This timer |
28f540f4 RM |
2279 | sends a @code{SIGVTALRM} signal to the process when it expires. |
2280 | @cindex virtual timer | |
2281 | @cindex timer, virtual | |
2282 | ||
01cdeca0 | 2283 | @item |
99a20616 UD |
2284 | A profiling timer that counts both processor time used by the process, |
2285 | and processor time spent in system calls on behalf of the process. This | |
2286 | timer sends a @code{SIGPROF} signal to the process when it expires. | |
28f540f4 RM |
2287 | @cindex profiling timer |
2288 | @cindex timer, profiling | |
2289 | ||
2290 | This timer is useful for profiling in interpreters. The interval timer | |
2291 | mechanism does not have the fine granularity necessary for profiling | |
2292 | native code. | |
2293 | @c @xref{profil} !!! | |
2294 | @end itemize | |
2295 | ||
2296 | You can only have one timer of each kind set at any given time. If you | |
2297 | set a timer that has not yet expired, that timer is simply reset to the | |
2298 | new value. | |
2299 | ||
2300 | You should establish a handler for the appropriate alarm signal using | |
76c23bac UD |
2301 | @code{signal} or @code{sigaction} before issuing a call to |
2302 | @code{setitimer} or @code{alarm}. Otherwise, an unusual chain of events | |
2303 | could cause the timer to expire before your program establishes the | |
2304 | handler. In this case it would be terminated, since termination is the | |
2305 | default action for the alarm signals. @xref{Signal Handling}. | |
28f540f4 | 2306 | |
f51dadcc UD |
2307 | To be able to use the alarm function to interrupt a system call which |
2308 | might block otherwise indefinitely it is important to @emph{not} set the | |
2309 | @code{SA_RESTART} flag when registering the signal handler using | |
2310 | @code{sigaction}. When not using @code{sigaction} things get even | |
2311 | uglier: the @code{signal} function has to fixed semantics with respect | |
2312 | to restarts. The BSD semantics for this function is to set the flag. | |
2313 | Therefore, if @code{sigaction} for whatever reason cannot be used, it is | |
2314 | necessary to use @code{sysv_signal} and not @code{signal}. | |
2315 | ||
28f540f4 RM |
2316 | The @code{setitimer} function is the primary means for setting an alarm. |
2317 | This facility is declared in the header file @file{sys/time.h}. The | |
2318 | @code{alarm} function, declared in @file{unistd.h}, provides a somewhat | |
2319 | simpler interface for setting the real-time timer. | |
2320 | @pindex unistd.h | |
2321 | @pindex sys/time.h | |
2322 | ||
2323 | @comment sys/time.h | |
2324 | @comment BSD | |
2325 | @deftp {Data Type} {struct itimerval} | |
2326 | This structure is used to specify when a timer should expire. It contains | |
2327 | the following members: | |
2328 | @table @code | |
2329 | @item struct timeval it_interval | |
99a20616 | 2330 | This is the period between successive timer interrupts. If zero, the |
28f540f4 RM |
2331 | alarm will only be sent once. |
2332 | ||
2333 | @item struct timeval it_value | |
99a20616 UD |
2334 | This is the period between now and the first timer interrupt. If zero, |
2335 | the alarm is disabled. | |
28f540f4 RM |
2336 | @end table |
2337 | ||
99a20616 | 2338 | The @code{struct timeval} data type is described in @ref{Elapsed Time}. |
28f540f4 RM |
2339 | @end deftp |
2340 | ||
2341 | @comment sys/time.h | |
2342 | @comment BSD | |
2343 | @deftypefun int setitimer (int @var{which}, struct itimerval *@var{new}, struct itimerval *@var{old}) | |
01cdeca0 | 2344 | The @code{setitimer} function sets the timer specified by @var{which} |
28f540f4 RM |
2345 | according to @var{new}. The @var{which} argument can have a value of |
2346 | @code{ITIMER_REAL}, @code{ITIMER_VIRTUAL}, or @code{ITIMER_PROF}. | |
2347 | ||
2348 | If @var{old} is not a null pointer, @code{setitimer} returns information | |
2349 | about any previous unexpired timer of the same kind in the structure it | |
2350 | points to. | |
2351 | ||
2352 | The return value is @code{0} on success and @code{-1} on failure. The | |
2353 | following @code{errno} error conditions are defined for this function: | |
2354 | ||
2355 | @table @code | |
2356 | @item EINVAL | |
99a20616 | 2357 | The timer period is too large. |
28f540f4 RM |
2358 | @end table |
2359 | @end deftypefun | |
2360 | ||
2361 | @comment sys/time.h | |
2362 | @comment BSD | |
2363 | @deftypefun int getitimer (int @var{which}, struct itimerval *@var{old}) | |
2364 | The @code{getitimer} function stores information about the timer specified | |
2365 | by @var{which} in the structure pointed at by @var{old}. | |
2366 | ||
2367 | The return value and error conditions are the same as for @code{setitimer}. | |
2368 | @end deftypefun | |
2369 | ||
2370 | @comment sys/time.h | |
2371 | @comment BSD | |
b642f101 | 2372 | @vtable @code |
28f540f4 | 2373 | @item ITIMER_REAL |
28f540f4 RM |
2374 | This constant can be used as the @var{which} argument to the |
2375 | @code{setitimer} and @code{getitimer} functions to specify the real-time | |
2376 | timer. | |
2377 | ||
2378 | @comment sys/time.h | |
2379 | @comment BSD | |
2380 | @item ITIMER_VIRTUAL | |
28f540f4 RM |
2381 | This constant can be used as the @var{which} argument to the |
2382 | @code{setitimer} and @code{getitimer} functions to specify the virtual | |
2383 | timer. | |
2384 | ||
2385 | @comment sys/time.h | |
2386 | @comment BSD | |
2387 | @item ITIMER_PROF | |
28f540f4 RM |
2388 | This constant can be used as the @var{which} argument to the |
2389 | @code{setitimer} and @code{getitimer} functions to specify the profiling | |
2390 | timer. | |
b642f101 | 2391 | @end vtable |
28f540f4 RM |
2392 | |
2393 | @comment unistd.h | |
2394 | @comment POSIX.1 | |
2395 | @deftypefun {unsigned int} alarm (unsigned int @var{seconds}) | |
2396 | The @code{alarm} function sets the real-time timer to expire in | |
2397 | @var{seconds} seconds. If you want to cancel any existing alarm, you | |
2398 | can do this by calling @code{alarm} with a @var{seconds} argument of | |
2399 | zero. | |
2400 | ||
2401 | The return value indicates how many seconds remain before the previous | |
2402 | alarm would have been sent. If there is no previous alarm, @code{alarm} | |
2403 | returns zero. | |
2404 | @end deftypefun | |
2405 | ||
2406 | The @code{alarm} function could be defined in terms of @code{setitimer} | |
2407 | like this: | |
2408 | ||
2409 | @smallexample | |
2410 | unsigned int | |
2411 | alarm (unsigned int seconds) | |
2412 | @{ | |
2413 | struct itimerval old, new; | |
2414 | new.it_interval.tv_usec = 0; | |
2415 | new.it_interval.tv_sec = 0; | |
2416 | new.it_value.tv_usec = 0; | |
2417 | new.it_value.tv_sec = (long int) seconds; | |
2418 | if (setitimer (ITIMER_REAL, &new, &old) < 0) | |
2419 | return 0; | |
2420 | else | |
2421 | return old.it_value.tv_sec; | |
2422 | @} | |
2423 | @end smallexample | |
2424 | ||
2425 | There is an example showing the use of the @code{alarm} function in | |
2426 | @ref{Handler Returns}. | |
2427 | ||
2428 | If you simply want your process to wait for a given number of seconds, | |
2429 | you should use the @code{sleep} function. @xref{Sleeping}. | |
2430 | ||
2431 | You shouldn't count on the signal arriving precisely when the timer | |
2432 | expires. In a multiprocessing environment there is typically some | |
2433 | amount of delay involved. | |
2434 | ||
2435 | @strong{Portability Note:} The @code{setitimer} and @code{getitimer} | |
2436 | functions are derived from BSD Unix, while the @code{alarm} function is | |
2437 | specified by the POSIX.1 standard. @code{setitimer} is more powerful than | |
2438 | @code{alarm}, but @code{alarm} is more widely used. | |
2439 | ||
2440 | @node Sleeping | |
2441 | @section Sleeping | |
2442 | ||
2443 | The function @code{sleep} gives a simple way to make the program wait | |
99a20616 UD |
2444 | for a short interval. If your program doesn't use signals (except to |
2445 | terminate), then you can expect @code{sleep} to wait reliably throughout | |
2446 | the specified interval. Otherwise, @code{sleep} can return sooner if a | |
2447 | signal arrives; if you want to wait for a given interval regardless of | |
2448 | signals, use @code{select} (@pxref{Waiting for I/O}) and don't specify | |
2449 | any descriptors to wait for. | |
28f540f4 RM |
2450 | @c !!! select can get EINTR; using SA_RESTART makes sleep win too. |
2451 | ||
2452 | @comment unistd.h | |
2453 | @comment POSIX.1 | |
2454 | @deftypefun {unsigned int} sleep (unsigned int @var{seconds}) | |
2455 | The @code{sleep} function waits for @var{seconds} or until a signal | |
01cdeca0 | 2456 | is delivered, whichever happens first. |
28f540f4 | 2457 | |
99a20616 UD |
2458 | If @code{sleep} function returns because the requested interval is over, |
2459 | it returns a value of zero. If it returns because of delivery of a | |
2460 | signal, its return value is the remaining time in the sleep interval. | |
28f540f4 RM |
2461 | |
2462 | The @code{sleep} function is declared in @file{unistd.h}. | |
2463 | @end deftypefun | |
2464 | ||
2465 | Resist the temptation to implement a sleep for a fixed amount of time by | |
2466 | using the return value of @code{sleep}, when nonzero, to call | |
2467 | @code{sleep} again. This will work with a certain amount of accuracy as | |
2468 | long as signals arrive infrequently. But each signal can cause the | |
2469 | eventual wakeup time to be off by an additional second or so. Suppose a | |
2470 | few signals happen to arrive in rapid succession by bad luck---there is | |
2471 | no limit on how much this could shorten or lengthen the wait. | |
2472 | ||
99a20616 UD |
2473 | Instead, compute the calendar time at which the program should stop |
2474 | waiting, and keep trying to wait until that calendar time. This won't | |
2475 | be off by more than a second. With just a little more work, you can use | |
2476 | @code{select} and make the waiting period quite accurate. (Of course, | |
2477 | heavy system load can cause additional unavoidable delays---unless the | |
2478 | machine is dedicated to one application, there is no way you can avoid | |
2479 | this.) | |
28f540f4 RM |
2480 | |
2481 | On some systems, @code{sleep} can do strange things if your program uses | |
2482 | @code{SIGALRM} explicitly. Even if @code{SIGALRM} signals are being | |
2483 | ignored or blocked when @code{sleep} is called, @code{sleep} might | |
2484 | return prematurely on delivery of a @code{SIGALRM} signal. If you have | |
2485 | established a handler for @code{SIGALRM} signals and a @code{SIGALRM} | |
2486 | signal is delivered while the process is sleeping, the action taken | |
2487 | might be just to cause @code{sleep} to return instead of invoking your | |
2488 | handler. And, if @code{sleep} is interrupted by delivery of a signal | |
2489 | whose handler requests an alarm or alters the handling of @code{SIGALRM}, | |
2490 | this handler and @code{sleep} will interfere. | |
2491 | ||
2492 | On the GNU system, it is safe to use @code{sleep} and @code{SIGALRM} in | |
2493 | the same program, because @code{sleep} does not work by means of | |
2494 | @code{SIGALRM}. | |
2495 | ||
dfd2257a UD |
2496 | @comment time.h |
2497 | @comment POSIX.1 | |
2498 | @deftypefun int nanosleep (const struct timespec *@var{requested_time}, struct timespec *@var{remaining}) | |
99a20616 UD |
2499 | If resolution to seconds is not enough the @code{nanosleep} function can |
2500 | be used. As the name suggests the sleep interval can be specified in | |
2501 | nanoseconds. The actual elapsed time of the sleep interval might be | |
2502 | longer since the system rounds the elapsed time you request up to the | |
2503 | next integer multiple of the actual resolution the system can deliver. | |
dfd2257a | 2504 | |
99a20616 UD |
2505 | *@code{requested_time} is the elapsed time of the interval you want to |
2506 | sleep. | |
2507 | ||
2508 | The function returns as *@code{remaining} the elapsed time left in the | |
2509 | interval for which you requested to sleep. If the interval completed | |
2510 | without getting interrupted by a signal, this is zero. | |
2511 | ||
b642f101 | 2512 | @code{struct timespec} is described in @xref{Elapsed Time}. |
99a20616 UD |
2513 | |
2514 | If the function returns because the interval is over the return value is | |
591e1ffb | 2515 | zero. If the function returns @math{-1} the global variable @var{errno} |
dfd2257a UD |
2516 | is set to the following values: |
2517 | ||
2518 | @table @code | |
2519 | @item EINTR | |
2520 | The call was interrupted because a signal was delivered to the thread. | |
2521 | If the @var{remaining} parameter is not the null pointer the structure | |
99a20616 UD |
2522 | pointed to by @var{remaining} is updated to contain the remaining |
2523 | elapsed time. | |
dfd2257a UD |
2524 | |
2525 | @item EINVAL | |
2526 | The nanosecond value in the @var{requested_time} parameter contains an | |
2527 | illegal value. Either the value is negative or greater than or equal to | |
2528 | 1000 million. | |
2529 | @end table | |
2530 | ||
76c23bac | 2531 | This function is a cancellation point in multi-threaded programs. This |
dfd2257a UD |
2532 | is a problem if the thread allocates some resources (like memory, file |
2533 | descriptors, semaphores or whatever) at the time @code{nanosleep} is | |
2534 | called. If the thread gets canceled these resources stay allocated | |
2535 | until the program ends. To avoid this calls to @code{nanosleep} should | |
76c23bac | 2536 | be protected using cancellation handlers. |
dfd2257a UD |
2537 | @c ref pthread_cleanup_push / pthread_cleanup_pop |
2538 | ||
2539 | The @code{nanosleep} function is declared in @file{time.h}. | |
2540 | @end deftypefun |