]>
Commit | Line | Data |
---|---|---|
28f540f4 | 1 | @node I/O on Streams, Low-Level I/O, I/O Overview, Top |
34992338 | 2 | @c %MENU% High-level, portable I/O facilities |
28f540f4 | 3 | @chapter Input/Output on Streams |
838e5ffe UD |
4 | @c fix an overfull: |
5 | @tex | |
6 | \hyphenation{which-ever} | |
7 | @end tex | |
28f540f4 RM |
8 | |
9 | This chapter describes the functions for creating streams and performing | |
10 | input and output operations on them. As discussed in @ref{I/O | |
11 | Overview}, a stream is a fairly abstract, high-level concept | |
12 | representing a communications channel to a file, device, or process. | |
13 | ||
14 | @menu | |
15 | * Streams:: About the data type representing a stream. | |
19c3f208 | 16 | * Standard Streams:: Streams to the standard input and output |
28f540f4 RM |
17 | devices are created for you. |
18 | * Opening Streams:: How to create a stream to talk to a file. | |
19 | * Closing Streams:: Close a stream when you are finished with it. | |
20 | * Simple Output:: Unformatted output by characters and lines. | |
21 | * Character Input:: Unformatted input by characters and words. | |
22 | * Line Input:: Reading a line or a record from a stream. | |
23 | * Unreading:: Peeking ahead/pushing back input just read. | |
24 | * Block Input/Output:: Input and output operations on blocks of data. | |
25 | * Formatted Output:: @code{printf} and related functions. | |
26 | * Customizing Printf:: You can define new conversion specifiers for | |
27 | @code{printf} and friends. | |
28 | * Formatted Input:: @code{scanf} and related functions. | |
29 | * EOF and Errors:: How you can tell if an I/O error happens. | |
30 | * Binary Streams:: Some systems distinguish between text files | |
31 | and binary files. | |
32 | * File Positioning:: About random-access streams. | |
f65fd747 | 33 | * Portable Positioning:: Random access on peculiar ISO C systems. |
28f540f4 RM |
34 | * Stream Buffering:: How to control buffering of streams. |
35 | * Other Kinds of Streams:: Streams that do not necessarily correspond | |
19c3f208 | 36 | to an open file. |
0501d603 | 37 | * Formatted Messages:: Print strictly formatted messages. |
28f540f4 RM |
38 | @end menu |
39 | ||
40 | @node Streams | |
41 | @section Streams | |
42 | ||
43 | For historical reasons, the type of the C data structure that represents | |
44 | a stream is called @code{FILE} rather than ``stream''. Since most of | |
45 | the library functions deal with objects of type @code{FILE *}, sometimes | |
46 | the term @dfn{file pointer} is also used to mean ``stream''. This leads | |
47 | to unfortunate confusion over terminology in many books on C. This | |
48 | manual, however, is careful to use the terms ``file'' and ``stream'' | |
49 | only in the technical sense. | |
50 | @cindex file pointer | |
51 | ||
52 | @pindex stdio.h | |
53 | The @code{FILE} type is declared in the header file @file{stdio.h}. | |
54 | ||
55 | @comment stdio.h | |
f65fd747 | 56 | @comment ISO |
28f540f4 RM |
57 | @deftp {Data Type} FILE |
58 | This is the data type used to represent stream objects. A @code{FILE} | |
59 | object holds all of the internal state information about the connection | |
60 | to the associated file, including such things as the file position | |
61 | indicator and buffering information. Each stream also has error and | |
62 | end-of-file status indicators that can be tested with the @code{ferror} | |
63 | and @code{feof} functions; see @ref{EOF and Errors}. | |
64 | @end deftp | |
65 | ||
66 | @code{FILE} objects are allocated and managed internally by the | |
67 | input/output library functions. Don't try to create your own objects of | |
68 | type @code{FILE}; let the library do it. Your programs should | |
69 | deal only with pointers to these objects (that is, @code{FILE *} values) | |
70 | rather than the objects themselves. | |
6d52618b | 71 | @c !!! should say that FILE's have "No user-serviceable parts inside." |
28f540f4 RM |
72 | |
73 | @node Standard Streams | |
74 | @section Standard Streams | |
75 | @cindex standard streams | |
76 | @cindex streams, standard | |
77 | ||
78 | When the @code{main} function of your program is invoked, it already has | |
79 | three predefined streams open and available for use. These represent | |
80 | the ``standard'' input and output channels that have been established | |
81 | for the process. | |
82 | ||
83 | These streams are declared in the header file @file{stdio.h}. | |
84 | @pindex stdio.h | |
85 | ||
86 | @comment stdio.h | |
f65fd747 | 87 | @comment ISO |
28f540f4 RM |
88 | @deftypevar {FILE *} stdin |
89 | The @dfn{standard input} stream, which is the normal source of input for the | |
90 | program. | |
91 | @end deftypevar | |
92 | @cindex standard input stream | |
93 | ||
94 | @comment stdio.h | |
f65fd747 | 95 | @comment ISO |
28f540f4 RM |
96 | @deftypevar {FILE *} stdout |
97 | The @dfn{standard output} stream, which is used for normal output from | |
98 | the program. | |
99 | @end deftypevar | |
100 | @cindex standard output stream | |
101 | ||
102 | @comment stdio.h | |
f65fd747 | 103 | @comment ISO |
28f540f4 RM |
104 | @deftypevar {FILE *} stderr |
105 | The @dfn{standard error} stream, which is used for error messages and | |
106 | diagnostics issued by the program. | |
107 | @end deftypevar | |
108 | @cindex standard error stream | |
109 | ||
110 | In the GNU system, you can specify what files or processes correspond to | |
111 | these streams using the pipe and redirection facilities provided by the | |
112 | shell. (The primitives shells use to implement these facilities are | |
113 | described in @ref{File System Interface}.) Most other operating systems | |
114 | provide similar mechanisms, but the details of how to use them can vary. | |
115 | ||
116 | In the GNU C library, @code{stdin}, @code{stdout}, and @code{stderr} are | |
117 | normal variables which you can set just like any others. For example, to redirect | |
118 | the standard output to a file, you could do: | |
119 | ||
120 | @smallexample | |
121 | fclose (stdout); | |
122 | stdout = fopen ("standard-output-file", "w"); | |
123 | @end smallexample | |
124 | ||
125 | Note however, that in other systems @code{stdin}, @code{stdout}, and | |
126 | @code{stderr} are macros that you cannot assign to in the normal way. | |
127 | But you can use @code{freopen} to get the effect of closing one and | |
128 | reopening it. @xref{Opening Streams}. | |
129 | ||
130 | @node Opening Streams | |
131 | @section Opening Streams | |
132 | ||
133 | @cindex opening a stream | |
134 | Opening a file with the @code{fopen} function creates a new stream and | |
135 | establishes a connection between the stream and a file. This may | |
19c3f208 | 136 | involve creating a new file. |
28f540f4 RM |
137 | |
138 | @pindex stdio.h | |
139 | Everything described in this section is declared in the header file | |
140 | @file{stdio.h}. | |
141 | ||
142 | @comment stdio.h | |
f65fd747 | 143 | @comment ISO |
28f540f4 RM |
144 | @deftypefun {FILE *} fopen (const char *@var{filename}, const char *@var{opentype}) |
145 | The @code{fopen} function opens a stream for I/O to the file | |
146 | @var{filename}, and returns a pointer to the stream. | |
147 | ||
148 | The @var{opentype} argument is a string that controls how the file is | |
149 | opened and specifies attributes of the resulting stream. It must begin | |
150 | with one of the following sequences of characters: | |
151 | ||
152 | @table @samp | |
153 | @item r | |
154 | Open an existing file for reading only. | |
155 | ||
156 | @item w | |
157 | Open the file for writing only. If the file already exists, it is | |
158 | truncated to zero length. Otherwise a new file is created. | |
159 | ||
160 | @item a | |
161 | Open a file for append access; that is, writing at the end of file only. | |
162 | If the file already exists, its initial contents are unchanged and | |
163 | output to the stream is appended to the end of the file. | |
164 | Otherwise, a new, empty file is created. | |
165 | ||
166 | @item r+ | |
167 | Open an existing file for both reading and writing. The initial contents | |
168 | of the file are unchanged and the initial file position is at the | |
169 | beginning of the file. | |
170 | ||
171 | @item w+ | |
172 | Open a file for both reading and writing. If the file already exists, it | |
173 | is truncated to zero length. Otherwise, a new file is created. | |
174 | ||
175 | @item a+ | |
176 | Open or create file for both reading and appending. If the file exists, | |
177 | its initial contents are unchanged. Otherwise, a new file is created. | |
178 | The initial file position for reading is at the beginning of the file, | |
179 | but output is always appended to the end of the file. | |
180 | @end table | |
181 | ||
182 | As you can see, @samp{+} requests a stream that can do both input and | |
f65fd747 | 183 | output. The ISO standard says that when using such a stream, you must |
28f540f4 RM |
184 | call @code{fflush} (@pxref{Stream Buffering}) or a file positioning |
185 | function such as @code{fseek} (@pxref{File Positioning}) when switching | |
186 | from reading to writing or vice versa. Otherwise, internal buffers | |
187 | might not be emptied properly. The GNU C library does not have this | |
188 | limitation; you can do arbitrary reading and writing operations on a | |
189 | stream in whatever order. | |
190 | ||
191 | Additional characters may appear after these to specify flags for the | |
192 | call. Always put the mode (@samp{r}, @samp{w+}, etc.) first; that is | |
193 | the only part you are guaranteed will be understood by all systems. | |
194 | ||
195 | The GNU C library defines one additional character for use in | |
196 | @var{opentype}: the character @samp{x} insists on creating a new | |
197 | file---if a file @var{filename} already exists, @code{fopen} fails | |
0be8752b | 198 | rather than opening it. If you use @samp{x} you are guaranteed that |
28f540f4 RM |
199 | you will not clobber an existing file. This is equivalent to the |
200 | @code{O_EXCL} option to the @code{open} function (@pxref{Opening and | |
201 | Closing Files}). | |
202 | ||
203 | The character @samp{b} in @var{opentype} has a standard meaning; it | |
204 | requests a binary stream rather than a text stream. But this makes no | |
205 | difference in POSIX systems (including the GNU system). If both | |
206 | @samp{+} and @samp{b} are specified, they can appear in either order. | |
207 | @xref{Binary Streams}. | |
208 | ||
4e2e9999 UD |
209 | @cindex stream orientation |
210 | @cindex orientation, stream | |
211 | If the @var{opentype} string contains the sequence | |
212 | @code{,ccs=@var{STRING}} then @var{STRING} is taken as the name of a | |
213 | coded character set and @code{fopen} will mark the stream as | |
214 | wide-oriented which appropriate conversion functions in place to convert | |
215 | from and to the character set @var{STRING} is place. Any other stream | |
216 | is opened initially unoriented and the orientation is decided with the | |
217 | first file operation. If the first operation is a wide character | |
218 | operation, the stream is not only marked as wide-oriented, also the | |
219 | conversion functions to convert to the coded character set used for the | |
220 | current locale are loaded. This will not change anymore from this point | |
221 | on even if the locale selected for the @code{LC_CTYPE} category is | |
222 | changed. | |
223 | ||
28f540f4 RM |
224 | Any other characters in @var{opentype} are simply ignored. They may be |
225 | meaningful in other systems. | |
226 | ||
227 | If the open fails, @code{fopen} returns a null pointer. | |
a3a4a74e UD |
228 | |
229 | When the sources are compiling with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 230 | 32 bit machine this function is in fact @code{fopen64} since the LFS |
a3a4a74e | 231 | interface replaces transparently the old interface. |
28f540f4 RM |
232 | @end deftypefun |
233 | ||
234 | You can have multiple streams (or file descriptors) pointing to the same | |
235 | file open at the same time. If you do only input, this works | |
236 | straightforwardly, but you must be careful if any output streams are | |
237 | included. @xref{Stream/Descriptor Precautions}. This is equally true | |
238 | whether the streams are in one program (not usual) or in several | |
239 | programs (which can easily happen). It may be advantageous to use the | |
240 | file locking facilities to avoid simultaneous access. @xref{File | |
241 | Locks}. | |
242 | ||
a3a4a74e UD |
243 | @comment stdio.h |
244 | @comment Unix98 | |
245 | @deftypefun {FILE *} fopen64 (const char *@var{filename}, const char *@var{opentype}) | |
246 | This function is similar to @code{fopen} but the stream it returns a | |
247 | pointer for is opened using @code{open64}. Therefore this stream can be | |
0be8752b | 248 | used even on files larger then @math{2^31} bytes on 32 bit machines. |
a3a4a74e UD |
249 | |
250 | Please note that the return type is still @code{FILE *}. There is no | |
251 | special @code{FILE} type for the LFS interface. | |
252 | ||
253 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
254 | bits machine this function is available under the name @code{fopen} | |
255 | and so transparently replaces the old interface. | |
256 | @end deftypefun | |
257 | ||
28f540f4 | 258 | @comment stdio.h |
f65fd747 | 259 | @comment ISO |
28f540f4 RM |
260 | @deftypevr Macro int FOPEN_MAX |
261 | The value of this macro is an integer constant expression that | |
262 | represents the minimum number of streams that the implementation | |
263 | guarantees can be open simultaneously. You might be able to open more | |
264 | than this many streams, but that is not guaranteed. The value of this | |
265 | constant is at least eight, which includes the three standard streams | |
266 | @code{stdin}, @code{stdout}, and @code{stderr}. In POSIX.1 systems this | |
267 | value is determined by the @code{OPEN_MAX} parameter; @pxref{General | |
268 | Limits}. In BSD and GNU, it is controlled by the @code{RLIMIT_NOFILE} | |
269 | resource limit; @pxref{Limits on Resources}. | |
270 | @end deftypevr | |
271 | ||
272 | @comment stdio.h | |
f65fd747 | 273 | @comment ISO |
28f540f4 RM |
274 | @deftypefun {FILE *} freopen (const char *@var{filename}, const char *@var{opentype}, FILE *@var{stream}) |
275 | This function is like a combination of @code{fclose} and @code{fopen}. | |
276 | It first closes the stream referred to by @var{stream}, ignoring any | |
277 | errors that are detected in the process. (Because errors are ignored, | |
278 | you should not use @code{freopen} on an output stream if you have | |
279 | actually done any output using the stream.) Then the file named by | |
280 | @var{filename} is opened with mode @var{opentype} as for @code{fopen}, | |
281 | and associated with the same stream object @var{stream}. | |
282 | ||
283 | If the operation fails, a null pointer is returned; otherwise, | |
284 | @code{freopen} returns @var{stream}. | |
285 | ||
286 | @code{freopen} has traditionally been used to connect a standard stream | |
287 | such as @code{stdin} with a file of your own choice. This is useful in | |
288 | programs in which use of a standard stream for certain purposes is | |
289 | hard-coded. In the GNU C library, you can simply close the standard | |
290 | streams and open new ones with @code{fopen}. But other systems lack | |
291 | this ability, so using @code{freopen} is more portable. | |
a3a4a74e UD |
292 | |
293 | When the sources are compiling with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 294 | 32 bit machine this function is in fact @code{freopen64} since the LFS |
a3a4a74e UD |
295 | interface replaces transparently the old interface. |
296 | @end deftypefun | |
297 | ||
298 | @comment stdio.h | |
299 | @comment Unix98 | |
300 | @deftypefun {FILE *} freopen64 (const char *@var{filename}, const char *@var{opentype}, FILE *@var{stream}) | |
301 | This function is similar to @code{freopen}. The only difference is that | |
0be8752b | 302 | on 32 bit machine the stream returned is able to read beyond the |
a3a4a74e UD |
303 | @math{2^31} bytes limits imposed by the normal interface. It should be |
304 | noted that the stream pointed to by @var{stream} need not be opened | |
305 | using @code{fopen64} or @code{freopen64} since its mode is not important | |
306 | for this function. | |
307 | ||
308 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
309 | bits machine this function is available under the name @code{freopen} | |
310 | and so transparently replaces the old interface. | |
28f540f4 RM |
311 | @end deftypefun |
312 | ||
313 | ||
314 | @node Closing Streams | |
315 | @section Closing Streams | |
316 | ||
317 | @cindex closing a stream | |
318 | When a stream is closed with @code{fclose}, the connection between the | |
319 | stream and the file is cancelled. After you have closed a stream, you | |
320 | cannot perform any additional operations on it. | |
321 | ||
322 | @comment stdio.h | |
f65fd747 | 323 | @comment ISO |
28f540f4 RM |
324 | @deftypefun int fclose (FILE *@var{stream}) |
325 | This function causes @var{stream} to be closed and the connection to | |
326 | the corresponding file to be broken. Any buffered output is written | |
327 | and any buffered input is discarded. The @code{fclose} function returns | |
328 | a value of @code{0} if the file was closed successfully, and @code{EOF} | |
19c3f208 | 329 | if an error was detected. |
28f540f4 RM |
330 | |
331 | It is important to check for errors when you call @code{fclose} to close | |
332 | an output stream, because real, everyday errors can be detected at this | |
333 | time. For example, when @code{fclose} writes the remaining buffered | |
334 | output, it might get an error because the disk is full. Even if you | |
335 | know the buffer is empty, errors can still occur when closing a file if | |
336 | you are using NFS. | |
337 | ||
338 | The function @code{fclose} is declared in @file{stdio.h}. | |
339 | @end deftypefun | |
340 | ||
6bc31da0 UD |
341 | To close all streams currently available the GNU C Library provides |
342 | another function. | |
343 | ||
344 | @comment stdio.h | |
345 | @comment GNU | |
346 | @deftypefun int fcloseall (void) | |
347 | This function causes all open streams of the process to be closed and | |
348 | the connection to corresponding files to be broken. All buffered data | |
f2ea0f5b | 349 | is written and any buffered input is discarded. The @code{fcloseall} |
6bc31da0 UD |
350 | function returns a value of @code{0} if all the files were closed |
351 | successfully, and @code{EOF} if an error was detected. | |
352 | ||
0be8752b | 353 | This function should be used only in special situations, e.g., when an |
6bc31da0 | 354 | error occurred and the program must be aborted. Normally each single |
0be8752b UD |
355 | stream should be closed separately so that problems with individual |
356 | streams can be identified. It is also problematic since the standard | |
357 | streams (@pxref{Standard Streams}) will also be closed. | |
6bc31da0 UD |
358 | |
359 | The function @code{fcloseall} is declared in @file{stdio.h}. | |
360 | @end deftypefun | |
361 | ||
28f540f4 RM |
362 | If the @code{main} function to your program returns, or if you call the |
363 | @code{exit} function (@pxref{Normal Termination}), all open streams are | |
364 | automatically closed properly. If your program terminates in any other | |
365 | manner, such as by calling the @code{abort} function (@pxref{Aborting a | |
366 | Program}) or from a fatal signal (@pxref{Signal Handling}), open streams | |
367 | might not be closed properly. Buffered output might not be flushed and | |
368 | files may be incomplete. For more information on buffering of streams, | |
369 | see @ref{Stream Buffering}. | |
370 | ||
371 | @node Simple Output | |
372 | @section Simple Output by Characters or Lines | |
373 | ||
374 | @cindex writing to a stream, by characters | |
375 | This section describes functions for performing character- and | |
376 | line-oriented output. | |
377 | ||
378 | These functions are declared in the header file @file{stdio.h}. | |
379 | @pindex stdio.h | |
380 | ||
381 | @comment stdio.h | |
f65fd747 | 382 | @comment ISO |
28f540f4 RM |
383 | @deftypefun int fputc (int @var{c}, FILE *@var{stream}) |
384 | The @code{fputc} function converts the character @var{c} to type | |
19c3f208 | 385 | @code{unsigned char}, and writes it to the stream @var{stream}. |
28f540f4 RM |
386 | @code{EOF} is returned if a write error occurs; otherwise the |
387 | character @var{c} is returned. | |
388 | @end deftypefun | |
389 | ||
390 | @comment stdio.h | |
f65fd747 | 391 | @comment ISO |
28f540f4 RM |
392 | @deftypefun int putc (int @var{c}, FILE *@var{stream}) |
393 | This is just like @code{fputc}, except that most systems implement it as | |
394 | a macro, making it faster. One consequence is that it may evaluate the | |
395 | @var{stream} argument more than once, which is an exception to the | |
396 | general rule for macros. @code{putc} is usually the best function to | |
397 | use for writing a single character. | |
398 | @end deftypefun | |
399 | ||
400 | @comment stdio.h | |
f65fd747 | 401 | @comment ISO |
28f540f4 RM |
402 | @deftypefun int putchar (int @var{c}) |
403 | The @code{putchar} function is equivalent to @code{putc} with | |
404 | @code{stdout} as the value of the @var{stream} argument. | |
405 | @end deftypefun | |
406 | ||
407 | @comment stdio.h | |
f65fd747 | 408 | @comment ISO |
28f540f4 RM |
409 | @deftypefun int fputs (const char *@var{s}, FILE *@var{stream}) |
410 | The function @code{fputs} writes the string @var{s} to the stream | |
411 | @var{stream}. The terminating null character is not written. | |
412 | This function does @emph{not} add a newline character, either. | |
413 | It outputs only the characters in the string. | |
414 | ||
415 | This function returns @code{EOF} if a write error occurs, and otherwise | |
416 | a non-negative value. | |
417 | ||
418 | For example: | |
419 | ||
420 | @smallexample | |
421 | fputs ("Are ", stdout); | |
422 | fputs ("you ", stdout); | |
423 | fputs ("hungry?\n", stdout); | |
424 | @end smallexample | |
425 | ||
426 | @noindent | |
427 | outputs the text @samp{Are you hungry?} followed by a newline. | |
428 | @end deftypefun | |
429 | ||
430 | @comment stdio.h | |
f65fd747 | 431 | @comment ISO |
28f540f4 RM |
432 | @deftypefun int puts (const char *@var{s}) |
433 | The @code{puts} function writes the string @var{s} to the stream | |
434 | @code{stdout} followed by a newline. The terminating null character of | |
435 | the string is not written. (Note that @code{fputs} does @emph{not} | |
436 | write a newline as this function does.) | |
437 | ||
438 | @code{puts} is the most convenient function for printing simple | |
439 | messages. For example: | |
440 | ||
441 | @smallexample | |
442 | puts ("This is a message."); | |
443 | @end smallexample | |
0be8752b UD |
444 | |
445 | @noindent | |
446 | outputs the text @samp{This is a message.} followed by a newline. | |
28f540f4 RM |
447 | @end deftypefun |
448 | ||
449 | @comment stdio.h | |
450 | @comment SVID | |
451 | @deftypefun int putw (int @var{w}, FILE *@var{stream}) | |
452 | This function writes the word @var{w} (that is, an @code{int}) to | |
453 | @var{stream}. It is provided for compatibility with SVID, but we | |
454 | recommend you use @code{fwrite} instead (@pxref{Block Input/Output}). | |
455 | @end deftypefun | |
456 | ||
457 | @node Character Input | |
458 | @section Character Input | |
459 | ||
460 | @cindex reading from a stream, by characters | |
461 | This section describes functions for performing character-oriented input. | |
462 | These functions are declared in the header file @file{stdio.h}. | |
463 | @pindex stdio.h | |
464 | ||
465 | These functions return an @code{int} value that is either a character of | |
466 | input, or the special value @code{EOF} (usually -1). It is important to | |
467 | store the result of these functions in a variable of type @code{int} | |
468 | instead of @code{char}, even when you plan to use it only as a | |
469 | character. Storing @code{EOF} in a @code{char} variable truncates its | |
470 | value to the size of a character, so that it is no longer | |
471 | distinguishable from the valid character @samp{(char) -1}. So always | |
472 | use an @code{int} for the result of @code{getc} and friends, and check | |
473 | for @code{EOF} after the call; once you've verified that the result is | |
474 | not @code{EOF}, you can be sure that it will fit in a @samp{char} | |
475 | variable without loss of information. | |
476 | ||
477 | @comment stdio.h | |
f65fd747 | 478 | @comment ISO |
28f540f4 RM |
479 | @deftypefun int fgetc (FILE *@var{stream}) |
480 | This function reads the next character as an @code{unsigned char} from | |
481 | the stream @var{stream} and returns its value, converted to an | |
482 | @code{int}. If an end-of-file condition or read error occurs, | |
19c3f208 | 483 | @code{EOF} is returned instead. |
28f540f4 RM |
484 | @end deftypefun |
485 | ||
486 | @comment stdio.h | |
f65fd747 | 487 | @comment ISO |
28f540f4 RM |
488 | @deftypefun int getc (FILE *@var{stream}) |
489 | This is just like @code{fgetc}, except that it is permissible (and | |
490 | typical) for it to be implemented as a macro that evaluates the | |
491 | @var{stream} argument more than once. @code{getc} is often highly | |
492 | optimized, so it is usually the best function to use to read a single | |
493 | character. | |
494 | @end deftypefun | |
495 | ||
496 | @comment stdio.h | |
f65fd747 | 497 | @comment ISO |
28f540f4 RM |
498 | @deftypefun int getchar (void) |
499 | The @code{getchar} function is equivalent to @code{getc} with @code{stdin} | |
500 | as the value of the @var{stream} argument. | |
501 | @end deftypefun | |
502 | ||
503 | Here is an example of a function that does input using @code{fgetc}. It | |
504 | would work just as well using @code{getc} instead, or using | |
505 | @code{getchar ()} instead of @w{@code{fgetc (stdin)}}. | |
506 | ||
507 | @smallexample | |
508 | int | |
509 | y_or_n_p (const char *question) | |
510 | @{ | |
511 | fputs (question, stdout); | |
512 | while (1) | |
513 | @{ | |
514 | int c, answer; | |
515 | /* @r{Write a space to separate answer from question.} */ | |
516 | fputc (' ', stdout); | |
517 | /* @r{Read the first character of the line.} | |
518 | @r{This should be the answer character, but might not be.} */ | |
519 | c = tolower (fgetc (stdin)); | |
520 | answer = c; | |
521 | /* @r{Discard rest of input line.} */ | |
522 | while (c != '\n' && c != EOF) | |
523 | c = fgetc (stdin); | |
524 | /* @r{Obey the answer if it was valid.} */ | |
525 | if (answer == 'y') | |
526 | return 1; | |
527 | if (answer == 'n') | |
528 | return 0; | |
529 | /* @r{Answer was invalid: ask for valid answer.} */ | |
530 | fputs ("Please answer y or n:", stdout); | |
531 | @} | |
532 | @} | |
533 | @end smallexample | |
534 | ||
535 | @comment stdio.h | |
536 | @comment SVID | |
537 | @deftypefun int getw (FILE *@var{stream}) | |
538 | This function reads a word (that is, an @code{int}) from @var{stream}. | |
539 | It's provided for compatibility with SVID. We recommend you use | |
540 | @code{fread} instead (@pxref{Block Input/Output}). Unlike @code{getc}, | |
541 | any @code{int} value could be a valid result. @code{getw} returns | |
542 | @code{EOF} when it encounters end-of-file or an error, but there is no | |
543 | way to distinguish this from an input word with value -1. | |
544 | @end deftypefun | |
545 | ||
546 | @node Line Input | |
547 | @section Line-Oriented Input | |
548 | ||
549 | Since many programs interpret input on the basis of lines, it's | |
550 | convenient to have functions to read a line of text from a stream. | |
551 | ||
552 | Standard C has functions to do this, but they aren't very safe: null | |
553 | characters and even (for @code{gets}) long lines can confuse them. So | |
554 | the GNU library provides the nonstandard @code{getline} function that | |
555 | makes it easy to read lines reliably. | |
556 | ||
557 | Another GNU extension, @code{getdelim}, generalizes @code{getline}. It | |
558 | reads a delimited record, defined as everything through the next | |
559 | occurrence of a specified delimiter character. | |
560 | ||
561 | All these functions are declared in @file{stdio.h}. | |
562 | ||
563 | @comment stdio.h | |
564 | @comment GNU | |
565 | @deftypefun ssize_t getline (char **@var{lineptr}, size_t *@var{n}, FILE *@var{stream}) | |
566 | This function reads an entire line from @var{stream}, storing the text | |
567 | (including the newline and a terminating null character) in a buffer | |
568 | and storing the buffer address in @code{*@var{lineptr}}. | |
569 | ||
570 | Before calling @code{getline}, you should place in @code{*@var{lineptr}} | |
571 | the address of a buffer @code{*@var{n}} bytes long, allocated with | |
572 | @code{malloc}. If this buffer is long enough to hold the line, | |
573 | @code{getline} stores the line in this buffer. Otherwise, | |
574 | @code{getline} makes the buffer bigger using @code{realloc}, storing the | |
575 | new buffer address back in @code{*@var{lineptr}} and the increased size | |
576 | back in @code{*@var{n}}. | |
577 | @xref{Unconstrained Allocation}. | |
578 | ||
579 | If you set @code{*@var{lineptr}} to a null pointer, and @code{*@var{n}} | |
580 | to zero, before the call, then @code{getline} allocates the initial | |
581 | buffer for you by calling @code{malloc}. | |
582 | ||
583 | In either case, when @code{getline} returns, @code{*@var{lineptr}} is | |
584 | a @code{char *} which points to the text of the line. | |
585 | ||
586 | When @code{getline} is successful, it returns the number of characters | |
587 | read (including the newline, but not including the terminating null). | |
588 | This value enables you to distinguish null characters that are part of | |
589 | the line from the null character inserted as a terminator. | |
590 | ||
591 | This function is a GNU extension, but it is the recommended way to read | |
592 | lines from a stream. The alternative standard functions are unreliable. | |
593 | ||
594 | If an error occurs or end of file is reached, @code{getline} returns | |
595 | @code{-1}. | |
596 | @end deftypefun | |
597 | ||
598 | @comment stdio.h | |
599 | @comment GNU | |
600 | @deftypefun ssize_t getdelim (char **@var{lineptr}, size_t *@var{n}, int @var{delimiter}, FILE *@var{stream}) | |
601 | This function is like @code{getline} except that the character which | |
602 | tells it to stop reading is not necessarily newline. The argument | |
603 | @var{delimiter} specifies the delimiter character; @code{getdelim} keeps | |
604 | reading until it sees that character (or end of file). | |
605 | ||
606 | The text is stored in @var{lineptr}, including the delimiter character | |
607 | and a terminating null. Like @code{getline}, @code{getdelim} makes | |
608 | @var{lineptr} bigger if it isn't big enough. | |
609 | ||
610 | @code{getline} is in fact implemented in terms of @code{getdelim}, just | |
611 | like this: | |
612 | ||
613 | @smallexample | |
614 | ssize_t | |
615 | getline (char **lineptr, size_t *n, FILE *stream) | |
616 | @{ | |
617 | return getdelim (lineptr, n, '\n', stream); | |
618 | @} | |
619 | @end smallexample | |
620 | @end deftypefun | |
621 | ||
622 | @comment stdio.h | |
f65fd747 | 623 | @comment ISO |
28f540f4 RM |
624 | @deftypefun {char *} fgets (char *@var{s}, int @var{count}, FILE *@var{stream}) |
625 | The @code{fgets} function reads characters from the stream @var{stream} | |
626 | up to and including a newline character and stores them in the string | |
627 | @var{s}, adding a null character to mark the end of the string. You | |
628 | must supply @var{count} characters worth of space in @var{s}, but the | |
629 | number of characters read is at most @var{count} @minus{} 1. The extra | |
630 | character space is used to hold the null character at the end of the | |
631 | string. | |
632 | ||
633 | If the system is already at end of file when you call @code{fgets}, then | |
634 | the contents of the array @var{s} are unchanged and a null pointer is | |
635 | returned. A null pointer is also returned if a read error occurs. | |
636 | Otherwise, the return value is the pointer @var{s}. | |
637 | ||
638 | @strong{Warning:} If the input data has a null character, you can't tell. | |
639 | So don't use @code{fgets} unless you know the data cannot contain a null. | |
640 | Don't use it to read files edited by the user because, if the user inserts | |
641 | a null character, you should either handle it properly or print a clear | |
642 | error message. We recommend using @code{getline} instead of @code{fgets}. | |
643 | @end deftypefun | |
644 | ||
645 | @comment stdio.h | |
f65fd747 | 646 | @comment ISO |
28f540f4 RM |
647 | @deftypefn {Deprecated function} {char *} gets (char *@var{s}) |
648 | The function @code{gets} reads characters from the stream @code{stdin} | |
649 | up to the next newline character, and stores them in the string @var{s}. | |
650 | The newline character is discarded (note that this differs from the | |
651 | behavior of @code{fgets}, which copies the newline character into the | |
652 | string). If @code{gets} encounters a read error or end-of-file, it | |
653 | returns a null pointer; otherwise it returns @var{s}. | |
654 | ||
655 | @strong{Warning:} The @code{gets} function is @strong{very dangerous} | |
656 | because it provides no protection against overflowing the string | |
657 | @var{s}. The GNU library includes it for compatibility only. You | |
658 | should @strong{always} use @code{fgets} or @code{getline} instead. To | |
659 | remind you of this, the linker (if using GNU @code{ld}) will issue a | |
660 | warning whenever you use @code{gets}. | |
661 | @end deftypefn | |
662 | ||
663 | @node Unreading | |
664 | @section Unreading | |
665 | @cindex peeking at input | |
666 | @cindex unreading characters | |
667 | @cindex pushing input back | |
668 | ||
669 | In parser programs it is often useful to examine the next character in | |
670 | the input stream without removing it from the stream. This is called | |
671 | ``peeking ahead'' at the input because your program gets a glimpse of | |
672 | the input it will read next. | |
673 | ||
674 | Using stream I/O, you can peek ahead at input by first reading it and | |
19c3f208 | 675 | then @dfn{unreading} it (also called @dfn{pushing it back} on the stream). |
28f540f4 RM |
676 | Unreading a character makes it available to be input again from the stream, |
677 | by the next call to @code{fgetc} or other input function on that stream. | |
678 | ||
679 | @menu | |
680 | * Unreading Idea:: An explanation of unreading with pictures. | |
681 | * How Unread:: How to call @code{ungetc} to do unreading. | |
682 | @end menu | |
683 | ||
684 | @node Unreading Idea | |
685 | @subsection What Unreading Means | |
686 | ||
687 | Here is a pictorial explanation of unreading. Suppose you have a | |
688 | stream reading a file that contains just six characters, the letters | |
689 | @samp{foobar}. Suppose you have read three characters so far. The | |
690 | situation looks like this: | |
691 | ||
692 | @smallexample | |
693 | f o o b a r | |
694 | ^ | |
695 | @end smallexample | |
696 | ||
697 | @noindent | |
698 | so the next input character will be @samp{b}. | |
699 | ||
700 | @c @group Invalid outside @example | |
701 | If instead of reading @samp{b} you unread the letter @samp{o}, you get a | |
702 | situation like this: | |
703 | ||
704 | @smallexample | |
705 | f o o b a r | |
706 | | | |
707 | o-- | |
708 | ^ | |
709 | @end smallexample | |
710 | ||
711 | @noindent | |
712 | so that the next input characters will be @samp{o} and @samp{b}. | |
713 | @c @end group | |
714 | ||
715 | @c @group | |
716 | If you unread @samp{9} instead of @samp{o}, you get this situation: | |
717 | ||
718 | @smallexample | |
719 | f o o b a r | |
720 | | | |
721 | 9-- | |
722 | ^ | |
723 | @end smallexample | |
724 | ||
725 | @noindent | |
726 | so that the next input characters will be @samp{9} and @samp{b}. | |
727 | @c @end group | |
728 | ||
729 | @node How Unread | |
730 | @subsection Using @code{ungetc} To Do Unreading | |
19c3f208 | 731 | |
28f540f4 RM |
732 | The function to unread a character is called @code{ungetc}, because it |
733 | reverses the action of @code{getc}. | |
734 | ||
735 | @comment stdio.h | |
f65fd747 | 736 | @comment ISO |
28f540f4 RM |
737 | @deftypefun int ungetc (int @var{c}, FILE *@var{stream}) |
738 | The @code{ungetc} function pushes back the character @var{c} onto the | |
739 | input stream @var{stream}. So the next input from @var{stream} will | |
740 | read @var{c} before anything else. | |
741 | ||
742 | If @var{c} is @code{EOF}, @code{ungetc} does nothing and just returns | |
743 | @code{EOF}. This lets you call @code{ungetc} with the return value of | |
744 | @code{getc} without needing to check for an error from @code{getc}. | |
745 | ||
746 | The character that you push back doesn't have to be the same as the last | |
747 | character that was actually read from the stream. In fact, it isn't | |
748 | necessary to actually read any characters from the stream before | |
749 | unreading them with @code{ungetc}! But that is a strange way to write | |
750 | a program; usually @code{ungetc} is used only to unread a character | |
751 | that was just read from the same stream. | |
752 | ||
753 | The GNU C library only supports one character of pushback---in other | |
754 | words, it does not work to call @code{ungetc} twice without doing input | |
755 | in between. Other systems might let you push back multiple characters; | |
756 | then reading from the stream retrieves the characters in the reverse | |
757 | order that they were pushed. | |
758 | ||
759 | Pushing back characters doesn't alter the file; only the internal | |
760 | buffering for the stream is affected. If a file positioning function | |
a5a0310d UD |
761 | (such as @code{fseek}, @code{fseeko} or @code{rewind}; @pxref{File |
762 | Positioning}) is called, any pending pushed-back characters are | |
763 | discarded. | |
28f540f4 RM |
764 | |
765 | Unreading a character on a stream that is at end of file clears the | |
766 | end-of-file indicator for the stream, because it makes the character of | |
767 | input available. After you read that character, trying to read again | |
768 | will encounter end of file. | |
769 | @end deftypefun | |
770 | ||
771 | Here is an example showing the use of @code{getc} and @code{ungetc} to | |
772 | skip over whitespace characters. When this function reaches a | |
773 | non-whitespace character, it unreads that character to be seen again on | |
774 | the next read operation on the stream. | |
775 | ||
776 | @smallexample | |
777 | #include <stdio.h> | |
778 | #include <ctype.h> | |
779 | ||
780 | void | |
781 | skip_whitespace (FILE *stream) | |
782 | @{ | |
783 | int c; | |
784 | do | |
785 | /* @r{No need to check for @code{EOF} because it is not} | |
786 | @r{@code{isspace}, and @code{ungetc} ignores @code{EOF}.} */ | |
787 | c = getc (stream); | |
788 | while (isspace (c)); | |
789 | ungetc (c, stream); | |
790 | @} | |
791 | @end smallexample | |
792 | ||
793 | @node Block Input/Output | |
794 | @section Block Input/Output | |
795 | ||
796 | This section describes how to do input and output operations on blocks | |
797 | of data. You can use these functions to read and write binary data, as | |
798 | well as to read and write text in fixed-size blocks instead of by | |
799 | characters or lines. | |
800 | @cindex binary I/O to a stream | |
801 | @cindex block I/O to a stream | |
802 | @cindex reading from a stream, by blocks | |
803 | @cindex writing to a stream, by blocks | |
804 | ||
805 | Binary files are typically used to read and write blocks of data in the | |
806 | same format as is used to represent the data in a running program. In | |
807 | other words, arbitrary blocks of memory---not just character or string | |
808 | objects---can be written to a binary file, and meaningfully read in | |
809 | again by the same program. | |
810 | ||
811 | Storing data in binary form is often considerably more efficient than | |
812 | using the formatted I/O functions. Also, for floating-point numbers, | |
813 | the binary form avoids possible loss of precision in the conversion | |
814 | process. On the other hand, binary files can't be examined or modified | |
815 | easily using many standard file utilities (such as text editors), and | |
816 | are not portable between different implementations of the language, or | |
817 | different kinds of computers. | |
818 | ||
819 | These functions are declared in @file{stdio.h}. | |
820 | @pindex stdio.h | |
821 | ||
822 | @comment stdio.h | |
f65fd747 | 823 | @comment ISO |
28f540f4 RM |
824 | @deftypefun size_t fread (void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) |
825 | This function reads up to @var{count} objects of size @var{size} into | |
826 | the array @var{data}, from the stream @var{stream}. It returns the | |
827 | number of objects actually read, which might be less than @var{count} if | |
828 | a read error occurs or the end of the file is reached. This function | |
829 | returns a value of zero (and doesn't read anything) if either @var{size} | |
830 | or @var{count} is zero. | |
831 | ||
832 | If @code{fread} encounters end of file in the middle of an object, it | |
833 | returns the number of complete objects read, and discards the partial | |
834 | object. Therefore, the stream remains at the actual end of the file. | |
835 | @end deftypefun | |
836 | ||
837 | @comment stdio.h | |
f65fd747 | 838 | @comment ISO |
28f540f4 RM |
839 | @deftypefun size_t fwrite (const void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) |
840 | This function writes up to @var{count} objects of size @var{size} from | |
841 | the array @var{data}, to the stream @var{stream}. The return value is | |
842 | normally @var{count}, if the call succeeds. Any other value indicates | |
843 | some sort of error, such as running out of space. | |
844 | @end deftypefun | |
845 | ||
846 | @node Formatted Output | |
847 | @section Formatted Output | |
848 | ||
849 | @cindex format string, for @code{printf} | |
850 | @cindex template, for @code{printf} | |
851 | @cindex formatted output to a stream | |
852 | @cindex writing to a stream, formatted | |
853 | The functions described in this section (@code{printf} and related | |
854 | functions) provide a convenient way to perform formatted output. You | |
855 | call @code{printf} with a @dfn{format string} or @dfn{template string} | |
856 | that specifies how to format the values of the remaining arguments. | |
857 | ||
858 | Unless your program is a filter that specifically performs line- or | |
859 | character-oriented processing, using @code{printf} or one of the other | |
860 | related functions described in this section is usually the easiest and | |
861 | most concise way to perform output. These functions are especially | |
862 | useful for printing error messages, tables of data, and the like. | |
863 | ||
864 | @menu | |
865 | * Formatted Output Basics:: Some examples to get you started. | |
866 | * Output Conversion Syntax:: General syntax of conversion | |
867 | specifications. | |
868 | * Table of Output Conversions:: Summary of output conversions and | |
869 | what they do. | |
870 | * Integer Conversions:: Details about formatting of integers. | |
871 | * Floating-Point Conversions:: Details about formatting of | |
872 | floating-point numbers. | |
873 | * Other Output Conversions:: Details about formatting of strings, | |
874 | characters, pointers, and the like. | |
875 | * Formatted Output Functions:: Descriptions of the actual functions. | |
876 | * Dynamic Output:: Functions that allocate memory for the output. | |
877 | * Variable Arguments Output:: @code{vprintf} and friends. | |
878 | * Parsing a Template String:: What kinds of args does a given template | |
19c3f208 | 879 | call for? |
28f540f4 RM |
880 | * Example of Parsing:: Sample program using @code{parse_printf_format}. |
881 | @end menu | |
882 | ||
883 | @node Formatted Output Basics | |
884 | @subsection Formatted Output Basics | |
885 | ||
886 | The @code{printf} function can be used to print any number of arguments. | |
887 | The template string argument you supply in a call provides | |
888 | information not only about the number of additional arguments, but also | |
889 | about their types and what style should be used for printing them. | |
890 | ||
891 | Ordinary characters in the template string are simply written to the | |
892 | output stream as-is, while @dfn{conversion specifications} introduced by | |
893 | a @samp{%} character in the template cause subsequent arguments to be | |
894 | formatted and written to the output stream. For example, | |
895 | @cindex conversion specifications (@code{printf}) | |
896 | ||
897 | @smallexample | |
898 | int pct = 37; | |
899 | char filename[] = "foo.txt"; | |
900 | printf ("Processing of `%s' is %d%% finished.\nPlease be patient.\n", | |
901 | filename, pct); | |
902 | @end smallexample | |
903 | ||
904 | @noindent | |
905 | produces output like | |
906 | ||
907 | @smallexample | |
908 | Processing of `foo.txt' is 37% finished. | |
909 | Please be patient. | |
910 | @end smallexample | |
911 | ||
912 | This example shows the use of the @samp{%d} conversion to specify that | |
913 | an @code{int} argument should be printed in decimal notation, the | |
914 | @samp{%s} conversion to specify printing of a string argument, and | |
915 | the @samp{%%} conversion to print a literal @samp{%} character. | |
916 | ||
917 | There are also conversions for printing an integer argument as an | |
918 | unsigned value in octal, decimal, or hexadecimal radix (@samp{%o}, | |
919 | @samp{%u}, or @samp{%x}, respectively); or as a character value | |
920 | (@samp{%c}). | |
921 | ||
922 | Floating-point numbers can be printed in normal, fixed-point notation | |
923 | using the @samp{%f} conversion or in exponential notation using the | |
924 | @samp{%e} conversion. The @samp{%g} conversion uses either @samp{%e} | |
925 | or @samp{%f} format, depending on what is more appropriate for the | |
926 | magnitude of the particular number. | |
927 | ||
928 | You can control formatting more precisely by writing @dfn{modifiers} | |
929 | between the @samp{%} and the character that indicates which conversion | |
930 | to apply. These slightly alter the ordinary behavior of the conversion. | |
931 | For example, most conversion specifications permit you to specify a | |
932 | minimum field width and a flag indicating whether you want the result | |
933 | left- or right-justified within the field. | |
934 | ||
935 | The specific flags and modifiers that are permitted and their | |
936 | interpretation vary depending on the particular conversion. They're all | |
937 | described in more detail in the following sections. Don't worry if this | |
938 | all seems excessively complicated at first; you can almost always get | |
939 | reasonable free-format output without using any of the modifiers at all. | |
940 | The modifiers are mostly used to make the output look ``prettier'' in | |
941 | tables. | |
942 | ||
943 | @node Output Conversion Syntax | |
944 | @subsection Output Conversion Syntax | |
945 | ||
946 | This section provides details about the precise syntax of conversion | |
947 | specifications that can appear in a @code{printf} template | |
948 | string. | |
949 | ||
390955cb UD |
950 | Characters in the template string that are not part of a conversion |
951 | specification are printed as-is to the output stream. Multibyte | |
952 | character sequences (@pxref{Character Set Handling}) are permitted in a | |
953 | template string. | |
28f540f4 RM |
954 | |
955 | The conversion specifications in a @code{printf} template string have | |
956 | the general form: | |
957 | ||
958 | @example | |
dfd2257a | 959 | % @r{[} @var{param-no} @r{$]} @var{flags} @var{width} @r{[} . @var{precision} @r{]} @var{type} @var{conversion} |
28f540f4 RM |
960 | @end example |
961 | ||
962 | For example, in the conversion specifier @samp{%-10.8ld}, the @samp{-} | |
963 | is a flag, @samp{10} specifies the field width, the precision is | |
964 | @samp{8}, the letter @samp{l} is a type modifier, and @samp{d} specifies | |
965 | the conversion style. (This particular type specifier says to | |
966 | print a @code{long int} argument in decimal notation, with a minimum of | |
967 | 8 digits left-justified in a field at least 10 characters wide.) | |
968 | ||
969 | In more detail, output conversion specifications consist of an | |
970 | initial @samp{%} character followed in sequence by: | |
971 | ||
972 | @itemize @bullet | |
dfd2257a UD |
973 | @item |
974 | An optional specification of the parameter used for this format. | |
e8b1163e | 975 | Normally the parameters to the @code{printf} function are assigned to the |
f2ea0f5b | 976 | formats in the order of appearance in the format string. But in some |
dfd2257a | 977 | situations (such as message translation) this is not desirable and this |
e8b1163e | 978 | extension allows an explicit parameter to be specified. |
dfd2257a UD |
979 | |
980 | The @var{param-no} part of the format must be an integer in the range of | |
981 | 1 to the maximum number of arguments present to the function call. Some | |
982 | implementations limit this number to a certainly upper bound. The exact | |
983 | limit can be retrieved by the following constant. | |
984 | ||
985 | @defvr Macro NL_ARGMAX | |
986 | The value of @code{ARGMAX} is the maximum value allowed for the | |
987 | specification of an positional parameter in a @code{printf} call. The | |
988 | actual value in effect at runtime can be retrieved by using | |
989 | @code{sysconf} using the @code{_SC_NL_ARGMAX} parameter @pxref{Sysconf | |
990 | Definition}. | |
991 | ||
992 | Some system have a quite low limit such as @math{9} for @w{System V} | |
993 | systems. The GNU C library has no real limit. | |
994 | @end defvr | |
995 | ||
996 | If any of the formats has a specification for the parameter position all | |
997 | of them in the format string shall have one. Otherwise the behaviour is | |
998 | undefined. | |
999 | ||
19c3f208 | 1000 | @item |
28f540f4 RM |
1001 | Zero or more @dfn{flag characters} that modify the normal behavior of |
1002 | the conversion specification. | |
1003 | @cindex flag character (@code{printf}) | |
1004 | ||
19c3f208 | 1005 | @item |
28f540f4 RM |
1006 | An optional decimal integer specifying the @dfn{minimum field width}. |
1007 | If the normal conversion produces fewer characters than this, the field | |
1008 | is padded with spaces to the specified width. This is a @emph{minimum} | |
1009 | value; if the normal conversion produces more characters than this, the | |
1010 | field is @emph{not} truncated. Normally, the output is right-justified | |
1011 | within the field. | |
1012 | @cindex minimum field width (@code{printf}) | |
1013 | ||
1014 | You can also specify a field width of @samp{*}. This means that the | |
1015 | next argument in the argument list (before the actual value to be | |
1016 | printed) is used as the field width. The value must be an @code{int}. | |
1017 | If the value is negative, this means to set the @samp{-} flag (see | |
1018 | below) and to use the absolute value as the field width. | |
1019 | ||
19c3f208 | 1020 | @item |
28f540f4 RM |
1021 | An optional @dfn{precision} to specify the number of digits to be |
1022 | written for the numeric conversions. If the precision is specified, it | |
1023 | consists of a period (@samp{.}) followed optionally by a decimal integer | |
1024 | (which defaults to zero if omitted). | |
1025 | @cindex precision (@code{printf}) | |
1026 | ||
1027 | You can also specify a precision of @samp{*}. This means that the next | |
1028 | argument in the argument list (before the actual value to be printed) is | |
1029 | used as the precision. The value must be an @code{int}, and is ignored | |
1030 | if it is negative. If you specify @samp{*} for both the field width and | |
1031 | precision, the field width argument precedes the precision argument. | |
1032 | Other C library versions may not recognize this syntax. | |
1033 | ||
1034 | @item | |
1035 | An optional @dfn{type modifier character}, which is used to specify the | |
1036 | data type of the corresponding argument if it differs from the default | |
1037 | type. (For example, the integer conversions assume a type of @code{int}, | |
1038 | but you can specify @samp{h}, @samp{l}, or @samp{L} for other integer | |
1039 | types.) | |
1040 | @cindex type modifier character (@code{printf}) | |
1041 | ||
1042 | @item | |
1043 | A character that specifies the conversion to be applied. | |
1044 | @end itemize | |
1045 | ||
19c3f208 | 1046 | The exact options that are permitted and how they are interpreted vary |
28f540f4 RM |
1047 | between the different conversion specifiers. See the descriptions of the |
1048 | individual conversions for information about the particular options that | |
1049 | they use. | |
1050 | ||
1051 | With the @samp{-Wformat} option, the GNU C compiler checks calls to | |
1052 | @code{printf} and related functions. It examines the format string and | |
1053 | verifies that the correct number and types of arguments are supplied. | |
1054 | There is also a GNU C syntax to tell the compiler that a function you | |
19c3f208 | 1055 | write uses a @code{printf}-style format string. |
28f540f4 RM |
1056 | @xref{Function Attributes, , Declaring Attributes of Functions, |
1057 | gcc.info, Using GNU CC}, for more information. | |
1058 | ||
1059 | @node Table of Output Conversions | |
1060 | @subsection Table of Output Conversions | |
1061 | @cindex output conversions, for @code{printf} | |
1062 | ||
1063 | Here is a table summarizing what all the different conversions do: | |
1064 | ||
1065 | @table @asis | |
1066 | @item @samp{%d}, @samp{%i} | |
1067 | Print an integer as a signed decimal number. @xref{Integer | |
1068 | Conversions}, for details. @samp{%d} and @samp{%i} are synonymous for | |
1069 | output, but are different when used with @code{scanf} for input | |
1070 | (@pxref{Table of Input Conversions}). | |
1071 | ||
1072 | @item @samp{%o} | |
1073 | Print an integer as an unsigned octal number. @xref{Integer | |
1074 | Conversions}, for details. | |
1075 | ||
1076 | @item @samp{%u} | |
1077 | Print an integer as an unsigned decimal number. @xref{Integer | |
1078 | Conversions}, for details. | |
1079 | ||
1080 | @item @samp{%x}, @samp{%X} | |
1081 | Print an integer as an unsigned hexadecimal number. @samp{%x} uses | |
1082 | lower-case letters and @samp{%X} uses upper-case. @xref{Integer | |
1083 | Conversions}, for details. | |
1084 | ||
1085 | @item @samp{%f} | |
1086 | Print a floating-point number in normal (fixed-point) notation. | |
1087 | @xref{Floating-Point Conversions}, for details. | |
1088 | ||
1089 | @item @samp{%e}, @samp{%E} | |
1090 | Print a floating-point number in exponential notation. @samp{%e} uses | |
1091 | lower-case letters and @samp{%E} uses upper-case. @xref{Floating-Point | |
1092 | Conversions}, for details. | |
1093 | ||
1094 | @item @samp{%g}, @samp{%G} | |
1095 | Print a floating-point number in either normal or exponential notation, | |
1096 | whichever is more appropriate for its magnitude. @samp{%g} uses | |
1097 | lower-case letters and @samp{%G} uses upper-case. @xref{Floating-Point | |
1098 | Conversions}, for details. | |
1099 | ||
2f6d1f1b UD |
1100 | @item @samp{%a}, @samp{%A} |
1101 | Print a floating-point number in a hexadecimal fractional notation which | |
1102 | the exponent to base 2 represented in decimal digits. @samp{%a} uses | |
1103 | lower-case letters and @samp{%A} uses upper-case. @xref{Floating-Point | |
1104 | Conversions}, for details. | |
1105 | ||
28f540f4 RM |
1106 | @item @samp{%c} |
1107 | Print a single character. @xref{Other Output Conversions}. | |
1108 | ||
1109 | @item @samp{%s} | |
1110 | Print a string. @xref{Other Output Conversions}. | |
1111 | ||
1112 | @item @samp{%p} | |
1113 | Print the value of a pointer. @xref{Other Output Conversions}. | |
1114 | ||
1115 | @item @samp{%n} | |
1116 | Get the number of characters printed so far. @xref{Other Output Conversions}. | |
1117 | Note that this conversion specification never produces any output. | |
1118 | ||
1119 | @item @samp{%m} | |
1120 | Print the string corresponding to the value of @code{errno}. | |
1121 | (This is a GNU extension.) | |
1122 | @xref{Other Output Conversions}. | |
1123 | ||
1124 | @item @samp{%%} | |
1125 | Print a literal @samp{%} character. @xref{Other Output Conversions}. | |
1126 | @end table | |
1127 | ||
1128 | If the syntax of a conversion specification is invalid, unpredictable | |
1129 | things will happen, so don't do this. If there aren't enough function | |
1130 | arguments provided to supply values for all the conversion | |
1131 | specifications in the template string, or if the arguments are not of | |
1132 | the correct types, the results are unpredictable. If you supply more | |
1133 | arguments than conversion specifications, the extra argument values are | |
1134 | simply ignored; this is sometimes useful. | |
1135 | ||
1136 | @node Integer Conversions | |
1137 | @subsection Integer Conversions | |
1138 | ||
1139 | This section describes the options for the @samp{%d}, @samp{%i}, | |
1140 | @samp{%o}, @samp{%u}, @samp{%x}, and @samp{%X} conversion | |
1141 | specifications. These conversions print integers in various formats. | |
1142 | ||
1143 | The @samp{%d} and @samp{%i} conversion specifications both print an | |
1144 | @code{int} argument as a signed decimal number; while @samp{%o}, | |
1145 | @samp{%u}, and @samp{%x} print the argument as an unsigned octal, | |
1146 | decimal, or hexadecimal number (respectively). The @samp{%X} conversion | |
1147 | specification is just like @samp{%x} except that it uses the characters | |
1148 | @samp{ABCDEF} as digits instead of @samp{abcdef}. | |
1149 | ||
1150 | The following flags are meaningful: | |
1151 | ||
1152 | @table @asis | |
1153 | @item @samp{-} | |
1154 | Left-justify the result in the field (instead of the normal | |
1155 | right-justification). | |
1156 | ||
1157 | @item @samp{+} | |
1158 | For the signed @samp{%d} and @samp{%i} conversions, print a | |
1159 | plus sign if the value is positive. | |
1160 | ||
1161 | @item @samp{ } | |
1162 | For the signed @samp{%d} and @samp{%i} conversions, if the result | |
1163 | doesn't start with a plus or minus sign, prefix it with a space | |
1164 | character instead. Since the @samp{+} flag ensures that the result | |
1165 | includes a sign, this flag is ignored if you supply both of them. | |
1166 | ||
1167 | @item @samp{#} | |
1168 | For the @samp{%o} conversion, this forces the leading digit to be | |
1169 | @samp{0}, as if by increasing the precision. For @samp{%x} or | |
1170 | @samp{%X}, this prefixes a leading @samp{0x} or @samp{0X} (respectively) | |
1171 | to the result. This doesn't do anything useful for the @samp{%d}, | |
1172 | @samp{%i}, or @samp{%u} conversions. Using this flag produces output | |
1173 | which can be parsed by the @code{strtoul} function (@pxref{Parsing of | |
1174 | Integers}) and @code{scanf} with the @samp{%i} conversion | |
1175 | (@pxref{Numeric Input Conversions}). | |
1176 | ||
1177 | @item @samp{'} | |
1178 | Separate the digits into groups as specified by the locale specified for | |
1179 | the @code{LC_NUMERIC} category; @pxref{General Numeric}. This flag is a | |
1180 | GNU extension. | |
1181 | ||
1182 | @item @samp{0} | |
1183 | Pad the field with zeros instead of spaces. The zeros are placed after | |
1184 | any indication of sign or base. This flag is ignored if the @samp{-} | |
1185 | flag is also specified, or if a precision is specified. | |
1186 | @end table | |
1187 | ||
1188 | If a precision is supplied, it specifies the minimum number of digits to | |
1189 | appear; leading zeros are produced if necessary. If you don't specify a | |
1190 | precision, the number is printed with as many digits as it needs. If | |
1191 | you convert a value of zero with an explicit precision of zero, then no | |
1192 | characters at all are produced. | |
1193 | ||
1194 | Without a type modifier, the corresponding argument is treated as an | |
1195 | @code{int} (for the signed conversions @samp{%i} and @samp{%d}) or | |
1196 | @code{unsigned int} (for the unsigned conversions @samp{%o}, @samp{%u}, | |
1197 | @samp{%x}, and @samp{%X}). Recall that since @code{printf} and friends | |
1198 | are variadic, any @code{char} and @code{short} arguments are | |
1199 | automatically converted to @code{int} by the default argument | |
1200 | promotions. For arguments of other integer types, you can use these | |
1201 | modifiers: | |
1202 | ||
1203 | @table @samp | |
cc3fa755 UD |
1204 | @item hh |
1205 | Specifies that the argument is a @code{signed char} or @code{unsigned | |
1206 | char}, as appropriate. A @code{char} argument is converted to an | |
1207 | @code{int} or @code{unsigned int} by the default argument promotions | |
1208 | anyway, but the @samp{h} modifier says to convert it back to a | |
1209 | @code{char} again. | |
1210 | ||
ec751a23 | 1211 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 1212 | |
28f540f4 RM |
1213 | @item h |
1214 | Specifies that the argument is a @code{short int} or @code{unsigned | |
1215 | short int}, as appropriate. A @code{short} argument is converted to an | |
1216 | @code{int} or @code{unsigned int} by the default argument promotions | |
1217 | anyway, but the @samp{h} modifier says to convert it back to a | |
1218 | @code{short} again. | |
1219 | ||
e852e889 UD |
1220 | @item j |
1221 | Specifies that the argument is a @code{intmax_t} or @code{uintmax_t}, as | |
1222 | appropriate. | |
1223 | ||
ec751a23 | 1224 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 1225 | |
28f540f4 RM |
1226 | @item l |
1227 | Specifies that the argument is a @code{long int} or @code{unsigned long | |
1228 | int}, as appropriate. Two @samp{l} characters is like the @samp{L} | |
1229 | modifier, below. | |
1230 | ||
1231 | @item L | |
1232 | @itemx ll | |
1233 | @itemx q | |
1234 | Specifies that the argument is a @code{long long int}. (This type is | |
1235 | an extension supported by the GNU C compiler. On systems that don't | |
1236 | support extra-long integers, this is the same as @code{long int}.) | |
1237 | ||
1238 | The @samp{q} modifier is another name for the same thing, which comes | |
1239 | from 4.4 BSD; a @w{@code{long long int}} is sometimes called a ``quad'' | |
1240 | @code{int}. | |
1241 | ||
e852e889 UD |
1242 | @item t |
1243 | Specifies that the argument is a @code{ptrdiff_t}. | |
1244 | ||
ec751a23 | 1245 | This modifier was introduced in @w{ISO C99}. |
e852e889 UD |
1246 | |
1247 | @item z | |
1248 | @itemx Z | |
1249 | Specifies that the argument is a @code{size_t}. | |
1250 | ||
ec751a23 | 1251 | @samp{z} was introduced in @w{ISO C99}. @samp{Z} is a GNU extension |
0be8752b | 1252 | predating this addition and should not be used in new code. |
28f540f4 RM |
1253 | @end table |
1254 | ||
1255 | Here is an example. Using the template string: | |
1256 | ||
1257 | @smallexample | |
1258 | "|%5d|%-5d|%+5d|%+-5d|% 5d|%05d|%5.0d|%5.2d|%d|\n" | |
1259 | @end smallexample | |
1260 | ||
1261 | @noindent | |
1262 | to print numbers using the different options for the @samp{%d} | |
1263 | conversion gives results like: | |
1264 | ||
1265 | @smallexample | |
1266 | | 0|0 | +0|+0 | 0|00000| | 00|0| | |
1267 | | 1|1 | +1|+1 | 1|00001| 1| 01|1| | |
1268 | | -1|-1 | -1|-1 | -1|-0001| -1| -01|-1| | |
1269 | |100000|100000|+100000| 100000|100000|100000|100000|100000| | |
1270 | @end smallexample | |
1271 | ||
1272 | In particular, notice what happens in the last case where the number | |
1273 | is too large to fit in the minimum field width specified. | |
1274 | ||
1275 | Here are some more examples showing how unsigned integers print under | |
1276 | various format options, using the template string: | |
1277 | ||
1278 | @smallexample | |
1279 | "|%5u|%5o|%5x|%5X|%#5o|%#5x|%#5X|%#10.8x|\n" | |
1280 | @end smallexample | |
1281 | ||
1282 | @smallexample | |
1283 | | 0| 0| 0| 0| 0| 0x0| 0X0|0x00000000| | |
1284 | | 1| 1| 1| 1| 01| 0x1| 0X1|0x00000001| | |
1285 | |100000|303240|186a0|186A0|0303240|0x186a0|0X186A0|0x000186a0| | |
1286 | @end smallexample | |
1287 | ||
1288 | ||
1289 | @node Floating-Point Conversions | |
1290 | @subsection Floating-Point Conversions | |
1291 | ||
1292 | This section discusses the conversion specifications for floating-point | |
1293 | numbers: the @samp{%f}, @samp{%e}, @samp{%E}, @samp{%g}, and @samp{%G} | |
1294 | conversions. | |
1295 | ||
1296 | The @samp{%f} conversion prints its argument in fixed-point notation, | |
1297 | producing output of the form | |
1298 | @w{[@code{-}]@var{ddd}@code{.}@var{ddd}}, | |
1299 | where the number of digits following the decimal point is controlled | |
1300 | by the precision you specify. | |
1301 | ||
1302 | The @samp{%e} conversion prints its argument in exponential notation, | |
1303 | producing output of the form | |
1304 | @w{[@code{-}]@var{d}@code{.}@var{ddd}@code{e}[@code{+}|@code{-}]@var{dd}}. | |
1305 | Again, the number of digits following the decimal point is controlled by | |
1306 | the precision. The exponent always contains at least two digits. The | |
1307 | @samp{%E} conversion is similar but the exponent is marked with the letter | |
1308 | @samp{E} instead of @samp{e}. | |
1309 | ||
1310 | The @samp{%g} and @samp{%G} conversions print the argument in the style | |
1311 | of @samp{%e} or @samp{%E} (respectively) if the exponent would be less | |
1312 | than -4 or greater than or equal to the precision; otherwise they use the | |
1313 | @samp{%f} style. Trailing zeros are removed from the fractional portion | |
1314 | of the result and a decimal-point character appears only if it is | |
1315 | followed by a digit. | |
1316 | ||
2f6d1f1b | 1317 | The @samp{%a} and @samp{%A} conversions are meant for representing |
0be8752b | 1318 | floating-point numbers exactly in textual form so that they can be |
2f6d1f1b UD |
1319 | exchanged as texts between different programs and/or machines. The |
1320 | numbers are represented is the form | |
1321 | @w{[@code{-}]@code{0x}@var{h}@code{.}@var{hhh}@code{p}[@code{+}|@code{-}]@var{dd}}. | |
1322 | At the left of the decimal-point character exactly one digit is print. | |
91ea72b7 | 1323 | This character is only @code{0} if the number is denormalized. |
49c091e5 | 1324 | Otherwise the value is unspecified; it is implementation dependent how many |
2f6d1f1b UD |
1325 | bits are used. The number of hexadecimal digits on the right side of |
1326 | the decimal-point character is equal to the precision. If the precision | |
1327 | is zero it is determined to be large enough to provide an exact | |
1328 | representation of the number (or it is large enough to distinguish two | |
1329 | adjacent values if the @code{FLT_RADIX} is not a power of 2, | |
91ea72b7 | 1330 | @pxref{Floating Point Parameters}). For the @samp{%a} conversion |
2f6d1f1b UD |
1331 | lower-case characters are used to represent the hexadecimal number and |
1332 | the prefix and exponent sign are printed as @code{0x} and @code{p} | |
1333 | respectively. Otherwise upper-case characters are used and @code{0X} | |
1334 | and @code{P} are used for the representation of prefix and exponent | |
1335 | string. The exponent to the base of two is printed as a decimal number | |
1336 | using at least one digit but at most as many digits as necessary to | |
1337 | represent the value exactly. | |
1338 | ||
1339 | If the value to be printed represents infinity or a NaN, the output is | |
1340 | @w{[@code{-}]@code{inf}} or @code{nan} respectively if the conversion | |
1341 | specifier is @samp{%a}, @samp{%e}, @samp{%f}, or @samp{%g} and it is | |
1342 | @w{[@code{-}]@code{INF}} or @code{NAN} respectively if the conversion is | |
1343 | @samp{%A}, @samp{%E}, or @samp{%G}. | |
1344 | ||
28f540f4 RM |
1345 | The following flags can be used to modify the behavior: |
1346 | ||
1347 | @comment We use @asis instead of @samp so we can have ` ' as an item. | |
1348 | @table @asis | |
1349 | @item @samp{-} | |
1350 | Left-justify the result in the field. Normally the result is | |
1351 | right-justified. | |
1352 | ||
1353 | @item @samp{+} | |
1354 | Always include a plus or minus sign in the result. | |
1355 | ||
1356 | @item @samp{ } | |
1357 | If the result doesn't start with a plus or minus sign, prefix it with a | |
1358 | space instead. Since the @samp{+} flag ensures that the result includes | |
1359 | a sign, this flag is ignored if you supply both of them. | |
1360 | ||
1361 | @item @samp{#} | |
1362 | Specifies that the result should always include a decimal point, even | |
1363 | if no digits follow it. For the @samp{%g} and @samp{%G} conversions, | |
1364 | this also forces trailing zeros after the decimal point to be left | |
1365 | in place where they would otherwise be removed. | |
1366 | ||
1367 | @item @samp{'} | |
1368 | Separate the digits of the integer part of the result into groups as | |
1369 | specified by the locale specified for the @code{LC_NUMERIC} category; | |
1370 | @pxref{General Numeric}. This flag is a GNU extension. | |
1371 | ||
1372 | @item @samp{0} | |
1373 | Pad the field with zeros instead of spaces; the zeros are placed | |
1374 | after any sign. This flag is ignored if the @samp{-} flag is also | |
1375 | specified. | |
1376 | @end table | |
1377 | ||
1378 | The precision specifies how many digits follow the decimal-point | |
1379 | character for the @samp{%f}, @samp{%e}, and @samp{%E} conversions. For | |
1380 | these conversions, the default precision is @code{6}. If the precision | |
1381 | is explicitly @code{0}, this suppresses the decimal point character | |
1382 | entirely. For the @samp{%g} and @samp{%G} conversions, the precision | |
1383 | specifies how many significant digits to print. Significant digits are | |
1384 | the first digit before the decimal point, and all the digits after it. | |
91ea72b7 | 1385 | If the precision is @code{0} or not specified for @samp{%g} or @samp{%G}, |
28f540f4 RM |
1386 | it is treated like a value of @code{1}. If the value being printed |
1387 | cannot be expressed accurately in the specified number of digits, the | |
1388 | value is rounded to the nearest number that fits. | |
1389 | ||
1390 | Without a type modifier, the floating-point conversions use an argument | |
1391 | of type @code{double}. (By the default argument promotions, any | |
1392 | @code{float} arguments are automatically converted to @code{double}.) | |
1393 | The following type modifier is supported: | |
1394 | ||
1395 | @table @samp | |
1396 | @item L | |
1397 | An uppercase @samp{L} specifies that the argument is a @code{long | |
1398 | double}. | |
1399 | @end table | |
1400 | ||
1401 | Here are some examples showing how numbers print using the various | |
1402 | floating-point conversions. All of the numbers were printed using | |
1403 | this template string: | |
1404 | ||
1405 | @smallexample | |
2f6d1f1b | 1406 | "|%13.4a|%13.4f|%13.4e|%13.4g|\n" |
28f540f4 RM |
1407 | @end smallexample |
1408 | ||
1409 | Here is the output: | |
1410 | ||
1411 | @smallexample | |
2f6d1f1b UD |
1412 | | 0x0.0000p+0| 0.0000| 0.0000e+00| 0| |
1413 | | 0x1.0000p-1| 0.5000| 5.0000e-01| 0.5| | |
1414 | | 0x1.0000p+0| 1.0000| 1.0000e+00| 1| | |
1415 | | -0x1.0000p+0| -1.0000| -1.0000e+00| -1| | |
1416 | | 0x1.9000p+6| 100.0000| 1.0000e+02| 100| | |
1417 | | 0x1.f400p+9| 1000.0000| 1.0000e+03| 1000| | |
1418 | | 0x1.3880p+13| 10000.0000| 1.0000e+04| 1e+04| | |
1419 | | 0x1.81c8p+13| 12345.0000| 1.2345e+04| 1.234e+04| | |
1420 | | 0x1.86a0p+16| 100000.0000| 1.0000e+05| 1e+05| | |
1421 | | 0x1.e240p+16| 123456.0000| 1.2346e+05| 1.235e+05| | |
28f540f4 RM |
1422 | @end smallexample |
1423 | ||
1424 | Notice how the @samp{%g} conversion drops trailing zeros. | |
1425 | ||
1426 | @node Other Output Conversions | |
1427 | @subsection Other Output Conversions | |
1428 | ||
1429 | This section describes miscellaneous conversions for @code{printf}. | |
1430 | ||
1431 | The @samp{%c} conversion prints a single character. The @code{int} | |
1432 | argument is first converted to an @code{unsigned char}. The @samp{-} | |
1433 | flag can be used to specify left-justification in the field, but no | |
1434 | other flags are defined, and no precision or type modifier can be given. | |
1435 | For example: | |
1436 | ||
1437 | @smallexample | |
1438 | printf ("%c%c%c%c%c", 'h', 'e', 'l', 'l', 'o'); | |
1439 | @end smallexample | |
1440 | ||
1441 | @noindent | |
1442 | prints @samp{hello}. | |
1443 | ||
1444 | The @samp{%s} conversion prints a string. The corresponding argument | |
1445 | must be of type @code{char *} (or @code{const char *}). A precision can | |
1446 | be specified to indicate the maximum number of characters to write; | |
1447 | otherwise characters in the string up to but not including the | |
1448 | terminating null character are written to the output stream. The | |
1449 | @samp{-} flag can be used to specify left-justification in the field, | |
1450 | but no other flags or type modifiers are defined for this conversion. | |
1451 | For example: | |
1452 | ||
1453 | @smallexample | |
1454 | printf ("%3s%-6s", "no", "where"); | |
1455 | @end smallexample | |
1456 | ||
1457 | @noindent | |
1458 | prints @samp{ nowhere }. | |
1459 | ||
1460 | If you accidentally pass a null pointer as the argument for a @samp{%s} | |
1461 | conversion, the GNU library prints it as @samp{(null)}. We think this | |
1462 | is more useful than crashing. But it's not good practice to pass a null | |
1463 | argument intentionally. | |
1464 | ||
1465 | The @samp{%m} conversion prints the string corresponding to the error | |
1466 | code in @code{errno}. @xref{Error Messages}. Thus: | |
1467 | ||
1468 | @smallexample | |
1469 | fprintf (stderr, "can't open `%s': %m\n", filename); | |
1470 | @end smallexample | |
1471 | ||
1472 | @noindent | |
1473 | is equivalent to: | |
1474 | ||
1475 | @smallexample | |
1476 | fprintf (stderr, "can't open `%s': %s\n", filename, strerror (errno)); | |
1477 | @end smallexample | |
1478 | ||
1479 | @noindent | |
1480 | The @samp{%m} conversion is a GNU C library extension. | |
1481 | ||
1482 | The @samp{%p} conversion prints a pointer value. The corresponding | |
1483 | argument must be of type @code{void *}. In practice, you can use any | |
1484 | type of pointer. | |
1485 | ||
1486 | In the GNU system, non-null pointers are printed as unsigned integers, | |
1487 | as if a @samp{%#x} conversion were used. Null pointers print as | |
1488 | @samp{(nil)}. (Pointers might print differently in other systems.) | |
1489 | ||
1490 | For example: | |
1491 | ||
1492 | @smallexample | |
1493 | printf ("%p", "testing"); | |
1494 | @end smallexample | |
1495 | ||
1496 | @noindent | |
1497 | prints @samp{0x} followed by a hexadecimal number---the address of the | |
1498 | string constant @code{"testing"}. It does not print the word | |
1499 | @samp{testing}. | |
1500 | ||
1501 | You can supply the @samp{-} flag with the @samp{%p} conversion to | |
1502 | specify left-justification, but no other flags, precision, or type | |
1503 | modifiers are defined. | |
1504 | ||
1505 | The @samp{%n} conversion is unlike any of the other output conversions. | |
1506 | It uses an argument which must be a pointer to an @code{int}, but | |
1507 | instead of printing anything it stores the number of characters printed | |
1508 | so far by this call at that location. The @samp{h} and @samp{l} type | |
1509 | modifiers are permitted to specify that the argument is of type | |
1510 | @code{short int *} or @code{long int *} instead of @code{int *}, but no | |
1511 | flags, field width, or precision are permitted. | |
1512 | ||
1513 | For example, | |
1514 | ||
1515 | @smallexample | |
1516 | int nchar; | |
1517 | printf ("%d %s%n\n", 3, "bears", &nchar); | |
1518 | @end smallexample | |
1519 | ||
1520 | @noindent | |
1521 | prints: | |
1522 | ||
1523 | @smallexample | |
1524 | 3 bears | |
1525 | @end smallexample | |
1526 | ||
1527 | @noindent | |
19c3f208 | 1528 | and sets @code{nchar} to @code{7}, because @samp{3 bears} is seven |
28f540f4 RM |
1529 | characters. |
1530 | ||
1531 | ||
1532 | The @samp{%%} conversion prints a literal @samp{%} character. This | |
1533 | conversion doesn't use an argument, and no flags, field width, | |
1534 | precision, or type modifiers are permitted. | |
1535 | ||
1536 | ||
1537 | @node Formatted Output Functions | |
1538 | @subsection Formatted Output Functions | |
1539 | ||
1540 | This section describes how to call @code{printf} and related functions. | |
1541 | Prototypes for these functions are in the header file @file{stdio.h}. | |
1542 | Because these functions take a variable number of arguments, you | |
1543 | @emph{must} declare prototypes for them before using them. Of course, | |
1544 | the easiest way to make sure you have all the right prototypes is to | |
1545 | just include @file{stdio.h}. | |
1546 | @pindex stdio.h | |
1547 | ||
1548 | @comment stdio.h | |
f65fd747 | 1549 | @comment ISO |
28f540f4 RM |
1550 | @deftypefun int printf (const char *@var{template}, @dots{}) |
1551 | The @code{printf} function prints the optional arguments under the | |
1552 | control of the template string @var{template} to the stream | |
1553 | @code{stdout}. It returns the number of characters printed, or a | |
1554 | negative value if there was an output error. | |
1555 | @end deftypefun | |
1556 | ||
1557 | @comment stdio.h | |
f65fd747 | 1558 | @comment ISO |
28f540f4 RM |
1559 | @deftypefun int fprintf (FILE *@var{stream}, const char *@var{template}, @dots{}) |
1560 | This function is just like @code{printf}, except that the output is | |
1561 | written to the stream @var{stream} instead of @code{stdout}. | |
1562 | @end deftypefun | |
1563 | ||
1564 | @comment stdio.h | |
f65fd747 | 1565 | @comment ISO |
28f540f4 RM |
1566 | @deftypefun int sprintf (char *@var{s}, const char *@var{template}, @dots{}) |
1567 | This is like @code{printf}, except that the output is stored in the character | |
1568 | array @var{s} instead of written to a stream. A null character is written | |
1569 | to mark the end of the string. | |
1570 | ||
1571 | The @code{sprintf} function returns the number of characters stored in | |
1572 | the array @var{s}, not including the terminating null character. | |
1573 | ||
1574 | The behavior of this function is undefined if copying takes place | |
1575 | between objects that overlap---for example, if @var{s} is also given | |
1576 | as an argument to be printed under control of the @samp{%s} conversion. | |
1577 | @xref{Copying and Concatenation}. | |
1578 | ||
1579 | @strong{Warning:} The @code{sprintf} function can be @strong{dangerous} | |
1580 | because it can potentially output more characters than can fit in the | |
1581 | allocation size of the string @var{s}. Remember that the field width | |
1582 | given in a conversion specification is only a @emph{minimum} value. | |
1583 | ||
1584 | To avoid this problem, you can use @code{snprintf} or @code{asprintf}, | |
1585 | described below. | |
1586 | @end deftypefun | |
1587 | ||
1588 | @comment stdio.h | |
1589 | @comment GNU | |
1590 | @deftypefun int snprintf (char *@var{s}, size_t @var{size}, const char *@var{template}, @dots{}) | |
1591 | The @code{snprintf} function is similar to @code{sprintf}, except that | |
1592 | the @var{size} argument specifies the maximum number of characters to | |
1593 | produce. The trailing null character is counted towards this limit, so | |
1594 | you should allocate at least @var{size} characters for the string @var{s}. | |
1595 | ||
fe7bdd63 | 1596 | The return value is the number of characters which would be generated |
da2d1bc5 UD |
1597 | for the given input, excluding the trailing null. If this value is |
1598 | greater or equal to @var{size}, not all characters from the result have | |
1599 | been stored in @var{s}. You should try again with a bigger output | |
1600 | string. Here is an example of doing this: | |
28f540f4 RM |
1601 | |
1602 | @smallexample | |
1603 | @group | |
1604 | /* @r{Construct a message describing the value of a variable} | |
1605 | @r{whose name is @var{name} and whose value is @var{value}.} */ | |
1606 | char * | |
1607 | make_message (char *name, char *value) | |
1608 | @{ | |
1609 | /* @r{Guess we need no more than 100 chars of space.} */ | |
1610 | int size = 100; | |
1611 | char *buffer = (char *) xmalloc (size); | |
4cca6b86 | 1612 | int nchars; |
28f540f4 RM |
1613 | @end group |
1614 | @group | |
4cca6b86 UD |
1615 | /* @r{Try to print in the allocated space.} */ |
1616 | nchars = snprintf (buffer, size, "value of %s is %s", | |
1617 | name, value); | |
1618 | @end group | |
1619 | @group | |
fe7bdd63 | 1620 | if (nchars >= size) |
28f540f4 | 1621 | @{ |
4b8f94d3 | 1622 | /* @r{Reallocate buffer now that we know |
838e5ffe | 1623 | how much space is needed.} */ |
4cca6b86 UD |
1624 | buffer = (char *) xrealloc (buffer, nchars + 1); |
1625 | ||
1626 | /* @r{Try again.} */ | |
4b8f94d3 | 1627 | snprintf (buffer, size, "value of %s is %s", |
838e5ffe | 1628 | name, value); |
28f540f4 | 1629 | @} |
4cca6b86 UD |
1630 | /* @r{The last call worked, return the string.} */ |
1631 | return buffer; | |
28f540f4 RM |
1632 | @} |
1633 | @end group | |
1634 | @end smallexample | |
1635 | ||
1636 | In practice, it is often easier just to use @code{asprintf}, below. | |
fb971363 UD |
1637 | |
1638 | @strong{Attention:} In the GNU C library version 2.0 the return value | |
1639 | is the number of characters stored, not including the terminating null. | |
1640 | If this value equals @code{@var{size} - 1}, then there was not enough | |
0ea554bf | 1641 | space in @var{s} for all the output. This change was necessary with |
ec751a23 | 1642 | the adoption of snprintf by ISO C99. |
28f540f4 RM |
1643 | @end deftypefun |
1644 | ||
1645 | @node Dynamic Output | |
1646 | @subsection Dynamically Allocating Formatted Output | |
1647 | ||
1648 | The functions in this section do formatted output and place the results | |
1649 | in dynamically allocated memory. | |
1650 | ||
1651 | @comment stdio.h | |
1652 | @comment GNU | |
1653 | @deftypefun int asprintf (char **@var{ptr}, const char *@var{template}, @dots{}) | |
1654 | This function is similar to @code{sprintf}, except that it dynamically | |
1655 | allocates a string (as with @code{malloc}; @pxref{Unconstrained | |
1656 | Allocation}) to hold the output, instead of putting the output in a | |
1657 | buffer you allocate in advance. The @var{ptr} argument should be the | |
1658 | address of a @code{char *} object, and @code{asprintf} stores a pointer | |
1659 | to the newly allocated string at that location. | |
1660 | ||
1661 | Here is how to use @code{asprintf} to get the same result as the | |
1662 | @code{snprintf} example, but more easily: | |
1663 | ||
1664 | @smallexample | |
1665 | /* @r{Construct a message describing the value of a variable} | |
1666 | @r{whose name is @var{name} and whose value is @var{value}.} */ | |
1667 | char * | |
1668 | make_message (char *name, char *value) | |
1669 | @{ | |
1670 | char *result; | |
1671 | asprintf (&result, "value of %s is %s", name, value); | |
1672 | return result; | |
1673 | @} | |
1674 | @end smallexample | |
1675 | @end deftypefun | |
1676 | ||
1677 | @comment stdio.h | |
1678 | @comment GNU | |
1679 | @deftypefun int obstack_printf (struct obstack *@var{obstack}, const char *@var{template}, @dots{}) | |
1680 | This function is similar to @code{asprintf}, except that it uses the | |
1681 | obstack @var{obstack} to allocate the space. @xref{Obstacks}. | |
1682 | ||
1683 | The characters are written onto the end of the current object. | |
1684 | To get at them, you must finish the object with @code{obstack_finish} | |
1685 | (@pxref{Growing Objects}).@refill | |
1686 | @end deftypefun | |
1687 | ||
1688 | @node Variable Arguments Output | |
1689 | @subsection Variable Arguments Output Functions | |
1690 | ||
1691 | The functions @code{vprintf} and friends are provided so that you can | |
1692 | define your own variadic @code{printf}-like functions that make use of | |
1693 | the same internals as the built-in formatted output functions. | |
1694 | ||
1695 | The most natural way to define such functions would be to use a language | |
1696 | construct to say, ``Call @code{printf} and pass this template plus all | |
1697 | of my arguments after the first five.'' But there is no way to do this | |
1698 | in C, and it would be hard to provide a way, since at the C language | |
1699 | level there is no way to tell how many arguments your function received. | |
1700 | ||
1701 | Since that method is impossible, we provide alternative functions, the | |
1702 | @code{vprintf} series, which lets you pass a @code{va_list} to describe | |
1703 | ``all of my arguments after the first five.'' | |
1704 | ||
19c3f208 | 1705 | When it is sufficient to define a macro rather than a real function, |
28f540f4 RM |
1706 | the GNU C compiler provides a way to do this much more easily with macros. |
1707 | For example: | |
1708 | ||
1709 | @smallexample | |
838e5ffe UD |
1710 | #define myprintf(a, b, c, d, e, rest...) \ |
1711 | printf (mytemplate , ## rest...) | |
28f540f4 RM |
1712 | @end smallexample |
1713 | ||
1714 | @noindent | |
1715 | @xref{Macro Varargs, , Macros with Variable Numbers of Arguments, | |
1716 | gcc.info, Using GNU CC}, for details. But this is limited to macros, | |
1717 | and does not apply to real functions at all. | |
1718 | ||
1719 | Before calling @code{vprintf} or the other functions listed in this | |
1720 | section, you @emph{must} call @code{va_start} (@pxref{Variadic | |
1721 | Functions}) to initialize a pointer to the variable arguments. Then you | |
1722 | can call @code{va_arg} to fetch the arguments that you want to handle | |
1723 | yourself. This advances the pointer past those arguments. | |
1724 | ||
1725 | Once your @code{va_list} pointer is pointing at the argument of your | |
1726 | choice, you are ready to call @code{vprintf}. That argument and all | |
1727 | subsequent arguments that were passed to your function are used by | |
1728 | @code{vprintf} along with the template that you specified separately. | |
1729 | ||
1730 | In some other systems, the @code{va_list} pointer may become invalid | |
1731 | after the call to @code{vprintf}, so you must not use @code{va_arg} | |
1732 | after you call @code{vprintf}. Instead, you should call @code{va_end} | |
1733 | to retire the pointer from service. However, you can safely call | |
1734 | @code{va_start} on another pointer variable and begin fetching the | |
1735 | arguments again through that pointer. Calling @code{vprintf} does not | |
1736 | destroy the argument list of your function, merely the particular | |
1737 | pointer that you passed to it. | |
1738 | ||
1739 | GNU C does not have such restrictions. You can safely continue to fetch | |
1740 | arguments from a @code{va_list} pointer after passing it to | |
1741 | @code{vprintf}, and @code{va_end} is a no-op. (Note, however, that | |
1742 | subsequent @code{va_arg} calls will fetch the same arguments which | |
1743 | @code{vprintf} previously used.) | |
1744 | ||
1745 | Prototypes for these functions are declared in @file{stdio.h}. | |
1746 | @pindex stdio.h | |
1747 | ||
1748 | @comment stdio.h | |
f65fd747 | 1749 | @comment ISO |
28f540f4 RM |
1750 | @deftypefun int vprintf (const char *@var{template}, va_list @var{ap}) |
1751 | This function is similar to @code{printf} except that, instead of taking | |
1752 | a variable number of arguments directly, it takes an argument list | |
1753 | pointer @var{ap}. | |
1754 | @end deftypefun | |
1755 | ||
1756 | @comment stdio.h | |
f65fd747 | 1757 | @comment ISO |
28f540f4 RM |
1758 | @deftypefun int vfprintf (FILE *@var{stream}, const char *@var{template}, va_list @var{ap}) |
1759 | This is the equivalent of @code{fprintf} with the variable argument list | |
1760 | specified directly as for @code{vprintf}. | |
1761 | @end deftypefun | |
1762 | ||
1763 | @comment stdio.h | |
f65fd747 | 1764 | @comment ISO |
28f540f4 RM |
1765 | @deftypefun int vsprintf (char *@var{s}, const char *@var{template}, va_list @var{ap}) |
1766 | This is the equivalent of @code{sprintf} with the variable argument list | |
1767 | specified directly as for @code{vprintf}. | |
1768 | @end deftypefun | |
1769 | ||
1770 | @comment stdio.h | |
1771 | @comment GNU | |
1772 | @deftypefun int vsnprintf (char *@var{s}, size_t @var{size}, const char *@var{template}, va_list @var{ap}) | |
1773 | This is the equivalent of @code{snprintf} with the variable argument list | |
1774 | specified directly as for @code{vprintf}. | |
1775 | @end deftypefun | |
1776 | ||
1777 | @comment stdio.h | |
1778 | @comment GNU | |
1779 | @deftypefun int vasprintf (char **@var{ptr}, const char *@var{template}, va_list @var{ap}) | |
1780 | The @code{vasprintf} function is the equivalent of @code{asprintf} with the | |
1781 | variable argument list specified directly as for @code{vprintf}. | |
1782 | @end deftypefun | |
1783 | ||
1784 | @comment stdio.h | |
1785 | @comment GNU | |
1786 | @deftypefun int obstack_vprintf (struct obstack *@var{obstack}, const char *@var{template}, va_list @var{ap}) | |
1787 | The @code{obstack_vprintf} function is the equivalent of | |
1788 | @code{obstack_printf} with the variable argument list specified directly | |
1789 | as for @code{vprintf}.@refill | |
1790 | @end deftypefun | |
1791 | ||
1792 | Here's an example showing how you might use @code{vfprintf}. This is a | |
1793 | function that prints error messages to the stream @code{stderr}, along | |
1794 | with a prefix indicating the name of the program | |
19c3f208 | 1795 | (@pxref{Error Messages}, for a description of |
28f540f4 RM |
1796 | @code{program_invocation_short_name}). |
1797 | ||
1798 | @smallexample | |
1799 | @group | |
1800 | #include <stdio.h> | |
1801 | #include <stdarg.h> | |
1802 | ||
1803 | void | |
1804 | eprintf (const char *template, ...) | |
1805 | @{ | |
1806 | va_list ap; | |
1807 | extern char *program_invocation_short_name; | |
1808 | ||
1809 | fprintf (stderr, "%s: ", program_invocation_short_name); | |
4b8f94d3 | 1810 | va_start (ap, template); |
28f540f4 RM |
1811 | vfprintf (stderr, template, ap); |
1812 | va_end (ap); | |
1813 | @} | |
1814 | @end group | |
1815 | @end smallexample | |
1816 | ||
1817 | @noindent | |
1818 | You could call @code{eprintf} like this: | |
1819 | ||
1820 | @smallexample | |
1821 | eprintf ("file `%s' does not exist\n", filename); | |
1822 | @end smallexample | |
1823 | ||
1824 | In GNU C, there is a special construct you can use to let the compiler | |
1825 | know that a function uses a @code{printf}-style format string. Then it | |
1826 | can check the number and types of arguments in each call to the | |
1827 | function, and warn you when they do not match the format string. | |
1828 | For example, take this declaration of @code{eprintf}: | |
1829 | ||
1830 | @smallexample | |
1831 | void eprintf (const char *template, ...) | |
1832 | __attribute__ ((format (printf, 1, 2))); | |
1833 | @end smallexample | |
1834 | ||
1835 | @noindent | |
1836 | This tells the compiler that @code{eprintf} uses a format string like | |
1837 | @code{printf} (as opposed to @code{scanf}; @pxref{Formatted Input}); | |
1838 | the format string appears as the first argument; | |
1839 | and the arguments to satisfy the format begin with the second. | |
1840 | @xref{Function Attributes, , Declaring Attributes of Functions, | |
1841 | gcc.info, Using GNU CC}, for more information. | |
1842 | ||
1843 | @node Parsing a Template String | |
1844 | @subsection Parsing a Template String | |
1845 | @cindex parsing a template string | |
1846 | ||
1847 | You can use the function @code{parse_printf_format} to obtain | |
1848 | information about the number and types of arguments that are expected by | |
1849 | a given template string. This function permits interpreters that | |
1850 | provide interfaces to @code{printf} to avoid passing along invalid | |
1851 | arguments from the user's program, which could cause a crash. | |
1852 | ||
1853 | All the symbols described in this section are declared in the header | |
1854 | file @file{printf.h}. | |
1855 | ||
1856 | @comment printf.h | |
1857 | @comment GNU | |
1858 | @deftypefun size_t parse_printf_format (const char *@var{template}, size_t @var{n}, int *@var{argtypes}) | |
1859 | This function returns information about the number and types of | |
1860 | arguments expected by the @code{printf} template string @var{template}. | |
1861 | The information is stored in the array @var{argtypes}; each element of | |
1862 | this array describes one argument. This information is encoded using | |
1863 | the various @samp{PA_} macros, listed below. | |
1864 | ||
04b9968b UD |
1865 | The argument @var{n} specifies the number of elements in the array |
1866 | @var{argtypes}. This is the maximum number of elements that | |
28f540f4 RM |
1867 | @code{parse_printf_format} will try to write. |
1868 | ||
1869 | @code{parse_printf_format} returns the total number of arguments required | |
1870 | by @var{template}. If this number is greater than @var{n}, then the | |
1871 | information returned describes only the first @var{n} arguments. If you | |
04b9968b | 1872 | want information about additional arguments, allocate a bigger |
28f540f4 RM |
1873 | array and call @code{parse_printf_format} again. |
1874 | @end deftypefun | |
1875 | ||
1876 | The argument types are encoded as a combination of a basic type and | |
1877 | modifier flag bits. | |
1878 | ||
1879 | @comment printf.h | |
1880 | @comment GNU | |
1881 | @deftypevr Macro int PA_FLAG_MASK | |
1882 | This macro is a bitmask for the type modifier flag bits. You can write | |
1883 | the expression @code{(argtypes[i] & PA_FLAG_MASK)} to extract just the | |
1884 | flag bits for an argument, or @code{(argtypes[i] & ~PA_FLAG_MASK)} to | |
1885 | extract just the basic type code. | |
1886 | @end deftypevr | |
1887 | ||
1888 | Here are symbolic constants that represent the basic types; they stand | |
1889 | for integer values. | |
1890 | ||
779ae82e | 1891 | @vtable @code |
28f540f4 RM |
1892 | @comment printf.h |
1893 | @comment GNU | |
1894 | @item PA_INT | |
28f540f4 RM |
1895 | This specifies that the base type is @code{int}. |
1896 | ||
1897 | @comment printf.h | |
1898 | @comment GNU | |
1899 | @item PA_CHAR | |
28f540f4 RM |
1900 | This specifies that the base type is @code{int}, cast to @code{char}. |
1901 | ||
1902 | @comment printf.h | |
1903 | @comment GNU | |
1904 | @item PA_STRING | |
28f540f4 RM |
1905 | This specifies that the base type is @code{char *}, a null-terminated string. |
1906 | ||
1907 | @comment printf.h | |
1908 | @comment GNU | |
1909 | @item PA_POINTER | |
28f540f4 RM |
1910 | This specifies that the base type is @code{void *}, an arbitrary pointer. |
1911 | ||
1912 | @comment printf.h | |
1913 | @comment GNU | |
1914 | @item PA_FLOAT | |
28f540f4 RM |
1915 | This specifies that the base type is @code{float}. |
1916 | ||
1917 | @comment printf.h | |
1918 | @comment GNU | |
1919 | @item PA_DOUBLE | |
28f540f4 RM |
1920 | This specifies that the base type is @code{double}. |
1921 | ||
1922 | @comment printf.h | |
1923 | @comment GNU | |
1924 | @item PA_LAST | |
28f540f4 RM |
1925 | You can define additional base types for your own programs as offsets |
1926 | from @code{PA_LAST}. For example, if you have data types @samp{foo} | |
1927 | and @samp{bar} with their own specialized @code{printf} conversions, | |
1928 | you could define encodings for these types as: | |
1929 | ||
1930 | @smallexample | |
1931 | #define PA_FOO PA_LAST | |
1932 | #define PA_BAR (PA_LAST + 1) | |
1933 | @end smallexample | |
779ae82e | 1934 | @end vtable |
28f540f4 RM |
1935 | |
1936 | Here are the flag bits that modify a basic type. They are combined with | |
1937 | the code for the basic type using inclusive-or. | |
1938 | ||
779ae82e | 1939 | @vtable @code |
28f540f4 RM |
1940 | @comment printf.h |
1941 | @comment GNU | |
1942 | @item PA_FLAG_PTR | |
28f540f4 RM |
1943 | If this bit is set, it indicates that the encoded type is a pointer to |
1944 | the base type, rather than an immediate value. | |
1945 | For example, @samp{PA_INT|PA_FLAG_PTR} represents the type @samp{int *}. | |
1946 | ||
1947 | @comment printf.h | |
1948 | @comment GNU | |
1949 | @item PA_FLAG_SHORT | |
28f540f4 RM |
1950 | If this bit is set, it indicates that the base type is modified with |
1951 | @code{short}. (This corresponds to the @samp{h} type modifier.) | |
1952 | ||
1953 | @comment printf.h | |
1954 | @comment GNU | |
1955 | @item PA_FLAG_LONG | |
28f540f4 RM |
1956 | If this bit is set, it indicates that the base type is modified with |
1957 | @code{long}. (This corresponds to the @samp{l} type modifier.) | |
1958 | ||
1959 | @comment printf.h | |
1960 | @comment GNU | |
1961 | @item PA_FLAG_LONG_LONG | |
28f540f4 RM |
1962 | If this bit is set, it indicates that the base type is modified with |
1963 | @code{long long}. (This corresponds to the @samp{L} type modifier.) | |
1964 | ||
1965 | @comment printf.h | |
1966 | @comment GNU | |
1967 | @item PA_FLAG_LONG_DOUBLE | |
28f540f4 RM |
1968 | This is a synonym for @code{PA_FLAG_LONG_LONG}, used by convention with |
1969 | a base type of @code{PA_DOUBLE} to indicate a type of @code{long double}. | |
779ae82e | 1970 | @end vtable |
28f540f4 RM |
1971 | |
1972 | @ifinfo | |
6d52618b | 1973 | For an example of using these facilities, see @ref{Example of Parsing}. |
28f540f4 RM |
1974 | @end ifinfo |
1975 | ||
1976 | @node Example of Parsing | |
1977 | @subsection Example of Parsing a Template String | |
1978 | ||
1979 | Here is an example of decoding argument types for a format string. We | |
1980 | assume this is part of an interpreter which contains arguments of type | |
1981 | @code{NUMBER}, @code{CHAR}, @code{STRING} and @code{STRUCTURE} (and | |
1982 | perhaps others which are not valid here). | |
1983 | ||
1984 | @smallexample | |
1985 | /* @r{Test whether the @var{nargs} specified objects} | |
1986 | @r{in the vector @var{args} are valid} | |
1987 | @r{for the format string @var{format}:} | |
1988 | @r{if so, return 1.} | |
1989 | @r{If not, return 0 after printing an error message.} */ | |
1990 | ||
1991 | int | |
1992 | validate_args (char *format, int nargs, OBJECT *args) | |
1993 | @{ | |
1994 | int *argtypes; | |
1995 | int nwanted; | |
1996 | ||
1997 | /* @r{Get the information about the arguments.} | |
1998 | @r{Each conversion specification must be at least two characters} | |
1999 | @r{long, so there cannot be more specifications than half the} | |
2000 | @r{length of the string.} */ | |
2001 | ||
2002 | argtypes = (int *) alloca (strlen (format) / 2 * sizeof (int)); | |
2003 | nwanted = parse_printf_format (string, nelts, argtypes); | |
2004 | ||
2005 | /* @r{Check the number of arguments.} */ | |
2006 | if (nwanted > nargs) | |
2007 | @{ | |
2008 | error ("too few arguments (at least %d required)", nwanted); | |
2009 | return 0; | |
2010 | @} | |
19c3f208 | 2011 | |
28f540f4 RM |
2012 | /* @r{Check the C type wanted for each argument} |
2013 | @r{and see if the object given is suitable.} */ | |
2014 | for (i = 0; i < nwanted; i++) | |
2015 | @{ | |
2016 | int wanted; | |
2017 | ||
2018 | if (argtypes[i] & PA_FLAG_PTR) | |
2019 | wanted = STRUCTURE; | |
2020 | else | |
2021 | switch (argtypes[i] & ~PA_FLAG_MASK) | |
2022 | @{ | |
2023 | case PA_INT: | |
2024 | case PA_FLOAT: | |
2025 | case PA_DOUBLE: | |
2026 | wanted = NUMBER; | |
2027 | break; | |
2028 | case PA_CHAR: | |
2029 | wanted = CHAR; | |
2030 | break; | |
2031 | case PA_STRING: | |
2032 | wanted = STRING; | |
2033 | break; | |
2034 | case PA_POINTER: | |
2035 | wanted = STRUCTURE; | |
2036 | break; | |
2037 | @} | |
2038 | if (TYPE (args[i]) != wanted) | |
2039 | @{ | |
2040 | error ("type mismatch for arg number %d", i); | |
2041 | return 0; | |
2042 | @} | |
2043 | @} | |
2044 | return 1; | |
2045 | @} | |
2046 | @end smallexample | |
2047 | ||
2048 | @node Customizing Printf | |
2049 | @section Customizing @code{printf} | |
2050 | @cindex customizing @code{printf} | |
2051 | @cindex defining new @code{printf} conversions | |
2052 | @cindex extending @code{printf} | |
2053 | ||
2054 | The GNU C library lets you define your own custom conversion specifiers | |
2055 | for @code{printf} template strings, to teach @code{printf} clever ways | |
2056 | to print the important data structures of your program. | |
2057 | ||
2058 | The way you do this is by registering the conversion with the function | |
2059 | @code{register_printf_function}; see @ref{Registering New Conversions}. | |
2060 | One of the arguments you pass to this function is a pointer to a handler | |
2061 | function that produces the actual output; see @ref{Defining the Output | |
2062 | Handler}, for information on how to write this function. | |
2063 | ||
2064 | You can also install a function that just returns information about the | |
2065 | number and type of arguments expected by the conversion specifier. | |
2066 | @xref{Parsing a Template String}, for information about this. | |
2067 | ||
2068 | The facilities of this section are declared in the header file | |
2069 | @file{printf.h}. | |
2070 | ||
2071 | @menu | |
19c3f208 | 2072 | * Registering New Conversions:: Using @code{register_printf_function} |
28f540f4 RM |
2073 | to register a new output conversion. |
2074 | * Conversion Specifier Options:: The handler must be able to get | |
2075 | the options specified in the | |
19c3f208 | 2076 | template when it is called. |
28f540f4 RM |
2077 | * Defining the Output Handler:: Defining the handler and arginfo |
2078 | functions that are passed as arguments | |
19c3f208 | 2079 | to @code{register_printf_function}. |
28f540f4 | 2080 | * Printf Extension Example:: How to define a @code{printf} |
19c3f208 | 2081 | handler function. |
29bb8719 | 2082 | * Predefined Printf Handlers:: Predefined @code{printf} handlers. |
28f540f4 RM |
2083 | @end menu |
2084 | ||
2085 | @strong{Portability Note:} The ability to extend the syntax of | |
f65fd747 | 2086 | @code{printf} template strings is a GNU extension. ISO standard C has |
28f540f4 RM |
2087 | nothing similar. |
2088 | ||
2089 | @node Registering New Conversions | |
2090 | @subsection Registering New Conversions | |
2091 | ||
2092 | The function to register a new output conversion is | |
2093 | @code{register_printf_function}, declared in @file{printf.h}. | |
2094 | @pindex printf.h | |
2095 | ||
2096 | @comment printf.h | |
2097 | @comment GNU | |
2098 | @deftypefun int register_printf_function (int @var{spec}, printf_function @var{handler-function}, printf_arginfo_function @var{arginfo-function}) | |
2099 | This function defines the conversion specifier character @var{spec}. | |
2100 | Thus, if @var{spec} is @code{'z'}, it defines the conversion @samp{%z}. | |
2101 | You can redefine the built-in conversions like @samp{%s}, but flag | |
2102 | characters like @samp{#} and type modifiers like @samp{l} can never be | |
2103 | used as conversions; calling @code{register_printf_function} for those | |
2104 | characters has no effect. | |
2105 | ||
2106 | The @var{handler-function} is the function called by @code{printf} and | |
2107 | friends when this conversion appears in a template string. | |
2108 | @xref{Defining the Output Handler}, for information about how to define | |
2109 | a function to pass as this argument. If you specify a null pointer, any | |
2110 | existing handler function for @var{spec} is removed. | |
2111 | ||
2112 | The @var{arginfo-function} is the function called by | |
2113 | @code{parse_printf_format} when this conversion appears in a | |
2114 | template string. @xref{Parsing a Template String}, for information | |
2115 | about this. | |
2116 | ||
54d79e99 UD |
2117 | @c The following is not true anymore. The `parse_printf_format' function |
2118 | @c is now also called from `vfprintf' via `parse_one_spec'. | |
2119 | @c --drepper@gnu, 1996/11/14 | |
2120 | @c | |
2121 | @c Normally, you install both functions for a conversion at the same time, | |
2122 | @c but if you are never going to call @code{parse_printf_format}, you do | |
2123 | @c not need to define an arginfo function. | |
2124 | ||
04b9968b | 2125 | @strong{Attention:} In the GNU C library versions before 2.0 the |
54d79e99 | 2126 | @var{arginfo-function} function did not need to be installed unless |
04b9968b | 2127 | the user used the @code{parse_printf_format} function. This has changed. |
54d79e99 UD |
2128 | Now a call to any of the @code{printf} functions will call this |
2129 | function when this format specifier appears in the format string. | |
28f540f4 RM |
2130 | |
2131 | The return value is @code{0} on success, and @code{-1} on failure | |
2132 | (which occurs if @var{spec} is out of range). | |
2133 | ||
2134 | You can redefine the standard output conversions, but this is probably | |
2135 | not a good idea because of the potential for confusion. Library routines | |
2136 | written by other people could break if you do this. | |
2137 | @end deftypefun | |
2138 | ||
2139 | @node Conversion Specifier Options | |
2140 | @subsection Conversion Specifier Options | |
2141 | ||
40deae08 RM |
2142 | If you define a meaning for @samp{%A}, what if the template contains |
2143 | @samp{%+23A} or @samp{%-#A}? To implement a sensible meaning for these, | |
28f540f4 RM |
2144 | the handler when called needs to be able to get the options specified in |
2145 | the template. | |
2146 | ||
838e5ffe UD |
2147 | Both the @var{handler-function} and @var{arginfo-function} accept an |
2148 | argument that points to a @code{struct printf_info}, which contains | |
2149 | information about the options appearing in an instance of the conversion | |
2150 | specifier. This data type is declared in the header file | |
2151 | @file{printf.h}. | |
28f540f4 RM |
2152 | @pindex printf.h |
2153 | ||
2154 | @comment printf.h | |
2155 | @comment GNU | |
2156 | @deftp {Type} {struct printf_info} | |
2157 | This structure is used to pass information about the options appearing | |
2158 | in an instance of a conversion specifier in a @code{printf} template | |
2159 | string to the handler and arginfo functions for that specifier. It | |
2160 | contains the following members: | |
2161 | ||
2162 | @table @code | |
2163 | @item int prec | |
2164 | This is the precision specified. The value is @code{-1} if no precision | |
2165 | was specified. If the precision was given as @samp{*}, the | |
2166 | @code{printf_info} structure passed to the handler function contains the | |
2167 | actual value retrieved from the argument list. But the structure passed | |
2168 | to the arginfo function contains a value of @code{INT_MIN}, since the | |
2169 | actual value is not known. | |
2170 | ||
2171 | @item int width | |
2172 | This is the minimum field width specified. The value is @code{0} if no | |
2173 | width was specified. If the field width was given as @samp{*}, the | |
2174 | @code{printf_info} structure passed to the handler function contains the | |
2175 | actual value retrieved from the argument list. But the structure passed | |
2176 | to the arginfo function contains a value of @code{INT_MIN}, since the | |
2177 | actual value is not known. | |
2178 | ||
54d79e99 | 2179 | @item wchar_t spec |
28f540f4 RM |
2180 | This is the conversion specifier character specified. It's stored in |
2181 | the structure so that you can register the same handler function for | |
2182 | multiple characters, but still have a way to tell them apart when the | |
2183 | handler function is called. | |
2184 | ||
2185 | @item unsigned int is_long_double | |
2186 | This is a boolean that is true if the @samp{L}, @samp{ll}, or @samp{q} | |
2187 | type modifier was specified. For integer conversions, this indicates | |
2188 | @code{long long int}, as opposed to @code{long double} for floating | |
2189 | point conversions. | |
2190 | ||
d64b6ad0 UD |
2191 | @item unsigned int is_char |
2192 | This is a boolean that is true if the @samp{hh} type modifier was specified. | |
2193 | ||
28f540f4 RM |
2194 | @item unsigned int is_short |
2195 | This is a boolean that is true if the @samp{h} type modifier was specified. | |
2196 | ||
2197 | @item unsigned int is_long | |
2198 | This is a boolean that is true if the @samp{l} type modifier was specified. | |
2199 | ||
2200 | @item unsigned int alt | |
2201 | This is a boolean that is true if the @samp{#} flag was specified. | |
2202 | ||
2203 | @item unsigned int space | |
2204 | This is a boolean that is true if the @samp{ } flag was specified. | |
2205 | ||
2206 | @item unsigned int left | |
2207 | This is a boolean that is true if the @samp{-} flag was specified. | |
2208 | ||
2209 | @item unsigned int showsign | |
2210 | This is a boolean that is true if the @samp{+} flag was specified. | |
2211 | ||
2212 | @item unsigned int group | |
2213 | This is a boolean that is true if the @samp{'} flag was specified. | |
2214 | ||
54d79e99 UD |
2215 | @item unsigned int extra |
2216 | This flag has a special meaning depending on the context. It could | |
2217 | be used freely by the user-defined handlers but when called from | |
2218 | the @code{printf} function this variable always contains the value | |
2219 | @code{0}. | |
2220 | ||
d64b6ad0 UD |
2221 | @item unsigned int wide |
2222 | This flag is set if the stream is wide oriented. | |
2223 | ||
54d79e99 | 2224 | @item wchar_t pad |
28f540f4 RM |
2225 | This is the character to use for padding the output to the minimum field |
2226 | width. The value is @code{'0'} if the @samp{0} flag was specified, and | |
2227 | @code{' '} otherwise. | |
2228 | @end table | |
2229 | @end deftp | |
2230 | ||
2231 | ||
2232 | @node Defining the Output Handler | |
2233 | @subsection Defining the Output Handler | |
2234 | ||
2235 | Now let's look at how to define the handler and arginfo functions | |
2236 | which are passed as arguments to @code{register_printf_function}. | |
2237 | ||
04b9968b | 2238 | @strong{Compatibility Note:} The interface changed in GNU libc |
54d79e99 UD |
2239 | version 2.0. Previously the third argument was of type |
2240 | @code{va_list *}. | |
2241 | ||
28f540f4 RM |
2242 | You should define your handler functions with a prototype like: |
2243 | ||
2244 | @smallexample | |
2245 | int @var{function} (FILE *stream, const struct printf_info *info, | |
54d79e99 | 2246 | const void *const *args) |
28f540f4 RM |
2247 | @end smallexample |
2248 | ||
54d79e99 | 2249 | The @var{stream} argument passed to the handler function is the stream to |
28f540f4 RM |
2250 | which it should write output. |
2251 | ||
54d79e99 | 2252 | The @var{info} argument is a pointer to a structure that contains |
28f540f4 RM |
2253 | information about the various options that were included with the |
2254 | conversion in the template string. You should not modify this structure | |
2255 | inside your handler function. @xref{Conversion Specifier Options}, for | |
2256 | a description of this data structure. | |
2257 | ||
54d79e99 UD |
2258 | @c The following changes some time back. --drepper@gnu, 1996/11/14 |
2259 | @c | |
2260 | @c The @code{ap_pointer} argument is used to pass the tail of the variable | |
2261 | @c argument list containing the values to be printed to your handler. | |
2262 | @c Unlike most other functions that can be passed an explicit variable | |
2263 | @c argument list, this is a @emph{pointer} to a @code{va_list}, rather than | |
2264 | @c the @code{va_list} itself. Thus, you should fetch arguments by | |
2265 | @c means of @code{va_arg (*ap_pointer, @var{type})}. | |
2266 | @c | |
2267 | @c (Passing a pointer here allows the function that calls your handler | |
2268 | @c function to update its own @code{va_list} variable to account for the | |
2269 | @c arguments that your handler processes. @xref{Variadic Functions}.) | |
2270 | ||
2271 | The @var{args} is a vector of pointers to the arguments data. | |
04b9968b | 2272 | The number of arguments was determined by calling the argument |
54d79e99 | 2273 | information function provided by the user. |
28f540f4 RM |
2274 | |
2275 | Your handler function should return a value just like @code{printf} | |
2276 | does: it should return the number of characters it has written, or a | |
2277 | negative value to indicate an error. | |
2278 | ||
2279 | @comment printf.h | |
2280 | @comment GNU | |
2281 | @deftp {Data Type} printf_function | |
2282 | This is the data type that a handler function should have. | |
2283 | @end deftp | |
2284 | ||
2285 | If you are going to use @w{@code{parse_printf_format}} in your | |
54d79e99 | 2286 | application, you must also define a function to pass as the |
28f540f4 | 2287 | @var{arginfo-function} argument for each new conversion you install with |
19c3f208 | 2288 | @code{register_printf_function}. |
28f540f4 | 2289 | |
54d79e99 | 2290 | You have to define these functions with a prototype like: |
28f540f4 RM |
2291 | |
2292 | @smallexample | |
2293 | int @var{function} (const struct printf_info *info, | |
2294 | size_t n, int *argtypes) | |
2295 | @end smallexample | |
2296 | ||
2297 | The return value from the function should be the number of arguments the | |
2298 | conversion expects. The function should also fill in no more than | |
2299 | @var{n} elements of the @var{argtypes} array with information about the | |
2300 | types of each of these arguments. This information is encoded using the | |
2301 | various @samp{PA_} macros. (You will notice that this is the same | |
2302 | calling convention @code{parse_printf_format} itself uses.) | |
2303 | ||
2304 | @comment printf.h | |
2305 | @comment GNU | |
2306 | @deftp {Data Type} printf_arginfo_function | |
2307 | This type is used to describe functions that return information about | |
2308 | the number and type of arguments used by a conversion specifier. | |
2309 | @end deftp | |
2310 | ||
2311 | @node Printf Extension Example | |
2312 | @subsection @code{printf} Extension Example | |
2313 | ||
2314 | Here is an example showing how to define a @code{printf} handler function. | |
19c3f208 | 2315 | This program defines a data structure called a @code{Widget} and |
28f540f4 RM |
2316 | defines the @samp{%W} conversion to print information about @w{@code{Widget *}} |
2317 | arguments, including the pointer value and the name stored in the data | |
2318 | structure. The @samp{%W} conversion supports the minimum field width and | |
2319 | left-justification options, but ignores everything else. | |
2320 | ||
2321 | @smallexample | |
2322 | @include rprintf.c.texi | |
2323 | @end smallexample | |
2324 | ||
2325 | The output produced by this program looks like: | |
2326 | ||
2327 | @smallexample | |
2328 | |<Widget 0xffeffb7c: mywidget>| | |
2329 | | <Widget 0xffeffb7c: mywidget>| | |
2330 | |<Widget 0xffeffb7c: mywidget> | | |
2331 | @end smallexample | |
2332 | ||
29bb8719 UD |
2333 | @node Predefined Printf Handlers |
2334 | @subsection Predefined @code{printf} Handlers | |
2335 | ||
2336 | The GNU libc also contains a concrete and useful application of the | |
2337 | @code{printf} handler extension. There are two functions available | |
2338 | which implement a special way to print floating-point numbers. | |
2339 | ||
2340 | @comment printf.h | |
2341 | @comment GNU | |
2342 | @deftypefun int printf_size (FILE *@var{fp}, const struct printf_info *@var{info}, const void *const *@var{args}) | |
2343 | Print a given floating point number as for the format @code{%f} except | |
2344 | that there is a postfix character indicating the divisor for the | |
2345 | number to make this less than 1000. There are two possible divisors: | |
04b9968b | 2346 | powers of 1024 or powers of 1000. Which one is used depends on the |
29bb8719 UD |
2347 | format character specified while registered this handler. If the |
2348 | character is of lower case, 1024 is used. For upper case characters, | |
2349 | 1000 is used. | |
2350 | ||
2351 | The postfix tag corresponds to bytes, kilobytes, megabytes, gigabytes, | |
2352 | etc. The full table is: | |
2353 | ||
779ae82e UD |
2354 | @ifinfo |
2355 | @multitable @hsep @vsep {' '} {2^10 (1024)} {zetta} {Upper} {10^24 (1000)} | |
29bb8719 UD |
2356 | @item low @tab Multiplier @tab From @tab Upper @tab Multiplier |
2357 | @item ' ' @tab 1 @tab @tab ' ' @tab 1 | |
2358 | @item k @tab 2^10 (1024) @tab kilo @tab K @tab 10^3 (1000) | |
2359 | @item m @tab 2^20 @tab mega @tab M @tab 10^6 | |
2360 | @item g @tab 2^30 @tab giga @tab G @tab 10^9 | |
2361 | @item t @tab 2^40 @tab tera @tab T @tab 10^12 | |
2362 | @item p @tab 2^50 @tab peta @tab P @tab 10^15 | |
2363 | @item e @tab 2^60 @tab exa @tab E @tab 10^18 | |
2364 | @item z @tab 2^70 @tab zetta @tab Z @tab 10^21 | |
2365 | @item y @tab 2^80 @tab yotta @tab Y @tab 10^24 | |
2366 | @end multitable | |
779ae82e UD |
2367 | @end ifinfo |
2368 | @iftex | |
2369 | @tex | |
2370 | \hbox to\hsize{\hfil\vbox{\offinterlineskip | |
2371 | \hrule | |
2372 | \halign{\strut#& \vrule#\tabskip=1em plus2em& {\tt#}\hfil& \vrule#& #\hfil& \vrule#& #\hfil& \vrule#& {\tt#}\hfil& \vrule#& #\hfil& \vrule#\tabskip=0pt\cr | |
2373 | \noalign{\hrule} | |
2374 | \omit&height2pt&\omit&&\omit&&\omit&&\omit&&\omit&\cr | |
2375 | && \omit low && Multiplier && From && \omit Upper && Multiplier &\cr | |
2376 | \omit&height2pt&\omit&&\omit&&\omit&&\omit&&\omit&\cr | |
2377 | \noalign{\hrule} | |
2378 | && {\tt\char32} && 1 && && {\tt\char32} && 1 &\cr | |
2379 | && k && $2^{10} = 1024$ && kilo && K && $10^3 = 1000$ &\cr | |
2380 | && m && $2^{20}$ && mega && M && $10^6$ &\cr | |
2381 | && g && $2^{30}$ && giga && G && $10^9$ &\cr | |
2382 | && t && $2^{40}$ && tera && T && $10^{12}$ &\cr | |
2383 | && p && $2^{50}$ && peta && P && $10^{15}$ &\cr | |
2384 | && e && $2^{60}$ && exa && E && $10^{18}$ &\cr | |
2385 | && z && $2^{70}$ && zetta && Z && $10^{21}$ &\cr | |
2386 | && y && $2^{80}$ && yotta && Y && $10^{24}$ &\cr | |
2387 | \noalign{\hrule}}}\hfil} | |
2388 | @end tex | |
2389 | @end iftex | |
29bb8719 UD |
2390 | |
2391 | The default precision is 3, i.e., 1024 is printed with a lower-case | |
2392 | format character as if it were @code{%.3fk} and will yield @code{1.000k}. | |
2393 | @end deftypefun | |
2394 | ||
2395 | Due to the requirements of @code{register_printf_function} we must also | |
04b9968b | 2396 | provide the function which returns information about the arguments. |
29bb8719 UD |
2397 | |
2398 | @comment printf.h | |
2399 | @comment GNU | |
2400 | @deftypefun int printf_size_info (const struct printf_info *@var{info}, size_t @var{n}, int *@var{argtypes}) | |
2401 | This function will return in @var{argtypes} the information about the | |
2402 | used parameters in the way the @code{vfprintf} implementation expects | |
2403 | it. The format always takes one argument. | |
2404 | @end deftypefun | |
2405 | ||
2406 | To use these functions both functions must be registered with a call like | |
2407 | ||
2408 | @smallexample | |
2409 | register_printf_function ('B', printf_size, printf_size_info); | |
2410 | @end smallexample | |
2411 | ||
2412 | Here we register the functions to print numbers as powers of 1000 since | |
f2ea0f5b | 2413 | the format character @code{'B'} is an upper-case character. If we |
29bb8719 UD |
2414 | would additionally use @code{'b'} in a line like |
2415 | ||
2416 | @smallexample | |
2417 | register_printf_function ('b', printf_size, printf_size_info); | |
2418 | @end smallexample | |
2419 | ||
2420 | @noindent | |
04b9968b UD |
2421 | we could also print using a power of 1024. Please note that all that is |
2422 | different in these two lines is the format specifier. The | |
2423 | @code{printf_size} function knows about the difference between lower and upper | |
29bb8719 UD |
2424 | case format specifiers. |
2425 | ||
2426 | The use of @code{'B'} and @code{'b'} is no coincidence. Rather it is | |
2427 | the preferred way to use this functionality since it is available on | |
04b9968b | 2428 | some other systems which also use format specifiers. |
29bb8719 | 2429 | |
28f540f4 RM |
2430 | @node Formatted Input |
2431 | @section Formatted Input | |
2432 | ||
2433 | @cindex formatted input from a stream | |
2434 | @cindex reading from a stream, formatted | |
2435 | @cindex format string, for @code{scanf} | |
2436 | @cindex template, for @code{scanf} | |
2437 | The functions described in this section (@code{scanf} and related | |
2438 | functions) provide facilities for formatted input analogous to the | |
2439 | formatted output facilities. These functions provide a mechanism for | |
2440 | reading arbitrary values under the control of a @dfn{format string} or | |
2441 | @dfn{template string}. | |
2442 | ||
2443 | @menu | |
2444 | * Formatted Input Basics:: Some basics to get you started. | |
2445 | * Input Conversion Syntax:: Syntax of conversion specifications. | |
2446 | * Table of Input Conversions:: Summary of input conversions and what they do. | |
2447 | * Numeric Input Conversions:: Details of conversions for reading numbers. | |
2448 | * String Input Conversions:: Details of conversions for reading strings. | |
2449 | * Dynamic String Input:: String conversions that @code{malloc} the buffer. | |
2450 | * Other Input Conversions:: Details of miscellaneous other conversions. | |
2451 | * Formatted Input Functions:: Descriptions of the actual functions. | |
2452 | * Variable Arguments Input:: @code{vscanf} and friends. | |
2453 | @end menu | |
2454 | ||
2455 | @node Formatted Input Basics | |
2456 | @subsection Formatted Input Basics | |
2457 | ||
2458 | Calls to @code{scanf} are superficially similar to calls to | |
2459 | @code{printf} in that arbitrary arguments are read under the control of | |
2460 | a template string. While the syntax of the conversion specifications in | |
2461 | the template is very similar to that for @code{printf}, the | |
2462 | interpretation of the template is oriented more towards free-format | |
2463 | input and simple pattern matching, rather than fixed-field formatting. | |
2464 | For example, most @code{scanf} conversions skip over any amount of | |
2465 | ``white space'' (including spaces, tabs, and newlines) in the input | |
2466 | file, and there is no concept of precision for the numeric input | |
2467 | conversions as there is for the corresponding output conversions. | |
2468 | Ordinarily, non-whitespace characters in the template are expected to | |
2469 | match characters in the input stream exactly, but a matching failure is | |
2470 | distinct from an input error on the stream. | |
2471 | @cindex conversion specifications (@code{scanf}) | |
2472 | ||
2473 | Another area of difference between @code{scanf} and @code{printf} is | |
2474 | that you must remember to supply pointers rather than immediate values | |
2475 | as the optional arguments to @code{scanf}; the values that are read are | |
2476 | stored in the objects that the pointers point to. Even experienced | |
2477 | programmers tend to forget this occasionally, so if your program is | |
2478 | getting strange errors that seem to be related to @code{scanf}, you | |
2479 | might want to double-check this. | |
2480 | ||
2481 | When a @dfn{matching failure} occurs, @code{scanf} returns immediately, | |
2482 | leaving the first non-matching character as the next character to be | |
2483 | read from the stream. The normal return value from @code{scanf} is the | |
2484 | number of values that were assigned, so you can use this to determine if | |
2485 | a matching error happened before all the expected values were read. | |
2486 | @cindex matching failure, in @code{scanf} | |
2487 | ||
2488 | The @code{scanf} function is typically used for things like reading in | |
2489 | the contents of tables. For example, here is a function that uses | |
2490 | @code{scanf} to initialize an array of @code{double}: | |
2491 | ||
2492 | @smallexample | |
2493 | void | |
2494 | readarray (double *array, int n) | |
2495 | @{ | |
2496 | int i; | |
2497 | for (i=0; i<n; i++) | |
2498 | if (scanf (" %lf", &(array[i])) != 1) | |
2499 | invalid_input_error (); | |
2500 | @} | |
2501 | @end smallexample | |
2502 | ||
2503 | The formatted input functions are not used as frequently as the | |
2504 | formatted output functions. Partly, this is because it takes some care | |
2505 | to use them properly. Another reason is that it is difficult to recover | |
2506 | from a matching error. | |
2507 | ||
2508 | If you are trying to read input that doesn't match a single, fixed | |
2509 | pattern, you may be better off using a tool such as Flex to generate a | |
2510 | lexical scanner, or Bison to generate a parser, rather than using | |
2511 | @code{scanf}. For more information about these tools, see @ref{, , , | |
2512 | flex.info, Flex: The Lexical Scanner Generator}, and @ref{, , , | |
2513 | bison.info, The Bison Reference Manual}. | |
2514 | ||
2515 | @node Input Conversion Syntax | |
2516 | @subsection Input Conversion Syntax | |
2517 | ||
2518 | A @code{scanf} template string is a string that contains ordinary | |
2519 | multibyte characters interspersed with conversion specifications that | |
2520 | start with @samp{%}. | |
2521 | ||
2522 | Any whitespace character (as defined by the @code{isspace} function; | |
2523 | @pxref{Classification of Characters}) in the template causes any number | |
2524 | of whitespace characters in the input stream to be read and discarded. | |
2525 | The whitespace characters that are matched need not be exactly the same | |
2526 | whitespace characters that appear in the template string. For example, | |
2527 | write @samp{ , } in the template to recognize a comma with optional | |
2528 | whitespace before and after. | |
2529 | ||
2530 | Other characters in the template string that are not part of conversion | |
2531 | specifications must match characters in the input stream exactly; if | |
2532 | this is not the case, a matching failure occurs. | |
2533 | ||
2534 | The conversion specifications in a @code{scanf} template string | |
2535 | have the general form: | |
2536 | ||
2537 | @smallexample | |
2538 | % @var{flags} @var{width} @var{type} @var{conversion} | |
2539 | @end smallexample | |
2540 | ||
2541 | In more detail, an input conversion specification consists of an initial | |
2542 | @samp{%} character followed in sequence by: | |
2543 | ||
2544 | @itemize @bullet | |
2545 | @item | |
2546 | An optional @dfn{flag character} @samp{*}, which says to ignore the text | |
2547 | read for this specification. When @code{scanf} finds a conversion | |
2548 | specification that uses this flag, it reads input as directed by the | |
2549 | rest of the conversion specification, but it discards this input, does | |
2550 | not use a pointer argument, and does not increment the count of | |
2551 | successful assignments. | |
2552 | @cindex flag character (@code{scanf}) | |
2553 | ||
2554 | @item | |
2555 | An optional flag character @samp{a} (valid with string conversions only) | |
2556 | which requests allocation of a buffer long enough to store the string in. | |
2557 | (This is a GNU extension.) | |
2558 | @xref{Dynamic String Input}. | |
2559 | ||
2560 | @item | |
2561 | An optional decimal integer that specifies the @dfn{maximum field | |
2562 | width}. Reading of characters from the input stream stops either when | |
2563 | this maximum is reached or when a non-matching character is found, | |
2564 | whichever happens first. Most conversions discard initial whitespace | |
2565 | characters (those that don't are explicitly documented), and these | |
2566 | discarded characters don't count towards the maximum field width. | |
2567 | String input conversions store a null character to mark the end of the | |
2568 | input; the maximum field width does not include this terminator. | |
2569 | @cindex maximum field width (@code{scanf}) | |
2570 | ||
2571 | @item | |
2572 | An optional @dfn{type modifier character}. For example, you can | |
2573 | specify a type modifier of @samp{l} with integer conversions such as | |
2574 | @samp{%d} to specify that the argument is a pointer to a @code{long int} | |
2575 | rather than a pointer to an @code{int}. | |
2576 | @cindex type modifier character (@code{scanf}) | |
2577 | ||
2578 | @item | |
2579 | A character that specifies the conversion to be applied. | |
2580 | @end itemize | |
2581 | ||
19c3f208 | 2582 | The exact options that are permitted and how they are interpreted vary |
28f540f4 RM |
2583 | between the different conversion specifiers. See the descriptions of the |
2584 | individual conversions for information about the particular options that | |
2585 | they allow. | |
2586 | ||
2587 | With the @samp{-Wformat} option, the GNU C compiler checks calls to | |
2588 | @code{scanf} and related functions. It examines the format string and | |
2589 | verifies that the correct number and types of arguments are supplied. | |
2590 | There is also a GNU C syntax to tell the compiler that a function you | |
19c3f208 | 2591 | write uses a @code{scanf}-style format string. |
28f540f4 RM |
2592 | @xref{Function Attributes, , Declaring Attributes of Functions, |
2593 | gcc.info, Using GNU CC}, for more information. | |
2594 | ||
2595 | @node Table of Input Conversions | |
2596 | @subsection Table of Input Conversions | |
2597 | @cindex input conversions, for @code{scanf} | |
2598 | ||
2599 | Here is a table that summarizes the various conversion specifications: | |
2600 | ||
2601 | @table @asis | |
2602 | @item @samp{%d} | |
2603 | Matches an optionally signed integer written in decimal. @xref{Numeric | |
2604 | Input Conversions}. | |
2605 | ||
2606 | @item @samp{%i} | |
2607 | Matches an optionally signed integer in any of the formats that the C | |
2608 | language defines for specifying an integer constant. @xref{Numeric | |
2609 | Input Conversions}. | |
2610 | ||
2611 | @item @samp{%o} | |
2612 | Matches an unsigned integer written in octal radix. | |
2613 | @xref{Numeric Input Conversions}. | |
2614 | ||
2615 | @item @samp{%u} | |
2616 | Matches an unsigned integer written in decimal radix. | |
2617 | @xref{Numeric Input Conversions}. | |
2618 | ||
2619 | @item @samp{%x}, @samp{%X} | |
2620 | Matches an unsigned integer written in hexadecimal radix. | |
2621 | @xref{Numeric Input Conversions}. | |
2622 | ||
2623 | @item @samp{%e}, @samp{%f}, @samp{%g}, @samp{%E}, @samp{%G} | |
2624 | Matches an optionally signed floating-point number. @xref{Numeric Input | |
2625 | Conversions}. | |
2626 | ||
2627 | @item @samp{%s} | |
2628 | Matches a string containing only non-whitespace characters. | |
2629 | @xref{String Input Conversions}. | |
2630 | ||
2631 | @item @samp{%[} | |
2632 | Matches a string of characters that belong to a specified set. | |
2633 | @xref{String Input Conversions}. | |
2634 | ||
2635 | @item @samp{%c} | |
2636 | Matches a string of one or more characters; the number of characters | |
2637 | read is controlled by the maximum field width given for the conversion. | |
2638 | @xref{String Input Conversions}. | |
2639 | ||
2640 | @item @samp{%p} | |
2641 | Matches a pointer value in the same implementation-defined format used | |
2642 | by the @samp{%p} output conversion for @code{printf}. @xref{Other Input | |
2643 | Conversions}. | |
2644 | ||
2645 | @item @samp{%n} | |
2646 | This conversion doesn't read any characters; it records the number of | |
2647 | characters read so far by this call. @xref{Other Input Conversions}. | |
2648 | ||
2649 | @item @samp{%%} | |
2650 | This matches a literal @samp{%} character in the input stream. No | |
2651 | corresponding argument is used. @xref{Other Input Conversions}. | |
2652 | @end table | |
2653 | ||
2654 | If the syntax of a conversion specification is invalid, the behavior is | |
2655 | undefined. If there aren't enough function arguments provided to supply | |
2656 | addresses for all the conversion specifications in the template strings | |
2657 | that perform assignments, or if the arguments are not of the correct | |
2658 | types, the behavior is also undefined. On the other hand, extra | |
2659 | arguments are simply ignored. | |
2660 | ||
2661 | @node Numeric Input Conversions | |
2662 | @subsection Numeric Input Conversions | |
2663 | ||
2664 | This section describes the @code{scanf} conversions for reading numeric | |
2665 | values. | |
2666 | ||
2667 | The @samp{%d} conversion matches an optionally signed integer in decimal | |
2668 | radix. The syntax that is recognized is the same as that for the | |
2669 | @code{strtol} function (@pxref{Parsing of Integers}) with the value | |
2670 | @code{10} for the @var{base} argument. | |
2671 | ||
2672 | The @samp{%i} conversion matches an optionally signed integer in any of | |
2673 | the formats that the C language defines for specifying an integer | |
2674 | constant. The syntax that is recognized is the same as that for the | |
2675 | @code{strtol} function (@pxref{Parsing of Integers}) with the value | |
2676 | @code{0} for the @var{base} argument. (You can print integers in this | |
2677 | syntax with @code{printf} by using the @samp{#} flag character with the | |
2678 | @samp{%x}, @samp{%o}, or @samp{%d} conversion. @xref{Integer Conversions}.) | |
2679 | ||
2680 | For example, any of the strings @samp{10}, @samp{0xa}, or @samp{012} | |
2681 | could be read in as integers under the @samp{%i} conversion. Each of | |
2682 | these specifies a number with decimal value @code{10}. | |
2683 | ||
2684 | The @samp{%o}, @samp{%u}, and @samp{%x} conversions match unsigned | |
2685 | integers in octal, decimal, and hexadecimal radices, respectively. The | |
2686 | syntax that is recognized is the same as that for the @code{strtoul} | |
2687 | function (@pxref{Parsing of Integers}) with the appropriate value | |
2688 | (@code{8}, @code{10}, or @code{16}) for the @var{base} argument. | |
2689 | ||
2690 | The @samp{%X} conversion is identical to the @samp{%x} conversion. They | |
2691 | both permit either uppercase or lowercase letters to be used as digits. | |
2692 | ||
2693 | The default type of the corresponding argument for the @code{%d} and | |
2694 | @code{%i} conversions is @code{int *}, and @code{unsigned int *} for the | |
2695 | other integer conversions. You can use the following type modifiers to | |
2696 | specify other sizes of integer: | |
2697 | ||
2698 | @table @samp | |
cc3fa755 UD |
2699 | @item hh |
2700 | Specifies that the argument is a @code{signed char *} or @code{unsigned | |
2701 | char *}. | |
2702 | ||
ec751a23 | 2703 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 2704 | |
28f540f4 RM |
2705 | @item h |
2706 | Specifies that the argument is a @code{short int *} or @code{unsigned | |
2707 | short int *}. | |
2708 | ||
e852e889 UD |
2709 | @item j |
2710 | Specifies that the argument is a @code{intmax_t *} or @code{uintmax_t *}. | |
2711 | ||
ec751a23 | 2712 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 2713 | |
28f540f4 RM |
2714 | @item l |
2715 | Specifies that the argument is a @code{long int *} or @code{unsigned | |
2716 | long int *}. Two @samp{l} characters is like the @samp{L} modifier, below. | |
2717 | ||
2718 | @need 100 | |
2719 | @item ll | |
2720 | @itemx L | |
2721 | @itemx q | |
2722 | Specifies that the argument is a @code{long long int *} or @code{unsigned long long int *}. (The @code{long long} type is an extension supported by the | |
2723 | GNU C compiler. For systems that don't provide extra-long integers, this | |
2724 | is the same as @code{long int}.) | |
2725 | ||
2726 | The @samp{q} modifier is another name for the same thing, which comes | |
2727 | from 4.4 BSD; a @w{@code{long long int}} is sometimes called a ``quad'' | |
2728 | @code{int}. | |
e852e889 UD |
2729 | |
2730 | @item t | |
2731 | Specifies that the argument is a @code{ptrdiff_t *}. | |
2732 | ||
ec751a23 | 2733 | This modifier was introduced in @w{ISO C99}. |
e852e889 UD |
2734 | |
2735 | @item z | |
2736 | Specifies that the argument is a @code{size_t *}. | |
2737 | ||
ec751a23 | 2738 | This modifier was introduced in @w{ISO C99}. |
28f540f4 RM |
2739 | @end table |
2740 | ||
2741 | All of the @samp{%e}, @samp{%f}, @samp{%g}, @samp{%E}, and @samp{%G} | |
2742 | input conversions are interchangeable. They all match an optionally | |
2743 | signed floating point number, in the same syntax as for the | |
2744 | @code{strtod} function (@pxref{Parsing of Floats}). | |
2745 | ||
2746 | For the floating-point input conversions, the default argument type is | |
2747 | @code{float *}. (This is different from the corresponding output | |
2748 | conversions, where the default type is @code{double}; remember that | |
2749 | @code{float} arguments to @code{printf} are converted to @code{double} | |
2750 | by the default argument promotions, but @code{float *} arguments are | |
2751 | not promoted to @code{double *}.) You can specify other sizes of float | |
2752 | using these type modifiers: | |
2753 | ||
2754 | @table @samp | |
2755 | @item l | |
2756 | Specifies that the argument is of type @code{double *}. | |
2757 | ||
2758 | @item L | |
2759 | Specifies that the argument is of type @code{long double *}. | |
2760 | @end table | |
2761 | ||
2c6fe0bd UD |
2762 | For all the above number parsing formats there is an additional optional |
2763 | flag @samp{'}. When this flag is given the @code{scanf} function | |
2764 | expects the number represented in the input string to be formatted | |
2765 | according to the grouping rules of the currently selected locale | |
2766 | (@pxref{General Numeric}). | |
2767 | ||
2768 | If the @code{"C"} or @code{"POSIX"} locale is selected there is no | |
2769 | difference. But for a locale which specifies values for the appropriate | |
2770 | fields in the locale the input must have the correct form in the input. | |
2771 | Otherwise the longest prefix with a correct form is processed. | |
2772 | ||
28f540f4 RM |
2773 | @node String Input Conversions |
2774 | @subsection String Input Conversions | |
2775 | ||
2776 | This section describes the @code{scanf} input conversions for reading | |
19c3f208 | 2777 | string and character values: @samp{%s}, @samp{%[}, and @samp{%c}. |
28f540f4 RM |
2778 | |
2779 | You have two options for how to receive the input from these | |
2780 | conversions: | |
2781 | ||
2782 | @itemize @bullet | |
2783 | @item | |
2784 | Provide a buffer to store it in. This is the default. You | |
2785 | should provide an argument of type @code{char *}. | |
2786 | ||
2787 | @strong{Warning:} To make a robust program, you must make sure that the | |
2788 | input (plus its terminating null) cannot possibly exceed the size of the | |
2789 | buffer you provide. In general, the only way to do this is to specify a | |
2790 | maximum field width one less than the buffer size. @strong{If you | |
2791 | provide the buffer, always specify a maximum field width to prevent | |
2792 | overflow.} | |
2793 | ||
2794 | @item | |
2795 | Ask @code{scanf} to allocate a big enough buffer, by specifying the | |
2796 | @samp{a} flag character. This is a GNU extension. You should provide | |
2797 | an argument of type @code{char **} for the buffer address to be stored | |
2798 | in. @xref{Dynamic String Input}. | |
2799 | @end itemize | |
2800 | ||
2801 | The @samp{%c} conversion is the simplest: it matches a fixed number of | |
04b9968b | 2802 | characters, always. The maximum field width says how many characters to |
28f540f4 RM |
2803 | read; if you don't specify the maximum, the default is 1. This |
2804 | conversion doesn't append a null character to the end of the text it | |
2805 | reads. It also does not skip over initial whitespace characters. It | |
2806 | reads precisely the next @var{n} characters, and fails if it cannot get | |
2807 | that many. Since there is always a maximum field width with @samp{%c} | |
2808 | (whether specified, or 1 by default), you can always prevent overflow by | |
2809 | making the buffer long enough. | |
2810 | ||
2811 | The @samp{%s} conversion matches a string of non-whitespace characters. | |
2812 | It skips and discards initial whitespace, but stops when it encounters | |
2813 | more whitespace after having read something. It stores a null character | |
2814 | at the end of the text that it reads. | |
2815 | ||
2816 | For example, reading the input: | |
2817 | ||
2818 | @smallexample | |
2819 | hello, world | |
2820 | @end smallexample | |
2821 | ||
2822 | @noindent | |
2823 | with the conversion @samp{%10c} produces @code{" hello, wo"}, but | |
2824 | reading the same input with the conversion @samp{%10s} produces | |
2825 | @code{"hello,"}. | |
2826 | ||
2827 | @strong{Warning:} If you do not specify a field width for @samp{%s}, | |
2828 | then the number of characters read is limited only by where the next | |
2829 | whitespace character appears. This almost certainly means that invalid | |
2830 | input can make your program crash---which is a bug. | |
2831 | ||
2832 | To read in characters that belong to an arbitrary set of your choice, | |
2833 | use the @samp{%[} conversion. You specify the set between the @samp{[} | |
2834 | character and a following @samp{]} character, using the same syntax used | |
2835 | in regular expressions. As special cases: | |
2836 | ||
2837 | @itemize @bullet | |
19c3f208 | 2838 | @item |
28f540f4 RM |
2839 | A literal @samp{]} character can be specified as the first character |
2840 | of the set. | |
2841 | ||
19c3f208 | 2842 | @item |
28f540f4 RM |
2843 | An embedded @samp{-} character (that is, one that is not the first or |
2844 | last character of the set) is used to specify a range of characters. | |
2845 | ||
19c3f208 | 2846 | @item |
28f540f4 RM |
2847 | If a caret character @samp{^} immediately follows the initial @samp{[}, |
2848 | then the set of allowed input characters is the everything @emph{except} | |
2849 | the characters listed. | |
2850 | @end itemize | |
2851 | ||
2852 | The @samp{%[} conversion does not skip over initial whitespace | |
2853 | characters. | |
2854 | ||
2855 | Here are some examples of @samp{%[} conversions and what they mean: | |
2856 | ||
2857 | @table @samp | |
2858 | @item %25[1234567890] | |
2859 | Matches a string of up to 25 digits. | |
2860 | ||
2861 | @item %25[][] | |
2862 | Matches a string of up to 25 square brackets. | |
2863 | ||
2864 | @item %25[^ \f\n\r\t\v] | |
2865 | Matches a string up to 25 characters long that doesn't contain any of | |
2866 | the standard whitespace characters. This is slightly different from | |
2867 | @samp{%s}, because if the input begins with a whitespace character, | |
2868 | @samp{%[} reports a matching failure while @samp{%s} simply discards the | |
2869 | initial whitespace. | |
2870 | ||
19c3f208 | 2871 | @item %25[a-z] |
28f540f4 RM |
2872 | Matches up to 25 lowercase characters. |
2873 | @end table | |
2874 | ||
2875 | One more reminder: the @samp{%s} and @samp{%[} conversions are | |
2876 | @strong{dangerous} if you don't specify a maximum width or use the | |
2877 | @samp{a} flag, because input too long would overflow whatever buffer you | |
2878 | have provided for it. No matter how long your buffer is, a user could | |
2879 | supply input that is longer. A well-written program reports invalid | |
2880 | input with a comprehensible error message, not with a crash. | |
2881 | ||
2882 | @node Dynamic String Input | |
2883 | @subsection Dynamically Allocating String Conversions | |
2884 | ||
2885 | A GNU extension to formatted input lets you safely read a string with no | |
2886 | maximum size. Using this feature, you don't supply a buffer; instead, | |
2887 | @code{scanf} allocates a buffer big enough to hold the data and gives | |
2888 | you its address. To use this feature, write @samp{a} as a flag | |
2889 | character, as in @samp{%as} or @samp{%a[0-9a-z]}. | |
2890 | ||
2891 | The pointer argument you supply for where to store the input should have | |
2892 | type @code{char **}. The @code{scanf} function allocates a buffer and | |
2893 | stores its address in the word that the argument points to. You should | |
2894 | free the buffer with @code{free} when you no longer need it. | |
2895 | ||
2896 | Here is an example of using the @samp{a} flag with the @samp{%[@dots{}]} | |
2897 | conversion specification to read a ``variable assignment'' of the form | |
2898 | @samp{@var{variable} = @var{value}}. | |
2899 | ||
2900 | @smallexample | |
2901 | @{ | |
2902 | char *variable, *value; | |
2903 | ||
2904 | if (2 > scanf ("%a[a-zA-Z0-9] = %a[^\n]\n", | |
2905 | &variable, &value)) | |
2906 | @{ | |
2907 | invalid_input_error (); | |
2908 | return 0; | |
2909 | @} | |
2910 | ||
2911 | @dots{} | |
2912 | @} | |
2913 | @end smallexample | |
2914 | ||
2915 | @node Other Input Conversions | |
2916 | @subsection Other Input Conversions | |
2917 | ||
2918 | This section describes the miscellaneous input conversions. | |
2919 | ||
2920 | The @samp{%p} conversion is used to read a pointer value. It recognizes | |
04b9968b | 2921 | the same syntax used by the @samp{%p} output conversion for |
28f540f4 RM |
2922 | @code{printf} (@pxref{Other Output Conversions}); that is, a hexadecimal |
2923 | number just as the @samp{%x} conversion accepts. The corresponding | |
2924 | argument should be of type @code{void **}; that is, the address of a | |
2925 | place to store a pointer. | |
2926 | ||
2927 | The resulting pointer value is not guaranteed to be valid if it was not | |
2928 | originally written during the same program execution that reads it in. | |
2929 | ||
2930 | The @samp{%n} conversion produces the number of characters read so far | |
2931 | by this call. The corresponding argument should be of type @code{int *}. | |
2932 | This conversion works in the same way as the @samp{%n} conversion for | |
2933 | @code{printf}; see @ref{Other Output Conversions}, for an example. | |
2934 | ||
2935 | The @samp{%n} conversion is the only mechanism for determining the | |
2936 | success of literal matches or conversions with suppressed assignments. | |
2937 | If the @samp{%n} follows the locus of a matching failure, then no value | |
2938 | is stored for it since @code{scanf} returns before processing the | |
2939 | @samp{%n}. If you store @code{-1} in that argument slot before calling | |
2940 | @code{scanf}, the presence of @code{-1} after @code{scanf} indicates an | |
2941 | error occurred before the @samp{%n} was reached. | |
2942 | ||
2943 | Finally, the @samp{%%} conversion matches a literal @samp{%} character | |
2944 | in the input stream, without using an argument. This conversion does | |
2945 | not permit any flags, field width, or type modifier to be specified. | |
2946 | ||
2947 | @node Formatted Input Functions | |
2948 | @subsection Formatted Input Functions | |
2949 | ||
2950 | Here are the descriptions of the functions for performing formatted | |
2951 | input. | |
2952 | Prototypes for these functions are in the header file @file{stdio.h}. | |
2953 | @pindex stdio.h | |
2954 | ||
2955 | @comment stdio.h | |
f65fd747 | 2956 | @comment ISO |
28f540f4 RM |
2957 | @deftypefun int scanf (const char *@var{template}, @dots{}) |
2958 | The @code{scanf} function reads formatted input from the stream | |
2959 | @code{stdin} under the control of the template string @var{template}. | |
2960 | The optional arguments are pointers to the places which receive the | |
2961 | resulting values. | |
2962 | ||
2963 | The return value is normally the number of successful assignments. If | |
04b9968b UD |
2964 | an end-of-file condition is detected before any matches are performed, |
2965 | including matches against whitespace and literal characters in the | |
2966 | template, then @code{EOF} is returned. | |
28f540f4 RM |
2967 | @end deftypefun |
2968 | ||
2969 | @comment stdio.h | |
f65fd747 | 2970 | @comment ISO |
28f540f4 RM |
2971 | @deftypefun int fscanf (FILE *@var{stream}, const char *@var{template}, @dots{}) |
2972 | This function is just like @code{scanf}, except that the input is read | |
2973 | from the stream @var{stream} instead of @code{stdin}. | |
2974 | @end deftypefun | |
2975 | ||
2976 | @comment stdio.h | |
f65fd747 | 2977 | @comment ISO |
28f540f4 RM |
2978 | @deftypefun int sscanf (const char *@var{s}, const char *@var{template}, @dots{}) |
2979 | This is like @code{scanf}, except that the characters are taken from the | |
2980 | null-terminated string @var{s} instead of from a stream. Reaching the | |
2981 | end of the string is treated as an end-of-file condition. | |
2982 | ||
2983 | The behavior of this function is undefined if copying takes place | |
2984 | between objects that overlap---for example, if @var{s} is also given | |
2985 | as an argument to receive a string read under control of the @samp{%s} | |
2986 | conversion. | |
2987 | @end deftypefun | |
2988 | ||
2989 | @node Variable Arguments Input | |
2990 | @subsection Variable Arguments Input Functions | |
2991 | ||
2992 | The functions @code{vscanf} and friends are provided so that you can | |
2993 | define your own variadic @code{scanf}-like functions that make use of | |
2994 | the same internals as the built-in formatted output functions. | |
2995 | These functions are analogous to the @code{vprintf} series of output | |
2996 | functions. @xref{Variable Arguments Output}, for important | |
2997 | information on how to use them. | |
2998 | ||
2999 | @strong{Portability Note:} The functions listed in this section are GNU | |
3000 | extensions. | |
3001 | ||
3002 | @comment stdio.h | |
3003 | @comment GNU | |
3004 | @deftypefun int vscanf (const char *@var{template}, va_list @var{ap}) | |
04b9968b | 3005 | This function is similar to @code{scanf}, but instead of taking |
28f540f4 RM |
3006 | a variable number of arguments directly, it takes an argument list |
3007 | pointer @var{ap} of type @code{va_list} (@pxref{Variadic Functions}). | |
3008 | @end deftypefun | |
3009 | ||
3010 | @comment stdio.h | |
3011 | @comment GNU | |
3012 | @deftypefun int vfscanf (FILE *@var{stream}, const char *@var{template}, va_list @var{ap}) | |
3013 | This is the equivalent of @code{fscanf} with the variable argument list | |
3014 | specified directly as for @code{vscanf}. | |
3015 | @end deftypefun | |
3016 | ||
3017 | @comment stdio.h | |
3018 | @comment GNU | |
3019 | @deftypefun int vsscanf (const char *@var{s}, const char *@var{template}, va_list @var{ap}) | |
3020 | This is the equivalent of @code{sscanf} with the variable argument list | |
3021 | specified directly as for @code{vscanf}. | |
3022 | @end deftypefun | |
3023 | ||
3024 | In GNU C, there is a special construct you can use to let the compiler | |
3025 | know that a function uses a @code{scanf}-style format string. Then it | |
3026 | can check the number and types of arguments in each call to the | |
3027 | function, and warn you when they do not match the format string. | |
04b9968b UD |
3028 | For details, @xref{Function Attributes, , Declaring Attributes of Functions, |
3029 | gcc.info, Using GNU CC}. | |
28f540f4 RM |
3030 | |
3031 | @node EOF and Errors | |
3032 | @section End-Of-File and Errors | |
3033 | ||
3034 | @cindex end of file, on a stream | |
3035 | Many of the functions described in this chapter return the value of the | |
3036 | macro @code{EOF} to indicate unsuccessful completion of the operation. | |
3037 | Since @code{EOF} is used to report both end of file and random errors, | |
3038 | it's often better to use the @code{feof} function to check explicitly | |
3039 | for end of file and @code{ferror} to check for errors. These functions | |
3040 | check indicators that are part of the internal state of the stream | |
3041 | object, indicators set if the appropriate condition was detected by a | |
3042 | previous I/O operation on that stream. | |
3043 | ||
3044 | These symbols are declared in the header file @file{stdio.h}. | |
3045 | @pindex stdio.h | |
3046 | ||
3047 | @comment stdio.h | |
f65fd747 | 3048 | @comment ISO |
28f540f4 RM |
3049 | @deftypevr Macro int EOF |
3050 | This macro is an integer value that is returned by a number of functions | |
3051 | to indicate an end-of-file condition, or some other error situation. | |
3052 | With the GNU library, @code{EOF} is @code{-1}. In other libraries, its | |
3053 | value may be some other negative number. | |
3054 | @end deftypevr | |
3055 | ||
3056 | @comment stdio.h | |
f65fd747 | 3057 | @comment ISO |
28f540f4 RM |
3058 | @deftypefun void clearerr (FILE *@var{stream}) |
3059 | This function clears the end-of-file and error indicators for the | |
3060 | stream @var{stream}. | |
3061 | ||
3062 | The file positioning functions (@pxref{File Positioning}) also clear the | |
3063 | end-of-file indicator for the stream. | |
3064 | @end deftypefun | |
3065 | ||
3066 | @comment stdio.h | |
f65fd747 | 3067 | @comment ISO |
28f540f4 RM |
3068 | @deftypefun int feof (FILE *@var{stream}) |
3069 | The @code{feof} function returns nonzero if and only if the end-of-file | |
3070 | indicator for the stream @var{stream} is set. | |
3071 | @end deftypefun | |
3072 | ||
3073 | @comment stdio.h | |
f65fd747 | 3074 | @comment ISO |
28f540f4 RM |
3075 | @deftypefun int ferror (FILE *@var{stream}) |
3076 | The @code{ferror} function returns nonzero if and only if the error | |
3077 | indicator for the stream @var{stream} is set, indicating that an error | |
3078 | has occurred on a previous operation on the stream. | |
3079 | @end deftypefun | |
3080 | ||
3081 | In addition to setting the error indicator associated with the stream, | |
3082 | the functions that operate on streams also set @code{errno} in the same | |
3083 | way as the corresponding low-level functions that operate on file | |
3084 | descriptors. For example, all of the functions that perform output to a | |
3085 | stream---such as @code{fputc}, @code{printf}, and @code{fflush}---are | |
3086 | implemented in terms of @code{write}, and all of the @code{errno} error | |
3087 | conditions defined for @code{write} are meaningful for these functions. | |
3088 | For more information about the descriptor-level I/O functions, see | |
3089 | @ref{Low-Level I/O}. | |
3090 | ||
3091 | @node Binary Streams | |
3092 | @section Text and Binary Streams | |
3093 | ||
3094 | The GNU system and other POSIX-compatible operating systems organize all | |
3095 | files as uniform sequences of characters. However, some other systems | |
3096 | make a distinction between files containing text and files containing | |
f65fd747 | 3097 | binary data, and the input and output facilities of @w{ISO C} provide for |
28f540f4 RM |
3098 | this distinction. This section tells you how to write programs portable |
3099 | to such systems. | |
3100 | ||
3101 | @cindex text stream | |
3102 | @cindex binary stream | |
3103 | When you open a stream, you can specify either a @dfn{text stream} or a | |
3104 | @dfn{binary stream}. You indicate that you want a binary stream by | |
3105 | specifying the @samp{b} modifier in the @var{opentype} argument to | |
3106 | @code{fopen}; see @ref{Opening Streams}. Without this | |
3107 | option, @code{fopen} opens the file as a text stream. | |
3108 | ||
3109 | Text and binary streams differ in several ways: | |
3110 | ||
3111 | @itemize @bullet | |
3112 | @item | |
3113 | The data read from a text stream is divided into @dfn{lines} which are | |
3114 | terminated by newline (@code{'\n'}) characters, while a binary stream is | |
3115 | simply a long series of characters. A text stream might on some systems | |
3116 | fail to handle lines more than 254 characters long (including the | |
3117 | terminating newline character). | |
3118 | @cindex lines (in a text file) | |
3119 | ||
3120 | @item | |
3121 | On some systems, text files can contain only printing characters, | |
3122 | horizontal tab characters, and newlines, and so text streams may not | |
3123 | support other characters. However, binary streams can handle any | |
3124 | character value. | |
3125 | ||
3126 | @item | |
3127 | Space characters that are written immediately preceding a newline | |
3128 | character in a text stream may disappear when the file is read in again. | |
3129 | ||
3130 | @item | |
3131 | More generally, there need not be a one-to-one mapping between | |
3132 | characters that are read from or written to a text stream, and the | |
3133 | characters in the actual file. | |
3134 | @end itemize | |
3135 | ||
3136 | Since a binary stream is always more capable and more predictable than a | |
3137 | text stream, you might wonder what purpose text streams serve. Why not | |
3138 | simply always use binary streams? The answer is that on these operating | |
3139 | systems, text and binary streams use different file formats, and the | |
3140 | only way to read or write ``an ordinary file of text'' that can work | |
3141 | with other text-oriented programs is through a text stream. | |
3142 | ||
3143 | In the GNU library, and on all POSIX systems, there is no difference | |
3144 | between text streams and binary streams. When you open a stream, you | |
3145 | get the same kind of stream regardless of whether you ask for binary. | |
3146 | This stream can handle any file content, and has none of the | |
3147 | restrictions that text streams sometimes have. | |
3148 | ||
3149 | @node File Positioning | |
3150 | @section File Positioning | |
3151 | @cindex file positioning on a stream | |
3152 | @cindex positioning a stream | |
3153 | @cindex seeking on a stream | |
3154 | ||
3155 | The @dfn{file position} of a stream describes where in the file the | |
3156 | stream is currently reading or writing. I/O on the stream advances the | |
3157 | file position through the file. In the GNU system, the file position is | |
3158 | represented as an integer, which counts the number of bytes from the | |
3159 | beginning of the file. @xref{File Position}. | |
3160 | ||
3161 | During I/O to an ordinary disk file, you can change the file position | |
3162 | whenever you wish, so as to read or write any portion of the file. Some | |
3163 | other kinds of files may also permit this. Files which support changing | |
3164 | the file position are sometimes referred to as @dfn{random-access} | |
3165 | files. | |
3166 | ||
3167 | You can use the functions in this section to examine or modify the file | |
3168 | position indicator associated with a stream. The symbols listed below | |
3169 | are declared in the header file @file{stdio.h}. | |
3170 | @pindex stdio.h | |
3171 | ||
3172 | @comment stdio.h | |
f65fd747 | 3173 | @comment ISO |
28f540f4 RM |
3174 | @deftypefun {long int} ftell (FILE *@var{stream}) |
3175 | This function returns the current file position of the stream | |
3176 | @var{stream}. | |
3177 | ||
3178 | This function can fail if the stream doesn't support file positioning, | |
3179 | or if the file position can't be represented in a @code{long int}, and | |
3180 | possibly for other reasons as well. If a failure occurs, a value of | |
3181 | @code{-1} is returned. | |
3182 | @end deftypefun | |
3183 | ||
a5a0310d UD |
3184 | @comment stdio.h |
3185 | @comment Unix98 | |
3186 | @deftypefun off_t ftello (FILE *@var{stream}) | |
04b9968b UD |
3187 | The @code{ftello} function is similar to @code{ftell}, except that it |
3188 | returns a value of type @code{off_t}. Systems which support this type | |
3189 | use it to describe all file positions, unlike the POSIX specification | |
3190 | which uses a long int. The two are not necessarily the same size. | |
3191 | Therefore, using ftell can lead to problems if the implementation is | |
3192 | written on top of a POSIX compliant low-level I/O implementation, and using | |
3193 | @code{ftello} is preferable whenever it is available. | |
3194 | ||
3195 | If this function fails it returns @code{(off_t) -1}. This can happen due | |
a5a0310d UD |
3196 | to missing support for file positioning or internal errors. Otherwise |
3197 | the return value is the current file position. | |
3198 | ||
3199 | The function is an extension defined in the Unix Single Specification | |
3200 | version 2. | |
a3a4a74e UD |
3201 | |
3202 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 3203 | 32 bit system this function is in fact @code{ftello64}. I.e., the |
a3a4a74e UD |
3204 | LFS interface transparently replaces the old interface. |
3205 | @end deftypefun | |
3206 | ||
3207 | @comment stdio.h | |
3208 | @comment Unix98 | |
3209 | @deftypefun off64_t ftello64 (FILE *@var{stream}) | |
3210 | This function is similar to @code{ftello} with the only difference that | |
3211 | the return value is of type @code{off64_t}. This also requires that the | |
3212 | stream @var{stream} was opened using either @code{fopen64}, | |
3213 | @code{freopen64}, or @code{tmpfile64} since otherwise the underlying | |
3214 | file operations to position the file pointer beyond the @math{2^31} | |
3215 | bytes limit might fail. | |
3216 | ||
3217 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
3218 | bits machine this function is available under the name @code{ftello} | |
3219 | and so transparently replaces the old interface. | |
a5a0310d UD |
3220 | @end deftypefun |
3221 | ||
28f540f4 | 3222 | @comment stdio.h |
f65fd747 | 3223 | @comment ISO |
28f540f4 RM |
3224 | @deftypefun int fseek (FILE *@var{stream}, long int @var{offset}, int @var{whence}) |
3225 | The @code{fseek} function is used to change the file position of the | |
3226 | stream @var{stream}. The value of @var{whence} must be one of the | |
3227 | constants @code{SEEK_SET}, @code{SEEK_CUR}, or @code{SEEK_END}, to | |
3228 | indicate whether the @var{offset} is relative to the beginning of the | |
3229 | file, the current file position, or the end of the file, respectively. | |
3230 | ||
3231 | This function returns a value of zero if the operation was successful, | |
3232 | and a nonzero value to indicate failure. A successful call also clears | |
3233 | the end-of-file indicator of @var{stream} and discards any characters | |
3234 | that were ``pushed back'' by the use of @code{ungetc}. | |
3235 | ||
3236 | @code{fseek} either flushes any buffered output before setting the file | |
3237 | position or else remembers it so it will be written later in its proper | |
3238 | place in the file. | |
3239 | @end deftypefun | |
3240 | ||
a5a0310d UD |
3241 | @comment stdio.h |
3242 | @comment Unix98 | |
3243 | @deftypefun int fseeko (FILE *@var{stream}, off_t @var{offset}, int @var{whence}) | |
3244 | This function is similar to @code{fseek} but it corrects a problem with | |
3245 | @code{fseek} in a system with POSIX types. Using a value of type | |
3246 | @code{long int} for the offset is not compatible with POSIX. | |
3247 | @code{fseeko} uses the correct type @code{off_t} for the @var{offset} | |
3248 | parameter. | |
3249 | ||
f2ea0f5b | 3250 | For this reason it is a good idea to prefer @code{ftello} whenever it is |
a5a0310d UD |
3251 | available since its functionality is (if different at all) closer the |
3252 | underlying definition. | |
3253 | ||
3254 | The functionality and return value is the same as for @code{fseek}. | |
3255 | ||
3256 | The function is an extension defined in the Unix Single Specification | |
3257 | version 2. | |
a3a4a74e UD |
3258 | |
3259 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 3260 | 32 bit system this function is in fact @code{fseeko64}. I.e., the |
a3a4a74e UD |
3261 | LFS interface transparently replaces the old interface. |
3262 | @end deftypefun | |
3263 | ||
3264 | @comment stdio.h | |
3265 | @comment Unix98 | |
3266 | @deftypefun int fseeko64 (FILE *@var{stream}, off64_t @var{offset}, int @var{whence}) | |
3267 | This function is similar to @code{fseeko} with the only difference that | |
3268 | the @var{offset} parameter is of type @code{off64_t}. This also | |
3269 | requires that the stream @var{stream} was opened using either | |
3270 | @code{fopen64}, @code{freopen64}, or @code{tmpfile64} since otherwise | |
3271 | the underlying file operations to position the file pointer beyond the | |
3272 | @math{2^31} bytes limit might fail. | |
3273 | ||
3274 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
3275 | bits machine this function is available under the name @code{fseeko} | |
3276 | and so transparently replaces the old interface. | |
a5a0310d UD |
3277 | @end deftypefun |
3278 | ||
3279 | @strong{Portability Note:} In non-POSIX systems, @code{ftell}, | |
3280 | @code{ftello}, @code{fseek} and @code{fseeko} might work reliably only | |
3281 | on binary streams. @xref{Binary Streams}. | |
28f540f4 RM |
3282 | |
3283 | The following symbolic constants are defined for use as the @var{whence} | |
3284 | argument to @code{fseek}. They are also used with the @code{lseek} | |
3285 | function (@pxref{I/O Primitives}) and to specify offsets for file locks | |
3286 | (@pxref{Control Operations}). | |
3287 | ||
3288 | @comment stdio.h | |
f65fd747 | 3289 | @comment ISO |
28f540f4 RM |
3290 | @deftypevr Macro int SEEK_SET |
3291 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
3292 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
3293 | the offset provided is relative to the beginning of the file. | |
28f540f4 RM |
3294 | @end deftypevr |
3295 | ||
3296 | @comment stdio.h | |
f65fd747 | 3297 | @comment ISO |
28f540f4 RM |
3298 | @deftypevr Macro int SEEK_CUR |
3299 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
3300 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
3301 | the offset provided is relative to the current file position. | |
28f540f4 RM |
3302 | @end deftypevr |
3303 | ||
3304 | @comment stdio.h | |
f65fd747 | 3305 | @comment ISO |
28f540f4 RM |
3306 | @deftypevr Macro int SEEK_END |
3307 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
3308 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
3309 | the offset provided is relative to the end of the file. | |
28f540f4 RM |
3310 | @end deftypevr |
3311 | ||
3312 | @comment stdio.h | |
f65fd747 | 3313 | @comment ISO |
28f540f4 RM |
3314 | @deftypefun void rewind (FILE *@var{stream}) |
3315 | The @code{rewind} function positions the stream @var{stream} at the | |
f2ea0f5b | 3316 | beginning of the file. It is equivalent to calling @code{fseek} or |
a5a0310d UD |
3317 | @code{fseeko} on the @var{stream} with an @var{offset} argument of |
3318 | @code{0L} and a @var{whence} argument of @code{SEEK_SET}, except that | |
3319 | the return value is discarded and the error indicator for the stream is | |
3320 | reset. | |
28f540f4 RM |
3321 | @end deftypefun |
3322 | ||
3323 | These three aliases for the @samp{SEEK_@dots{}} constants exist for the | |
3324 | sake of compatibility with older BSD systems. They are defined in two | |
3325 | different header files: @file{fcntl.h} and @file{sys/file.h}. | |
3326 | ||
3327 | @table @code | |
3328 | @comment sys/file.h | |
3329 | @comment BSD | |
3330 | @item L_SET | |
3331 | @vindex L_SET | |
3332 | An alias for @code{SEEK_SET}. | |
3333 | ||
3334 | @comment sys/file.h | |
3335 | @comment BSD | |
3336 | @item L_INCR | |
3337 | @vindex L_INCR | |
3338 | An alias for @code{SEEK_CUR}. | |
3339 | ||
3340 | @comment sys/file.h | |
3341 | @comment BSD | |
3342 | @item L_XTND | |
3343 | @vindex L_XTND | |
3344 | An alias for @code{SEEK_END}. | |
3345 | @end table | |
3346 | ||
3347 | @node Portable Positioning | |
3348 | @section Portable File-Position Functions | |
3349 | ||
3350 | On the GNU system, the file position is truly a character count. You | |
a5a0310d UD |
3351 | can specify any character count value as an argument to @code{fseek} or |
3352 | @code{fseeko} and get reliable results for any random access file. | |
3353 | However, some @w{ISO C} systems do not represent file positions in this | |
3354 | way. | |
28f540f4 RM |
3355 | |
3356 | On some systems where text streams truly differ from binary streams, it | |
3357 | is impossible to represent the file position of a text stream as a count | |
3358 | of characters from the beginning of the file. For example, the file | |
3359 | position on some systems must encode both a record offset within the | |
3360 | file, and a character offset within the record. | |
3361 | ||
3362 | As a consequence, if you want your programs to be portable to these | |
3363 | systems, you must observe certain rules: | |
3364 | ||
3365 | @itemize @bullet | |
3366 | @item | |
3367 | The value returned from @code{ftell} on a text stream has no predictable | |
3368 | relationship to the number of characters you have read so far. The only | |
3369 | thing you can rely on is that you can use it subsequently as the | |
a5a0310d UD |
3370 | @var{offset} argument to @code{fseek} or @code{fseeko} to move back to |
3371 | the same file position. | |
28f540f4 | 3372 | |
19c3f208 | 3373 | @item |
a5a0310d | 3374 | In a call to @code{fseek} or @code{fseeko} on a text stream, either the |
04b9968b UD |
3375 | @var{offset} must be zero, or @var{whence} must be @code{SEEK_SET} and |
3376 | and the @var{offset} must be the result of an earlier call to @code{ftell} | |
3377 | on the same stream. | |
28f540f4 RM |
3378 | |
3379 | @item | |
3380 | The value of the file position indicator of a text stream is undefined | |
3381 | while there are characters that have been pushed back with @code{ungetc} | |
3382 | that haven't been read or discarded. @xref{Unreading}. | |
3383 | @end itemize | |
3384 | ||
3385 | But even if you observe these rules, you may still have trouble for long | |
3386 | files, because @code{ftell} and @code{fseek} use a @code{long int} value | |
3387 | to represent the file position. This type may not have room to encode | |
a5a0310d UD |
3388 | all the file positions in a large file. Using the @code{ftello} and |
3389 | @code{fseeko} functions might help here since the @code{off_t} type is | |
3390 | expected to be able to hold all file position values but this still does | |
3391 | not help to handle additional information which must be associated with | |
3392 | a file position. | |
28f540f4 RM |
3393 | |
3394 | So if you do want to support systems with peculiar encodings for the | |
3395 | file positions, it is better to use the functions @code{fgetpos} and | |
3396 | @code{fsetpos} instead. These functions represent the file position | |
3397 | using the data type @code{fpos_t}, whose internal representation varies | |
3398 | from system to system. | |
3399 | ||
3400 | These symbols are declared in the header file @file{stdio.h}. | |
3401 | @pindex stdio.h | |
3402 | ||
3403 | @comment stdio.h | |
f65fd747 | 3404 | @comment ISO |
28f540f4 RM |
3405 | @deftp {Data Type} fpos_t |
3406 | This is the type of an object that can encode information about the | |
3407 | file position of a stream, for use by the functions @code{fgetpos} and | |
3408 | @code{fsetpos}. | |
3409 | ||
3410 | In the GNU system, @code{fpos_t} is equivalent to @code{off_t} or | |
3411 | @code{long int}. In other systems, it might have a different internal | |
3412 | representation. | |
a3a4a74e | 3413 | |
0be8752b | 3414 | When compiling with @code{_FILE_OFFSET_BITS == 64} on a 32 bit machine |
a3a4a74e UD |
3415 | this type is in fact equivalent to @code{off64_t} since the LFS |
3416 | interface transparently replaced the old interface. | |
3417 | @end deftp | |
3418 | ||
3419 | @comment stdio.h | |
3420 | @comment Unix98 | |
3421 | @deftp {Data Type} fpos64_t | |
3422 | This is the type of an object that can encode information about the | |
3423 | file position of a stream, for use by the functions @code{fgetpos64} and | |
3424 | @code{fsetpos64}. | |
3425 | ||
3426 | In the GNU system, @code{fpos64_t} is equivalent to @code{off64_t} or | |
3427 | @code{long long int}. In other systems, it might have a different internal | |
3428 | representation. | |
28f540f4 RM |
3429 | @end deftp |
3430 | ||
3431 | @comment stdio.h | |
f65fd747 | 3432 | @comment ISO |
28f540f4 RM |
3433 | @deftypefun int fgetpos (FILE *@var{stream}, fpos_t *@var{position}) |
3434 | This function stores the value of the file position indicator for the | |
3435 | stream @var{stream} in the @code{fpos_t} object pointed to by | |
3436 | @var{position}. If successful, @code{fgetpos} returns zero; otherwise | |
3437 | it returns a nonzero value and stores an implementation-defined positive | |
3438 | value in @code{errno}. | |
a3a4a74e UD |
3439 | |
3440 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 3441 | 32 bit system the function is in fact @code{fgetpos64}. I.e., the LFS |
a3a4a74e UD |
3442 | interface transparently replaced the old interface. |
3443 | @end deftypefun | |
3444 | ||
3445 | @comment stdio.h | |
3446 | @comment Unix98 | |
3447 | @deftypefun int fgetpos64 (FILE *@var{stream}, fpos64_t *@var{position}) | |
3448 | This function is similar to @code{fgetpos} but the file position is | |
3449 | returned in a variable of type @code{fpos64_t} to which @var{position} | |
3450 | points. | |
3451 | ||
3452 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
3453 | bits machine this function is available under the name @code{fgetpos} | |
3454 | and so transparently replaces the old interface. | |
28f540f4 RM |
3455 | @end deftypefun |
3456 | ||
3457 | @comment stdio.h | |
f65fd747 | 3458 | @comment ISO |
a3a4a74e | 3459 | @deftypefun int fsetpos (FILE *@var{stream}, const fpos_t *@var{position}) |
28f540f4 RM |
3460 | This function sets the file position indicator for the stream @var{stream} |
3461 | to the position @var{position}, which must have been set by a previous | |
3462 | call to @code{fgetpos} on the same stream. If successful, @code{fsetpos} | |
3463 | clears the end-of-file indicator on the stream, discards any characters | |
3464 | that were ``pushed back'' by the use of @code{ungetc}, and returns a value | |
3465 | of zero. Otherwise, @code{fsetpos} returns a nonzero value and stores | |
3466 | an implementation-defined positive value in @code{errno}. | |
a3a4a74e UD |
3467 | |
3468 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 3469 | 32 bit system the function is in fact @code{fsetpos64}. I.e., the LFS |
a3a4a74e UD |
3470 | interface transparently replaced the old interface. |
3471 | @end deftypefun | |
3472 | ||
3473 | @comment stdio.h | |
3474 | @comment Unix98 | |
3475 | @deftypefun int fsetpos64 (FILE *@var{stream}, const fpos64_t *@var{position}) | |
3476 | This function is similar to @code{fsetpos} but the file position used | |
3477 | for positioning is provided in a variable of type @code{fpos64_t} to | |
3478 | which @var{position} points. | |
3479 | ||
3480 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
3481 | bits machine this function is available under the name @code{fsetpos} | |
3482 | and so transparently replaces the old interface. | |
28f540f4 RM |
3483 | @end deftypefun |
3484 | ||
3485 | @node Stream Buffering | |
3486 | @section Stream Buffering | |
3487 | ||
3488 | @cindex buffering of streams | |
3489 | Characters that are written to a stream are normally accumulated and | |
3490 | transmitted asynchronously to the file in a block, instead of appearing | |
3491 | as soon as they are output by the application program. Similarly, | |
3492 | streams often retrieve input from the host environment in blocks rather | |
3493 | than on a character-by-character basis. This is called @dfn{buffering}. | |
3494 | ||
3495 | If you are writing programs that do interactive input and output using | |
3496 | streams, you need to understand how buffering works when you design the | |
3497 | user interface to your program. Otherwise, you might find that output | |
3498 | (such as progress or prompt messages) doesn't appear when you intended | |
04b9968b | 3499 | it to, or displays some other unexpected behavior. |
28f540f4 RM |
3500 | |
3501 | This section deals only with controlling when characters are transmitted | |
3502 | between the stream and the file or device, and @emph{not} with how | |
3503 | things like echoing, flow control, and the like are handled on specific | |
3504 | classes of devices. For information on common control operations on | |
3505 | terminal devices, see @ref{Low-Level Terminal Interface}. | |
3506 | ||
3507 | You can bypass the stream buffering facilities altogether by using the | |
3508 | low-level input and output functions that operate on file descriptors | |
3509 | instead. @xref{Low-Level I/O}. | |
3510 | ||
3511 | @menu | |
3512 | * Buffering Concepts:: Terminology is defined here. | |
3513 | * Flushing Buffers:: How to ensure that output buffers are flushed. | |
3514 | * Controlling Buffering:: How to specify what kind of buffering to use. | |
3515 | @end menu | |
3516 | ||
3517 | @node Buffering Concepts | |
3518 | @subsection Buffering Concepts | |
3519 | ||
3520 | There are three different kinds of buffering strategies: | |
3521 | ||
3522 | @itemize @bullet | |
3523 | @item | |
3524 | Characters written to or read from an @dfn{unbuffered} stream are | |
3525 | transmitted individually to or from the file as soon as possible. | |
3526 | @cindex unbuffered stream | |
3527 | ||
3528 | @item | |
3529 | Characters written to a @dfn{line buffered} stream are transmitted to | |
3530 | the file in blocks when a newline character is encountered. | |
3531 | @cindex line buffered stream | |
3532 | ||
3533 | @item | |
3534 | Characters written to or read from a @dfn{fully buffered} stream are | |
3535 | transmitted to or from the file in blocks of arbitrary size. | |
3536 | @cindex fully buffered stream | |
3537 | @end itemize | |
3538 | ||
3539 | Newly opened streams are normally fully buffered, with one exception: a | |
3540 | stream connected to an interactive device such as a terminal is | |
3541 | initially line buffered. @xref{Controlling Buffering}, for information | |
3542 | on how to select a different kind of buffering. Usually the automatic | |
3543 | selection gives you the most convenient kind of buffering for the file | |
3544 | or device you open. | |
3545 | ||
3546 | The use of line buffering for interactive devices implies that output | |
3547 | messages ending in a newline will appear immediately---which is usually | |
3548 | what you want. Output that doesn't end in a newline might or might not | |
3549 | show up immediately, so if you want them to appear immediately, you | |
3550 | should flush buffered output explicitly with @code{fflush}, as described | |
3551 | in @ref{Flushing Buffers}. | |
3552 | ||
3553 | @node Flushing Buffers | |
3554 | @subsection Flushing Buffers | |
3555 | ||
3556 | @cindex flushing a stream | |
3557 | @dfn{Flushing} output on a buffered stream means transmitting all | |
3558 | accumulated characters to the file. There are many circumstances when | |
3559 | buffered output on a stream is flushed automatically: | |
3560 | ||
3561 | @itemize @bullet | |
3562 | @item | |
3563 | When you try to do output and the output buffer is full. | |
3564 | ||
3565 | @item | |
3566 | When the stream is closed. @xref{Closing Streams}. | |
3567 | ||
19c3f208 | 3568 | @item |
28f540f4 RM |
3569 | When the program terminates by calling @code{exit}. |
3570 | @xref{Normal Termination}. | |
3571 | ||
3572 | @item | |
3573 | When a newline is written, if the stream is line buffered. | |
3574 | ||
3575 | @item | |
3576 | Whenever an input operation on @emph{any} stream actually reads data | |
3577 | from its file. | |
3578 | @end itemize | |
3579 | ||
3580 | If you want to flush the buffered output at another time, call | |
3581 | @code{fflush}, which is declared in the header file @file{stdio.h}. | |
3582 | @pindex stdio.h | |
3583 | ||
3584 | @comment stdio.h | |
f65fd747 | 3585 | @comment ISO |
28f540f4 RM |
3586 | @deftypefun int fflush (FILE *@var{stream}) |
3587 | This function causes any buffered output on @var{stream} to be delivered | |
3588 | to the file. If @var{stream} is a null pointer, then | |
3589 | @code{fflush} causes buffered output on @emph{all} open output streams | |
3590 | to be flushed. | |
3591 | ||
3592 | This function returns @code{EOF} if a write error occurs, or zero | |
3593 | otherwise. | |
3594 | @end deftypefun | |
3595 | ||
3596 | @strong{Compatibility Note:} Some brain-damaged operating systems have | |
3597 | been known to be so thoroughly fixated on line-oriented input and output | |
3598 | that flushing a line buffered stream causes a newline to be written! | |
3599 | Fortunately, this ``feature'' seems to be becoming less common. You do | |
3600 | not need to worry about this in the GNU system. | |
3601 | ||
3602 | ||
3603 | @node Controlling Buffering | |
3604 | @subsection Controlling Which Kind of Buffering | |
3605 | ||
3606 | After opening a stream (but before any other operations have been | |
3607 | performed on it), you can explicitly specify what kind of buffering you | |
3608 | want it to have using the @code{setvbuf} function. | |
3609 | @cindex buffering, controlling | |
3610 | ||
3611 | The facilities listed in this section are declared in the header | |
3612 | file @file{stdio.h}. | |
3613 | @pindex stdio.h | |
3614 | ||
3615 | @comment stdio.h | |
f65fd747 | 3616 | @comment ISO |
28f540f4 RM |
3617 | @deftypefun int setvbuf (FILE *@var{stream}, char *@var{buf}, int @var{mode}, size_t @var{size}) |
3618 | This function is used to specify that the stream @var{stream} should | |
3619 | have the buffering mode @var{mode}, which can be either @code{_IOFBF} | |
3620 | (for full buffering), @code{_IOLBF} (for line buffering), or | |
3621 | @code{_IONBF} (for unbuffered input/output). | |
3622 | ||
3623 | If you specify a null pointer as the @var{buf} argument, then @code{setvbuf} | |
3624 | allocates a buffer itself using @code{malloc}. This buffer will be freed | |
3625 | when you close the stream. | |
3626 | ||
3627 | Otherwise, @var{buf} should be a character array that can hold at least | |
3628 | @var{size} characters. You should not free the space for this array as | |
3629 | long as the stream remains open and this array remains its buffer. You | |
3630 | should usually either allocate it statically, or @code{malloc} | |
3631 | (@pxref{Unconstrained Allocation}) the buffer. Using an automatic array | |
3632 | is not a good idea unless you close the file before exiting the block | |
3633 | that declares the array. | |
3634 | ||
3635 | While the array remains a stream buffer, the stream I/O functions will | |
3636 | use the buffer for their internal purposes. You shouldn't try to access | |
3637 | the values in the array directly while the stream is using it for | |
3638 | buffering. | |
3639 | ||
3640 | The @code{setvbuf} function returns zero on success, or a nonzero value | |
3641 | if the value of @var{mode} is not valid or if the request could not | |
3642 | be honored. | |
3643 | @end deftypefun | |
3644 | ||
3645 | @comment stdio.h | |
f65fd747 | 3646 | @comment ISO |
28f540f4 RM |
3647 | @deftypevr Macro int _IOFBF |
3648 | The value of this macro is an integer constant expression that can be | |
3649 | used as the @var{mode} argument to the @code{setvbuf} function to | |
3650 | specify that the stream should be fully buffered. | |
3651 | @end deftypevr | |
3652 | ||
3653 | @comment stdio.h | |
f65fd747 | 3654 | @comment ISO |
28f540f4 RM |
3655 | @deftypevr Macro int _IOLBF |
3656 | The value of this macro is an integer constant expression that can be | |
3657 | used as the @var{mode} argument to the @code{setvbuf} function to | |
3658 | specify that the stream should be line buffered. | |
3659 | @end deftypevr | |
3660 | ||
3661 | @comment stdio.h | |
f65fd747 | 3662 | @comment ISO |
28f540f4 RM |
3663 | @deftypevr Macro int _IONBF |
3664 | The value of this macro is an integer constant expression that can be | |
3665 | used as the @var{mode} argument to the @code{setvbuf} function to | |
3666 | specify that the stream should be unbuffered. | |
3667 | @end deftypevr | |
3668 | ||
3669 | @comment stdio.h | |
f65fd747 | 3670 | @comment ISO |
28f540f4 RM |
3671 | @deftypevr Macro int BUFSIZ |
3672 | The value of this macro is an integer constant expression that is good | |
3673 | to use for the @var{size} argument to @code{setvbuf}. This value is | |
3674 | guaranteed to be at least @code{256}. | |
3675 | ||
3676 | The value of @code{BUFSIZ} is chosen on each system so as to make stream | |
19c3f208 | 3677 | I/O efficient. So it is a good idea to use @code{BUFSIZ} as the size |
28f540f4 RM |
3678 | for the buffer when you call @code{setvbuf}. |
3679 | ||
3680 | Actually, you can get an even better value to use for the buffer size | |
3681 | by means of the @code{fstat} system call: it is found in the | |
3682 | @code{st_blksize} field of the file attributes. @xref{Attribute Meanings}. | |
3683 | ||
3684 | Sometimes people also use @code{BUFSIZ} as the allocation size of | |
3685 | buffers used for related purposes, such as strings used to receive a | |
3686 | line of input with @code{fgets} (@pxref{Character Input}). There is no | |
3687 | particular reason to use @code{BUFSIZ} for this instead of any other | |
3688 | integer, except that it might lead to doing I/O in chunks of an | |
3689 | efficient size. | |
3690 | @end deftypevr | |
3691 | ||
3692 | @comment stdio.h | |
f65fd747 | 3693 | @comment ISO |
28f540f4 RM |
3694 | @deftypefun void setbuf (FILE *@var{stream}, char *@var{buf}) |
3695 | If @var{buf} is a null pointer, the effect of this function is | |
3696 | equivalent to calling @code{setvbuf} with a @var{mode} argument of | |
3697 | @code{_IONBF}. Otherwise, it is equivalent to calling @code{setvbuf} | |
3698 | with @var{buf}, and a @var{mode} of @code{_IOFBF} and a @var{size} | |
3699 | argument of @code{BUFSIZ}. | |
3700 | ||
3701 | The @code{setbuf} function is provided for compatibility with old code; | |
3702 | use @code{setvbuf} in all new programs. | |
3703 | @end deftypefun | |
3704 | ||
3705 | @comment stdio.h | |
3706 | @comment BSD | |
3707 | @deftypefun void setbuffer (FILE *@var{stream}, char *@var{buf}, size_t @var{size}) | |
3708 | If @var{buf} is a null pointer, this function makes @var{stream} unbuffered. | |
3709 | Otherwise, it makes @var{stream} fully buffered using @var{buf} as the | |
3710 | buffer. The @var{size} argument specifies the length of @var{buf}. | |
3711 | ||
3712 | This function is provided for compatibility with old BSD code. Use | |
3713 | @code{setvbuf} instead. | |
3714 | @end deftypefun | |
3715 | ||
3716 | @comment stdio.h | |
3717 | @comment BSD | |
3718 | @deftypefun void setlinebuf (FILE *@var{stream}) | |
3719 | This function makes @var{stream} be line buffered, and allocates the | |
3720 | buffer for you. | |
3721 | ||
3722 | This function is provided for compatibility with old BSD code. Use | |
3723 | @code{setvbuf} instead. | |
3724 | @end deftypefun | |
3725 | ||
3726 | @node Other Kinds of Streams | |
3727 | @section Other Kinds of Streams | |
3728 | ||
3729 | The GNU library provides ways for you to define additional kinds of | |
3730 | streams that do not necessarily correspond to an open file. | |
3731 | ||
3732 | One such type of stream takes input from or writes output to a string. | |
3733 | These kinds of streams are used internally to implement the | |
3734 | @code{sprintf} and @code{sscanf} functions. You can also create such a | |
3735 | stream explicitly, using the functions described in @ref{String Streams}. | |
3736 | ||
3737 | More generally, you can define streams that do input/output to arbitrary | |
3738 | objects using functions supplied by your program. This protocol is | |
3739 | discussed in @ref{Custom Streams}. | |
3740 | ||
3741 | @strong{Portability Note:} The facilities described in this section are | |
3742 | specific to GNU. Other systems or C implementations might or might not | |
3743 | provide equivalent functionality. | |
3744 | ||
3745 | @menu | |
19c3f208 | 3746 | * String Streams:: Streams that get data from or put data in |
28f540f4 RM |
3747 | a string or memory buffer. |
3748 | * Obstack Streams:: Streams that store data in an obstack. | |
3749 | * Custom Streams:: Defining your own streams with an arbitrary | |
3750 | input data source and/or output data sink. | |
3751 | @end menu | |
3752 | ||
3753 | @node String Streams | |
3754 | @subsection String Streams | |
3755 | ||
3756 | @cindex stream, for I/O to a string | |
3757 | @cindex string stream | |
3758 | The @code{fmemopen} and @code{open_memstream} functions allow you to do | |
3759 | I/O to a string or memory buffer. These facilities are declared in | |
3760 | @file{stdio.h}. | |
3761 | @pindex stdio.h | |
3762 | ||
3763 | @comment stdio.h | |
3764 | @comment GNU | |
3765 | @deftypefun {FILE *} fmemopen (void *@var{buf}, size_t @var{size}, const char *@var{opentype}) | |
3766 | This function opens a stream that allows the access specified by the | |
3767 | @var{opentype} argument, that reads from or writes to the buffer specified | |
3768 | by the argument @var{buf}. This array must be at least @var{size} bytes long. | |
3769 | ||
3770 | If you specify a null pointer as the @var{buf} argument, @code{fmemopen} | |
04b9968b UD |
3771 | dynamically allocates an array @var{size} bytes long (as with @code{malloc}; |
3772 | @pxref{Unconstrained Allocation}). This is really only useful | |
28f540f4 RM |
3773 | if you are going to write things to the buffer and then read them back |
3774 | in again, because you have no way of actually getting a pointer to the | |
3775 | buffer (for this, try @code{open_memstream}, below). The buffer is | |
3776 | freed when the stream is open. | |
3777 | ||
3778 | The argument @var{opentype} is the same as in @code{fopen} | |
8b7fb588 | 3779 | (@pxref{Opening Streams}). If the @var{opentype} specifies |
28f540f4 RM |
3780 | append mode, then the initial file position is set to the first null |
3781 | character in the buffer. Otherwise the initial file position is at the | |
3782 | beginning of the buffer. | |
3783 | ||
3784 | When a stream open for writing is flushed or closed, a null character | |
3785 | (zero byte) is written at the end of the buffer if it fits. You | |
3786 | should add an extra byte to the @var{size} argument to account for this. | |
3787 | Attempts to write more than @var{size} bytes to the buffer result | |
3788 | in an error. | |
3789 | ||
3790 | For a stream open for reading, null characters (zero bytes) in the | |
3791 | buffer do not count as ``end of file''. Read operations indicate end of | |
3792 | file only when the file position advances past @var{size} bytes. So, if | |
3793 | you want to read characters from a null-terminated string, you should | |
3794 | supply the length of the string as the @var{size} argument. | |
3795 | @end deftypefun | |
3796 | ||
3797 | Here is an example of using @code{fmemopen} to create a stream for | |
3798 | reading from a string: | |
3799 | ||
3800 | @smallexample | |
3801 | @include memopen.c.texi | |
3802 | @end smallexample | |
3803 | ||
3804 | This program produces the following output: | |
3805 | ||
3806 | @smallexample | |
3807 | Got f | |
3808 | Got o | |
3809 | Got o | |
3810 | Got b | |
3811 | Got a | |
3812 | Got r | |
3813 | @end smallexample | |
3814 | ||
3815 | @comment stdio.h | |
3816 | @comment GNU | |
3817 | @deftypefun {FILE *} open_memstream (char **@var{ptr}, size_t *@var{sizeloc}) | |
3818 | This function opens a stream for writing to a buffer. The buffer is | |
3819 | allocated dynamically (as with @code{malloc}; @pxref{Unconstrained | |
3820 | Allocation}) and grown as necessary. | |
3821 | ||
3822 | When the stream is closed with @code{fclose} or flushed with | |
3823 | @code{fflush}, the locations @var{ptr} and @var{sizeloc} are updated to | |
3824 | contain the pointer to the buffer and its size. The values thus stored | |
3825 | remain valid only as long as no further output on the stream takes | |
3826 | place. If you do more output, you must flush the stream again to store | |
3827 | new values before you use them again. | |
3828 | ||
3829 | A null character is written at the end of the buffer. This null character | |
3830 | is @emph{not} included in the size value stored at @var{sizeloc}. | |
3831 | ||
a5a0310d UD |
3832 | You can move the stream's file position with @code{fseek} or |
3833 | @code{fseeko} (@pxref{File Positioning}). Moving the file position past | |
3834 | the end of the data already written fills the intervening space with | |
3835 | zeroes. | |
28f540f4 RM |
3836 | @end deftypefun |
3837 | ||
3838 | Here is an example of using @code{open_memstream}: | |
3839 | ||
3840 | @smallexample | |
3841 | @include memstrm.c.texi | |
3842 | @end smallexample | |
3843 | ||
3844 | This program produces the following output: | |
3845 | ||
3846 | @smallexample | |
3847 | buf = `hello', size = 5 | |
3848 | buf = `hello, world', size = 12 | |
3849 | @end smallexample | |
3850 | ||
3851 | @c @group Invalid outside @example. | |
3852 | @node Obstack Streams | |
3853 | @subsection Obstack Streams | |
3854 | ||
3855 | You can open an output stream that puts it data in an obstack. | |
3856 | @xref{Obstacks}. | |
3857 | ||
3858 | @comment stdio.h | |
3859 | @comment GNU | |
3860 | @deftypefun {FILE *} open_obstack_stream (struct obstack *@var{obstack}) | |
3861 | This function opens a stream for writing data into the obstack @var{obstack}. | |
3862 | This starts an object in the obstack and makes it grow as data is | |
3863 | written (@pxref{Growing Objects}). | |
3864 | @c @end group Doubly invalid because not nested right. | |
3865 | ||
3866 | Calling @code{fflush} on this stream updates the current size of the | |
3867 | object to match the amount of data that has been written. After a call | |
3868 | to @code{fflush}, you can examine the object temporarily. | |
3869 | ||
a5a0310d UD |
3870 | You can move the file position of an obstack stream with @code{fseek} or |
3871 | @code{fseeko} (@pxref{File Positioning}). Moving the file position past | |
3872 | the end of the data written fills the intervening space with zeros. | |
28f540f4 RM |
3873 | |
3874 | To make the object permanent, update the obstack with @code{fflush}, and | |
3875 | then use @code{obstack_finish} to finalize the object and get its address. | |
3876 | The following write to the stream starts a new object in the obstack, | |
3877 | and later writes add to that object until you do another @code{fflush} | |
3878 | and @code{obstack_finish}. | |
3879 | ||
3880 | But how do you find out how long the object is? You can get the length | |
3881 | in bytes by calling @code{obstack_object_size} (@pxref{Status of an | |
3882 | Obstack}), or you can null-terminate the object like this: | |
3883 | ||
3884 | @smallexample | |
3885 | obstack_1grow (@var{obstack}, 0); | |
3886 | @end smallexample | |
3887 | ||
3888 | Whichever one you do, you must do it @emph{before} calling | |
3889 | @code{obstack_finish}. (You can do both if you wish.) | |
3890 | @end deftypefun | |
3891 | ||
3892 | Here is a sample function that uses @code{open_obstack_stream}: | |
3893 | ||
3894 | @smallexample | |
3895 | char * | |
3896 | make_message_string (const char *a, int b) | |
3897 | @{ | |
3898 | FILE *stream = open_obstack_stream (&message_obstack); | |
3899 | output_task (stream); | |
3900 | fprintf (stream, ": "); | |
3901 | fprintf (stream, a, b); | |
3902 | fprintf (stream, "\n"); | |
3903 | fclose (stream); | |
3904 | obstack_1grow (&message_obstack, 0); | |
3905 | return obstack_finish (&message_obstack); | |
3906 | @} | |
3907 | @end smallexample | |
3908 | ||
3909 | @node Custom Streams | |
3910 | @subsection Programming Your Own Custom Streams | |
3911 | @cindex custom streams | |
3912 | @cindex programming your own streams | |
3913 | ||
3914 | This section describes how you can make a stream that gets input from an | |
3915 | arbitrary data source or writes output to an arbitrary data sink | |
52cea457 UD |
3916 | programmed by you. We call these @dfn{custom streams}. The functions |
3917 | and types described here are all GNU extensions. | |
28f540f4 RM |
3918 | |
3919 | @c !!! this does not talk at all about the higher-level hooks | |
3920 | ||
3921 | @menu | |
3922 | * Streams and Cookies:: The @dfn{cookie} records where to fetch or | |
19c3f208 | 3923 | store data that is read or written. |
28f540f4 | 3924 | * Hook Functions:: How you should define the four @dfn{hook |
19c3f208 | 3925 | functions} that a custom stream needs. |
28f540f4 RM |
3926 | @end menu |
3927 | ||
3928 | @node Streams and Cookies | |
3929 | @subsubsection Custom Streams and Cookies | |
3930 | @cindex cookie, for custom stream | |
3931 | ||
3932 | Inside every custom stream is a special object called the @dfn{cookie}. | |
3933 | This is an object supplied by you which records where to fetch or store | |
3934 | the data read or written. It is up to you to define a data type to use | |
3935 | for the cookie. The stream functions in the library never refer | |
3936 | directly to its contents, and they don't even know what the type is; | |
3937 | they record its address with type @code{void *}. | |
3938 | ||
3939 | To implement a custom stream, you must specify @emph{how} to fetch or | |
3940 | store the data in the specified place. You do this by defining | |
3941 | @dfn{hook functions} to read, write, change ``file position'', and close | |
3942 | the stream. All four of these functions will be passed the stream's | |
3943 | cookie so they can tell where to fetch or store the data. The library | |
3944 | functions don't know what's inside the cookie, but your functions will | |
3945 | know. | |
3946 | ||
3947 | When you create a custom stream, you must specify the cookie pointer, | |
19c3f208 | 3948 | and also the four hook functions stored in a structure of type |
28f540f4 RM |
3949 | @code{cookie_io_functions_t}. |
3950 | ||
3951 | These facilities are declared in @file{stdio.h}. | |
3952 | @pindex stdio.h | |
3953 | ||
3954 | @comment stdio.h | |
3955 | @comment GNU | |
3956 | @deftp {Data Type} {cookie_io_functions_t} | |
19c3f208 | 3957 | This is a structure type that holds the functions that define the |
28f540f4 RM |
3958 | communications protocol between the stream and its cookie. It has |
3959 | the following members: | |
3960 | ||
3961 | @table @code | |
3962 | @item cookie_read_function_t *read | |
3963 | This is the function that reads data from the cookie. If the value is a | |
f2ea0f5b | 3964 | null pointer instead of a function, then read operations on this stream |
28f540f4 RM |
3965 | always return @code{EOF}. |
3966 | ||
3967 | @item cookie_write_function_t *write | |
3968 | This is the function that writes data to the cookie. If the value is a | |
3969 | null pointer instead of a function, then data written to the stream is | |
3970 | discarded. | |
3971 | ||
3972 | @item cookie_seek_function_t *seek | |
3973 | This is the function that performs the equivalent of file positioning on | |
3974 | the cookie. If the value is a null pointer instead of a function, calls | |
a5a0310d UD |
3975 | to @code{fseek} or @code{fseeko} on this stream can only seek to |
3976 | locations within the buffer; any attempt to seek outside the buffer will | |
3977 | return an @code{ESPIPE} error. | |
28f540f4 RM |
3978 | |
3979 | @item cookie_close_function_t *close | |
3980 | This function performs any appropriate cleanup on the cookie when | |
3981 | closing the stream. If the value is a null pointer instead of a | |
3982 | function, nothing special is done to close the cookie when the stream is | |
3983 | closed. | |
3984 | @end table | |
3985 | @end deftp | |
3986 | ||
3987 | @comment stdio.h | |
3988 | @comment GNU | |
3989 | @deftypefun {FILE *} fopencookie (void *@var{cookie}, const char *@var{opentype}, cookie_io_functions_t @var{io-functions}) | |
3990 | This function actually creates the stream for communicating with the | |
3991 | @var{cookie} using the functions in the @var{io-functions} argument. | |
3992 | The @var{opentype} argument is interpreted as for @code{fopen}; | |
3993 | see @ref{Opening Streams}. (But note that the ``truncate on | |
3994 | open'' option is ignored.) The new stream is fully buffered. | |
3995 | ||
3996 | The @code{fopencookie} function returns the newly created stream, or a null | |
3997 | pointer in case of an error. | |
3998 | @end deftypefun | |
3999 | ||
4000 | @node Hook Functions | |
4001 | @subsubsection Custom Stream Hook Functions | |
4002 | @cindex hook functions (of custom streams) | |
4003 | ||
4004 | Here are more details on how you should define the four hook functions | |
4005 | that a custom stream needs. | |
4006 | ||
4007 | You should define the function to read data from the cookie as: | |
4008 | ||
4009 | @smallexample | |
4010 | ssize_t @var{reader} (void *@var{cookie}, void *@var{buffer}, size_t @var{size}) | |
4011 | @end smallexample | |
4012 | ||
4013 | This is very similar to the @code{read} function; see @ref{I/O | |
4014 | Primitives}. Your function should transfer up to @var{size} bytes into | |
4015 | the @var{buffer}, and return the number of bytes read, or zero to | |
4016 | indicate end-of-file. You can return a value of @code{-1} to indicate | |
4017 | an error. | |
4018 | ||
4019 | You should define the function to write data to the cookie as: | |
4020 | ||
4021 | @smallexample | |
4022 | ssize_t @var{writer} (void *@var{cookie}, const void *@var{buffer}, size_t @var{size}) | |
4023 | @end smallexample | |
4024 | ||
4025 | This is very similar to the @code{write} function; see @ref{I/O | |
4026 | Primitives}. Your function should transfer up to @var{size} bytes from | |
4027 | the buffer, and return the number of bytes written. You can return a | |
4028 | value of @code{-1} to indicate an error. | |
4029 | ||
4030 | You should define the function to perform seek operations on the cookie | |
4031 | as: | |
4032 | ||
4033 | @smallexample | |
4034 | int @var{seeker} (void *@var{cookie}, fpos_t *@var{position}, int @var{whence}) | |
4035 | @end smallexample | |
4036 | ||
4037 | For this function, the @var{position} and @var{whence} arguments are | |
4038 | interpreted as for @code{fgetpos}; see @ref{Portable Positioning}. In | |
4039 | the GNU library, @code{fpos_t} is equivalent to @code{off_t} or | |
4040 | @code{long int}, and simply represents the number of bytes from the | |
4041 | beginning of the file. | |
4042 | ||
19c3f208 | 4043 | After doing the seek operation, your function should store the resulting |
28f540f4 RM |
4044 | file position relative to the beginning of the file in @var{position}. |
4045 | Your function should return a value of @code{0} on success and @code{-1} | |
4046 | to indicate an error. | |
4047 | ||
4048 | You should define the function to do cleanup operations on the cookie | |
4049 | appropriate for closing the stream as: | |
4050 | ||
4051 | @smallexample | |
4052 | int @var{cleaner} (void *@var{cookie}) | |
4053 | @end smallexample | |
4054 | ||
4055 | Your function should return @code{-1} to indicate an error, and @code{0} | |
4056 | otherwise. | |
4057 | ||
4058 | @comment stdio.h | |
4059 | @comment GNU | |
4060 | @deftp {Data Type} cookie_read_function | |
4061 | This is the data type that the read function for a custom stream should have. | |
4062 | If you declare the function as shown above, this is the type it will have. | |
4063 | @end deftp | |
4064 | ||
4065 | @comment stdio.h | |
4066 | @comment GNU | |
4067 | @deftp {Data Type} cookie_write_function | |
4068 | The data type of the write function for a custom stream. | |
4069 | @end deftp | |
4070 | ||
4071 | @comment stdio.h | |
4072 | @comment GNU | |
4073 | @deftp {Data Type} cookie_seek_function | |
4074 | The data type of the seek function for a custom stream. | |
4075 | @end deftp | |
4076 | ||
4077 | @comment stdio.h | |
4078 | @comment GNU | |
4079 | @deftp {Data Type} cookie_close_function | |
4080 | The data type of the close function for a custom stream. | |
4081 | @end deftp | |
4082 | ||
4083 | @ignore | |
4084 | Roland says: | |
4085 | ||
4086 | @quotation | |
4087 | There is another set of functions one can give a stream, the | |
4088 | input-room and output-room functions. These functions must | |
4089 | understand stdio internals. To describe how to use these | |
4090 | functions, you also need to document lots of how stdio works | |
4091 | internally (which isn't relevant for other uses of stdio). | |
4092 | Perhaps I can write an interface spec from which you can write | |
4093 | good documentation. But it's pretty complex and deals with lots | |
4094 | of nitty-gritty details. I think it might be better to let this | |
4095 | wait until the rest of the manual is more done and polished. | |
4096 | @end quotation | |
4097 | @end ignore | |
4098 | ||
4099 | @c ??? This section could use an example. | |
0501d603 UD |
4100 | |
4101 | ||
4102 | @node Formatted Messages | |
4103 | @section Formatted Messages | |
4104 | @cindex formatted messages | |
4105 | ||
4106 | On systems which are based on System V messages of programs (especially | |
4107 | the system tools) are printed in a strict form using the @code{fmtmsg} | |
4108 | function. The uniformity sometimes helps the user to interpret messages | |
cf29ffbe | 4109 | and the strictness tests of the @code{fmtmsg} function ensure that the |
0501d603 UD |
4110 | programmer follows some minimal requirements. |
4111 | ||
4112 | @menu | |
4113 | * Printing Formatted Messages:: The @code{fmtmsg} function. | |
4114 | * Adding Severity Classes:: Add more severity classes. | |
4115 | * Example:: How to use @code{fmtmsg} and @code{addseverity}. | |
4116 | @end menu | |
4117 | ||
4118 | ||
4119 | @node Printing Formatted Messages | |
4120 | @subsection Printing Formatted Messages | |
4121 | ||
4122 | Messages can be printed to standard error and/or to the console. To | |
cf29ffbe | 4123 | select the destination the programmer can use the following two values, |
0501d603 UD |
4124 | bitwise OR combined if wanted, for the @var{classification} parameter of |
4125 | @code{fmtmsg}: | |
4126 | ||
4127 | @vtable @code | |
4128 | @item MM_PRINT | |
4129 | Display the message in standard error. | |
4130 | @item MM_CONSOLE | |
4131 | Display the message on the system console. | |
4132 | @end vtable | |
4133 | ||
f2ea0f5b | 4134 | The erroneous piece of the system can be signalled by exactly one of the |
0501d603 UD |
4135 | following values which also is bitwise ORed with the |
4136 | @var{classification} parameter to @code{fmtmsg}: | |
4137 | ||
4138 | @vtable @code | |
4139 | @item MM_HARD | |
4140 | The source of the condition is some hardware. | |
4141 | @item MM_SOFT | |
4142 | The source of the condition is some software. | |
4143 | @item MM_FIRM | |
4144 | The source of the condition is some firmware. | |
4145 | @end vtable | |
4146 | ||
4147 | A third component of the @var{classification} parameter to @code{fmtmsg} | |
4148 | can describe the part of the system which detects the problem. This is | |
4149 | done by using exactly one of the following values: | |
4150 | ||
4151 | @vtable @code | |
4152 | @item MM_APPL | |
f2ea0f5b | 4153 | The erroneous condition is detected by the application. |
0501d603 | 4154 | @item MM_UTIL |
f2ea0f5b | 4155 | The erroneous condition is detected by a utility. |
0501d603 | 4156 | @item MM_OPSYS |
f2ea0f5b | 4157 | The erroneous condition is detected by the operating system. |
0501d603 UD |
4158 | @end vtable |
4159 | ||
4160 | A last component of @var{classification} can signal the results of this | |
4161 | message. Exactly one of the following values can be used: | |
4162 | ||
4163 | @vtable @code | |
4164 | @item MM_RECOVER | |
4165 | It is a recoverable error. | |
4166 | @item MM_NRECOV | |
4167 | It is a non-recoverable error. | |
4168 | @end vtable | |
4169 | ||
4170 | @comment fmtmsg.h | |
4171 | @comment XPG | |
4172 | @deftypefun int fmtmsg (long int @var{classification}, const char *@var{label}, int @var{severity}, const char *@var{text}, const char *@var{action}, const char *@var{tag}) | |
4173 | Display a message described by its parameters on the device(s) specified | |
4174 | in the @var{classification} parameter. The @var{label} parameter | |
4175 | identifies the source of the message. The string should consist of two | |
4176 | colon separated parts where the first part has not more than 10 and the | |
04b9968b | 4177 | second part not more than 14 characters. The @var{text} parameter |
f2ea0f5b | 4178 | describes the condition of the error, the @var{action} parameter possible |
0501d603 UD |
4179 | steps to recover from the error and the @var{tag} parameter is a |
4180 | reference to the online documentation where more information can be | |
4181 | found. It should contain the @var{label} value and a unique | |
4182 | identification number. | |
4183 | ||
cf29ffbe | 4184 | Each of the parameters can be a special value which means this value |
0501d603 UD |
4185 | is to be omitted. The symbolic names for these values are: |
4186 | ||
4187 | @vtable @code | |
4188 | @item MM_NULLLBL | |
4189 | Ignore @var{label} parameter. | |
4190 | @item MM_NULLSEV | |
4191 | Ignore @var{severity} parameter. | |
4192 | @item MM_NULLMC | |
4193 | Ignore @var{classification} parameter. This implies that nothing is | |
4194 | actually printed. | |
4195 | @item MM_NULLTXT | |
4196 | Ignore @var{text} parameter. | |
4197 | @item MM_NULLACT | |
4198 | Ignore @var{action} parameter. | |
4199 | @item MM_NULLTAG | |
4200 | Ignore @var{tag} parameter. | |
4201 | @end vtable | |
4202 | ||
cf29ffbe | 4203 | There is another way certain fields can be omitted from the output to |
0501d603 UD |
4204 | standard error. This is described below in the description of |
4205 | environment variables influencing the behaviour. | |
4206 | ||
4207 | The @var{severity} parameter can have one of the values in the following | |
4208 | table: | |
4209 | @cindex severity class | |
4210 | ||
4211 | @vtable @code | |
4212 | @item MM_NOSEV | |
4213 | Nothing is printed, this value is the same as @code{MM_NULLSEV}. | |
4214 | @item MM_HALT | |
4215 | This value is printed as @code{HALT}. | |
4216 | @item MM_ERROR | |
4217 | This value is printed as @code{ERROR}. | |
4218 | @item MM_WARNING | |
4219 | This value is printed as @code{WARNING}. | |
4220 | @item MM_INFO | |
4221 | This value is printed as @code{INFO}. | |
4222 | @end vtable | |
4223 | ||
4224 | The numeric value of these five macros are between @code{0} and | |
4225 | @code{4}. Using the environment variable @code{SEV_LEVEL} or using the | |
4226 | @code{addseverity} function one can add more severity levels with their | |
4227 | corresponding string to print. This is described below | |
4228 | (@pxref{Adding Severity Classes}). | |
4229 | ||
4230 | @noindent | |
4231 | If no parameter is ignored the output looks like this: | |
4232 | ||
4233 | @smallexample | |
4234 | @var{label}: @var{severity-string}: @var{text} | |
4235 | TO FIX: @var{action} @var{tag} | |
4236 | @end smallexample | |
4237 | ||
4238 | The colons, new line characters and the @code{TO FIX} string are | |
4239 | inserted if necessary, i.e., if the corresponding parameter is not | |
4240 | ignored. | |
4241 | ||
4242 | This function is specified in the X/Open Portability Guide. It is also | |
04b9968b | 4243 | available on all systems derived from System V. |
0501d603 | 4244 | |
cf29ffbe | 4245 | The function returns the value @code{MM_OK} if no error occurred. If |
0501d603 UD |
4246 | only the printing to standard error failed, it returns @code{MM_NOMSG}. |
4247 | If printing to the console fails, it returns @code{MM_NOCON}. If | |
cf29ffbe | 4248 | nothing is printed @code{MM_NOTOK} is returned. Among situations where |
0501d603 UD |
4249 | all outputs fail this last value is also returned if a parameter value |
4250 | is incorrect. | |
4251 | @end deftypefun | |
4252 | ||
4253 | There are two environment variables which influence the behaviour of | |
4254 | @code{fmtmsg}. The first is @code{MSGVERB}. It is used to control the | |
4255 | output actually happening on standard error (@emph{not} the console | |
04b9968b | 4256 | output). Each of the five fields can explicitly be enabled. To do |
0501d603 | 4257 | this the user has to put the @code{MSGVERB} variable with a format like |
cf29ffbe | 4258 | the following in the environment before calling the @code{fmtmsg} function |
0501d603 UD |
4259 | the first time: |
4260 | ||
4261 | @smallexample | |
4262 | MSGVERB=@var{keyword}[:@var{keyword}[:...]] | |
4263 | @end smallexample | |
4264 | ||
4265 | Valid @var{keyword}s are @code{label}, @code{severity}, @code{text}, | |
4266 | @code{action}, and @code{tag}. If the environment variable is not given | |
4267 | or is the empty string, a not supported keyword is given or the value is | |
4268 | somehow else invalid, no part of the message is masked out. | |
4269 | ||
4270 | The second environment variable which influences the behaviour of | |
4271 | @code{fmtmsg} is @code{SEV_LEVEL}. This variable and the change in the | |
4272 | behaviour of @code{fmtmsg} is not specified in the X/Open Portability | |
4273 | Guide. It is available in System V systems, though. It can be used to | |
cf29ffbe | 4274 | introduce new severity levels. By default, only the five severity levels |
0501d603 UD |
4275 | described above are available. Any other numeric value would make |
4276 | @code{fmtmsg} print nothing. | |
4277 | ||
4278 | If the user puts @code{SEV_LEVEL} with a format like | |
4279 | ||
4280 | @smallexample | |
4281 | SEV_LEVEL=[@var{description}[:@var{description}[:...]]] | |
4282 | @end smallexample | |
4283 | ||
4284 | @noindent | |
4285 | in the environment of the process before the first call to | |
4286 | @code{fmtmsg}, where @var{description} has a value of the form | |
4287 | ||
4288 | @smallexample | |
4289 | @var{severity-keyword},@var{level},@var{printstring} | |
4290 | @end smallexample | |
4291 | ||
4292 | The @var{severity-keyword} part is not used by @code{fmtmsg} but it has | |
4293 | to be present. The @var{level} part is a string representation of a | |
4294 | number. The numeric value must be a number greater than 4. This value | |
4295 | must be used in the @var{severity} parameter of @code{fmtmsg} to select | |
4296 | this class. It is not possible to overwrite any of the predefined | |
4297 | classes. The @var{printstring} is the string printed when a message of | |
4298 | this class is processed by @code{fmtmsg} (see above, @code{fmtsmg} does | |
4299 | not print the numeric value but instead the string representation). | |
4300 | ||
4301 | ||
4302 | @node Adding Severity Classes | |
4303 | @subsection Adding Severity Classes | |
4304 | @cindex severity class | |
4305 | ||
04b9968b | 4306 | There is another possibility to introduce severity classes besides using |
0501d603 UD |
4307 | the environment variable @code{SEV_LEVEL}. This simplifies the task of |
4308 | introducing new classes in a running program. One could use the | |
4309 | @code{setenv} or @code{putenv} function to set the environment variable, | |
cf29ffbe | 4310 | but this is toilsome. |
0501d603 UD |
4311 | |
4312 | @deftypefun int addseverity (int @var{severity}, const char *@var{string}) | |
04b9968b | 4313 | This function allows the introduction of new severity classes which can be |
0501d603 UD |
4314 | addressed by the @var{severity} parameter of the @code{fmtmsg} function. |
4315 | The @var{severity} parameter of @code{addseverity} must match the value | |
04b9968b | 4316 | for the parameter with the same name of @code{fmtmsg}, and @var{string} |
0501d603 UD |
4317 | is the string printed in the actual messages instead of the numeric |
4318 | value. | |
4319 | ||
4320 | If @var{string} is @code{NULL} the severity class with the numeric value | |
4321 | according to @var{severity} is removed. | |
4322 | ||
cf29ffbe UD |
4323 | It is not possible to overwrite or remove one of the default severity |
4324 | classes. All calls to @code{addseverity} with @var{severity} set to one | |
4325 | of the values for the default classes will fail. | |
4326 | ||
0501d603 UD |
4327 | The return value is @code{MM_OK} if the task was successfully performed. |
4328 | If the return value is @code{MM_NOTOK} something went wrong. This could | |
4329 | mean that no more memory is available or a class is not available when | |
4330 | it has to be removed. | |
4331 | ||
4332 | This function is not specified in the X/Open Portability Guide although | |
cf29ffbe | 4333 | the @code{fmtsmg} function is. It is available on System V systems. |
0501d603 UD |
4334 | @end deftypefun |
4335 | ||
4336 | ||
4337 | @node Example | |
4338 | @subsection How to use @code{fmtmsg} and @code{addseverity} | |
4339 | ||
4340 | Here is a simple example program to illustrate the use of the both | |
4341 | functions described in this section. | |
4342 | ||
4343 | @smallexample | |
4344 | @include fmtmsgexpl.c.texi | |
4345 | @end smallexample | |
4346 | ||
04b9968b UD |
4347 | The second call to @code{fmtmsg} illustrates a use of this function as |
4348 | it usually occurs on System V systems, which heavily use this function. | |
4349 | It seems worthwhile to give a short explanation here of how this system | |
4350 | works on System V. The value of the | |
0501d603 UD |
4351 | @var{label} field (@code{UX:cat}) says that the error occured in the |
4352 | Unix program @code{cat}. The explanation of the error follows and the | |
4353 | value for the @var{action} parameter is @code{"refer to manual"}. One | |
04b9968b | 4354 | could be more specific here, if necessary. The @var{tag} field contains, |
0501d603 UD |
4355 | as proposed above, the value of the string given for the @var{label} |
4356 | parameter, and additionally a unique ID (@code{001} in this case). For | |
4357 | a GNU environment this string could contain a reference to the | |
4358 | corresponding node in the Info page for the program. | |
4359 | ||
4360 | @noindent | |
4361 | Running this program without specifying the @code{MSGVERB} and | |
4362 | @code{SEV_LEVEL} function produces the following output: | |
4363 | ||
4364 | @smallexample | |
4365 | UX:cat: NOTE2: invalid syntax | |
4366 | TO FIX: refer to manual UX:cat:001 | |
4367 | @end smallexample | |
4368 | ||
4369 | We see the different fields of the message and how the extra glue (the | |
4370 | colons and the @code{TO FIX} string) are printed. But only one of the | |
4371 | three calls to @code{fmtmsg} produced output. The first call does not | |
4372 | print anything because the @var{label} parameter is not in the correct | |
8b7fb588 UD |
4373 | form. The string must contain two fields, separated by a colon |
4374 | (@pxref{Printing Formatted Messages}). The third @code{fmtmsg} call | |
0501d603 UD |
4375 | produced no output since the class with the numeric value @code{6} is |
4376 | not defined. Although a class with numeric value @code{5} is also not | |
04b9968b | 4377 | defined by default, the call to @code{addseverity} introduces it and |
0ea554bf | 4378 | the second call to @code{fmtmsg} produces the above output. |
0501d603 UD |
4379 | |
4380 | When we change the environment of the program to contain | |
4381 | @code{SEV_LEVEL=XXX,6,NOTE} when running it we get a different result: | |
4382 | ||
4383 | @smallexample | |
4384 | UX:cat: NOTE2: invalid syntax | |
4385 | TO FIX: refer to manual UX:cat:001 | |
4386 | label:foo: NOTE: text | |
4387 | TO FIX: action tag | |
4388 | @end smallexample | |
4389 | ||
04b9968b | 4390 | Now the third call to @code{fmtmsg} produced some output and we see how |
0501d603 UD |
4391 | the string @code{NOTE} from the environment variable appears in the |
4392 | message. | |
4393 | ||
04b9968b | 4394 | Now we can reduce the output by specifying which fields we are |
0501d603 UD |
4395 | interested in. If we additionally set the environment variable |
4396 | @code{MSGVERB} to the value @code{severity:label:action} we get the | |
4397 | following output: | |
4398 | ||
4399 | @smallexample | |
4400 | UX:cat: NOTE2 | |
4401 | TO FIX: refer to manual | |
4402 | label:foo: NOTE | |
4403 | TO FIX: action | |
4404 | @end smallexample | |
4405 | ||
4406 | @noindent | |
4407 | I.e., the output produced by the @var{text} and the @var{tag} parameters | |
cf29ffbe | 4408 | to @code{fmtmsg} vanished. Please also note that now there is no colon |
0501d603 | 4409 | after the @code{NOTE} and @code{NOTE2} strings in the output. This is |
04b9968b | 4410 | not necessary since there is no more output on this line because the text |
0501d603 | 4411 | is missing. |