]>
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 |
162ba701 | 17 | devices are created for you. |
28f540f4 RM |
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. | |
7b4161bb | 20 | * Streams and Threads:: Issues with streams in threaded programs. |
b5e73f56 | 21 | * Streams and I18N:: Streams in internationalized applications. |
28f540f4 RM |
22 | * Simple Output:: Unformatted output by characters and lines. |
23 | * Character Input:: Unformatted input by characters and words. | |
24 | * Line Input:: Reading a line or a record from a stream. | |
25 | * Unreading:: Peeking ahead/pushing back input just read. | |
26 | * Block Input/Output:: Input and output operations on blocks of data. | |
27 | * Formatted Output:: @code{printf} and related functions. | |
28 | * Customizing Printf:: You can define new conversion specifiers for | |
162ba701 | 29 | @code{printf} and friends. |
28f540f4 RM |
30 | * Formatted Input:: @code{scanf} and related functions. |
31 | * EOF and Errors:: How you can tell if an I/O error happens. | |
460e040a | 32 | * Error Recovery:: What you can do about errors. |
28f540f4 | 33 | * Binary Streams:: Some systems distinguish between text files |
162ba701 | 34 | and binary files. |
28f540f4 | 35 | * File Positioning:: About random-access streams. |
f65fd747 | 36 | * Portable Positioning:: Random access on peculiar ISO C systems. |
28f540f4 RM |
37 | * Stream Buffering:: How to control buffering of streams. |
38 | * Other Kinds of Streams:: Streams that do not necessarily correspond | |
162ba701 | 39 | to an open file. |
0501d603 | 40 | * Formatted Messages:: Print strictly formatted messages. |
28f540f4 RM |
41 | @end menu |
42 | ||
43 | @node Streams | |
44 | @section Streams | |
45 | ||
46 | For historical reasons, the type of the C data structure that represents | |
47 | a stream is called @code{FILE} rather than ``stream''. Since most of | |
48 | the library functions deal with objects of type @code{FILE *}, sometimes | |
49 | the term @dfn{file pointer} is also used to mean ``stream''. This leads | |
50 | to unfortunate confusion over terminology in many books on C. This | |
51 | manual, however, is careful to use the terms ``file'' and ``stream'' | |
52 | only in the technical sense. | |
53 | @cindex file pointer | |
54 | ||
55 | @pindex stdio.h | |
56 | The @code{FILE} type is declared in the header file @file{stdio.h}. | |
57 | ||
58 | @comment stdio.h | |
f65fd747 | 59 | @comment ISO |
28f540f4 RM |
60 | @deftp {Data Type} FILE |
61 | This is the data type used to represent stream objects. A @code{FILE} | |
62 | object holds all of the internal state information about the connection | |
63 | to the associated file, including such things as the file position | |
64 | indicator and buffering information. Each stream also has error and | |
65 | end-of-file status indicators that can be tested with the @code{ferror} | |
66 | and @code{feof} functions; see @ref{EOF and Errors}. | |
67 | @end deftp | |
68 | ||
69 | @code{FILE} objects are allocated and managed internally by the | |
70 | input/output library functions. Don't try to create your own objects of | |
71 | type @code{FILE}; let the library do it. Your programs should | |
72 | deal only with pointers to these objects (that is, @code{FILE *} values) | |
73 | rather than the objects themselves. | |
6d52618b | 74 | @c !!! should say that FILE's have "No user-serviceable parts inside." |
28f540f4 RM |
75 | |
76 | @node Standard Streams | |
77 | @section Standard Streams | |
78 | @cindex standard streams | |
79 | @cindex streams, standard | |
80 | ||
81 | When the @code{main} function of your program is invoked, it already has | |
82 | three predefined streams open and available for use. These represent | |
83 | the ``standard'' input and output channels that have been established | |
84 | for the process. | |
85 | ||
86 | These streams are declared in the header file @file{stdio.h}. | |
87 | @pindex stdio.h | |
88 | ||
89 | @comment stdio.h | |
f65fd747 | 90 | @comment ISO |
28f540f4 RM |
91 | @deftypevar {FILE *} stdin |
92 | The @dfn{standard input} stream, which is the normal source of input for the | |
93 | program. | |
94 | @end deftypevar | |
95 | @cindex standard input stream | |
96 | ||
97 | @comment stdio.h | |
f65fd747 | 98 | @comment ISO |
28f540f4 RM |
99 | @deftypevar {FILE *} stdout |
100 | The @dfn{standard output} stream, which is used for normal output from | |
101 | the program. | |
102 | @end deftypevar | |
103 | @cindex standard output stream | |
104 | ||
105 | @comment stdio.h | |
f65fd747 | 106 | @comment ISO |
28f540f4 RM |
107 | @deftypevar {FILE *} stderr |
108 | The @dfn{standard error} stream, which is used for error messages and | |
109 | diagnostics issued by the program. | |
110 | @end deftypevar | |
111 | @cindex standard error stream | |
112 | ||
a7a93d50 | 113 | On @gnusystems{}, you can specify what files or processes correspond to |
28f540f4 RM |
114 | these streams using the pipe and redirection facilities provided by the |
115 | shell. (The primitives shells use to implement these facilities are | |
116 | described in @ref{File System Interface}.) Most other operating systems | |
117 | provide similar mechanisms, but the details of how to use them can vary. | |
118 | ||
1f77f049 | 119 | In @theglibc{}, @code{stdin}, @code{stdout}, and @code{stderr} are |
b5e73f56 UD |
120 | normal variables which you can set just like any others. For example, |
121 | to redirect the standard output to a file, you could do: | |
28f540f4 RM |
122 | |
123 | @smallexample | |
124 | fclose (stdout); | |
125 | stdout = fopen ("standard-output-file", "w"); | |
126 | @end smallexample | |
127 | ||
128 | Note however, that in other systems @code{stdin}, @code{stdout}, and | |
129 | @code{stderr} are macros that you cannot assign to in the normal way. | |
130 | But you can use @code{freopen} to get the effect of closing one and | |
131 | reopening it. @xref{Opening Streams}. | |
132 | ||
b5e73f56 UD |
133 | The three streams @code{stdin}, @code{stdout}, and @code{stderr} are not |
134 | unoriented at program start (@pxref{Streams and I18N}). | |
135 | ||
28f540f4 RM |
136 | @node Opening Streams |
137 | @section Opening Streams | |
138 | ||
139 | @cindex opening a stream | |
140 | Opening a file with the @code{fopen} function creates a new stream and | |
141 | establishes a connection between the stream and a file. This may | |
19c3f208 | 142 | involve creating a new file. |
28f540f4 RM |
143 | |
144 | @pindex stdio.h | |
145 | Everything described in this section is declared in the header file | |
146 | @file{stdio.h}. | |
147 | ||
148 | @comment stdio.h | |
f65fd747 | 149 | @comment ISO |
28f540f4 RM |
150 | @deftypefun {FILE *} fopen (const char *@var{filename}, const char *@var{opentype}) |
151 | The @code{fopen} function opens a stream for I/O to the file | |
152 | @var{filename}, and returns a pointer to the stream. | |
153 | ||
154 | The @var{opentype} argument is a string that controls how the file is | |
155 | opened and specifies attributes of the resulting stream. It must begin | |
156 | with one of the following sequences of characters: | |
157 | ||
158 | @table @samp | |
159 | @item r | |
160 | Open an existing file for reading only. | |
161 | ||
162 | @item w | |
163 | Open the file for writing only. If the file already exists, it is | |
164 | truncated to zero length. Otherwise a new file is created. | |
165 | ||
166 | @item a | |
167 | Open a file for append access; that is, writing at the end of file only. | |
168 | If the file already exists, its initial contents are unchanged and | |
169 | output to the stream is appended to the end of the file. | |
170 | Otherwise, a new, empty file is created. | |
171 | ||
172 | @item r+ | |
173 | Open an existing file for both reading and writing. The initial contents | |
174 | of the file are unchanged and the initial file position is at the | |
175 | beginning of the file. | |
176 | ||
177 | @item w+ | |
178 | Open a file for both reading and writing. If the file already exists, it | |
179 | is truncated to zero length. Otherwise, a new file is created. | |
180 | ||
181 | @item a+ | |
182 | Open or create file for both reading and appending. If the file exists, | |
183 | its initial contents are unchanged. Otherwise, a new file is created. | |
184 | The initial file position for reading is at the beginning of the file, | |
185 | but output is always appended to the end of the file. | |
186 | @end table | |
187 | ||
188 | As you can see, @samp{+} requests a stream that can do both input and | |
162ba701 PIM |
189 | output. When using such a stream, you must call @code{fflush} |
190 | (@pxref{Stream Buffering}) or a file positioning function such as | |
191 | @code{fseek} (@pxref{File Positioning}) when switching from reading | |
192 | to writing or vice versa. Otherwise, internal buffers might not be | |
193 | emptied properly. | |
28f540f4 RM |
194 | |
195 | Additional characters may appear after these to specify flags for the | |
196 | call. Always put the mode (@samp{r}, @samp{w+}, etc.) first; that is | |
197 | the only part you are guaranteed will be understood by all systems. | |
198 | ||
823fbbb4 AJ |
199 | @Theglibc{} defines additional characters for use in @var{opentype}: |
200 | ||
201 | @table @samp | |
202 | @item c | |
203 | The file is opened with cancellation in the I/O functions disabled. | |
204 | ||
205 | @item e | |
206 | The underlying file descriptor will be closed if you use any of the | |
207 | @code{exec@dots{}} functions (@pxref{Executing a File}). (This is | |
208 | equivalent to having set @code{FD_CLOEXEC} on that descriptor. | |
209 | @xref{Descriptor Flags}.) | |
210 | ||
211 | @item m | |
212 | The file is opened and accessed using @code{mmap}. This is only | |
213 | supported with files opened for reading. | |
214 | ||
215 | @item x | |
216 | Insist on creating a new file---if a file @var{filename} already | |
217 | exists, @code{fopen} fails rather than opening it. If you use | |
218 | @samp{x} you are guaranteed that you will not clobber an existing | |
219 | file. This is equivalent to the @code{O_EXCL} option to the | |
220 | @code{open} function (@pxref{Opening and Closing Files}). | |
221 | ||
222 | The @samp{x} modifier is part of @w{ISO C11}. | |
223 | @end table | |
28f540f4 RM |
224 | |
225 | The character @samp{b} in @var{opentype} has a standard meaning; it | |
226 | requests a binary stream rather than a text stream. But this makes no | |
a7a93d50 | 227 | difference in POSIX systems (including @gnusystems{}). If both |
28f540f4 RM |
228 | @samp{+} and @samp{b} are specified, they can appear in either order. |
229 | @xref{Binary Streams}. | |
230 | ||
4e2e9999 UD |
231 | @cindex stream orientation |
232 | @cindex orientation, stream | |
233 | If the @var{opentype} string contains the sequence | |
234 | @code{,ccs=@var{STRING}} then @var{STRING} is taken as the name of a | |
235 | coded character set and @code{fopen} will mark the stream as | |
5b23062f JM |
236 | wide-oriented with appropriate conversion functions in place to convert |
237 | from and to the character set @var{STRING}. Any other stream | |
4e2e9999 UD |
238 | is opened initially unoriented and the orientation is decided with the |
239 | first file operation. If the first operation is a wide character | |
240 | operation, the stream is not only marked as wide-oriented, also the | |
241 | conversion functions to convert to the coded character set used for the | |
242 | current locale are loaded. This will not change anymore from this point | |
243 | on even if the locale selected for the @code{LC_CTYPE} category is | |
244 | changed. | |
245 | ||
28f540f4 RM |
246 | Any other characters in @var{opentype} are simply ignored. They may be |
247 | meaningful in other systems. | |
248 | ||
249 | If the open fails, @code{fopen} returns a null pointer. | |
a3a4a74e UD |
250 | |
251 | When the sources are compiling with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 252 | 32 bit machine this function is in fact @code{fopen64} since the LFS |
a3a4a74e | 253 | interface replaces transparently the old interface. |
28f540f4 RM |
254 | @end deftypefun |
255 | ||
256 | You can have multiple streams (or file descriptors) pointing to the same | |
257 | file open at the same time. If you do only input, this works | |
258 | straightforwardly, but you must be careful if any output streams are | |
259 | included. @xref{Stream/Descriptor Precautions}. This is equally true | |
260 | whether the streams are in one program (not usual) or in several | |
261 | programs (which can easily happen). It may be advantageous to use the | |
262 | file locking facilities to avoid simultaneous access. @xref{File | |
263 | Locks}. | |
264 | ||
a3a4a74e UD |
265 | @comment stdio.h |
266 | @comment Unix98 | |
267 | @deftypefun {FILE *} fopen64 (const char *@var{filename}, const char *@var{opentype}) | |
268 | This function is similar to @code{fopen} but the stream it returns a | |
269 | pointer for is opened using @code{open64}. Therefore this stream can be | |
f24a6d08 | 270 | used even on files larger than @math{2^31} bytes on 32 bit machines. |
a3a4a74e UD |
271 | |
272 | Please note that the return type is still @code{FILE *}. There is no | |
273 | special @code{FILE} type for the LFS interface. | |
274 | ||
275 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
276 | bits machine this function is available under the name @code{fopen} | |
277 | and so transparently replaces the old interface. | |
278 | @end deftypefun | |
279 | ||
28f540f4 | 280 | @comment stdio.h |
f65fd747 | 281 | @comment ISO |
28f540f4 RM |
282 | @deftypevr Macro int FOPEN_MAX |
283 | The value of this macro is an integer constant expression that | |
284 | represents the minimum number of streams that the implementation | |
285 | guarantees can be open simultaneously. You might be able to open more | |
286 | than this many streams, but that is not guaranteed. The value of this | |
287 | constant is at least eight, which includes the three standard streams | |
288 | @code{stdin}, @code{stdout}, and @code{stderr}. In POSIX.1 systems this | |
289 | value is determined by the @code{OPEN_MAX} parameter; @pxref{General | |
290 | Limits}. In BSD and GNU, it is controlled by the @code{RLIMIT_NOFILE} | |
291 | resource limit; @pxref{Limits on Resources}. | |
292 | @end deftypevr | |
293 | ||
294 | @comment stdio.h | |
f65fd747 | 295 | @comment ISO |
28f540f4 RM |
296 | @deftypefun {FILE *} freopen (const char *@var{filename}, const char *@var{opentype}, FILE *@var{stream}) |
297 | This function is like a combination of @code{fclose} and @code{fopen}. | |
298 | It first closes the stream referred to by @var{stream}, ignoring any | |
299 | errors that are detected in the process. (Because errors are ignored, | |
300 | you should not use @code{freopen} on an output stream if you have | |
301 | actually done any output using the stream.) Then the file named by | |
302 | @var{filename} is opened with mode @var{opentype} as for @code{fopen}, | |
303 | and associated with the same stream object @var{stream}. | |
304 | ||
305 | If the operation fails, a null pointer is returned; otherwise, | |
306 | @code{freopen} returns @var{stream}. | |
307 | ||
308 | @code{freopen} has traditionally been used to connect a standard stream | |
309 | such as @code{stdin} with a file of your own choice. This is useful in | |
310 | programs in which use of a standard stream for certain purposes is | |
1f77f049 | 311 | hard-coded. In @theglibc{}, you can simply close the standard |
28f540f4 RM |
312 | streams and open new ones with @code{fopen}. But other systems lack |
313 | this ability, so using @code{freopen} is more portable. | |
a3a4a74e UD |
314 | |
315 | When the sources are compiling with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 316 | 32 bit machine this function is in fact @code{freopen64} since the LFS |
a3a4a74e UD |
317 | interface replaces transparently the old interface. |
318 | @end deftypefun | |
319 | ||
320 | @comment stdio.h | |
321 | @comment Unix98 | |
322 | @deftypefun {FILE *} freopen64 (const char *@var{filename}, const char *@var{opentype}, FILE *@var{stream}) | |
323 | This function is similar to @code{freopen}. The only difference is that | |
0be8752b | 324 | on 32 bit machine the stream returned is able to read beyond the |
a3a4a74e UD |
325 | @math{2^31} bytes limits imposed by the normal interface. It should be |
326 | noted that the stream pointed to by @var{stream} need not be opened | |
327 | using @code{fopen64} or @code{freopen64} since its mode is not important | |
328 | for this function. | |
329 | ||
330 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
331 | bits machine this function is available under the name @code{freopen} | |
332 | and so transparently replaces the old interface. | |
28f540f4 RM |
333 | @end deftypefun |
334 | ||
f8b23b32 UD |
335 | In some situations it is useful to know whether a given stream is |
336 | available for reading or writing. This information is normally not | |
337 | available and would have to be remembered separately. Solaris | |
338 | introduced a few functions to get this information from the stream | |
1f77f049 | 339 | descriptor and these functions are also available in @theglibc{}. |
f8b23b32 UD |
340 | |
341 | @comment stdio_ext.h | |
342 | @comment GNU | |
343 | @deftypefun int __freadable (FILE *@var{stream}) | |
344 | The @code{__freadable} function determines whether the stream | |
345 | @var{stream} was opened to allow reading. In this case the return value | |
346 | is nonzero. For write-only streams the function returns zero. | |
347 | ||
348 | This function is declared in @file{stdio_ext.h}. | |
349 | @end deftypefun | |
350 | ||
351 | @comment stdio_ext.h | |
352 | @comment GNU | |
353 | @deftypefun int __fwritable (FILE *@var{stream}) | |
354 | The @code{__fwritable} function determines whether the stream | |
355 | @var{stream} was opened to allow writing. In this case the return value | |
356 | is nonzero. For read-only streams the function returns zero. | |
357 | ||
358 | This function is declared in @file{stdio_ext.h}. | |
359 | @end deftypefun | |
360 | ||
361 | For slightly different kind of problems there are two more functions. | |
362 | They provide even finer-grained information. | |
363 | ||
364 | @comment stdio_ext.h | |
365 | @comment GNU | |
366 | @deftypefun int __freading (FILE *@var{stream}) | |
367 | The @code{__freading} function determines whether the stream | |
368 | @var{stream} was last read from or whether it is opened read-only. In | |
369 | this case the return value is nonzero, otherwise it is zero. | |
370 | Determining whether a stream opened for reading and writing was last | |
371 | used for writing allows to draw conclusions about the content about the | |
372 | buffer, among other things. | |
373 | ||
374 | This function is declared in @file{stdio_ext.h}. | |
375 | @end deftypefun | |
376 | ||
377 | @comment stdio_ext.h | |
378 | @comment GNU | |
379 | @deftypefun int __fwriting (FILE *@var{stream}) | |
380 | The @code{__fwriting} function determines whether the stream | |
381 | @var{stream} was last written to or whether it is opened write-only. In | |
382 | this case the return value is nonzero, otherwise it is zero. | |
383 | ||
384 | This function is declared in @file{stdio_ext.h}. | |
385 | @end deftypefun | |
386 | ||
28f540f4 RM |
387 | |
388 | @node Closing Streams | |
389 | @section Closing Streams | |
390 | ||
391 | @cindex closing a stream | |
392 | When a stream is closed with @code{fclose}, the connection between the | |
0bc93a2f | 393 | stream and the file is canceled. After you have closed a stream, you |
28f540f4 RM |
394 | cannot perform any additional operations on it. |
395 | ||
396 | @comment stdio.h | |
f65fd747 | 397 | @comment ISO |
28f540f4 RM |
398 | @deftypefun int fclose (FILE *@var{stream}) |
399 | This function causes @var{stream} to be closed and the connection to | |
400 | the corresponding file to be broken. Any buffered output is written | |
401 | and any buffered input is discarded. The @code{fclose} function returns | |
402 | a value of @code{0} if the file was closed successfully, and @code{EOF} | |
19c3f208 | 403 | if an error was detected. |
28f540f4 RM |
404 | |
405 | It is important to check for errors when you call @code{fclose} to close | |
406 | an output stream, because real, everyday errors can be detected at this | |
407 | time. For example, when @code{fclose} writes the remaining buffered | |
408 | output, it might get an error because the disk is full. Even if you | |
409 | know the buffer is empty, errors can still occur when closing a file if | |
410 | you are using NFS. | |
411 | ||
412 | The function @code{fclose} is declared in @file{stdio.h}. | |
413 | @end deftypefun | |
414 | ||
1f77f049 | 415 | To close all streams currently available @theglibc{} provides |
6bc31da0 UD |
416 | another function. |
417 | ||
418 | @comment stdio.h | |
419 | @comment GNU | |
420 | @deftypefun int fcloseall (void) | |
421 | This function causes all open streams of the process to be closed and | |
422 | the connection to corresponding files to be broken. All buffered data | |
f2ea0f5b | 423 | is written and any buffered input is discarded. The @code{fcloseall} |
6bc31da0 UD |
424 | function returns a value of @code{0} if all the files were closed |
425 | successfully, and @code{EOF} if an error was detected. | |
426 | ||
0be8752b | 427 | This function should be used only in special situations, e.g., when an |
6bc31da0 | 428 | error occurred and the program must be aborted. Normally each single |
0be8752b UD |
429 | stream should be closed separately so that problems with individual |
430 | streams can be identified. It is also problematic since the standard | |
431 | streams (@pxref{Standard Streams}) will also be closed. | |
6bc31da0 UD |
432 | |
433 | The function @code{fcloseall} is declared in @file{stdio.h}. | |
434 | @end deftypefun | |
435 | ||
28f540f4 RM |
436 | If the @code{main} function to your program returns, or if you call the |
437 | @code{exit} function (@pxref{Normal Termination}), all open streams are | |
438 | automatically closed properly. If your program terminates in any other | |
439 | manner, such as by calling the @code{abort} function (@pxref{Aborting a | |
440 | Program}) or from a fatal signal (@pxref{Signal Handling}), open streams | |
441 | might not be closed properly. Buffered output might not be flushed and | |
442 | files may be incomplete. For more information on buffering of streams, | |
443 | see @ref{Stream Buffering}. | |
444 | ||
7b4161bb UD |
445 | @node Streams and Threads |
446 | @section Streams and Threads | |
447 | ||
448 | @cindex threads | |
449 | @cindex multi-threaded application | |
450 | Streams can be used in multi-threaded applications in the same way they | |
451 | are used in single-threaded applications. But the programmer must be | |
11bf311e | 452 | aware of the possible complications. It is important to know about |
7b4161bb UD |
453 | these also if the program one writes never use threads since the design |
454 | and implementation of many stream functions is heavily influenced by the | |
f6af7428 | 455 | requirements added by multi-threaded programming. |
7b4161bb UD |
456 | |
457 | The POSIX standard requires that by default the stream operations are | |
0bc93a2f | 458 | atomic. I.e., issuing two stream operations for the same stream in two |
7b4161bb UD |
459 | threads at the same time will cause the operations to be executed as if |
460 | they were issued sequentially. The buffer operations performed while | |
461 | reading or writing are protected from other uses of the same stream. To | |
462 | do this each stream has an internal lock object which has to be | |
463 | (implicitly) acquired before any work can be done. | |
464 | ||
465 | But there are situations where this is not enough and there are also | |
466 | situations where this is not wanted. The implicit locking is not enough | |
467 | if the program requires more than one stream function call to happen | |
468 | atomically. One example would be if an output line a program wants to | |
469 | generate is created by several function calls. The functions by | |
470 | themselves would ensure only atomicity of their own operation, but not | |
471 | atomicity over all the function calls. For this it is necessary to | |
472 | perform the stream locking in the application code. | |
473 | ||
474 | @comment stdio.h | |
475 | @comment POSIX | |
476 | @deftypefun void flockfile (FILE *@var{stream}) | |
477 | The @code{flockfile} function acquires the internal locking object | |
f6af7428 | 478 | associated with the stream @var{stream}. This ensures that no other |
7b4161bb UD |
479 | thread can explicitly through @code{flockfile}/@code{ftrylockfile} or |
480 | implicit through a call of a stream function lock the stream. The | |
481 | thread will block until the lock is acquired. An explicit call to | |
482 | @code{funlockfile} has to be used to release the lock. | |
483 | @end deftypefun | |
484 | ||
485 | @comment stdio.h | |
486 | @comment POSIX | |
487 | @deftypefun int ftrylockfile (FILE *@var{stream}) | |
f6af7428 | 488 | The @code{ftrylockfile} function tries to acquire the internal locking |
7b4161bb UD |
489 | object associated with the stream @var{stream} just like |
490 | @code{flockfile}. But unlike @code{flockfile} this function does not | |
491 | block if the lock is not available. @code{ftrylockfile} returns zero if | |
492 | the lock was successfully acquired. Otherwise the stream is locked by | |
493 | another thread. | |
494 | @end deftypefun | |
495 | ||
496 | @comment stdio.h | |
497 | @comment POSIX | |
498 | @deftypefun void funlockfile (FILE *@var{stream}) | |
499 | The @code{funlockfile} function releases the internal locking object of | |
500 | the stream @var{stream}. The stream must have been locked before by a | |
501 | call to @code{flockfile} or a successful call of @code{ftrylockfile}. | |
502 | The implicit locking performed by the stream operations do not count. | |
503 | The @code{funlockfile} function does not return an error status and the | |
504 | behavior of a call for a stream which is not locked by the current | |
505 | thread is undefined. | |
506 | @end deftypefun | |
507 | ||
508 | The following example shows how the functions above can be used to | |
509 | generate an output line atomically even in multi-threaded applications | |
510 | (yes, the same job could be done with one @code{fprintf} call but it is | |
511 | sometimes not possible): | |
512 | ||
513 | @smallexample | |
514 | FILE *fp; | |
515 | @{ | |
95fdc6a0 | 516 | @dots{} |
7b4161bb UD |
517 | flockfile (fp); |
518 | fputs ("This is test number ", fp); | |
519 | fprintf (fp, "%d\n", test); | |
520 | funlockfile (fp) | |
521 | @} | |
522 | @end smallexample | |
523 | ||
524 | Without the explicit locking it would be possible for another thread to | |
525 | use the stream @var{fp} after the @code{fputs} call return and before | |
526 | @code{fprintf} was called with the result that the number does not | |
527 | follow the word @samp{number}. | |
528 | ||
529 | From this description it might already be clear that the locking objects | |
530 | in streams are no simple mutexes. Since locking the same stream twice | |
531 | in the same thread is allowed the locking objects must be equivalent to | |
532 | recursive mutexes. These mutexes keep track of the owner and the number | |
533 | of times the lock is acquired. The same number of @code{funlockfile} | |
534 | calls by the same threads is necessary to unlock the stream completely. | |
535 | For instance: | |
536 | ||
537 | @smallexample | |
538 | void | |
539 | foo (FILE *fp) | |
540 | @{ | |
541 | ftrylockfile (fp); | |
542 | fputs ("in foo\n", fp); | |
543 | /* @r{This is very wrong!!!} */ | |
544 | funlockfile (fp); | |
545 | @} | |
546 | @end smallexample | |
547 | ||
548 | It is important here that the @code{funlockfile} function is only called | |
549 | if the @code{ftrylockfile} function succeeded in locking the stream. It | |
550 | is therefore always wrong to ignore the result of @code{ftrylockfile}. | |
551 | And it makes no sense since otherwise one would use @code{flockfile}. | |
552 | The result of code like that above is that either @code{funlockfile} | |
f6af7428 | 553 | tries to free a stream that hasn't been locked by the current thread or it |
7b4161bb UD |
554 | frees the stream prematurely. The code should look like this: |
555 | ||
556 | @smallexample | |
557 | void | |
558 | foo (FILE *fp) | |
559 | @{ | |
560 | if (ftrylockfile (fp) == 0) | |
561 | @{ | |
562 | fputs ("in foo\n", fp); | |
563 | funlockfile (fp); | |
564 | @} | |
565 | @} | |
566 | @end smallexample | |
567 | ||
568 | Now that we covered why it is necessary to have these locking it is | |
f6af7428 | 569 | necessary to talk about situations when locking is unwanted and what can |
7b4161bb UD |
570 | be done. The locking operations (explicit or implicit) don't come for |
571 | free. Even if a lock is not taken the cost is not zero. The operations | |
3ca9e670 | 572 | which have to be performed require memory operations that are safe in |
7b4161bb UD |
573 | multi-processor environments. With the many local caches involved in |
574 | such systems this is quite costly. So it is best to avoid the locking | |
3ca9e670 UD |
575 | completely if it is not needed -- because the code in question is never |
576 | used in a context where two or more threads may use a stream at a time. | |
7b4161bb UD |
577 | This can be determined most of the time for application code; for |
578 | library code which can be used in many contexts one should default to be | |
579 | conservative and use locking. | |
580 | ||
581 | There are two basic mechanisms to avoid locking. The first is to use | |
582 | the @code{_unlocked} variants of the stream operations. The POSIX | |
1f77f049 | 583 | standard defines quite a few of those and @theglibc{} adds a few |
7b4161bb | 584 | more. These variants of the functions behave just like the functions |
3ca9e670 | 585 | with the name without the suffix except that they do not lock the |
7b4161bb | 586 | stream. Using these functions is very desirable since they are |
f6af7428 | 587 | potentially much faster. This is not only because the locking |
7b4161bb | 588 | operation itself is avoided. More importantly, functions like |
0bc93a2f | 589 | @code{putc} and @code{getc} are very simple and traditionally (before the |
7b4161bb | 590 | introduction of threads) were implemented as macros which are very fast |
3ca9e670 UD |
591 | if the buffer is not empty. With the addition of locking requirements |
592 | these functions are no longer implemented as macros since they would | |
bbf70ae9 | 593 | expand to too much code. |
3ca9e670 | 594 | But these macros are still available with the same functionality under the new |
7b4161bb UD |
595 | names @code{putc_unlocked} and @code{getc_unlocked}. This possibly huge |
596 | difference of speed also suggests the use of the @code{_unlocked} | |
597 | functions even if locking is required. The difference is that the | |
598 | locking then has to be performed in the program: | |
599 | ||
600 | @smallexample | |
601 | void | |
602 | foo (FILE *fp, char *buf) | |
603 | @{ | |
604 | flockfile (fp); | |
605 | while (*buf != '/') | |
606 | putc_unlocked (*buf++, fp); | |
607 | funlockfile (fp); | |
608 | @} | |
609 | @end smallexample | |
610 | ||
611 | If in this example the @code{putc} function would be used and the | |
612 | explicit locking would be missing the @code{putc} function would have to | |
613 | acquire the lock in every call, potentially many times depending on when | |
614 | the loop terminates. Writing it the way illustrated above allows the | |
615 | @code{putc_unlocked} macro to be used which means no locking and direct | |
616 | manipulation of the buffer of the stream. | |
617 | ||
618 | A second way to avoid locking is by using a non-standard function which | |
1f77f049 | 619 | was introduced in Solaris and is available in @theglibc{} as well. |
7b4161bb UD |
620 | |
621 | @comment stdio_ext.h | |
622 | @comment GNU | |
623 | @deftypefun int __fsetlocking (FILE *@var{stream}, int @var{type}) | |
624 | ||
625 | The @code{__fsetlocking} function can be used to select whether the | |
626 | stream operations will implicitly acquire the locking object of the | |
627 | stream @var{stream}. By default this is done but it can be disabled and | |
628 | reinstated using this function. There are three values defined for the | |
629 | @var{type} parameter. | |
630 | ||
631 | @vtable @code | |
632 | @item FSETLOCKING_INTERNAL | |
633 | The stream @code{stream} will from now on use the default internal | |
634 | locking. Every stream operation with exception of the @code{_unlocked} | |
635 | variants will implicitly lock the stream. | |
636 | ||
637 | @item FSETLOCKING_BYCALLER | |
638 | After the @code{__fsetlocking} function returns the user is responsible | |
639 | for locking the stream. None of the stream operations will implicitly | |
640 | do this anymore until the state is set back to | |
641 | @code{FSETLOCKING_INTERNAL}. | |
642 | ||
643 | @item FSETLOCKING_QUERY | |
644 | @code{__fsetlocking} only queries the current locking state of the | |
645 | stream. The return value will be @code{FSETLOCKING_INTERNAL} or | |
646 | @code{FSETLOCKING_BYCALLER} depending on the state. | |
647 | @end vtable | |
648 | ||
649 | The return value of @code{__fsetlocking} is either | |
650 | @code{FSETLOCKING_INTERNAL} or @code{FSETLOCKING_BYCALLER} depending on | |
651 | the state of the stream before the call. | |
652 | ||
653 | This function and the values for the @var{type} parameter are declared | |
654 | in @file{stdio_ext.h}. | |
655 | @end deftypefun | |
656 | ||
657 | This function is especially useful when program code has to be used | |
658 | which is written without knowledge about the @code{_unlocked} functions | |
3ca9e670 | 659 | (or if the programmer was too lazy to use them). |
7b4161bb | 660 | |
b5e73f56 UD |
661 | @node Streams and I18N |
662 | @section Streams in Internationalized Applications | |
663 | ||
664 | @w{ISO C90} introduced the new type @code{wchar_t} to allow handling | |
665 | larger character sets. What was missing was a possibility to output | |
666 | strings of @code{wchar_t} directly. One had to convert them into | |
667 | multibyte strings using @code{mbstowcs} (there was no @code{mbsrtowcs} | |
668 | yet) and then use the normal stream functions. While this is doable it | |
669 | is very cumbersome since performing the conversions is not trivial and | |
670 | greatly increases program complexity and size. | |
671 | ||
672 | The Unix standard early on (I think in XPG4.2) introduced two additional | |
673 | format specifiers for the @code{printf} and @code{scanf} families of | |
674 | functions. Printing and reading of single wide characters was made | |
675 | possible using the @code{%C} specifier and wide character strings can be | |
676 | handled with @code{%S}. These modifiers behave just like @code{%c} and | |
677 | @code{%s} only that they expect the corresponding argument to have the | |
678 | wide character type and that the wide character and string are | |
679 | transformed into/from multibyte strings before being used. | |
680 | ||
681 | This was a beginning but it is still not good enough. Not always is it | |
682 | desirable to use @code{printf} and @code{scanf}. The other, smaller and | |
683 | faster functions cannot handle wide characters. Second, it is not | |
684 | possible to have a format string for @code{printf} and @code{scanf} | |
685 | consisting of wide characters. The result is that format strings would | |
686 | have to be generated if they have to contain non-basic characters. | |
687 | ||
688 | @cindex C++ streams | |
689 | @cindex streams, C++ | |
690 | In the @w{Amendment 1} to @w{ISO C90} a whole new set of functions was | |
691 | added to solve the problem. Most of the stream functions got a | |
692 | counterpart which take a wide character or wide character string instead | |
693 | of a character or string respectively. The new functions operate on the | |
694 | same streams (like @code{stdout}). This is different from the model of | |
695 | the C++ runtime library where separate streams for wide and normal I/O | |
696 | are used. | |
697 | ||
698 | @cindex orientation, stream | |
699 | @cindex stream orientation | |
700 | Being able to use the same stream for wide and normal operations comes | |
701 | with a restriction: a stream can be used either for wide operations or | |
702 | for normal operations. Once it is decided there is no way back. Only a | |
703 | call to @code{freopen} or @code{freopen64} can reset the | |
704 | @dfn{orientation}. The orientation can be decided in three ways: | |
705 | ||
706 | @itemize @bullet | |
707 | @item | |
708 | If any of the normal character functions is used (this includes the | |
22298d18 | 709 | @code{fread} and @code{fwrite} functions) the stream is marked as not |
b5e73f56 UD |
710 | wide oriented. |
711 | ||
712 | @item | |
713 | If any of the wide character functions is used the stream is marked as | |
22298d18 | 714 | wide oriented. |
b5e73f56 UD |
715 | |
716 | @item | |
717 | The @code{fwide} function can be used to set the orientation either way. | |
718 | @end itemize | |
719 | ||
720 | It is important to never mix the use of wide and not wide operations on | |
721 | a stream. There are no diagnostics issued. The application behavior | |
722 | will simply be strange or the application will simply crash. The | |
723 | @code{fwide} function can help avoiding this. | |
724 | ||
725 | @comment wchar.h | |
726 | @comment ISO | |
727 | @deftypefun int fwide (FILE *@var{stream}, int @var{mode}) | |
728 | ||
22298d18 | 729 | The @code{fwide} function can be used to set and query the state of the |
b5e73f56 UD |
730 | orientation of the stream @var{stream}. If the @var{mode} parameter has |
731 | a positive value the streams get wide oriented, for negative values | |
732 | narrow oriented. It is not possible to overwrite previous orientations | |
733 | with @code{fwide}. I.e., if the stream @var{stream} was already | |
734 | oriented before the call nothing is done. | |
735 | ||
736 | If @var{mode} is zero the current orientation state is queried and | |
737 | nothing is changed. | |
738 | ||
739 | The @code{fwide} function returns a negative value, zero, or a positive | |
740 | value if the stream is narrow, not at all, or wide oriented | |
741 | respectively. | |
742 | ||
743 | This function was introduced in @w{Amendment 1} to @w{ISO C90} and is | |
744 | declared in @file{wchar.h}. | |
745 | @end deftypefun | |
746 | ||
747 | It is generally a good idea to orient a stream as early as possible. | |
748 | This can prevent surprise especially for the standard streams | |
749 | @code{stdin}, @code{stdout}, and @code{stderr}. If some library | |
750 | function in some situations uses one of these streams and this use | |
751 | orients the stream in a different way the rest of the application | |
752 | expects it one might end up with hard to reproduce errors. Remember | |
753 | that no errors are signal if the streams are used incorrectly. Leaving | |
754 | a stream unoriented after creation is normally only necessary for | |
755 | library functions which create streams which can be used in different | |
756 | contexts. | |
757 | ||
758 | When writing code which uses streams and which can be used in different | |
759 | contexts it is important to query the orientation of the stream before | |
760 | using it (unless the rules of the library interface demand a specific | |
761 | orientation). The following little, silly function illustrates this. | |
762 | ||
763 | @smallexample | |
764 | void | |
765 | print_f (FILE *fp) | |
766 | @{ | |
767 | if (fwide (fp, 0) > 0) | |
768 | /* @r{Positive return value means wide orientation.} */ | |
769 | fputwc (L'f', fp); | |
770 | else | |
771 | fputc ('f', fp); | |
772 | @} | |
773 | @end smallexample | |
774 | ||
775 | Note that in this case the function @code{print_f} decides about the | |
776 | orientation of the stream if it was unoriented before (will not happen | |
777 | if the advise above is followed). | |
778 | ||
779 | The encoding used for the @code{wchar_t} values is unspecified and the | |
780 | user must not make any assumptions about it. For I/O of @code{wchar_t} | |
781 | values this means that it is impossible to write these values directly | |
782 | to the stream. This is not what follows from the @w{ISO C} locale model | |
783 | either. What happens instead is that the bytes read from or written to | |
784 | the underlying media are first converted into the internal encoding | |
785 | chosen by the implementation for @code{wchar_t}. The external encoding | |
786 | is determined by the @code{LC_CTYPE} category of the current locale or | |
787 | by the @samp{ccs} part of the mode specification given to @code{fopen}, | |
788 | @code{fopen64}, @code{freopen}, or @code{freopen64}. How and when the | |
789 | conversion happens is unspecified and it happens invisible to the user. | |
790 | ||
791 | Since a stream is created in the unoriented state it has at that point | |
792 | no conversion associated with it. The conversion which will be used is | |
793 | determined by the @code{LC_CTYPE} category selected at the time the | |
794 | stream is oriented. If the locales are changed at the runtime this | |
795 | might produce surprising results unless one pays attention. This is | |
796 | just another good reason to orient the stream explicitly as soon as | |
797 | possible, perhaps with a call to @code{fwide}. | |
798 | ||
28f540f4 RM |
799 | @node Simple Output |
800 | @section Simple Output by Characters or Lines | |
801 | ||
802 | @cindex writing to a stream, by characters | |
803 | This section describes functions for performing character- and | |
804 | line-oriented output. | |
805 | ||
b5e73f56 UD |
806 | These narrow streams functions are declared in the header file |
807 | @file{stdio.h} and the wide stream functions in @file{wchar.h}. | |
28f540f4 | 808 | @pindex stdio.h |
b5e73f56 | 809 | @pindex wchar.h |
28f540f4 RM |
810 | |
811 | @comment stdio.h | |
f65fd747 | 812 | @comment ISO |
28f540f4 RM |
813 | @deftypefun int fputc (int @var{c}, FILE *@var{stream}) |
814 | The @code{fputc} function converts the character @var{c} to type | |
19c3f208 | 815 | @code{unsigned char}, and writes it to the stream @var{stream}. |
28f540f4 RM |
816 | @code{EOF} is returned if a write error occurs; otherwise the |
817 | character @var{c} is returned. | |
818 | @end deftypefun | |
819 | ||
b5e73f56 UD |
820 | @comment wchar.h |
821 | @comment ISO | |
822 | @deftypefun wint_t fputwc (wchar_t @var{wc}, FILE *@var{stream}) | |
823 | The @code{fputwc} function writes the wide character @var{wc} to the | |
824 | stream @var{stream}. @code{WEOF} is returned if a write error occurs; | |
825 | otherwise the character @var{wc} is returned. | |
826 | @end deftypefun | |
827 | ||
7b4161bb UD |
828 | @comment stdio.h |
829 | @comment POSIX | |
830 | @deftypefun int fputc_unlocked (int @var{c}, FILE *@var{stream}) | |
831 | The @code{fputc_unlocked} function is equivalent to the @code{fputc} | |
1dc843f7 | 832 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
833 | @end deftypefun |
834 | ||
b5e73f56 UD |
835 | @comment wchar.h |
836 | @comment POSIX | |
8ded91fb | 837 | @deftypefun wint_t fputwc_unlocked (wchar_t @var{wc}, FILE *@var{stream}) |
b5e73f56 | 838 | The @code{fputwc_unlocked} function is equivalent to the @code{fputwc} |
1dc843f7 | 839 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
840 | |
841 | This function is a GNU extension. | |
842 | @end deftypefun | |
843 | ||
28f540f4 | 844 | @comment stdio.h |
f65fd747 | 845 | @comment ISO |
28f540f4 RM |
846 | @deftypefun int putc (int @var{c}, FILE *@var{stream}) |
847 | This is just like @code{fputc}, except that most systems implement it as | |
848 | a macro, making it faster. One consequence is that it may evaluate the | |
849 | @var{stream} argument more than once, which is an exception to the | |
850 | general rule for macros. @code{putc} is usually the best function to | |
851 | use for writing a single character. | |
852 | @end deftypefun | |
853 | ||
b5e73f56 UD |
854 | @comment wchar.h |
855 | @comment ISO | |
856 | @deftypefun wint_t putwc (wchar_t @var{wc}, FILE *@var{stream}) | |
857 | This is just like @code{fputwc}, except that it can be implement as | |
858 | a macro, making it faster. One consequence is that it may evaluate the | |
859 | @var{stream} argument more than once, which is an exception to the | |
860 | general rule for macros. @code{putwc} is usually the best function to | |
861 | use for writing a single wide character. | |
862 | @end deftypefun | |
863 | ||
7b4161bb UD |
864 | @comment stdio.h |
865 | @comment POSIX | |
866 | @deftypefun int putc_unlocked (int @var{c}, FILE *@var{stream}) | |
867 | The @code{putc_unlocked} function is equivalent to the @code{putc} | |
1dc843f7 | 868 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
869 | @end deftypefun |
870 | ||
b5e73f56 UD |
871 | @comment wchar.h |
872 | @comment GNU | |
873 | @deftypefun wint_t putwc_unlocked (wchar_t @var{wc}, FILE *@var{stream}) | |
874 | The @code{putwc_unlocked} function is equivalent to the @code{putwc} | |
1dc843f7 | 875 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
876 | |
877 | This function is a GNU extension. | |
878 | @end deftypefun | |
879 | ||
28f540f4 | 880 | @comment stdio.h |
f65fd747 | 881 | @comment ISO |
28f540f4 RM |
882 | @deftypefun int putchar (int @var{c}) |
883 | The @code{putchar} function is equivalent to @code{putc} with | |
884 | @code{stdout} as the value of the @var{stream} argument. | |
885 | @end deftypefun | |
886 | ||
b5e73f56 UD |
887 | @comment wchar.h |
888 | @comment ISO | |
c44a663d | 889 | @deftypefun wint_t putwchar (wchar_t @var{wc}) |
b5e73f56 UD |
890 | The @code{putwchar} function is equivalent to @code{putwc} with |
891 | @code{stdout} as the value of the @var{stream} argument. | |
892 | @end deftypefun | |
893 | ||
7b4161bb UD |
894 | @comment stdio.h |
895 | @comment POSIX | |
896 | @deftypefun int putchar_unlocked (int @var{c}) | |
897 | The @code{putchar_unlocked} function is equivalent to the @code{putchar} | |
1dc843f7 | 898 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
899 | @end deftypefun |
900 | ||
b5e73f56 UD |
901 | @comment wchar.h |
902 | @comment GNU | |
903 | @deftypefun wint_t putwchar_unlocked (wchar_t @var{wc}) | |
904 | The @code{putwchar_unlocked} function is equivalent to the @code{putwchar} | |
1dc843f7 | 905 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
906 | |
907 | This function is a GNU extension. | |
908 | @end deftypefun | |
909 | ||
28f540f4 | 910 | @comment stdio.h |
f65fd747 | 911 | @comment ISO |
28f540f4 RM |
912 | @deftypefun int fputs (const char *@var{s}, FILE *@var{stream}) |
913 | The function @code{fputs} writes the string @var{s} to the stream | |
914 | @var{stream}. The terminating null character is not written. | |
915 | This function does @emph{not} add a newline character, either. | |
916 | It outputs only the characters in the string. | |
917 | ||
918 | This function returns @code{EOF} if a write error occurs, and otherwise | |
919 | a non-negative value. | |
920 | ||
921 | For example: | |
922 | ||
923 | @smallexample | |
924 | fputs ("Are ", stdout); | |
925 | fputs ("you ", stdout); | |
926 | fputs ("hungry?\n", stdout); | |
927 | @end smallexample | |
928 | ||
929 | @noindent | |
930 | outputs the text @samp{Are you hungry?} followed by a newline. | |
931 | @end deftypefun | |
932 | ||
b5e73f56 UD |
933 | @comment wchar.h |
934 | @comment ISO | |
935 | @deftypefun int fputws (const wchar_t *@var{ws}, FILE *@var{stream}) | |
936 | The function @code{fputws} writes the wide character string @var{ws} to | |
937 | the stream @var{stream}. The terminating null character is not written. | |
938 | This function does @emph{not} add a newline character, either. It | |
939 | outputs only the characters in the string. | |
940 | ||
941 | This function returns @code{WEOF} if a write error occurs, and otherwise | |
942 | a non-negative value. | |
943 | @end deftypefun | |
944 | ||
7b4161bb UD |
945 | @comment stdio.h |
946 | @comment GNU | |
947 | @deftypefun int fputs_unlocked (const char *@var{s}, FILE *@var{stream}) | |
948 | The @code{fputs_unlocked} function is equivalent to the @code{fputs} | |
1dc843f7 | 949 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
950 | |
951 | This function is a GNU extension. | |
952 | @end deftypefun | |
953 | ||
b5e73f56 UD |
954 | @comment wchar.h |
955 | @comment GNU | |
956 | @deftypefun int fputws_unlocked (const wchar_t *@var{ws}, FILE *@var{stream}) | |
957 | The @code{fputws_unlocked} function is equivalent to the @code{fputws} | |
1dc843f7 | 958 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
959 | |
960 | This function is a GNU extension. | |
961 | @end deftypefun | |
962 | ||
28f540f4 | 963 | @comment stdio.h |
f65fd747 | 964 | @comment ISO |
28f540f4 RM |
965 | @deftypefun int puts (const char *@var{s}) |
966 | The @code{puts} function writes the string @var{s} to the stream | |
967 | @code{stdout} followed by a newline. The terminating null character of | |
968 | the string is not written. (Note that @code{fputs} does @emph{not} | |
969 | write a newline as this function does.) | |
970 | ||
971 | @code{puts} is the most convenient function for printing simple | |
972 | messages. For example: | |
973 | ||
974 | @smallexample | |
975 | puts ("This is a message."); | |
976 | @end smallexample | |
0be8752b UD |
977 | |
978 | @noindent | |
979 | outputs the text @samp{This is a message.} followed by a newline. | |
28f540f4 RM |
980 | @end deftypefun |
981 | ||
982 | @comment stdio.h | |
983 | @comment SVID | |
984 | @deftypefun int putw (int @var{w}, FILE *@var{stream}) | |
985 | This function writes the word @var{w} (that is, an @code{int}) to | |
986 | @var{stream}. It is provided for compatibility with SVID, but we | |
987 | recommend you use @code{fwrite} instead (@pxref{Block Input/Output}). | |
988 | @end deftypefun | |
989 | ||
990 | @node Character Input | |
991 | @section Character Input | |
992 | ||
993 | @cindex reading from a stream, by characters | |
b5e73f56 UD |
994 | This section describes functions for performing character-oriented |
995 | input. These narrow streams functions are declared in the header file | |
996 | @file{stdio.h} and the wide character functions are declared in | |
997 | @file{wchar.h}. | |
28f540f4 | 998 | @pindex stdio.h |
b5e73f56 UD |
999 | @pindex wchar.h |
1000 | ||
1001 | These functions return an @code{int} or @code{wint_t} value (for narrow | |
1002 | and wide stream functions respectively) that is either a character of | |
1003 | input, or the special value @code{EOF}/@code{WEOF} (usually -1). For | |
1004 | the narrow stream functions it is important to store the result of these | |
1005 | functions in a variable of type @code{int} instead of @code{char}, even | |
1006 | when you plan to use it only as a character. Storing @code{EOF} in a | |
1007 | @code{char} variable truncates its value to the size of a character, so | |
1008 | that it is no longer distinguishable from the valid character | |
1009 | @samp{(char) -1}. So always use an @code{int} for the result of | |
1010 | @code{getc} and friends, and check for @code{EOF} after the call; once | |
1011 | you've verified that the result is not @code{EOF}, you can be sure that | |
1012 | it will fit in a @samp{char} variable without loss of information. | |
28f540f4 RM |
1013 | |
1014 | @comment stdio.h | |
f65fd747 | 1015 | @comment ISO |
28f540f4 RM |
1016 | @deftypefun int fgetc (FILE *@var{stream}) |
1017 | This function reads the next character as an @code{unsigned char} from | |
1018 | the stream @var{stream} and returns its value, converted to an | |
1019 | @code{int}. If an end-of-file condition or read error occurs, | |
19c3f208 | 1020 | @code{EOF} is returned instead. |
28f540f4 RM |
1021 | @end deftypefun |
1022 | ||
b5e73f56 UD |
1023 | @comment wchar.h |
1024 | @comment ISO | |
1025 | @deftypefun wint_t fgetwc (FILE *@var{stream}) | |
1026 | This function reads the next wide character from the stream @var{stream} | |
1027 | and returns its value. If an end-of-file condition or read error | |
1028 | occurs, @code{WEOF} is returned instead. | |
1029 | @end deftypefun | |
1030 | ||
7b4161bb UD |
1031 | @comment stdio.h |
1032 | @comment POSIX | |
1033 | @deftypefun int fgetc_unlocked (FILE *@var{stream}) | |
1034 | The @code{fgetc_unlocked} function is equivalent to the @code{fgetc} | |
1dc843f7 | 1035 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1036 | @end deftypefun |
1037 | ||
b5e73f56 UD |
1038 | @comment wchar.h |
1039 | @comment GNU | |
1040 | @deftypefun wint_t fgetwc_unlocked (FILE *@var{stream}) | |
1041 | The @code{fgetwc_unlocked} function is equivalent to the @code{fgetwc} | |
1dc843f7 | 1042 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
1043 | |
1044 | This function is a GNU extension. | |
1045 | @end deftypefun | |
1046 | ||
28f540f4 | 1047 | @comment stdio.h |
f65fd747 | 1048 | @comment ISO |
28f540f4 RM |
1049 | @deftypefun int getc (FILE *@var{stream}) |
1050 | This is just like @code{fgetc}, except that it is permissible (and | |
1051 | typical) for it to be implemented as a macro that evaluates the | |
1052 | @var{stream} argument more than once. @code{getc} is often highly | |
1053 | optimized, so it is usually the best function to use to read a single | |
1054 | character. | |
1055 | @end deftypefun | |
1056 | ||
b5e73f56 UD |
1057 | @comment wchar.h |
1058 | @comment ISO | |
1059 | @deftypefun wint_t getwc (FILE *@var{stream}) | |
1060 | This is just like @code{fgetwc}, except that it is permissible for it to | |
1061 | be implemented as a macro that evaluates the @var{stream} argument more | |
1062 | than once. @code{getwc} can be highly optimized, so it is usually the | |
1063 | best function to use to read a single wide character. | |
1064 | @end deftypefun | |
1065 | ||
7b4161bb UD |
1066 | @comment stdio.h |
1067 | @comment POSIX | |
1068 | @deftypefun int getc_unlocked (FILE *@var{stream}) | |
1069 | The @code{getc_unlocked} function is equivalent to the @code{getc} | |
1dc843f7 | 1070 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1071 | @end deftypefun |
1072 | ||
b5e73f56 UD |
1073 | @comment wchar.h |
1074 | @comment GNU | |
1075 | @deftypefun wint_t getwc_unlocked (FILE *@var{stream}) | |
1076 | The @code{getwc_unlocked} function is equivalent to the @code{getwc} | |
1dc843f7 | 1077 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
1078 | |
1079 | This function is a GNU extension. | |
1080 | @end deftypefun | |
1081 | ||
28f540f4 | 1082 | @comment stdio.h |
f65fd747 | 1083 | @comment ISO |
28f540f4 RM |
1084 | @deftypefun int getchar (void) |
1085 | The @code{getchar} function is equivalent to @code{getc} with @code{stdin} | |
1086 | as the value of the @var{stream} argument. | |
1087 | @end deftypefun | |
1088 | ||
b5e73f56 UD |
1089 | @comment wchar.h |
1090 | @comment ISO | |
1091 | @deftypefun wint_t getwchar (void) | |
1092 | The @code{getwchar} function is equivalent to @code{getwc} with @code{stdin} | |
1093 | as the value of the @var{stream} argument. | |
1094 | @end deftypefun | |
1095 | ||
7b4161bb UD |
1096 | @comment stdio.h |
1097 | @comment POSIX | |
1098 | @deftypefun int getchar_unlocked (void) | |
1099 | The @code{getchar_unlocked} function is equivalent to the @code{getchar} | |
1dc843f7 | 1100 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1101 | @end deftypefun |
1102 | ||
b5e73f56 UD |
1103 | @comment wchar.h |
1104 | @comment GNU | |
1105 | @deftypefun wint_t getwchar_unlocked (void) | |
1106 | The @code{getwchar_unlocked} function is equivalent to the @code{getwchar} | |
1dc843f7 | 1107 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
1108 | |
1109 | This function is a GNU extension. | |
1110 | @end deftypefun | |
1111 | ||
28f540f4 RM |
1112 | Here is an example of a function that does input using @code{fgetc}. It |
1113 | would work just as well using @code{getc} instead, or using | |
b5e73f56 UD |
1114 | @code{getchar ()} instead of @w{@code{fgetc (stdin)}}. The code would |
1115 | also work the same for the wide character stream functions. | |
28f540f4 RM |
1116 | |
1117 | @smallexample | |
1118 | int | |
1119 | y_or_n_p (const char *question) | |
1120 | @{ | |
1121 | fputs (question, stdout); | |
1122 | while (1) | |
1123 | @{ | |
1124 | int c, answer; | |
1125 | /* @r{Write a space to separate answer from question.} */ | |
1126 | fputc (' ', stdout); | |
1127 | /* @r{Read the first character of the line.} | |
162ba701 | 1128 | @r{This should be the answer character, but might not be.} */ |
28f540f4 RM |
1129 | c = tolower (fgetc (stdin)); |
1130 | answer = c; | |
1131 | /* @r{Discard rest of input line.} */ | |
1132 | while (c != '\n' && c != EOF) | |
162ba701 | 1133 | c = fgetc (stdin); |
28f540f4 RM |
1134 | /* @r{Obey the answer if it was valid.} */ |
1135 | if (answer == 'y') | |
162ba701 | 1136 | return 1; |
28f540f4 | 1137 | if (answer == 'n') |
162ba701 | 1138 | return 0; |
28f540f4 RM |
1139 | /* @r{Answer was invalid: ask for valid answer.} */ |
1140 | fputs ("Please answer y or n:", stdout); | |
1141 | @} | |
1142 | @} | |
1143 | @end smallexample | |
1144 | ||
1145 | @comment stdio.h | |
1146 | @comment SVID | |
1147 | @deftypefun int getw (FILE *@var{stream}) | |
1148 | This function reads a word (that is, an @code{int}) from @var{stream}. | |
1149 | It's provided for compatibility with SVID. We recommend you use | |
1150 | @code{fread} instead (@pxref{Block Input/Output}). Unlike @code{getc}, | |
1151 | any @code{int} value could be a valid result. @code{getw} returns | |
1152 | @code{EOF} when it encounters end-of-file or an error, but there is no | |
1153 | way to distinguish this from an input word with value -1. | |
1154 | @end deftypefun | |
1155 | ||
1156 | @node Line Input | |
1157 | @section Line-Oriented Input | |
1158 | ||
b5e73f56 | 1159 | Since many programs interpret input on the basis of lines, it is |
28f540f4 RM |
1160 | convenient to have functions to read a line of text from a stream. |
1161 | ||
1162 | Standard C has functions to do this, but they aren't very safe: null | |
1163 | characters and even (for @code{gets}) long lines can confuse them. So | |
1f77f049 | 1164 | @theglibc{} provides the nonstandard @code{getline} function that |
28f540f4 RM |
1165 | makes it easy to read lines reliably. |
1166 | ||
1167 | Another GNU extension, @code{getdelim}, generalizes @code{getline}. It | |
1168 | reads a delimited record, defined as everything through the next | |
1169 | occurrence of a specified delimiter character. | |
1170 | ||
1171 | All these functions are declared in @file{stdio.h}. | |
1172 | ||
1173 | @comment stdio.h | |
1174 | @comment GNU | |
1175 | @deftypefun ssize_t getline (char **@var{lineptr}, size_t *@var{n}, FILE *@var{stream}) | |
1176 | This function reads an entire line from @var{stream}, storing the text | |
1177 | (including the newline and a terminating null character) in a buffer | |
1178 | and storing the buffer address in @code{*@var{lineptr}}. | |
1179 | ||
1180 | Before calling @code{getline}, you should place in @code{*@var{lineptr}} | |
1181 | the address of a buffer @code{*@var{n}} bytes long, allocated with | |
1182 | @code{malloc}. If this buffer is long enough to hold the line, | |
1183 | @code{getline} stores the line in this buffer. Otherwise, | |
1184 | @code{getline} makes the buffer bigger using @code{realloc}, storing the | |
1185 | new buffer address back in @code{*@var{lineptr}} and the increased size | |
1186 | back in @code{*@var{n}}. | |
1187 | @xref{Unconstrained Allocation}. | |
1188 | ||
1189 | If you set @code{*@var{lineptr}} to a null pointer, and @code{*@var{n}} | |
1190 | to zero, before the call, then @code{getline} allocates the initial | |
1191 | buffer for you by calling @code{malloc}. | |
1192 | ||
1193 | In either case, when @code{getline} returns, @code{*@var{lineptr}} is | |
1194 | a @code{char *} which points to the text of the line. | |
1195 | ||
1196 | When @code{getline} is successful, it returns the number of characters | |
1197 | read (including the newline, but not including the terminating null). | |
1198 | This value enables you to distinguish null characters that are part of | |
1199 | the line from the null character inserted as a terminator. | |
1200 | ||
1201 | This function is a GNU extension, but it is the recommended way to read | |
1202 | lines from a stream. The alternative standard functions are unreliable. | |
1203 | ||
ec28fc7c UD |
1204 | If an error occurs or end of file is reached without any bytes read, |
1205 | @code{getline} returns @code{-1}. | |
28f540f4 RM |
1206 | @end deftypefun |
1207 | ||
1208 | @comment stdio.h | |
1209 | @comment GNU | |
1210 | @deftypefun ssize_t getdelim (char **@var{lineptr}, size_t *@var{n}, int @var{delimiter}, FILE *@var{stream}) | |
1211 | This function is like @code{getline} except that the character which | |
1212 | tells it to stop reading is not necessarily newline. The argument | |
1213 | @var{delimiter} specifies the delimiter character; @code{getdelim} keeps | |
1214 | reading until it sees that character (or end of file). | |
1215 | ||
1216 | The text is stored in @var{lineptr}, including the delimiter character | |
1217 | and a terminating null. Like @code{getline}, @code{getdelim} makes | |
1218 | @var{lineptr} bigger if it isn't big enough. | |
1219 | ||
1220 | @code{getline} is in fact implemented in terms of @code{getdelim}, just | |
1221 | like this: | |
1222 | ||
1223 | @smallexample | |
1224 | ssize_t | |
1225 | getline (char **lineptr, size_t *n, FILE *stream) | |
1226 | @{ | |
1227 | return getdelim (lineptr, n, '\n', stream); | |
1228 | @} | |
1229 | @end smallexample | |
1230 | @end deftypefun | |
1231 | ||
1232 | @comment stdio.h | |
f65fd747 | 1233 | @comment ISO |
28f540f4 RM |
1234 | @deftypefun {char *} fgets (char *@var{s}, int @var{count}, FILE *@var{stream}) |
1235 | The @code{fgets} function reads characters from the stream @var{stream} | |
1236 | up to and including a newline character and stores them in the string | |
1237 | @var{s}, adding a null character to mark the end of the string. You | |
1238 | must supply @var{count} characters worth of space in @var{s}, but the | |
1239 | number of characters read is at most @var{count} @minus{} 1. The extra | |
1240 | character space is used to hold the null character at the end of the | |
1241 | string. | |
1242 | ||
1243 | If the system is already at end of file when you call @code{fgets}, then | |
1244 | the contents of the array @var{s} are unchanged and a null pointer is | |
1245 | returned. A null pointer is also returned if a read error occurs. | |
1246 | Otherwise, the return value is the pointer @var{s}. | |
1247 | ||
1248 | @strong{Warning:} If the input data has a null character, you can't tell. | |
1249 | So don't use @code{fgets} unless you know the data cannot contain a null. | |
1250 | Don't use it to read files edited by the user because, if the user inserts | |
1251 | a null character, you should either handle it properly or print a clear | |
1252 | error message. We recommend using @code{getline} instead of @code{fgets}. | |
1253 | @end deftypefun | |
1254 | ||
b5e73f56 UD |
1255 | @comment wchar.h |
1256 | @comment ISO | |
1257 | @deftypefun {wchar_t *} fgetws (wchar_t *@var{ws}, int @var{count}, FILE *@var{stream}) | |
1258 | The @code{fgetws} function reads wide characters from the stream | |
1259 | @var{stream} up to and including a newline character and stores them in | |
1260 | the string @var{ws}, adding a null wide character to mark the end of the | |
1261 | string. You must supply @var{count} wide characters worth of space in | |
1262 | @var{ws}, but the number of characters read is at most @var{count} | |
1263 | @minus{} 1. The extra character space is used to hold the null wide | |
1264 | character at the end of the string. | |
1265 | ||
1266 | If the system is already at end of file when you call @code{fgetws}, then | |
1267 | the contents of the array @var{ws} are unchanged and a null pointer is | |
1268 | returned. A null pointer is also returned if a read error occurs. | |
1269 | Otherwise, the return value is the pointer @var{ws}. | |
1270 | ||
1271 | @strong{Warning:} If the input data has a null wide character (which are | |
1272 | null bytes in the input stream), you can't tell. So don't use | |
1273 | @code{fgetws} unless you know the data cannot contain a null. Don't use | |
1274 | it to read files edited by the user because, if the user inserts a null | |
1275 | character, you should either handle it properly or print a clear error | |
1276 | message. | |
1277 | @comment XXX We need getwline!!! | |
1278 | @end deftypefun | |
1279 | ||
7b4161bb UD |
1280 | @comment stdio.h |
1281 | @comment GNU | |
1282 | @deftypefun {char *} fgets_unlocked (char *@var{s}, int @var{count}, FILE *@var{stream}) | |
1283 | The @code{fgets_unlocked} function is equivalent to the @code{fgets} | |
1dc843f7 | 1284 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1285 | |
1286 | This function is a GNU extension. | |
1287 | @end deftypefun | |
1288 | ||
b5e73f56 UD |
1289 | @comment wchar.h |
1290 | @comment GNU | |
1291 | @deftypefun {wchar_t *} fgetws_unlocked (wchar_t *@var{ws}, int @var{count}, FILE *@var{stream}) | |
1292 | The @code{fgetws_unlocked} function is equivalent to the @code{fgetws} | |
1dc843f7 | 1293 | function except that it does not implicitly lock the stream. |
b5e73f56 UD |
1294 | |
1295 | This function is a GNU extension. | |
1296 | @end deftypefun | |
1297 | ||
28f540f4 | 1298 | @comment stdio.h |
f65fd747 | 1299 | @comment ISO |
28f540f4 RM |
1300 | @deftypefn {Deprecated function} {char *} gets (char *@var{s}) |
1301 | The function @code{gets} reads characters from the stream @code{stdin} | |
1302 | up to the next newline character, and stores them in the string @var{s}. | |
1303 | The newline character is discarded (note that this differs from the | |
1304 | behavior of @code{fgets}, which copies the newline character into the | |
1305 | string). If @code{gets} encounters a read error or end-of-file, it | |
1306 | returns a null pointer; otherwise it returns @var{s}. | |
1307 | ||
1308 | @strong{Warning:} The @code{gets} function is @strong{very dangerous} | |
1309 | because it provides no protection against overflowing the string | |
1f77f049 | 1310 | @var{s}. @Theglibc{} includes it for compatibility only. You |
28f540f4 RM |
1311 | should @strong{always} use @code{fgets} or @code{getline} instead. To |
1312 | remind you of this, the linker (if using GNU @code{ld}) will issue a | |
1313 | warning whenever you use @code{gets}. | |
1314 | @end deftypefn | |
1315 | ||
1316 | @node Unreading | |
1317 | @section Unreading | |
1318 | @cindex peeking at input | |
1319 | @cindex unreading characters | |
1320 | @cindex pushing input back | |
1321 | ||
1322 | In parser programs it is often useful to examine the next character in | |
1323 | the input stream without removing it from the stream. This is called | |
1324 | ``peeking ahead'' at the input because your program gets a glimpse of | |
1325 | the input it will read next. | |
1326 | ||
1327 | Using stream I/O, you can peek ahead at input by first reading it and | |
19c3f208 | 1328 | then @dfn{unreading} it (also called @dfn{pushing it back} on the stream). |
28f540f4 RM |
1329 | Unreading a character makes it available to be input again from the stream, |
1330 | by the next call to @code{fgetc} or other input function on that stream. | |
1331 | ||
1332 | @menu | |
1333 | * Unreading Idea:: An explanation of unreading with pictures. | |
1334 | * How Unread:: How to call @code{ungetc} to do unreading. | |
1335 | @end menu | |
1336 | ||
1337 | @node Unreading Idea | |
1338 | @subsection What Unreading Means | |
1339 | ||
1340 | Here is a pictorial explanation of unreading. Suppose you have a | |
1341 | stream reading a file that contains just six characters, the letters | |
1342 | @samp{foobar}. Suppose you have read three characters so far. The | |
1343 | situation looks like this: | |
1344 | ||
1345 | @smallexample | |
1346 | f o o b a r | |
162ba701 | 1347 | ^ |
28f540f4 RM |
1348 | @end smallexample |
1349 | ||
1350 | @noindent | |
1351 | so the next input character will be @samp{b}. | |
1352 | ||
1353 | @c @group Invalid outside @example | |
1354 | If instead of reading @samp{b} you unread the letter @samp{o}, you get a | |
1355 | situation like this: | |
1356 | ||
1357 | @smallexample | |
1358 | f o o b a r | |
162ba701 | 1359 | | |
28f540f4 RM |
1360 | o-- |
1361 | ^ | |
1362 | @end smallexample | |
1363 | ||
1364 | @noindent | |
1365 | so that the next input characters will be @samp{o} and @samp{b}. | |
1366 | @c @end group | |
1367 | ||
1368 | @c @group | |
1369 | If you unread @samp{9} instead of @samp{o}, you get this situation: | |
1370 | ||
1371 | @smallexample | |
1372 | f o o b a r | |
162ba701 | 1373 | | |
28f540f4 RM |
1374 | 9-- |
1375 | ^ | |
1376 | @end smallexample | |
1377 | ||
1378 | @noindent | |
1379 | so that the next input characters will be @samp{9} and @samp{b}. | |
1380 | @c @end group | |
1381 | ||
1382 | @node How Unread | |
1383 | @subsection Using @code{ungetc} To Do Unreading | |
19c3f208 | 1384 | |
28f540f4 RM |
1385 | The function to unread a character is called @code{ungetc}, because it |
1386 | reverses the action of @code{getc}. | |
1387 | ||
1388 | @comment stdio.h | |
f65fd747 | 1389 | @comment ISO |
28f540f4 RM |
1390 | @deftypefun int ungetc (int @var{c}, FILE *@var{stream}) |
1391 | The @code{ungetc} function pushes back the character @var{c} onto the | |
1392 | input stream @var{stream}. So the next input from @var{stream} will | |
1393 | read @var{c} before anything else. | |
1394 | ||
1395 | If @var{c} is @code{EOF}, @code{ungetc} does nothing and just returns | |
1396 | @code{EOF}. This lets you call @code{ungetc} with the return value of | |
1397 | @code{getc} without needing to check for an error from @code{getc}. | |
1398 | ||
1399 | The character that you push back doesn't have to be the same as the last | |
1400 | character that was actually read from the stream. In fact, it isn't | |
1401 | necessary to actually read any characters from the stream before | |
f4efd068 UD |
1402 | unreading them with @code{ungetc}! But that is a strange way to write a |
1403 | program; usually @code{ungetc} is used only to unread a character that | |
1f77f049 | 1404 | was just read from the same stream. @Theglibc{} supports this |
f4efd068 | 1405 | even on files opened in binary mode, but other systems might not. |
28f540f4 | 1406 | |
1f77f049 | 1407 | @Theglibc{} only supports one character of pushback---in other |
28f540f4 RM |
1408 | words, it does not work to call @code{ungetc} twice without doing input |
1409 | in between. Other systems might let you push back multiple characters; | |
1410 | then reading from the stream retrieves the characters in the reverse | |
1411 | order that they were pushed. | |
1412 | ||
1413 | Pushing back characters doesn't alter the file; only the internal | |
1414 | buffering for the stream is affected. If a file positioning function | |
a5a0310d UD |
1415 | (such as @code{fseek}, @code{fseeko} or @code{rewind}; @pxref{File |
1416 | Positioning}) is called, any pending pushed-back characters are | |
1417 | discarded. | |
28f540f4 RM |
1418 | |
1419 | Unreading a character on a stream that is at end of file clears the | |
1420 | end-of-file indicator for the stream, because it makes the character of | |
1421 | input available. After you read that character, trying to read again | |
1422 | will encounter end of file. | |
1423 | @end deftypefun | |
1424 | ||
b5e73f56 UD |
1425 | @comment wchar.h |
1426 | @comment ISO | |
1427 | @deftypefun wint_t ungetwc (wint_t @var{wc}, FILE *@var{stream}) | |
1428 | The @code{ungetwc} function behaves just like @code{ungetc} just that it | |
1429 | pushes back a wide character. | |
1430 | @end deftypefun | |
1431 | ||
28f540f4 RM |
1432 | Here is an example showing the use of @code{getc} and @code{ungetc} to |
1433 | skip over whitespace characters. When this function reaches a | |
1434 | non-whitespace character, it unreads that character to be seen again on | |
1435 | the next read operation on the stream. | |
1436 | ||
1437 | @smallexample | |
1438 | #include <stdio.h> | |
1439 | #include <ctype.h> | |
1440 | ||
1441 | void | |
1442 | skip_whitespace (FILE *stream) | |
1443 | @{ | |
1444 | int c; | |
1445 | do | |
1446 | /* @r{No need to check for @code{EOF} because it is not} | |
1447 | @r{@code{isspace}, and @code{ungetc} ignores @code{EOF}.} */ | |
1448 | c = getc (stream); | |
1449 | while (isspace (c)); | |
1450 | ungetc (c, stream); | |
1451 | @} | |
1452 | @end smallexample | |
1453 | ||
1454 | @node Block Input/Output | |
1455 | @section Block Input/Output | |
1456 | ||
1457 | This section describes how to do input and output operations on blocks | |
1458 | of data. You can use these functions to read and write binary data, as | |
1459 | well as to read and write text in fixed-size blocks instead of by | |
1460 | characters or lines. | |
1461 | @cindex binary I/O to a stream | |
1462 | @cindex block I/O to a stream | |
1463 | @cindex reading from a stream, by blocks | |
1464 | @cindex writing to a stream, by blocks | |
1465 | ||
1466 | Binary files are typically used to read and write blocks of data in the | |
1467 | same format as is used to represent the data in a running program. In | |
1468 | other words, arbitrary blocks of memory---not just character or string | |
1469 | objects---can be written to a binary file, and meaningfully read in | |
1470 | again by the same program. | |
1471 | ||
1472 | Storing data in binary form is often considerably more efficient than | |
1473 | using the formatted I/O functions. Also, for floating-point numbers, | |
1474 | the binary form avoids possible loss of precision in the conversion | |
1475 | process. On the other hand, binary files can't be examined or modified | |
1476 | easily using many standard file utilities (such as text editors), and | |
1477 | are not portable between different implementations of the language, or | |
1478 | different kinds of computers. | |
1479 | ||
1480 | These functions are declared in @file{stdio.h}. | |
1481 | @pindex stdio.h | |
1482 | ||
1483 | @comment stdio.h | |
f65fd747 | 1484 | @comment ISO |
28f540f4 RM |
1485 | @deftypefun size_t fread (void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) |
1486 | This function reads up to @var{count} objects of size @var{size} into | |
1487 | the array @var{data}, from the stream @var{stream}. It returns the | |
1488 | number of objects actually read, which might be less than @var{count} if | |
1489 | a read error occurs or the end of the file is reached. This function | |
1490 | returns a value of zero (and doesn't read anything) if either @var{size} | |
1491 | or @var{count} is zero. | |
1492 | ||
1493 | If @code{fread} encounters end of file in the middle of an object, it | |
1494 | returns the number of complete objects read, and discards the partial | |
1495 | object. Therefore, the stream remains at the actual end of the file. | |
1496 | @end deftypefun | |
1497 | ||
7b4161bb UD |
1498 | @comment stdio.h |
1499 | @comment GNU | |
1500 | @deftypefun size_t fread_unlocked (void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) | |
1501 | The @code{fread_unlocked} function is equivalent to the @code{fread} | |
1dc843f7 | 1502 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1503 | |
1504 | This function is a GNU extension. | |
1505 | @end deftypefun | |
1506 | ||
28f540f4 | 1507 | @comment stdio.h |
f65fd747 | 1508 | @comment ISO |
28f540f4 RM |
1509 | @deftypefun size_t fwrite (const void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) |
1510 | This function writes up to @var{count} objects of size @var{size} from | |
1511 | the array @var{data}, to the stream @var{stream}. The return value is | |
1512 | normally @var{count}, if the call succeeds. Any other value indicates | |
1513 | some sort of error, such as running out of space. | |
1514 | @end deftypefun | |
1515 | ||
7b4161bb UD |
1516 | @comment stdio.h |
1517 | @comment GNU | |
1518 | @deftypefun size_t fwrite_unlocked (const void *@var{data}, size_t @var{size}, size_t @var{count}, FILE *@var{stream}) | |
1519 | The @code{fwrite_unlocked} function is equivalent to the @code{fwrite} | |
1dc843f7 | 1520 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
1521 | |
1522 | This function is a GNU extension. | |
1523 | @end deftypefun | |
1524 | ||
28f540f4 RM |
1525 | @node Formatted Output |
1526 | @section Formatted Output | |
1527 | ||
1528 | @cindex format string, for @code{printf} | |
1529 | @cindex template, for @code{printf} | |
1530 | @cindex formatted output to a stream | |
1531 | @cindex writing to a stream, formatted | |
1532 | The functions described in this section (@code{printf} and related | |
1533 | functions) provide a convenient way to perform formatted output. You | |
1534 | call @code{printf} with a @dfn{format string} or @dfn{template string} | |
1535 | that specifies how to format the values of the remaining arguments. | |
1536 | ||
1537 | Unless your program is a filter that specifically performs line- or | |
1538 | character-oriented processing, using @code{printf} or one of the other | |
1539 | related functions described in this section is usually the easiest and | |
1540 | most concise way to perform output. These functions are especially | |
1541 | useful for printing error messages, tables of data, and the like. | |
1542 | ||
1543 | @menu | |
1544 | * Formatted Output Basics:: Some examples to get you started. | |
1545 | * Output Conversion Syntax:: General syntax of conversion | |
162ba701 | 1546 | specifications. |
28f540f4 | 1547 | * Table of Output Conversions:: Summary of output conversions and |
162ba701 | 1548 | what they do. |
28f540f4 RM |
1549 | * Integer Conversions:: Details about formatting of integers. |
1550 | * Floating-Point Conversions:: Details about formatting of | |
162ba701 | 1551 | floating-point numbers. |
28f540f4 | 1552 | * Other Output Conversions:: Details about formatting of strings, |
162ba701 | 1553 | characters, pointers, and the like. |
28f540f4 RM |
1554 | * Formatted Output Functions:: Descriptions of the actual functions. |
1555 | * Dynamic Output:: Functions that allocate memory for the output. | |
1556 | * Variable Arguments Output:: @code{vprintf} and friends. | |
1557 | * Parsing a Template String:: What kinds of args does a given template | |
162ba701 | 1558 | call for? |
28f540f4 RM |
1559 | * Example of Parsing:: Sample program using @code{parse_printf_format}. |
1560 | @end menu | |
1561 | ||
1562 | @node Formatted Output Basics | |
1563 | @subsection Formatted Output Basics | |
1564 | ||
1565 | The @code{printf} function can be used to print any number of arguments. | |
1566 | The template string argument you supply in a call provides | |
1567 | information not only about the number of additional arguments, but also | |
1568 | about their types and what style should be used for printing them. | |
1569 | ||
1570 | Ordinary characters in the template string are simply written to the | |
1571 | output stream as-is, while @dfn{conversion specifications} introduced by | |
1572 | a @samp{%} character in the template cause subsequent arguments to be | |
1573 | formatted and written to the output stream. For example, | |
1574 | @cindex conversion specifications (@code{printf}) | |
1575 | ||
1576 | @smallexample | |
1577 | int pct = 37; | |
1578 | char filename[] = "foo.txt"; | |
1579 | printf ("Processing of `%s' is %d%% finished.\nPlease be patient.\n", | |
162ba701 | 1580 | filename, pct); |
28f540f4 RM |
1581 | @end smallexample |
1582 | ||
1583 | @noindent | |
1584 | produces output like | |
1585 | ||
1586 | @smallexample | |
1587 | Processing of `foo.txt' is 37% finished. | |
1588 | Please be patient. | |
1589 | @end smallexample | |
1590 | ||
1591 | This example shows the use of the @samp{%d} conversion to specify that | |
1592 | an @code{int} argument should be printed in decimal notation, the | |
1593 | @samp{%s} conversion to specify printing of a string argument, and | |
1594 | the @samp{%%} conversion to print a literal @samp{%} character. | |
1595 | ||
1596 | There are also conversions for printing an integer argument as an | |
1597 | unsigned value in octal, decimal, or hexadecimal radix (@samp{%o}, | |
1598 | @samp{%u}, or @samp{%x}, respectively); or as a character value | |
1599 | (@samp{%c}). | |
1600 | ||
1601 | Floating-point numbers can be printed in normal, fixed-point notation | |
1602 | using the @samp{%f} conversion or in exponential notation using the | |
1603 | @samp{%e} conversion. The @samp{%g} conversion uses either @samp{%e} | |
1604 | or @samp{%f} format, depending on what is more appropriate for the | |
1605 | magnitude of the particular number. | |
1606 | ||
1607 | You can control formatting more precisely by writing @dfn{modifiers} | |
1608 | between the @samp{%} and the character that indicates which conversion | |
1609 | to apply. These slightly alter the ordinary behavior of the conversion. | |
1610 | For example, most conversion specifications permit you to specify a | |
1611 | minimum field width and a flag indicating whether you want the result | |
1612 | left- or right-justified within the field. | |
1613 | ||
1614 | The specific flags and modifiers that are permitted and their | |
1615 | interpretation vary depending on the particular conversion. They're all | |
1616 | described in more detail in the following sections. Don't worry if this | |
1617 | all seems excessively complicated at first; you can almost always get | |
1618 | reasonable free-format output without using any of the modifiers at all. | |
1619 | The modifiers are mostly used to make the output look ``prettier'' in | |
1620 | tables. | |
1621 | ||
1622 | @node Output Conversion Syntax | |
1623 | @subsection Output Conversion Syntax | |
1624 | ||
1625 | This section provides details about the precise syntax of conversion | |
1626 | specifications that can appear in a @code{printf} template | |
1627 | string. | |
1628 | ||
390955cb UD |
1629 | Characters in the template string that are not part of a conversion |
1630 | specification are printed as-is to the output stream. Multibyte | |
1631 | character sequences (@pxref{Character Set Handling}) are permitted in a | |
1632 | template string. | |
28f540f4 RM |
1633 | |
1634 | The conversion specifications in a @code{printf} template string have | |
1635 | the general form: | |
1636 | ||
f89829da | 1637 | @smallexample |
dfd2257a | 1638 | % @r{[} @var{param-no} @r{$]} @var{flags} @var{width} @r{[} . @var{precision} @r{]} @var{type} @var{conversion} |
f89829da UD |
1639 | @end smallexample |
1640 | ||
1641 | @noindent | |
1642 | or | |
1643 | ||
1644 | @smallexample | |
1645 | % @r{[} @var{param-no} @r{$]} @var{flags} @var{width} . @r{*} @r{[} @var{param-no} @r{$]} @var{type} @var{conversion} | |
1646 | @end smallexample | |
28f540f4 RM |
1647 | |
1648 | For example, in the conversion specifier @samp{%-10.8ld}, the @samp{-} | |
1649 | is a flag, @samp{10} specifies the field width, the precision is | |
1650 | @samp{8}, the letter @samp{l} is a type modifier, and @samp{d} specifies | |
1651 | the conversion style. (This particular type specifier says to | |
1652 | print a @code{long int} argument in decimal notation, with a minimum of | |
1653 | 8 digits left-justified in a field at least 10 characters wide.) | |
1654 | ||
1655 | In more detail, output conversion specifications consist of an | |
1656 | initial @samp{%} character followed in sequence by: | |
1657 | ||
1658 | @itemize @bullet | |
dfd2257a UD |
1659 | @item |
1660 | An optional specification of the parameter used for this format. | |
e8b1163e | 1661 | Normally the parameters to the @code{printf} function are assigned to the |
f2ea0f5b | 1662 | formats in the order of appearance in the format string. But in some |
dfd2257a | 1663 | situations (such as message translation) this is not desirable and this |
e8b1163e | 1664 | extension allows an explicit parameter to be specified. |
dfd2257a | 1665 | |
f89829da | 1666 | The @var{param-no} parts of the format must be integers in the range of |
dfd2257a UD |
1667 | 1 to the maximum number of arguments present to the function call. Some |
1668 | implementations limit this number to a certainly upper bound. The exact | |
1669 | limit can be retrieved by the following constant. | |
1670 | ||
1671 | @defvr Macro NL_ARGMAX | |
f89829da | 1672 | The value of @code{NL_ARGMAX} is the maximum value allowed for the |
dfd2257a UD |
1673 | specification of an positional parameter in a @code{printf} call. The |
1674 | actual value in effect at runtime can be retrieved by using | |
1675 | @code{sysconf} using the @code{_SC_NL_ARGMAX} parameter @pxref{Sysconf | |
1676 | Definition}. | |
1677 | ||
1678 | Some system have a quite low limit such as @math{9} for @w{System V} | |
1f77f049 | 1679 | systems. @Theglibc{} has no real limit. |
dfd2257a UD |
1680 | @end defvr |
1681 | ||
1682 | If any of the formats has a specification for the parameter position all | |
0bc93a2f | 1683 | of them in the format string shall have one. Otherwise the behavior is |
dfd2257a UD |
1684 | undefined. |
1685 | ||
19c3f208 | 1686 | @item |
28f540f4 RM |
1687 | Zero or more @dfn{flag characters} that modify the normal behavior of |
1688 | the conversion specification. | |
1689 | @cindex flag character (@code{printf}) | |
1690 | ||
19c3f208 | 1691 | @item |
28f540f4 RM |
1692 | An optional decimal integer specifying the @dfn{minimum field width}. |
1693 | If the normal conversion produces fewer characters than this, the field | |
1694 | is padded with spaces to the specified width. This is a @emph{minimum} | |
1695 | value; if the normal conversion produces more characters than this, the | |
1696 | field is @emph{not} truncated. Normally, the output is right-justified | |
1697 | within the field. | |
1698 | @cindex minimum field width (@code{printf}) | |
1699 | ||
1700 | You can also specify a field width of @samp{*}. This means that the | |
1701 | next argument in the argument list (before the actual value to be | |
1702 | printed) is used as the field width. The value must be an @code{int}. | |
1703 | If the value is negative, this means to set the @samp{-} flag (see | |
1704 | below) and to use the absolute value as the field width. | |
1705 | ||
19c3f208 | 1706 | @item |
28f540f4 RM |
1707 | An optional @dfn{precision} to specify the number of digits to be |
1708 | written for the numeric conversions. If the precision is specified, it | |
1709 | consists of a period (@samp{.}) followed optionally by a decimal integer | |
1710 | (which defaults to zero if omitted). | |
1711 | @cindex precision (@code{printf}) | |
1712 | ||
1713 | You can also specify a precision of @samp{*}. This means that the next | |
1714 | argument in the argument list (before the actual value to be printed) is | |
1715 | used as the precision. The value must be an @code{int}, and is ignored | |
1716 | if it is negative. If you specify @samp{*} for both the field width and | |
1717 | precision, the field width argument precedes the precision argument. | |
1718 | Other C library versions may not recognize this syntax. | |
1719 | ||
1720 | @item | |
1721 | An optional @dfn{type modifier character}, which is used to specify the | |
1722 | data type of the corresponding argument if it differs from the default | |
1723 | type. (For example, the integer conversions assume a type of @code{int}, | |
1724 | but you can specify @samp{h}, @samp{l}, or @samp{L} for other integer | |
1725 | types.) | |
1726 | @cindex type modifier character (@code{printf}) | |
1727 | ||
1728 | @item | |
1729 | A character that specifies the conversion to be applied. | |
1730 | @end itemize | |
1731 | ||
19c3f208 | 1732 | The exact options that are permitted and how they are interpreted vary |
28f540f4 RM |
1733 | between the different conversion specifiers. See the descriptions of the |
1734 | individual conversions for information about the particular options that | |
1735 | they use. | |
1736 | ||
1737 | With the @samp{-Wformat} option, the GNU C compiler checks calls to | |
1738 | @code{printf} and related functions. It examines the format string and | |
1739 | verifies that the correct number and types of arguments are supplied. | |
1740 | There is also a GNU C syntax to tell the compiler that a function you | |
19c3f208 | 1741 | write uses a @code{printf}-style format string. |
28f540f4 RM |
1742 | @xref{Function Attributes, , Declaring Attributes of Functions, |
1743 | gcc.info, Using GNU CC}, for more information. | |
1744 | ||
1745 | @node Table of Output Conversions | |
1746 | @subsection Table of Output Conversions | |
1747 | @cindex output conversions, for @code{printf} | |
1748 | ||
1749 | Here is a table summarizing what all the different conversions do: | |
1750 | ||
1751 | @table @asis | |
1752 | @item @samp{%d}, @samp{%i} | |
1753 | Print an integer as a signed decimal number. @xref{Integer | |
1754 | Conversions}, for details. @samp{%d} and @samp{%i} are synonymous for | |
1755 | output, but are different when used with @code{scanf} for input | |
1756 | (@pxref{Table of Input Conversions}). | |
1757 | ||
1758 | @item @samp{%o} | |
1759 | Print an integer as an unsigned octal number. @xref{Integer | |
1760 | Conversions}, for details. | |
1761 | ||
1762 | @item @samp{%u} | |
1763 | Print an integer as an unsigned decimal number. @xref{Integer | |
1764 | Conversions}, for details. | |
1765 | ||
1766 | @item @samp{%x}, @samp{%X} | |
1767 | Print an integer as an unsigned hexadecimal number. @samp{%x} uses | |
1768 | lower-case letters and @samp{%X} uses upper-case. @xref{Integer | |
1769 | Conversions}, for details. | |
1770 | ||
1771 | @item @samp{%f} | |
1772 | Print a floating-point number in normal (fixed-point) notation. | |
1773 | @xref{Floating-Point Conversions}, for details. | |
1774 | ||
1775 | @item @samp{%e}, @samp{%E} | |
1776 | Print a floating-point number in exponential notation. @samp{%e} uses | |
1777 | lower-case letters and @samp{%E} uses upper-case. @xref{Floating-Point | |
1778 | Conversions}, for details. | |
1779 | ||
1780 | @item @samp{%g}, @samp{%G} | |
1781 | Print a floating-point number in either normal or exponential notation, | |
1782 | whichever is more appropriate for its magnitude. @samp{%g} uses | |
1783 | lower-case letters and @samp{%G} uses upper-case. @xref{Floating-Point | |
1784 | Conversions}, for details. | |
1785 | ||
2f6d1f1b UD |
1786 | @item @samp{%a}, @samp{%A} |
1787 | Print a floating-point number in a hexadecimal fractional notation which | |
1788 | the exponent to base 2 represented in decimal digits. @samp{%a} uses | |
1789 | lower-case letters and @samp{%A} uses upper-case. @xref{Floating-Point | |
1790 | Conversions}, for details. | |
1791 | ||
28f540f4 RM |
1792 | @item @samp{%c} |
1793 | Print a single character. @xref{Other Output Conversions}. | |
1794 | ||
b5e73f56 UD |
1795 | @item @samp{%C} |
1796 | This is an alias for @samp{%lc} which is supported for compatibility | |
1797 | with the Unix standard. | |
1798 | ||
28f540f4 RM |
1799 | @item @samp{%s} |
1800 | Print a string. @xref{Other Output Conversions}. | |
1801 | ||
b5e73f56 UD |
1802 | @item @samp{%S} |
1803 | This is an alias for @samp{%ls} which is supported for compatibility | |
1804 | with the Unix standard. | |
1805 | ||
28f540f4 RM |
1806 | @item @samp{%p} |
1807 | Print the value of a pointer. @xref{Other Output Conversions}. | |
1808 | ||
1809 | @item @samp{%n} | |
1810 | Get the number of characters printed so far. @xref{Other Output Conversions}. | |
1811 | Note that this conversion specification never produces any output. | |
1812 | ||
1813 | @item @samp{%m} | |
1814 | Print the string corresponding to the value of @code{errno}. | |
1815 | (This is a GNU extension.) | |
1816 | @xref{Other Output Conversions}. | |
1817 | ||
1818 | @item @samp{%%} | |
1819 | Print a literal @samp{%} character. @xref{Other Output Conversions}. | |
1820 | @end table | |
1821 | ||
1822 | If the syntax of a conversion specification is invalid, unpredictable | |
1823 | things will happen, so don't do this. If there aren't enough function | |
1824 | arguments provided to supply values for all the conversion | |
1825 | specifications in the template string, or if the arguments are not of | |
1826 | the correct types, the results are unpredictable. If you supply more | |
1827 | arguments than conversion specifications, the extra argument values are | |
1828 | simply ignored; this is sometimes useful. | |
1829 | ||
1830 | @node Integer Conversions | |
1831 | @subsection Integer Conversions | |
1832 | ||
1833 | This section describes the options for the @samp{%d}, @samp{%i}, | |
1834 | @samp{%o}, @samp{%u}, @samp{%x}, and @samp{%X} conversion | |
1835 | specifications. These conversions print integers in various formats. | |
1836 | ||
1837 | The @samp{%d} and @samp{%i} conversion specifications both print an | |
1838 | @code{int} argument as a signed decimal number; while @samp{%o}, | |
1839 | @samp{%u}, and @samp{%x} print the argument as an unsigned octal, | |
1840 | decimal, or hexadecimal number (respectively). The @samp{%X} conversion | |
1841 | specification is just like @samp{%x} except that it uses the characters | |
1842 | @samp{ABCDEF} as digits instead of @samp{abcdef}. | |
1843 | ||
1844 | The following flags are meaningful: | |
1845 | ||
1846 | @table @asis | |
1847 | @item @samp{-} | |
1848 | Left-justify the result in the field (instead of the normal | |
1849 | right-justification). | |
1850 | ||
1851 | @item @samp{+} | |
1852 | For the signed @samp{%d} and @samp{%i} conversions, print a | |
1853 | plus sign if the value is positive. | |
1854 | ||
1855 | @item @samp{ } | |
1856 | For the signed @samp{%d} and @samp{%i} conversions, if the result | |
1857 | doesn't start with a plus or minus sign, prefix it with a space | |
1858 | character instead. Since the @samp{+} flag ensures that the result | |
1859 | includes a sign, this flag is ignored if you supply both of them. | |
1860 | ||
1861 | @item @samp{#} | |
1862 | For the @samp{%o} conversion, this forces the leading digit to be | |
1863 | @samp{0}, as if by increasing the precision. For @samp{%x} or | |
1864 | @samp{%X}, this prefixes a leading @samp{0x} or @samp{0X} (respectively) | |
1865 | to the result. This doesn't do anything useful for the @samp{%d}, | |
1866 | @samp{%i}, or @samp{%u} conversions. Using this flag produces output | |
1867 | which can be parsed by the @code{strtoul} function (@pxref{Parsing of | |
1868 | Integers}) and @code{scanf} with the @samp{%i} conversion | |
1869 | (@pxref{Numeric Input Conversions}). | |
1870 | ||
1871 | @item @samp{'} | |
1872 | Separate the digits into groups as specified by the locale specified for | |
1873 | the @code{LC_NUMERIC} category; @pxref{General Numeric}. This flag is a | |
1874 | GNU extension. | |
1875 | ||
1876 | @item @samp{0} | |
1877 | Pad the field with zeros instead of spaces. The zeros are placed after | |
1878 | any indication of sign or base. This flag is ignored if the @samp{-} | |
1879 | flag is also specified, or if a precision is specified. | |
1880 | @end table | |
1881 | ||
1882 | If a precision is supplied, it specifies the minimum number of digits to | |
1883 | appear; leading zeros are produced if necessary. If you don't specify a | |
1884 | precision, the number is printed with as many digits as it needs. If | |
1885 | you convert a value of zero with an explicit precision of zero, then no | |
1886 | characters at all are produced. | |
1887 | ||
1888 | Without a type modifier, the corresponding argument is treated as an | |
1889 | @code{int} (for the signed conversions @samp{%i} and @samp{%d}) or | |
1890 | @code{unsigned int} (for the unsigned conversions @samp{%o}, @samp{%u}, | |
1891 | @samp{%x}, and @samp{%X}). Recall that since @code{printf} and friends | |
1892 | are variadic, any @code{char} and @code{short} arguments are | |
1893 | automatically converted to @code{int} by the default argument | |
1894 | promotions. For arguments of other integer types, you can use these | |
1895 | modifiers: | |
1896 | ||
1897 | @table @samp | |
cc3fa755 UD |
1898 | @item hh |
1899 | Specifies that the argument is a @code{signed char} or @code{unsigned | |
1900 | char}, as appropriate. A @code{char} argument is converted to an | |
1901 | @code{int} or @code{unsigned int} by the default argument promotions | |
1902 | anyway, but the @samp{h} modifier says to convert it back to a | |
1903 | @code{char} again. | |
1904 | ||
ec751a23 | 1905 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 1906 | |
28f540f4 RM |
1907 | @item h |
1908 | Specifies that the argument is a @code{short int} or @code{unsigned | |
1909 | short int}, as appropriate. A @code{short} argument is converted to an | |
1910 | @code{int} or @code{unsigned int} by the default argument promotions | |
1911 | anyway, but the @samp{h} modifier says to convert it back to a | |
1912 | @code{short} again. | |
1913 | ||
e852e889 UD |
1914 | @item j |
1915 | Specifies that the argument is a @code{intmax_t} or @code{uintmax_t}, as | |
1916 | appropriate. | |
1917 | ||
ec751a23 | 1918 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 1919 | |
28f540f4 RM |
1920 | @item l |
1921 | Specifies that the argument is a @code{long int} or @code{unsigned long | |
1922 | int}, as appropriate. Two @samp{l} characters is like the @samp{L} | |
1923 | modifier, below. | |
1924 | ||
b5e73f56 UD |
1925 | If used with @samp{%c} or @samp{%s} the corresponding parameter is |
1926 | considered as a wide character or wide character string respectively. | |
1927 | This use of @samp{l} was introduced in @w{Amendment 1} to @w{ISO C90}. | |
1928 | ||
28f540f4 RM |
1929 | @item L |
1930 | @itemx ll | |
1931 | @itemx q | |
1932 | Specifies that the argument is a @code{long long int}. (This type is | |
1933 | an extension supported by the GNU C compiler. On systems that don't | |
1934 | support extra-long integers, this is the same as @code{long int}.) | |
1935 | ||
1936 | The @samp{q} modifier is another name for the same thing, which comes | |
1937 | from 4.4 BSD; a @w{@code{long long int}} is sometimes called a ``quad'' | |
1938 | @code{int}. | |
1939 | ||
e852e889 UD |
1940 | @item t |
1941 | Specifies that the argument is a @code{ptrdiff_t}. | |
1942 | ||
ec751a23 | 1943 | This modifier was introduced in @w{ISO C99}. |
e852e889 UD |
1944 | |
1945 | @item z | |
1946 | @itemx Z | |
1947 | Specifies that the argument is a @code{size_t}. | |
1948 | ||
ec751a23 | 1949 | @samp{z} was introduced in @w{ISO C99}. @samp{Z} is a GNU extension |
0be8752b | 1950 | predating this addition and should not be used in new code. |
28f540f4 RM |
1951 | @end table |
1952 | ||
1953 | Here is an example. Using the template string: | |
1954 | ||
1955 | @smallexample | |
1956 | "|%5d|%-5d|%+5d|%+-5d|% 5d|%05d|%5.0d|%5.2d|%d|\n" | |
1957 | @end smallexample | |
1958 | ||
1959 | @noindent | |
1960 | to print numbers using the different options for the @samp{%d} | |
1961 | conversion gives results like: | |
1962 | ||
1963 | @smallexample | |
1964 | | 0|0 | +0|+0 | 0|00000| | 00|0| | |
1965 | | 1|1 | +1|+1 | 1|00001| 1| 01|1| | |
1966 | | -1|-1 | -1|-1 | -1|-0001| -1| -01|-1| | |
776e8492 | 1967 | |100000|100000|+100000|+100000| 100000|100000|100000|100000|100000| |
28f540f4 RM |
1968 | @end smallexample |
1969 | ||
1970 | In particular, notice what happens in the last case where the number | |
1971 | is too large to fit in the minimum field width specified. | |
1972 | ||
1973 | Here are some more examples showing how unsigned integers print under | |
1974 | various format options, using the template string: | |
1975 | ||
1976 | @smallexample | |
1977 | "|%5u|%5o|%5x|%5X|%#5o|%#5x|%#5X|%#10.8x|\n" | |
1978 | @end smallexample | |
1979 | ||
1980 | @smallexample | |
776e8492 | 1981 | | 0| 0| 0| 0| 0| 0| 0| 00000000| |
28f540f4 RM |
1982 | | 1| 1| 1| 1| 01| 0x1| 0X1|0x00000001| |
1983 | |100000|303240|186a0|186A0|0303240|0x186a0|0X186A0|0x000186a0| | |
1984 | @end smallexample | |
1985 | ||
1986 | ||
1987 | @node Floating-Point Conversions | |
1988 | @subsection Floating-Point Conversions | |
1989 | ||
1990 | This section discusses the conversion specifications for floating-point | |
1991 | numbers: the @samp{%f}, @samp{%e}, @samp{%E}, @samp{%g}, and @samp{%G} | |
1992 | conversions. | |
1993 | ||
1994 | The @samp{%f} conversion prints its argument in fixed-point notation, | |
1995 | producing output of the form | |
1996 | @w{[@code{-}]@var{ddd}@code{.}@var{ddd}}, | |
1997 | where the number of digits following the decimal point is controlled | |
1998 | by the precision you specify. | |
1999 | ||
2000 | The @samp{%e} conversion prints its argument in exponential notation, | |
2001 | producing output of the form | |
2002 | @w{[@code{-}]@var{d}@code{.}@var{ddd}@code{e}[@code{+}|@code{-}]@var{dd}}. | |
2003 | Again, the number of digits following the decimal point is controlled by | |
2004 | the precision. The exponent always contains at least two digits. The | |
2005 | @samp{%E} conversion is similar but the exponent is marked with the letter | |
2006 | @samp{E} instead of @samp{e}. | |
2007 | ||
2008 | The @samp{%g} and @samp{%G} conversions print the argument in the style | |
2009 | of @samp{%e} or @samp{%E} (respectively) if the exponent would be less | |
776e8492 | 2010 | than -4 or greater than or equal to the precision; otherwise they use |
11bf311e | 2011 | the @samp{%f} style. A precision of @code{0}, is taken as 1. |
776e8492 AJ |
2012 | Trailing zeros are removed from the fractional portion of the result and |
2013 | a decimal-point character appears only if it is followed by a digit. | |
28f540f4 | 2014 | |
2f6d1f1b | 2015 | The @samp{%a} and @samp{%A} conversions are meant for representing |
0be8752b | 2016 | floating-point numbers exactly in textual form so that they can be |
2f6d1f1b UD |
2017 | exchanged as texts between different programs and/or machines. The |
2018 | numbers are represented is the form | |
2019 | @w{[@code{-}]@code{0x}@var{h}@code{.}@var{hhh}@code{p}[@code{+}|@code{-}]@var{dd}}. | |
2020 | At the left of the decimal-point character exactly one digit is print. | |
91ea72b7 | 2021 | This character is only @code{0} if the number is denormalized. |
49c091e5 | 2022 | Otherwise the value is unspecified; it is implementation dependent how many |
2f6d1f1b UD |
2023 | bits are used. The number of hexadecimal digits on the right side of |
2024 | the decimal-point character is equal to the precision. If the precision | |
2025 | is zero it is determined to be large enough to provide an exact | |
2026 | representation of the number (or it is large enough to distinguish two | |
2027 | adjacent values if the @code{FLT_RADIX} is not a power of 2, | |
91ea72b7 | 2028 | @pxref{Floating Point Parameters}). For the @samp{%a} conversion |
2f6d1f1b UD |
2029 | lower-case characters are used to represent the hexadecimal number and |
2030 | the prefix and exponent sign are printed as @code{0x} and @code{p} | |
2031 | respectively. Otherwise upper-case characters are used and @code{0X} | |
2032 | and @code{P} are used for the representation of prefix and exponent | |
2033 | string. The exponent to the base of two is printed as a decimal number | |
2034 | using at least one digit but at most as many digits as necessary to | |
2035 | represent the value exactly. | |
2036 | ||
2037 | If the value to be printed represents infinity or a NaN, the output is | |
2038 | @w{[@code{-}]@code{inf}} or @code{nan} respectively if the conversion | |
2039 | specifier is @samp{%a}, @samp{%e}, @samp{%f}, or @samp{%g} and it is | |
2040 | @w{[@code{-}]@code{INF}} or @code{NAN} respectively if the conversion is | |
2041 | @samp{%A}, @samp{%E}, or @samp{%G}. | |
2042 | ||
28f540f4 RM |
2043 | The following flags can be used to modify the behavior: |
2044 | ||
2045 | @comment We use @asis instead of @samp so we can have ` ' as an item. | |
2046 | @table @asis | |
2047 | @item @samp{-} | |
2048 | Left-justify the result in the field. Normally the result is | |
2049 | right-justified. | |
2050 | ||
2051 | @item @samp{+} | |
2052 | Always include a plus or minus sign in the result. | |
2053 | ||
2054 | @item @samp{ } | |
2055 | If the result doesn't start with a plus or minus sign, prefix it with a | |
2056 | space instead. Since the @samp{+} flag ensures that the result includes | |
2057 | a sign, this flag is ignored if you supply both of them. | |
2058 | ||
2059 | @item @samp{#} | |
2060 | Specifies that the result should always include a decimal point, even | |
2061 | if no digits follow it. For the @samp{%g} and @samp{%G} conversions, | |
2062 | this also forces trailing zeros after the decimal point to be left | |
2063 | in place where they would otherwise be removed. | |
2064 | ||
2065 | @item @samp{'} | |
2066 | Separate the digits of the integer part of the result into groups as | |
2067 | specified by the locale specified for the @code{LC_NUMERIC} category; | |
2068 | @pxref{General Numeric}. This flag is a GNU extension. | |
2069 | ||
2070 | @item @samp{0} | |
2071 | Pad the field with zeros instead of spaces; the zeros are placed | |
2072 | after any sign. This flag is ignored if the @samp{-} flag is also | |
2073 | specified. | |
2074 | @end table | |
2075 | ||
2076 | The precision specifies how many digits follow the decimal-point | |
2077 | character for the @samp{%f}, @samp{%e}, and @samp{%E} conversions. For | |
2078 | these conversions, the default precision is @code{6}. If the precision | |
2079 | is explicitly @code{0}, this suppresses the decimal point character | |
2080 | entirely. For the @samp{%g} and @samp{%G} conversions, the precision | |
2081 | specifies how many significant digits to print. Significant digits are | |
2082 | the first digit before the decimal point, and all the digits after it. | |
91ea72b7 | 2083 | If the precision is @code{0} or not specified for @samp{%g} or @samp{%G}, |
28f540f4 RM |
2084 | it is treated like a value of @code{1}. If the value being printed |
2085 | cannot be expressed accurately in the specified number of digits, the | |
2086 | value is rounded to the nearest number that fits. | |
2087 | ||
2088 | Without a type modifier, the floating-point conversions use an argument | |
2089 | of type @code{double}. (By the default argument promotions, any | |
2090 | @code{float} arguments are automatically converted to @code{double}.) | |
2091 | The following type modifier is supported: | |
2092 | ||
2093 | @table @samp | |
2094 | @item L | |
2095 | An uppercase @samp{L} specifies that the argument is a @code{long | |
2096 | double}. | |
2097 | @end table | |
2098 | ||
2099 | Here are some examples showing how numbers print using the various | |
2100 | floating-point conversions. All of the numbers were printed using | |
2101 | this template string: | |
2102 | ||
2103 | @smallexample | |
2f6d1f1b | 2104 | "|%13.4a|%13.4f|%13.4e|%13.4g|\n" |
28f540f4 RM |
2105 | @end smallexample |
2106 | ||
2107 | Here is the output: | |
2108 | ||
2109 | @smallexample | |
2f6d1f1b UD |
2110 | | 0x0.0000p+0| 0.0000| 0.0000e+00| 0| |
2111 | | 0x1.0000p-1| 0.5000| 5.0000e-01| 0.5| | |
2112 | | 0x1.0000p+0| 1.0000| 1.0000e+00| 1| | |
2113 | | -0x1.0000p+0| -1.0000| -1.0000e+00| -1| | |
2114 | | 0x1.9000p+6| 100.0000| 1.0000e+02| 100| | |
2115 | | 0x1.f400p+9| 1000.0000| 1.0000e+03| 1000| | |
2116 | | 0x1.3880p+13| 10000.0000| 1.0000e+04| 1e+04| | |
2117 | | 0x1.81c8p+13| 12345.0000| 1.2345e+04| 1.234e+04| | |
2118 | | 0x1.86a0p+16| 100000.0000| 1.0000e+05| 1e+05| | |
2119 | | 0x1.e240p+16| 123456.0000| 1.2346e+05| 1.235e+05| | |
28f540f4 RM |
2120 | @end smallexample |
2121 | ||
2122 | Notice how the @samp{%g} conversion drops trailing zeros. | |
2123 | ||
2124 | @node Other Output Conversions | |
2125 | @subsection Other Output Conversions | |
2126 | ||
2127 | This section describes miscellaneous conversions for @code{printf}. | |
2128 | ||
b5e73f56 UD |
2129 | The @samp{%c} conversion prints a single character. In case there is no |
2130 | @samp{l} modifier the @code{int} argument is first converted to an | |
2131 | @code{unsigned char}. Then, if used in a wide stream function, the | |
2132 | character is converted into the corresponding wide character. The | |
2133 | @samp{-} flag can be used to specify left-justification in the field, | |
2134 | but no other flags are defined, and no precision or type modifier can be | |
2135 | given. For example: | |
28f540f4 RM |
2136 | |
2137 | @smallexample | |
2138 | printf ("%c%c%c%c%c", 'h', 'e', 'l', 'l', 'o'); | |
2139 | @end smallexample | |
2140 | ||
2141 | @noindent | |
2142 | prints @samp{hello}. | |
2143 | ||
b5e73f56 UD |
2144 | If there is a @samp{l} modifier present the argument is expected to be |
2145 | of type @code{wint_t}. If used in a multibyte function the wide | |
2146 | character is converted into a multibyte character before being added to | |
2147 | the output. In this case more than one output byte can be produced. | |
2148 | ||
2149 | The @samp{%s} conversion prints a string. If no @samp{l} modifier is | |
2150 | present the corresponding argument must be of type @code{char *} (or | |
2151 | @code{const char *}). If used in a wide stream function the string is | |
2152 | first converted in a wide character string. A precision can be | |
2153 | specified to indicate the maximum number of characters to write; | |
28f540f4 RM |
2154 | otherwise characters in the string up to but not including the |
2155 | terminating null character are written to the output stream. The | |
2156 | @samp{-} flag can be used to specify left-justification in the field, | |
2157 | but no other flags or type modifiers are defined for this conversion. | |
2158 | For example: | |
2159 | ||
2160 | @smallexample | |
2161 | printf ("%3s%-6s", "no", "where"); | |
2162 | @end smallexample | |
2163 | ||
2164 | @noindent | |
2165 | prints @samp{ nowhere }. | |
2166 | ||
b5e73f56 UD |
2167 | If there is a @samp{l} modifier present the argument is expected to be of type @code{wchar_t} (or @code{const wchar_t *}). |
2168 | ||
28f540f4 | 2169 | If you accidentally pass a null pointer as the argument for a @samp{%s} |
1f77f049 | 2170 | conversion, @theglibc{} prints it as @samp{(null)}. We think this |
28f540f4 RM |
2171 | is more useful than crashing. But it's not good practice to pass a null |
2172 | argument intentionally. | |
2173 | ||
2174 | The @samp{%m} conversion prints the string corresponding to the error | |
2175 | code in @code{errno}. @xref{Error Messages}. Thus: | |
2176 | ||
2177 | @smallexample | |
2178 | fprintf (stderr, "can't open `%s': %m\n", filename); | |
2179 | @end smallexample | |
2180 | ||
2181 | @noindent | |
2182 | is equivalent to: | |
2183 | ||
2184 | @smallexample | |
2185 | fprintf (stderr, "can't open `%s': %s\n", filename, strerror (errno)); | |
2186 | @end smallexample | |
2187 | ||
2188 | @noindent | |
1f77f049 | 2189 | The @samp{%m} conversion is a @glibcadj{} extension. |
28f540f4 RM |
2190 | |
2191 | The @samp{%p} conversion prints a pointer value. The corresponding | |
2192 | argument must be of type @code{void *}. In practice, you can use any | |
2193 | type of pointer. | |
2194 | ||
a7a93d50 | 2195 | In @theglibc{}, non-null pointers are printed as unsigned integers, |
28f540f4 RM |
2196 | as if a @samp{%#x} conversion were used. Null pointers print as |
2197 | @samp{(nil)}. (Pointers might print differently in other systems.) | |
2198 | ||
2199 | For example: | |
2200 | ||
2201 | @smallexample | |
2202 | printf ("%p", "testing"); | |
2203 | @end smallexample | |
2204 | ||
2205 | @noindent | |
2206 | prints @samp{0x} followed by a hexadecimal number---the address of the | |
2207 | string constant @code{"testing"}. It does not print the word | |
2208 | @samp{testing}. | |
2209 | ||
2210 | You can supply the @samp{-} flag with the @samp{%p} conversion to | |
2211 | specify left-justification, but no other flags, precision, or type | |
2212 | modifiers are defined. | |
2213 | ||
2214 | The @samp{%n} conversion is unlike any of the other output conversions. | |
2215 | It uses an argument which must be a pointer to an @code{int}, but | |
2216 | instead of printing anything it stores the number of characters printed | |
2217 | so far by this call at that location. The @samp{h} and @samp{l} type | |
2218 | modifiers are permitted to specify that the argument is of type | |
2219 | @code{short int *} or @code{long int *} instead of @code{int *}, but no | |
2220 | flags, field width, or precision are permitted. | |
2221 | ||
2222 | For example, | |
2223 | ||
2224 | @smallexample | |
2225 | int nchar; | |
2226 | printf ("%d %s%n\n", 3, "bears", &nchar); | |
2227 | @end smallexample | |
2228 | ||
2229 | @noindent | |
2230 | prints: | |
2231 | ||
2232 | @smallexample | |
2233 | 3 bears | |
2234 | @end smallexample | |
2235 | ||
2236 | @noindent | |
19c3f208 | 2237 | and sets @code{nchar} to @code{7}, because @samp{3 bears} is seven |
28f540f4 RM |
2238 | characters. |
2239 | ||
2240 | ||
2241 | The @samp{%%} conversion prints a literal @samp{%} character. This | |
2242 | conversion doesn't use an argument, and no flags, field width, | |
2243 | precision, or type modifiers are permitted. | |
2244 | ||
2245 | ||
2246 | @node Formatted Output Functions | |
2247 | @subsection Formatted Output Functions | |
2248 | ||
2249 | This section describes how to call @code{printf} and related functions. | |
2250 | Prototypes for these functions are in the header file @file{stdio.h}. | |
2251 | Because these functions take a variable number of arguments, you | |
2252 | @emph{must} declare prototypes for them before using them. Of course, | |
2253 | the easiest way to make sure you have all the right prototypes is to | |
2254 | just include @file{stdio.h}. | |
2255 | @pindex stdio.h | |
2256 | ||
2257 | @comment stdio.h | |
f65fd747 | 2258 | @comment ISO |
28f540f4 RM |
2259 | @deftypefun int printf (const char *@var{template}, @dots{}) |
2260 | The @code{printf} function prints the optional arguments under the | |
2261 | control of the template string @var{template} to the stream | |
2262 | @code{stdout}. It returns the number of characters printed, or a | |
2263 | negative value if there was an output error. | |
2264 | @end deftypefun | |
2265 | ||
b5e73f56 UD |
2266 | @comment wchar.h |
2267 | @comment ISO | |
2268 | @deftypefun int wprintf (const wchar_t *@var{template}, @dots{}) | |
2269 | The @code{wprintf} function prints the optional arguments under the | |
2270 | control of the wide template string @var{template} to the stream | |
2271 | @code{stdout}. It returns the number of wide characters printed, or a | |
2272 | negative value if there was an output error. | |
2273 | @end deftypefun | |
2274 | ||
28f540f4 | 2275 | @comment stdio.h |
f65fd747 | 2276 | @comment ISO |
28f540f4 RM |
2277 | @deftypefun int fprintf (FILE *@var{stream}, const char *@var{template}, @dots{}) |
2278 | This function is just like @code{printf}, except that the output is | |
2279 | written to the stream @var{stream} instead of @code{stdout}. | |
2280 | @end deftypefun | |
2281 | ||
b5e73f56 UD |
2282 | @comment wchar.h |
2283 | @comment ISO | |
2284 | @deftypefun int fwprintf (FILE *@var{stream}, const wchar_t *@var{template}, @dots{}) | |
2285 | This function is just like @code{wprintf}, except that the output is | |
2286 | written to the stream @var{stream} instead of @code{stdout}. | |
2287 | @end deftypefun | |
2288 | ||
28f540f4 | 2289 | @comment stdio.h |
f65fd747 | 2290 | @comment ISO |
28f540f4 RM |
2291 | @deftypefun int sprintf (char *@var{s}, const char *@var{template}, @dots{}) |
2292 | This is like @code{printf}, except that the output is stored in the character | |
2293 | array @var{s} instead of written to a stream. A null character is written | |
2294 | to mark the end of the string. | |
2295 | ||
2296 | The @code{sprintf} function returns the number of characters stored in | |
2297 | the array @var{s}, not including the terminating null character. | |
2298 | ||
2299 | The behavior of this function is undefined if copying takes place | |
2300 | between objects that overlap---for example, if @var{s} is also given | |
2301 | as an argument to be printed under control of the @samp{%s} conversion. | |
2302 | @xref{Copying and Concatenation}. | |
2303 | ||
2304 | @strong{Warning:} The @code{sprintf} function can be @strong{dangerous} | |
2305 | because it can potentially output more characters than can fit in the | |
2306 | allocation size of the string @var{s}. Remember that the field width | |
2307 | given in a conversion specification is only a @emph{minimum} value. | |
2308 | ||
2309 | To avoid this problem, you can use @code{snprintf} or @code{asprintf}, | |
2310 | described below. | |
2311 | @end deftypefun | |
2312 | ||
b5e73f56 UD |
2313 | @comment wchar.h |
2314 | @comment GNU | |
2315 | @deftypefun int swprintf (wchar_t *@var{s}, size_t @var{size}, const wchar_t *@var{template}, @dots{}) | |
2316 | This is like @code{wprintf}, except that the output is stored in the | |
2317 | wide character array @var{ws} instead of written to a stream. A null | |
2318 | wide character is written to mark the end of the string. The @var{size} | |
2319 | argument specifies the maximum number of characters to produce. The | |
2320 | trailing null character is counted towards this limit, so you should | |
2321 | allocate at least @var{size} wide characters for the string @var{ws}. | |
2322 | ||
2f278c94 UD |
2323 | The return value is the number of characters generated for the given |
2324 | input, excluding the trailing null. If not all output fits into the | |
2325 | provided buffer a negative value is returned. You should try again with | |
2326 | a bigger output string. @emph{Note:} this is different from how | |
2327 | @code{snprintf} handles this situation. | |
b5e73f56 UD |
2328 | |
2329 | Note that the corresponding narrow stream function takes fewer | |
2330 | parameters. @code{swprintf} in fact corresponds to the @code{snprintf} | |
2331 | function. Since the @code{sprintf} function can be dangerous and should | |
2332 | be avoided the @w{ISO C} committee refused to make the same mistake | |
2333 | again and decided to not define an function exactly corresponding to | |
2334 | @code{sprintf}. | |
2335 | @end deftypefun | |
2336 | ||
28f540f4 RM |
2337 | @comment stdio.h |
2338 | @comment GNU | |
2339 | @deftypefun int snprintf (char *@var{s}, size_t @var{size}, const char *@var{template}, @dots{}) | |
2340 | The @code{snprintf} function is similar to @code{sprintf}, except that | |
2341 | the @var{size} argument specifies the maximum number of characters to | |
2342 | produce. The trailing null character is counted towards this limit, so | |
2343 | you should allocate at least @var{size} characters for the string @var{s}. | |
42443a47 JL |
2344 | If @var{size} is zero, nothing, not even the null byte, shall be written and |
2345 | @var{s} may be a null pointer. | |
28f540f4 | 2346 | |
fe7bdd63 | 2347 | The return value is the number of characters which would be generated |
da2d1bc5 UD |
2348 | for the given input, excluding the trailing null. If this value is |
2349 | greater or equal to @var{size}, not all characters from the result have | |
2350 | been stored in @var{s}. You should try again with a bigger output | |
2351 | string. Here is an example of doing this: | |
28f540f4 RM |
2352 | |
2353 | @smallexample | |
2354 | @group | |
2355 | /* @r{Construct a message describing the value of a variable} | |
2356 | @r{whose name is @var{name} and whose value is @var{value}.} */ | |
2357 | char * | |
2358 | make_message (char *name, char *value) | |
2359 | @{ | |
2360 | /* @r{Guess we need no more than 100 chars of space.} */ | |
2361 | int size = 100; | |
2362 | char *buffer = (char *) xmalloc (size); | |
4cca6b86 | 2363 | int nchars; |
28f540f4 RM |
2364 | @end group |
2365 | @group | |
7ba73c63 UD |
2366 | if (buffer == NULL) |
2367 | return NULL; | |
2368 | ||
4cca6b86 UD |
2369 | /* @r{Try to print in the allocated space.} */ |
2370 | nchars = snprintf (buffer, size, "value of %s is %s", | |
162ba701 | 2371 | name, value); |
4cca6b86 UD |
2372 | @end group |
2373 | @group | |
fe7bdd63 | 2374 | if (nchars >= size) |
28f540f4 | 2375 | @{ |
4b8f94d3 | 2376 | /* @r{Reallocate buffer now that we know |
162ba701 | 2377 | how much space is needed.} */ |
53631fa8 UD |
2378 | size = nchars + 1; |
2379 | buffer = (char *) xrealloc (buffer, size); | |
4cca6b86 | 2380 | |
7ba73c63 | 2381 | if (buffer != NULL) |
162ba701 PIM |
2382 | /* @r{Try again.} */ |
2383 | snprintf (buffer, size, "value of %s is %s", | |
2384 | name, value); | |
28f540f4 | 2385 | @} |
4cca6b86 UD |
2386 | /* @r{The last call worked, return the string.} */ |
2387 | return buffer; | |
28f540f4 RM |
2388 | @} |
2389 | @end group | |
2390 | @end smallexample | |
2391 | ||
2392 | In practice, it is often easier just to use @code{asprintf}, below. | |
fb971363 | 2393 | |
1f77f049 | 2394 | @strong{Attention:} In versions of @theglibc{} prior to 2.1 the |
5a74e68a UD |
2395 | return value is the number of characters stored, not including the |
2396 | terminating null; unless there was not enough space in @var{s} to | |
2397 | store the result in which case @code{-1} is returned. This was | |
2398 | changed in order to comply with the @w{ISO C99} standard. | |
28f540f4 RM |
2399 | @end deftypefun |
2400 | ||
2401 | @node Dynamic Output | |
2402 | @subsection Dynamically Allocating Formatted Output | |
2403 | ||
2404 | The functions in this section do formatted output and place the results | |
2405 | in dynamically allocated memory. | |
2406 | ||
2407 | @comment stdio.h | |
2408 | @comment GNU | |
2409 | @deftypefun int asprintf (char **@var{ptr}, const char *@var{template}, @dots{}) | |
2410 | This function is similar to @code{sprintf}, except that it dynamically | |
2411 | allocates a string (as with @code{malloc}; @pxref{Unconstrained | |
2412 | Allocation}) to hold the output, instead of putting the output in a | |
2413 | buffer you allocate in advance. The @var{ptr} argument should be the | |
3a4e0609 UD |
2414 | address of a @code{char *} object, and a successful call to |
2415 | @code{asprintf} stores a pointer to the newly allocated string at that | |
2416 | location. | |
28f540f4 | 2417 | |
7ba73c63 | 2418 | The return value is the number of characters allocated for the buffer, or |
0bc93a2f | 2419 | less than zero if an error occurred. Usually this means that the buffer |
7ba73c63 UD |
2420 | could not be allocated. |
2421 | ||
28f540f4 RM |
2422 | Here is how to use @code{asprintf} to get the same result as the |
2423 | @code{snprintf} example, but more easily: | |
2424 | ||
2425 | @smallexample | |
2426 | /* @r{Construct a message describing the value of a variable} | |
2427 | @r{whose name is @var{name} and whose value is @var{value}.} */ | |
2428 | char * | |
2429 | make_message (char *name, char *value) | |
2430 | @{ | |
2431 | char *result; | |
7ba73c63 UD |
2432 | if (asprintf (&result, "value of %s is %s", name, value) < 0) |
2433 | return NULL; | |
28f540f4 RM |
2434 | return result; |
2435 | @} | |
2436 | @end smallexample | |
2437 | @end deftypefun | |
2438 | ||
2439 | @comment stdio.h | |
2440 | @comment GNU | |
2441 | @deftypefun int obstack_printf (struct obstack *@var{obstack}, const char *@var{template}, @dots{}) | |
2442 | This function is similar to @code{asprintf}, except that it uses the | |
2443 | obstack @var{obstack} to allocate the space. @xref{Obstacks}. | |
2444 | ||
2445 | The characters are written onto the end of the current object. | |
2446 | To get at them, you must finish the object with @code{obstack_finish} | |
2447 | (@pxref{Growing Objects}).@refill | |
2448 | @end deftypefun | |
2449 | ||
2450 | @node Variable Arguments Output | |
2451 | @subsection Variable Arguments Output Functions | |
2452 | ||
2453 | The functions @code{vprintf} and friends are provided so that you can | |
2454 | define your own variadic @code{printf}-like functions that make use of | |
2455 | the same internals as the built-in formatted output functions. | |
2456 | ||
2457 | The most natural way to define such functions would be to use a language | |
2458 | construct to say, ``Call @code{printf} and pass this template plus all | |
2459 | of my arguments after the first five.'' But there is no way to do this | |
2460 | in C, and it would be hard to provide a way, since at the C language | |
2461 | level there is no way to tell how many arguments your function received. | |
2462 | ||
2463 | Since that method is impossible, we provide alternative functions, the | |
2464 | @code{vprintf} series, which lets you pass a @code{va_list} to describe | |
2465 | ``all of my arguments after the first five.'' | |
2466 | ||
19c3f208 | 2467 | When it is sufficient to define a macro rather than a real function, |
28f540f4 RM |
2468 | the GNU C compiler provides a way to do this much more easily with macros. |
2469 | For example: | |
2470 | ||
2471 | @smallexample | |
838e5ffe | 2472 | #define myprintf(a, b, c, d, e, rest...) \ |
162ba701 | 2473 | printf (mytemplate , ## rest) |
28f540f4 RM |
2474 | @end smallexample |
2475 | ||
2476 | @noindent | |
9229b83e RM |
2477 | @xref{Variadic Macros,,, cpp, The C preprocessor}, for details. |
2478 | But this is limited to macros, and does not apply to real functions at all. | |
28f540f4 RM |
2479 | |
2480 | Before calling @code{vprintf} or the other functions listed in this | |
2481 | section, you @emph{must} call @code{va_start} (@pxref{Variadic | |
2482 | Functions}) to initialize a pointer to the variable arguments. Then you | |
2483 | can call @code{va_arg} to fetch the arguments that you want to handle | |
2484 | yourself. This advances the pointer past those arguments. | |
2485 | ||
2486 | Once your @code{va_list} pointer is pointing at the argument of your | |
2487 | choice, you are ready to call @code{vprintf}. That argument and all | |
2488 | subsequent arguments that were passed to your function are used by | |
2489 | @code{vprintf} along with the template that you specified separately. | |
2490 | ||
2491 | In some other systems, the @code{va_list} pointer may become invalid | |
2492 | after the call to @code{vprintf}, so you must not use @code{va_arg} | |
2493 | after you call @code{vprintf}. Instead, you should call @code{va_end} | |
2494 | to retire the pointer from service. However, you can safely call | |
2495 | @code{va_start} on another pointer variable and begin fetching the | |
2496 | arguments again through that pointer. Calling @code{vprintf} does not | |
2497 | destroy the argument list of your function, merely the particular | |
2498 | pointer that you passed to it. | |
2499 | ||
2500 | GNU C does not have such restrictions. You can safely continue to fetch | |
2501 | arguments from a @code{va_list} pointer after passing it to | |
2502 | @code{vprintf}, and @code{va_end} is a no-op. (Note, however, that | |
2503 | subsequent @code{va_arg} calls will fetch the same arguments which | |
2504 | @code{vprintf} previously used.) | |
2505 | ||
2506 | Prototypes for these functions are declared in @file{stdio.h}. | |
2507 | @pindex stdio.h | |
2508 | ||
2509 | @comment stdio.h | |
f65fd747 | 2510 | @comment ISO |
28f540f4 RM |
2511 | @deftypefun int vprintf (const char *@var{template}, va_list @var{ap}) |
2512 | This function is similar to @code{printf} except that, instead of taking | |
2513 | a variable number of arguments directly, it takes an argument list | |
2514 | pointer @var{ap}. | |
2515 | @end deftypefun | |
2516 | ||
b5e73f56 UD |
2517 | @comment wchar.h |
2518 | @comment ISO | |
2519 | @deftypefun int vwprintf (const wchar_t *@var{template}, va_list @var{ap}) | |
2520 | This function is similar to @code{wprintf} except that, instead of taking | |
2521 | a variable number of arguments directly, it takes an argument list | |
2522 | pointer @var{ap}. | |
2523 | @end deftypefun | |
2524 | ||
28f540f4 | 2525 | @comment stdio.h |
f65fd747 | 2526 | @comment ISO |
28f540f4 RM |
2527 | @deftypefun int vfprintf (FILE *@var{stream}, const char *@var{template}, va_list @var{ap}) |
2528 | This is the equivalent of @code{fprintf} with the variable argument list | |
2529 | specified directly as for @code{vprintf}. | |
2530 | @end deftypefun | |
2531 | ||
b5e73f56 UD |
2532 | @comment wchar.h |
2533 | @comment ISO | |
2534 | @deftypefun int vfwprintf (FILE *@var{stream}, const wchar_t *@var{template}, va_list @var{ap}) | |
2535 | This is the equivalent of @code{fwprintf} with the variable argument list | |
2536 | specified directly as for @code{vwprintf}. | |
2537 | @end deftypefun | |
2538 | ||
28f540f4 | 2539 | @comment stdio.h |
f65fd747 | 2540 | @comment ISO |
28f540f4 RM |
2541 | @deftypefun int vsprintf (char *@var{s}, const char *@var{template}, va_list @var{ap}) |
2542 | This is the equivalent of @code{sprintf} with the variable argument list | |
2543 | specified directly as for @code{vprintf}. | |
2544 | @end deftypefun | |
2545 | ||
b5e73f56 UD |
2546 | @comment wchar.h |
2547 | @comment GNU | |
2548 | @deftypefun int vswprintf (wchar_t *@var{s}, size_t @var{size}, const wchar_t *@var{template}, va_list @var{ap}) | |
2549 | This is the equivalent of @code{swprintf} with the variable argument list | |
2550 | specified directly as for @code{vwprintf}. | |
2551 | @end deftypefun | |
2552 | ||
28f540f4 RM |
2553 | @comment stdio.h |
2554 | @comment GNU | |
2555 | @deftypefun int vsnprintf (char *@var{s}, size_t @var{size}, const char *@var{template}, va_list @var{ap}) | |
2556 | This is the equivalent of @code{snprintf} with the variable argument list | |
2557 | specified directly as for @code{vprintf}. | |
2558 | @end deftypefun | |
2559 | ||
2560 | @comment stdio.h | |
2561 | @comment GNU | |
2562 | @deftypefun int vasprintf (char **@var{ptr}, const char *@var{template}, va_list @var{ap}) | |
2563 | The @code{vasprintf} function is the equivalent of @code{asprintf} with the | |
2564 | variable argument list specified directly as for @code{vprintf}. | |
2565 | @end deftypefun | |
2566 | ||
2567 | @comment stdio.h | |
2568 | @comment GNU | |
2569 | @deftypefun int obstack_vprintf (struct obstack *@var{obstack}, const char *@var{template}, va_list @var{ap}) | |
2570 | The @code{obstack_vprintf} function is the equivalent of | |
2571 | @code{obstack_printf} with the variable argument list specified directly | |
2572 | as for @code{vprintf}.@refill | |
2573 | @end deftypefun | |
2574 | ||
2575 | Here's an example showing how you might use @code{vfprintf}. This is a | |
2576 | function that prints error messages to the stream @code{stderr}, along | |
2577 | with a prefix indicating the name of the program | |
19c3f208 | 2578 | (@pxref{Error Messages}, for a description of |
28f540f4 RM |
2579 | @code{program_invocation_short_name}). |
2580 | ||
2581 | @smallexample | |
2582 | @group | |
2583 | #include <stdio.h> | |
2584 | #include <stdarg.h> | |
2585 | ||
2586 | void | |
2587 | eprintf (const char *template, ...) | |
2588 | @{ | |
2589 | va_list ap; | |
2590 | extern char *program_invocation_short_name; | |
2591 | ||
2592 | fprintf (stderr, "%s: ", program_invocation_short_name); | |
4b8f94d3 | 2593 | va_start (ap, template); |
28f540f4 RM |
2594 | vfprintf (stderr, template, ap); |
2595 | va_end (ap); | |
2596 | @} | |
2597 | @end group | |
2598 | @end smallexample | |
2599 | ||
2600 | @noindent | |
2601 | You could call @code{eprintf} like this: | |
2602 | ||
2603 | @smallexample | |
2604 | eprintf ("file `%s' does not exist\n", filename); | |
2605 | @end smallexample | |
2606 | ||
2607 | In GNU C, there is a special construct you can use to let the compiler | |
2608 | know that a function uses a @code{printf}-style format string. Then it | |
2609 | can check the number and types of arguments in each call to the | |
2610 | function, and warn you when they do not match the format string. | |
2611 | For example, take this declaration of @code{eprintf}: | |
2612 | ||
2613 | @smallexample | |
2614 | void eprintf (const char *template, ...) | |
162ba701 | 2615 | __attribute__ ((format (printf, 1, 2))); |
28f540f4 RM |
2616 | @end smallexample |
2617 | ||
2618 | @noindent | |
2619 | This tells the compiler that @code{eprintf} uses a format string like | |
2620 | @code{printf} (as opposed to @code{scanf}; @pxref{Formatted Input}); | |
2621 | the format string appears as the first argument; | |
2622 | and the arguments to satisfy the format begin with the second. | |
2623 | @xref{Function Attributes, , Declaring Attributes of Functions, | |
2624 | gcc.info, Using GNU CC}, for more information. | |
2625 | ||
2626 | @node Parsing a Template String | |
2627 | @subsection Parsing a Template String | |
2628 | @cindex parsing a template string | |
2629 | ||
2630 | You can use the function @code{parse_printf_format} to obtain | |
2631 | information about the number and types of arguments that are expected by | |
2632 | a given template string. This function permits interpreters that | |
2633 | provide interfaces to @code{printf} to avoid passing along invalid | |
2634 | arguments from the user's program, which could cause a crash. | |
2635 | ||
2636 | All the symbols described in this section are declared in the header | |
2637 | file @file{printf.h}. | |
2638 | ||
2639 | @comment printf.h | |
2640 | @comment GNU | |
2641 | @deftypefun size_t parse_printf_format (const char *@var{template}, size_t @var{n}, int *@var{argtypes}) | |
2642 | This function returns information about the number and types of | |
2643 | arguments expected by the @code{printf} template string @var{template}. | |
2644 | The information is stored in the array @var{argtypes}; each element of | |
2645 | this array describes one argument. This information is encoded using | |
2646 | the various @samp{PA_} macros, listed below. | |
2647 | ||
04b9968b UD |
2648 | The argument @var{n} specifies the number of elements in the array |
2649 | @var{argtypes}. This is the maximum number of elements that | |
28f540f4 RM |
2650 | @code{parse_printf_format} will try to write. |
2651 | ||
2652 | @code{parse_printf_format} returns the total number of arguments required | |
2653 | by @var{template}. If this number is greater than @var{n}, then the | |
2654 | information returned describes only the first @var{n} arguments. If you | |
04b9968b | 2655 | want information about additional arguments, allocate a bigger |
28f540f4 RM |
2656 | array and call @code{parse_printf_format} again. |
2657 | @end deftypefun | |
2658 | ||
2659 | The argument types are encoded as a combination of a basic type and | |
2660 | modifier flag bits. | |
2661 | ||
2662 | @comment printf.h | |
2663 | @comment GNU | |
2664 | @deftypevr Macro int PA_FLAG_MASK | |
2665 | This macro is a bitmask for the type modifier flag bits. You can write | |
2666 | the expression @code{(argtypes[i] & PA_FLAG_MASK)} to extract just the | |
2667 | flag bits for an argument, or @code{(argtypes[i] & ~PA_FLAG_MASK)} to | |
2668 | extract just the basic type code. | |
2669 | @end deftypevr | |
2670 | ||
2671 | Here are symbolic constants that represent the basic types; they stand | |
2672 | for integer values. | |
2673 | ||
779ae82e | 2674 | @vtable @code |
28f540f4 RM |
2675 | @comment printf.h |
2676 | @comment GNU | |
2677 | @item PA_INT | |
28f540f4 RM |
2678 | This specifies that the base type is @code{int}. |
2679 | ||
2680 | @comment printf.h | |
2681 | @comment GNU | |
2682 | @item PA_CHAR | |
28f540f4 RM |
2683 | This specifies that the base type is @code{int}, cast to @code{char}. |
2684 | ||
2685 | @comment printf.h | |
2686 | @comment GNU | |
2687 | @item PA_STRING | |
28f540f4 RM |
2688 | This specifies that the base type is @code{char *}, a null-terminated string. |
2689 | ||
2690 | @comment printf.h | |
2691 | @comment GNU | |
2692 | @item PA_POINTER | |
28f540f4 RM |
2693 | This specifies that the base type is @code{void *}, an arbitrary pointer. |
2694 | ||
2695 | @comment printf.h | |
2696 | @comment GNU | |
2697 | @item PA_FLOAT | |
28f540f4 RM |
2698 | This specifies that the base type is @code{float}. |
2699 | ||
2700 | @comment printf.h | |
2701 | @comment GNU | |
2702 | @item PA_DOUBLE | |
28f540f4 RM |
2703 | This specifies that the base type is @code{double}. |
2704 | ||
2705 | @comment printf.h | |
2706 | @comment GNU | |
2707 | @item PA_LAST | |
28f540f4 RM |
2708 | You can define additional base types for your own programs as offsets |
2709 | from @code{PA_LAST}. For example, if you have data types @samp{foo} | |
2710 | and @samp{bar} with their own specialized @code{printf} conversions, | |
2711 | you could define encodings for these types as: | |
2712 | ||
2713 | @smallexample | |
2714 | #define PA_FOO PA_LAST | |
2715 | #define PA_BAR (PA_LAST + 1) | |
2716 | @end smallexample | |
779ae82e | 2717 | @end vtable |
28f540f4 RM |
2718 | |
2719 | Here are the flag bits that modify a basic type. They are combined with | |
2720 | the code for the basic type using inclusive-or. | |
2721 | ||
779ae82e | 2722 | @vtable @code |
28f540f4 RM |
2723 | @comment printf.h |
2724 | @comment GNU | |
2725 | @item PA_FLAG_PTR | |
28f540f4 RM |
2726 | If this bit is set, it indicates that the encoded type is a pointer to |
2727 | the base type, rather than an immediate value. | |
2728 | For example, @samp{PA_INT|PA_FLAG_PTR} represents the type @samp{int *}. | |
2729 | ||
2730 | @comment printf.h | |
2731 | @comment GNU | |
2732 | @item PA_FLAG_SHORT | |
28f540f4 RM |
2733 | If this bit is set, it indicates that the base type is modified with |
2734 | @code{short}. (This corresponds to the @samp{h} type modifier.) | |
2735 | ||
2736 | @comment printf.h | |
2737 | @comment GNU | |
2738 | @item PA_FLAG_LONG | |
28f540f4 RM |
2739 | If this bit is set, it indicates that the base type is modified with |
2740 | @code{long}. (This corresponds to the @samp{l} type modifier.) | |
2741 | ||
2742 | @comment printf.h | |
2743 | @comment GNU | |
2744 | @item PA_FLAG_LONG_LONG | |
28f540f4 RM |
2745 | If this bit is set, it indicates that the base type is modified with |
2746 | @code{long long}. (This corresponds to the @samp{L} type modifier.) | |
2747 | ||
2748 | @comment printf.h | |
2749 | @comment GNU | |
2750 | @item PA_FLAG_LONG_DOUBLE | |
28f540f4 RM |
2751 | This is a synonym for @code{PA_FLAG_LONG_LONG}, used by convention with |
2752 | a base type of @code{PA_DOUBLE} to indicate a type of @code{long double}. | |
779ae82e | 2753 | @end vtable |
28f540f4 RM |
2754 | |
2755 | @ifinfo | |
6d52618b | 2756 | For an example of using these facilities, see @ref{Example of Parsing}. |
28f540f4 RM |
2757 | @end ifinfo |
2758 | ||
2759 | @node Example of Parsing | |
2760 | @subsection Example of Parsing a Template String | |
2761 | ||
2762 | Here is an example of decoding argument types for a format string. We | |
2763 | assume this is part of an interpreter which contains arguments of type | |
2764 | @code{NUMBER}, @code{CHAR}, @code{STRING} and @code{STRUCTURE} (and | |
2765 | perhaps others which are not valid here). | |
2766 | ||
2767 | @smallexample | |
2768 | /* @r{Test whether the @var{nargs} specified objects} | |
2769 | @r{in the vector @var{args} are valid} | |
2770 | @r{for the format string @var{format}:} | |
2771 | @r{if so, return 1.} | |
2772 | @r{If not, return 0 after printing an error message.} */ | |
2773 | ||
2774 | int | |
2775 | validate_args (char *format, int nargs, OBJECT *args) | |
2776 | @{ | |
2777 | int *argtypes; | |
2778 | int nwanted; | |
2779 | ||
2780 | /* @r{Get the information about the arguments.} | |
2781 | @r{Each conversion specification must be at least two characters} | |
2782 | @r{long, so there cannot be more specifications than half the} | |
2783 | @r{length of the string.} */ | |
2784 | ||
2785 | argtypes = (int *) alloca (strlen (format) / 2 * sizeof (int)); | |
2786 | nwanted = parse_printf_format (string, nelts, argtypes); | |
2787 | ||
2788 | /* @r{Check the number of arguments.} */ | |
2789 | if (nwanted > nargs) | |
2790 | @{ | |
2791 | error ("too few arguments (at least %d required)", nwanted); | |
2792 | return 0; | |
2793 | @} | |
19c3f208 | 2794 | |
28f540f4 RM |
2795 | /* @r{Check the C type wanted for each argument} |
2796 | @r{and see if the object given is suitable.} */ | |
2797 | for (i = 0; i < nwanted; i++) | |
2798 | @{ | |
2799 | int wanted; | |
2800 | ||
2801 | if (argtypes[i] & PA_FLAG_PTR) | |
162ba701 | 2802 | wanted = STRUCTURE; |
28f540f4 | 2803 | else |
162ba701 PIM |
2804 | switch (argtypes[i] & ~PA_FLAG_MASK) |
2805 | @{ | |
2806 | case PA_INT: | |
2807 | case PA_FLOAT: | |
2808 | case PA_DOUBLE: | |
2809 | wanted = NUMBER; | |
2810 | break; | |
2811 | case PA_CHAR: | |
2812 | wanted = CHAR; | |
2813 | break; | |
2814 | case PA_STRING: | |
2815 | wanted = STRING; | |
2816 | break; | |
2817 | case PA_POINTER: | |
2818 | wanted = STRUCTURE; | |
2819 | break; | |
2820 | @} | |
28f540f4 | 2821 | if (TYPE (args[i]) != wanted) |
162ba701 PIM |
2822 | @{ |
2823 | error ("type mismatch for arg number %d", i); | |
2824 | return 0; | |
2825 | @} | |
28f540f4 RM |
2826 | @} |
2827 | return 1; | |
2828 | @} | |
2829 | @end smallexample | |
2830 | ||
2831 | @node Customizing Printf | |
2832 | @section Customizing @code{printf} | |
2833 | @cindex customizing @code{printf} | |
2834 | @cindex defining new @code{printf} conversions | |
2835 | @cindex extending @code{printf} | |
2836 | ||
1f77f049 | 2837 | @Theglibc{} lets you define your own custom conversion specifiers |
28f540f4 RM |
2838 | for @code{printf} template strings, to teach @code{printf} clever ways |
2839 | to print the important data structures of your program. | |
2840 | ||
2841 | The way you do this is by registering the conversion with the function | |
2842 | @code{register_printf_function}; see @ref{Registering New Conversions}. | |
2843 | One of the arguments you pass to this function is a pointer to a handler | |
2844 | function that produces the actual output; see @ref{Defining the Output | |
2845 | Handler}, for information on how to write this function. | |
2846 | ||
2847 | You can also install a function that just returns information about the | |
2848 | number and type of arguments expected by the conversion specifier. | |
2849 | @xref{Parsing a Template String}, for information about this. | |
2850 | ||
2851 | The facilities of this section are declared in the header file | |
2852 | @file{printf.h}. | |
2853 | ||
2854 | @menu | |
19c3f208 | 2855 | * Registering New Conversions:: Using @code{register_printf_function} |
162ba701 | 2856 | to register a new output conversion. |
28f540f4 | 2857 | * Conversion Specifier Options:: The handler must be able to get |
162ba701 PIM |
2858 | the options specified in the |
2859 | template when it is called. | |
28f540f4 | 2860 | * Defining the Output Handler:: Defining the handler and arginfo |
162ba701 PIM |
2861 | functions that are passed as arguments |
2862 | to @code{register_printf_function}. | |
28f540f4 | 2863 | * Printf Extension Example:: How to define a @code{printf} |
162ba701 | 2864 | handler function. |
29bb8719 | 2865 | * Predefined Printf Handlers:: Predefined @code{printf} handlers. |
28f540f4 RM |
2866 | @end menu |
2867 | ||
2868 | @strong{Portability Note:} The ability to extend the syntax of | |
f65fd747 | 2869 | @code{printf} template strings is a GNU extension. ISO standard C has |
28f540f4 RM |
2870 | nothing similar. |
2871 | ||
2872 | @node Registering New Conversions | |
2873 | @subsection Registering New Conversions | |
2874 | ||
2875 | The function to register a new output conversion is | |
2876 | @code{register_printf_function}, declared in @file{printf.h}. | |
2877 | @pindex printf.h | |
2878 | ||
2879 | @comment printf.h | |
2880 | @comment GNU | |
2881 | @deftypefun int register_printf_function (int @var{spec}, printf_function @var{handler-function}, printf_arginfo_function @var{arginfo-function}) | |
2882 | This function defines the conversion specifier character @var{spec}. | |
42be70d4 | 2883 | Thus, if @var{spec} is @code{'Y'}, it defines the conversion @samp{%Y}. |
28f540f4 RM |
2884 | You can redefine the built-in conversions like @samp{%s}, but flag |
2885 | characters like @samp{#} and type modifiers like @samp{l} can never be | |
2886 | used as conversions; calling @code{register_printf_function} for those | |
42be70d4 UD |
2887 | characters has no effect. It is advisable not to use lowercase letters, |
2888 | since the ISO C standard warns that additional lowercase letters may be | |
2889 | standardized in future editions of the standard. | |
28f540f4 RM |
2890 | |
2891 | The @var{handler-function} is the function called by @code{printf} and | |
2892 | friends when this conversion appears in a template string. | |
2893 | @xref{Defining the Output Handler}, for information about how to define | |
2894 | a function to pass as this argument. If you specify a null pointer, any | |
2895 | existing handler function for @var{spec} is removed. | |
2896 | ||
2897 | The @var{arginfo-function} is the function called by | |
2898 | @code{parse_printf_format} when this conversion appears in a | |
2899 | template string. @xref{Parsing a Template String}, for information | |
2900 | about this. | |
2901 | ||
54d79e99 UD |
2902 | @c The following is not true anymore. The `parse_printf_format' function |
2903 | @c is now also called from `vfprintf' via `parse_one_spec'. | |
2904 | @c --drepper@gnu, 1996/11/14 | |
2905 | @c | |
2906 | @c Normally, you install both functions for a conversion at the same time, | |
2907 | @c but if you are never going to call @code{parse_printf_format}, you do | |
2908 | @c not need to define an arginfo function. | |
2909 | ||
1f77f049 | 2910 | @strong{Attention:} In @theglibc{} versions before 2.0 the |
54d79e99 | 2911 | @var{arginfo-function} function did not need to be installed unless |
04b9968b | 2912 | the user used the @code{parse_printf_format} function. This has changed. |
54d79e99 UD |
2913 | Now a call to any of the @code{printf} functions will call this |
2914 | function when this format specifier appears in the format string. | |
28f540f4 RM |
2915 | |
2916 | The return value is @code{0} on success, and @code{-1} on failure | |
2917 | (which occurs if @var{spec} is out of range). | |
2918 | ||
2919 | You can redefine the standard output conversions, but this is probably | |
2920 | not a good idea because of the potential for confusion. Library routines | |
2921 | written by other people could break if you do this. | |
2922 | @end deftypefun | |
2923 | ||
2924 | @node Conversion Specifier Options | |
2925 | @subsection Conversion Specifier Options | |
2926 | ||
40deae08 RM |
2927 | If you define a meaning for @samp{%A}, what if the template contains |
2928 | @samp{%+23A} or @samp{%-#A}? To implement a sensible meaning for these, | |
28f540f4 RM |
2929 | the handler when called needs to be able to get the options specified in |
2930 | the template. | |
2931 | ||
838e5ffe UD |
2932 | Both the @var{handler-function} and @var{arginfo-function} accept an |
2933 | argument that points to a @code{struct printf_info}, which contains | |
2934 | information about the options appearing in an instance of the conversion | |
2935 | specifier. This data type is declared in the header file | |
2936 | @file{printf.h}. | |
28f540f4 RM |
2937 | @pindex printf.h |
2938 | ||
2939 | @comment printf.h | |
2940 | @comment GNU | |
2941 | @deftp {Type} {struct printf_info} | |
2942 | This structure is used to pass information about the options appearing | |
2943 | in an instance of a conversion specifier in a @code{printf} template | |
2944 | string to the handler and arginfo functions for that specifier. It | |
2945 | contains the following members: | |
2946 | ||
2947 | @table @code | |
2948 | @item int prec | |
2949 | This is the precision specified. The value is @code{-1} if no precision | |
2950 | was specified. If the precision was given as @samp{*}, the | |
2951 | @code{printf_info} structure passed to the handler function contains the | |
2952 | actual value retrieved from the argument list. But the structure passed | |
2953 | to the arginfo function contains a value of @code{INT_MIN}, since the | |
2954 | actual value is not known. | |
2955 | ||
2956 | @item int width | |
2957 | This is the minimum field width specified. The value is @code{0} if no | |
2958 | width was specified. If the field width was given as @samp{*}, the | |
2959 | @code{printf_info} structure passed to the handler function contains the | |
2960 | actual value retrieved from the argument list. But the structure passed | |
2961 | to the arginfo function contains a value of @code{INT_MIN}, since the | |
2962 | actual value is not known. | |
2963 | ||
54d79e99 | 2964 | @item wchar_t spec |
28f540f4 RM |
2965 | This is the conversion specifier character specified. It's stored in |
2966 | the structure so that you can register the same handler function for | |
2967 | multiple characters, but still have a way to tell them apart when the | |
2968 | handler function is called. | |
2969 | ||
2970 | @item unsigned int is_long_double | |
2971 | This is a boolean that is true if the @samp{L}, @samp{ll}, or @samp{q} | |
2972 | type modifier was specified. For integer conversions, this indicates | |
2973 | @code{long long int}, as opposed to @code{long double} for floating | |
2974 | point conversions. | |
2975 | ||
d64b6ad0 UD |
2976 | @item unsigned int is_char |
2977 | This is a boolean that is true if the @samp{hh} type modifier was specified. | |
2978 | ||
28f540f4 RM |
2979 | @item unsigned int is_short |
2980 | This is a boolean that is true if the @samp{h} type modifier was specified. | |
2981 | ||
2982 | @item unsigned int is_long | |
2983 | This is a boolean that is true if the @samp{l} type modifier was specified. | |
2984 | ||
2985 | @item unsigned int alt | |
2986 | This is a boolean that is true if the @samp{#} flag was specified. | |
2987 | ||
2988 | @item unsigned int space | |
2989 | This is a boolean that is true if the @samp{ } flag was specified. | |
2990 | ||
2991 | @item unsigned int left | |
2992 | This is a boolean that is true if the @samp{-} flag was specified. | |
2993 | ||
2994 | @item unsigned int showsign | |
2995 | This is a boolean that is true if the @samp{+} flag was specified. | |
2996 | ||
2997 | @item unsigned int group | |
2998 | This is a boolean that is true if the @samp{'} flag was specified. | |
2999 | ||
54d79e99 UD |
3000 | @item unsigned int extra |
3001 | This flag has a special meaning depending on the context. It could | |
3002 | be used freely by the user-defined handlers but when called from | |
3003 | the @code{printf} function this variable always contains the value | |
3004 | @code{0}. | |
3005 | ||
d64b6ad0 UD |
3006 | @item unsigned int wide |
3007 | This flag is set if the stream is wide oriented. | |
3008 | ||
54d79e99 | 3009 | @item wchar_t pad |
28f540f4 RM |
3010 | This is the character to use for padding the output to the minimum field |
3011 | width. The value is @code{'0'} if the @samp{0} flag was specified, and | |
3012 | @code{' '} otherwise. | |
3013 | @end table | |
3014 | @end deftp | |
3015 | ||
3016 | ||
3017 | @node Defining the Output Handler | |
3018 | @subsection Defining the Output Handler | |
3019 | ||
3020 | Now let's look at how to define the handler and arginfo functions | |
3021 | which are passed as arguments to @code{register_printf_function}. | |
3022 | ||
1f77f049 | 3023 | @strong{Compatibility Note:} The interface changed in @theglibc{} |
54d79e99 UD |
3024 | version 2.0. Previously the third argument was of type |
3025 | @code{va_list *}. | |
3026 | ||
28f540f4 RM |
3027 | You should define your handler functions with a prototype like: |
3028 | ||
3029 | @smallexample | |
3030 | int @var{function} (FILE *stream, const struct printf_info *info, | |
162ba701 | 3031 | const void *const *args) |
28f540f4 RM |
3032 | @end smallexample |
3033 | ||
54d79e99 | 3034 | The @var{stream} argument passed to the handler function is the stream to |
28f540f4 RM |
3035 | which it should write output. |
3036 | ||
54d79e99 | 3037 | The @var{info} argument is a pointer to a structure that contains |
28f540f4 RM |
3038 | information about the various options that were included with the |
3039 | conversion in the template string. You should not modify this structure | |
3040 | inside your handler function. @xref{Conversion Specifier Options}, for | |
3041 | a description of this data structure. | |
3042 | ||
54d79e99 UD |
3043 | @c The following changes some time back. --drepper@gnu, 1996/11/14 |
3044 | @c | |
3045 | @c The @code{ap_pointer} argument is used to pass the tail of the variable | |
3046 | @c argument list containing the values to be printed to your handler. | |
3047 | @c Unlike most other functions that can be passed an explicit variable | |
3048 | @c argument list, this is a @emph{pointer} to a @code{va_list}, rather than | |
3049 | @c the @code{va_list} itself. Thus, you should fetch arguments by | |
3050 | @c means of @code{va_arg (*ap_pointer, @var{type})}. | |
3051 | @c | |
3052 | @c (Passing a pointer here allows the function that calls your handler | |
3053 | @c function to update its own @code{va_list} variable to account for the | |
3054 | @c arguments that your handler processes. @xref{Variadic Functions}.) | |
3055 | ||
3056 | The @var{args} is a vector of pointers to the arguments data. | |
04b9968b | 3057 | The number of arguments was determined by calling the argument |
54d79e99 | 3058 | information function provided by the user. |
28f540f4 RM |
3059 | |
3060 | Your handler function should return a value just like @code{printf} | |
3061 | does: it should return the number of characters it has written, or a | |
3062 | negative value to indicate an error. | |
3063 | ||
3064 | @comment printf.h | |
3065 | @comment GNU | |
3066 | @deftp {Data Type} printf_function | |
3067 | This is the data type that a handler function should have. | |
3068 | @end deftp | |
3069 | ||
3070 | If you are going to use @w{@code{parse_printf_format}} in your | |
54d79e99 | 3071 | application, you must also define a function to pass as the |
28f540f4 | 3072 | @var{arginfo-function} argument for each new conversion you install with |
19c3f208 | 3073 | @code{register_printf_function}. |
28f540f4 | 3074 | |
54d79e99 | 3075 | You have to define these functions with a prototype like: |
28f540f4 RM |
3076 | |
3077 | @smallexample | |
3078 | int @var{function} (const struct printf_info *info, | |
162ba701 | 3079 | size_t n, int *argtypes) |
28f540f4 RM |
3080 | @end smallexample |
3081 | ||
3082 | The return value from the function should be the number of arguments the | |
3083 | conversion expects. The function should also fill in no more than | |
3084 | @var{n} elements of the @var{argtypes} array with information about the | |
3085 | types of each of these arguments. This information is encoded using the | |
3086 | various @samp{PA_} macros. (You will notice that this is the same | |
3087 | calling convention @code{parse_printf_format} itself uses.) | |
3088 | ||
3089 | @comment printf.h | |
3090 | @comment GNU | |
3091 | @deftp {Data Type} printf_arginfo_function | |
3092 | This type is used to describe functions that return information about | |
3093 | the number and type of arguments used by a conversion specifier. | |
3094 | @end deftp | |
3095 | ||
3096 | @node Printf Extension Example | |
3097 | @subsection @code{printf} Extension Example | |
3098 | ||
3099 | Here is an example showing how to define a @code{printf} handler function. | |
19c3f208 | 3100 | This program defines a data structure called a @code{Widget} and |
28f540f4 RM |
3101 | defines the @samp{%W} conversion to print information about @w{@code{Widget *}} |
3102 | arguments, including the pointer value and the name stored in the data | |
3103 | structure. The @samp{%W} conversion supports the minimum field width and | |
3104 | left-justification options, but ignores everything else. | |
3105 | ||
3106 | @smallexample | |
3107 | @include rprintf.c.texi | |
3108 | @end smallexample | |
3109 | ||
3110 | The output produced by this program looks like: | |
3111 | ||
3112 | @smallexample | |
3113 | |<Widget 0xffeffb7c: mywidget>| | |
3114 | | <Widget 0xffeffb7c: mywidget>| | |
3115 | |<Widget 0xffeffb7c: mywidget> | | |
3116 | @end smallexample | |
3117 | ||
29bb8719 UD |
3118 | @node Predefined Printf Handlers |
3119 | @subsection Predefined @code{printf} Handlers | |
3120 | ||
1f77f049 | 3121 | @Theglibc{} also contains a concrete and useful application of the |
29bb8719 UD |
3122 | @code{printf} handler extension. There are two functions available |
3123 | which implement a special way to print floating-point numbers. | |
3124 | ||
3125 | @comment printf.h | |
3126 | @comment GNU | |
3127 | @deftypefun int printf_size (FILE *@var{fp}, const struct printf_info *@var{info}, const void *const *@var{args}) | |
3128 | Print a given floating point number as for the format @code{%f} except | |
3129 | that there is a postfix character indicating the divisor for the | |
3130 | number to make this less than 1000. There are two possible divisors: | |
04b9968b | 3131 | powers of 1024 or powers of 1000. Which one is used depends on the |
29bb8719 UD |
3132 | format character specified while registered this handler. If the |
3133 | character is of lower case, 1024 is used. For upper case characters, | |
3134 | 1000 is used. | |
3135 | ||
3136 | The postfix tag corresponds to bytes, kilobytes, megabytes, gigabytes, | |
3137 | etc. The full table is: | |
3138 | ||
779ae82e | 3139 | @ifinfo |
029264d1 | 3140 | @multitable {' '} {2^10 (1024)} {zetta} {Upper} {10^24 (1000)} |
29bb8719 UD |
3141 | @item low @tab Multiplier @tab From @tab Upper @tab Multiplier |
3142 | @item ' ' @tab 1 @tab @tab ' ' @tab 1 | |
3143 | @item k @tab 2^10 (1024) @tab kilo @tab K @tab 10^3 (1000) | |
3144 | @item m @tab 2^20 @tab mega @tab M @tab 10^6 | |
3145 | @item g @tab 2^30 @tab giga @tab G @tab 10^9 | |
3146 | @item t @tab 2^40 @tab tera @tab T @tab 10^12 | |
3147 | @item p @tab 2^50 @tab peta @tab P @tab 10^15 | |
3148 | @item e @tab 2^60 @tab exa @tab E @tab 10^18 | |
3149 | @item z @tab 2^70 @tab zetta @tab Z @tab 10^21 | |
3150 | @item y @tab 2^80 @tab yotta @tab Y @tab 10^24 | |
3151 | @end multitable | |
779ae82e UD |
3152 | @end ifinfo |
3153 | @iftex | |
3154 | @tex | |
3155 | \hbox to\hsize{\hfil\vbox{\offinterlineskip | |
3156 | \hrule | |
3157 | \halign{\strut#& \vrule#\tabskip=1em plus2em& {\tt#}\hfil& \vrule#& #\hfil& \vrule#& #\hfil& \vrule#& {\tt#}\hfil& \vrule#& #\hfil& \vrule#\tabskip=0pt\cr | |
3158 | \noalign{\hrule} | |
3159 | \omit&height2pt&\omit&&\omit&&\omit&&\omit&&\omit&\cr | |
3160 | && \omit low && Multiplier && From && \omit Upper && Multiplier &\cr | |
3161 | \omit&height2pt&\omit&&\omit&&\omit&&\omit&&\omit&\cr | |
3162 | \noalign{\hrule} | |
3163 | && {\tt\char32} && 1 && && {\tt\char32} && 1 &\cr | |
3164 | && k && $2^{10} = 1024$ && kilo && K && $10^3 = 1000$ &\cr | |
3165 | && m && $2^{20}$ && mega && M && $10^6$ &\cr | |
3166 | && g && $2^{30}$ && giga && G && $10^9$ &\cr | |
3167 | && t && $2^{40}$ && tera && T && $10^{12}$ &\cr | |
3168 | && p && $2^{50}$ && peta && P && $10^{15}$ &\cr | |
3169 | && e && $2^{60}$ && exa && E && $10^{18}$ &\cr | |
3170 | && z && $2^{70}$ && zetta && Z && $10^{21}$ &\cr | |
3171 | && y && $2^{80}$ && yotta && Y && $10^{24}$ &\cr | |
3172 | \noalign{\hrule}}}\hfil} | |
3173 | @end tex | |
3174 | @end iftex | |
29bb8719 UD |
3175 | |
3176 | The default precision is 3, i.e., 1024 is printed with a lower-case | |
3177 | format character as if it were @code{%.3fk} and will yield @code{1.000k}. | |
3178 | @end deftypefun | |
3179 | ||
3180 | Due to the requirements of @code{register_printf_function} we must also | |
04b9968b | 3181 | provide the function which returns information about the arguments. |
29bb8719 UD |
3182 | |
3183 | @comment printf.h | |
3184 | @comment GNU | |
3185 | @deftypefun int printf_size_info (const struct printf_info *@var{info}, size_t @var{n}, int *@var{argtypes}) | |
3186 | This function will return in @var{argtypes} the information about the | |
3187 | used parameters in the way the @code{vfprintf} implementation expects | |
3188 | it. The format always takes one argument. | |
3189 | @end deftypefun | |
3190 | ||
3191 | To use these functions both functions must be registered with a call like | |
3192 | ||
3193 | @smallexample | |
3194 | register_printf_function ('B', printf_size, printf_size_info); | |
3195 | @end smallexample | |
3196 | ||
3197 | Here we register the functions to print numbers as powers of 1000 since | |
f2ea0f5b | 3198 | the format character @code{'B'} is an upper-case character. If we |
29bb8719 UD |
3199 | would additionally use @code{'b'} in a line like |
3200 | ||
3201 | @smallexample | |
3202 | register_printf_function ('b', printf_size, printf_size_info); | |
3203 | @end smallexample | |
3204 | ||
3205 | @noindent | |
04b9968b UD |
3206 | we could also print using a power of 1024. Please note that all that is |
3207 | different in these two lines is the format specifier. The | |
3208 | @code{printf_size} function knows about the difference between lower and upper | |
29bb8719 UD |
3209 | case format specifiers. |
3210 | ||
3211 | The use of @code{'B'} and @code{'b'} is no coincidence. Rather it is | |
3212 | the preferred way to use this functionality since it is available on | |
04b9968b | 3213 | some other systems which also use format specifiers. |
29bb8719 | 3214 | |
28f540f4 RM |
3215 | @node Formatted Input |
3216 | @section Formatted Input | |
3217 | ||
3218 | @cindex formatted input from a stream | |
3219 | @cindex reading from a stream, formatted | |
3220 | @cindex format string, for @code{scanf} | |
3221 | @cindex template, for @code{scanf} | |
3222 | The functions described in this section (@code{scanf} and related | |
3223 | functions) provide facilities for formatted input analogous to the | |
3224 | formatted output facilities. These functions provide a mechanism for | |
3225 | reading arbitrary values under the control of a @dfn{format string} or | |
3226 | @dfn{template string}. | |
3227 | ||
3228 | @menu | |
3229 | * Formatted Input Basics:: Some basics to get you started. | |
3230 | * Input Conversion Syntax:: Syntax of conversion specifications. | |
3231 | * Table of Input Conversions:: Summary of input conversions and what they do. | |
3232 | * Numeric Input Conversions:: Details of conversions for reading numbers. | |
3233 | * String Input Conversions:: Details of conversions for reading strings. | |
3234 | * Dynamic String Input:: String conversions that @code{malloc} the buffer. | |
3235 | * Other Input Conversions:: Details of miscellaneous other conversions. | |
3236 | * Formatted Input Functions:: Descriptions of the actual functions. | |
3237 | * Variable Arguments Input:: @code{vscanf} and friends. | |
3238 | @end menu | |
3239 | ||
3240 | @node Formatted Input Basics | |
3241 | @subsection Formatted Input Basics | |
3242 | ||
3243 | Calls to @code{scanf} are superficially similar to calls to | |
3244 | @code{printf} in that arbitrary arguments are read under the control of | |
3245 | a template string. While the syntax of the conversion specifications in | |
3246 | the template is very similar to that for @code{printf}, the | |
3247 | interpretation of the template is oriented more towards free-format | |
3248 | input and simple pattern matching, rather than fixed-field formatting. | |
3249 | For example, most @code{scanf} conversions skip over any amount of | |
3250 | ``white space'' (including spaces, tabs, and newlines) in the input | |
3251 | file, and there is no concept of precision for the numeric input | |
3252 | conversions as there is for the corresponding output conversions. | |
3253 | Ordinarily, non-whitespace characters in the template are expected to | |
3254 | match characters in the input stream exactly, but a matching failure is | |
3255 | distinct from an input error on the stream. | |
3256 | @cindex conversion specifications (@code{scanf}) | |
3257 | ||
3258 | Another area of difference between @code{scanf} and @code{printf} is | |
3259 | that you must remember to supply pointers rather than immediate values | |
3260 | as the optional arguments to @code{scanf}; the values that are read are | |
3261 | stored in the objects that the pointers point to. Even experienced | |
3262 | programmers tend to forget this occasionally, so if your program is | |
3263 | getting strange errors that seem to be related to @code{scanf}, you | |
3264 | might want to double-check this. | |
3265 | ||
3266 | When a @dfn{matching failure} occurs, @code{scanf} returns immediately, | |
3267 | leaving the first non-matching character as the next character to be | |
3268 | read from the stream. The normal return value from @code{scanf} is the | |
3269 | number of values that were assigned, so you can use this to determine if | |
3270 | a matching error happened before all the expected values were read. | |
3271 | @cindex matching failure, in @code{scanf} | |
3272 | ||
3273 | The @code{scanf} function is typically used for things like reading in | |
3274 | the contents of tables. For example, here is a function that uses | |
3275 | @code{scanf} to initialize an array of @code{double}: | |
3276 | ||
3277 | @smallexample | |
3278 | void | |
3279 | readarray (double *array, int n) | |
3280 | @{ | |
3281 | int i; | |
3282 | for (i=0; i<n; i++) | |
3283 | if (scanf (" %lf", &(array[i])) != 1) | |
3284 | invalid_input_error (); | |
3285 | @} | |
3286 | @end smallexample | |
3287 | ||
3288 | The formatted input functions are not used as frequently as the | |
3289 | formatted output functions. Partly, this is because it takes some care | |
3290 | to use them properly. Another reason is that it is difficult to recover | |
3291 | from a matching error. | |
3292 | ||
3293 | If you are trying to read input that doesn't match a single, fixed | |
3294 | pattern, you may be better off using a tool such as Flex to generate a | |
3295 | lexical scanner, or Bison to generate a parser, rather than using | |
9f508673 UD |
3296 | @code{scanf}. For more information about these tools, see @ref{Top, , , |
3297 | flex.info, Flex: The Lexical Scanner Generator}, and @ref{Top, , , | |
28f540f4 RM |
3298 | bison.info, The Bison Reference Manual}. |
3299 | ||
3300 | @node Input Conversion Syntax | |
3301 | @subsection Input Conversion Syntax | |
3302 | ||
3303 | A @code{scanf} template string is a string that contains ordinary | |
3304 | multibyte characters interspersed with conversion specifications that | |
3305 | start with @samp{%}. | |
3306 | ||
3307 | Any whitespace character (as defined by the @code{isspace} function; | |
3308 | @pxref{Classification of Characters}) in the template causes any number | |
3309 | of whitespace characters in the input stream to be read and discarded. | |
3310 | The whitespace characters that are matched need not be exactly the same | |
3311 | whitespace characters that appear in the template string. For example, | |
3312 | write @samp{ , } in the template to recognize a comma with optional | |
3313 | whitespace before and after. | |
3314 | ||
3315 | Other characters in the template string that are not part of conversion | |
3316 | specifications must match characters in the input stream exactly; if | |
3317 | this is not the case, a matching failure occurs. | |
3318 | ||
3319 | The conversion specifications in a @code{scanf} template string | |
3320 | have the general form: | |
3321 | ||
3322 | @smallexample | |
3323 | % @var{flags} @var{width} @var{type} @var{conversion} | |
3324 | @end smallexample | |
3325 | ||
3326 | In more detail, an input conversion specification consists of an initial | |
3327 | @samp{%} character followed in sequence by: | |
3328 | ||
3329 | @itemize @bullet | |
3330 | @item | |
3331 | An optional @dfn{flag character} @samp{*}, which says to ignore the text | |
3332 | read for this specification. When @code{scanf} finds a conversion | |
3333 | specification that uses this flag, it reads input as directed by the | |
3334 | rest of the conversion specification, but it discards this input, does | |
3335 | not use a pointer argument, and does not increment the count of | |
3336 | successful assignments. | |
3337 | @cindex flag character (@code{scanf}) | |
3338 | ||
3339 | @item | |
3340 | An optional flag character @samp{a} (valid with string conversions only) | |
3341 | which requests allocation of a buffer long enough to store the string in. | |
3342 | (This is a GNU extension.) | |
3343 | @xref{Dynamic String Input}. | |
3344 | ||
3345 | @item | |
3346 | An optional decimal integer that specifies the @dfn{maximum field | |
3347 | width}. Reading of characters from the input stream stops either when | |
3348 | this maximum is reached or when a non-matching character is found, | |
3349 | whichever happens first. Most conversions discard initial whitespace | |
3350 | characters (those that don't are explicitly documented), and these | |
3351 | discarded characters don't count towards the maximum field width. | |
3352 | String input conversions store a null character to mark the end of the | |
3353 | input; the maximum field width does not include this terminator. | |
3354 | @cindex maximum field width (@code{scanf}) | |
3355 | ||
3356 | @item | |
3357 | An optional @dfn{type modifier character}. For example, you can | |
3358 | specify a type modifier of @samp{l} with integer conversions such as | |
3359 | @samp{%d} to specify that the argument is a pointer to a @code{long int} | |
3360 | rather than a pointer to an @code{int}. | |
3361 | @cindex type modifier character (@code{scanf}) | |
3362 | ||
3363 | @item | |
3364 | A character that specifies the conversion to be applied. | |
3365 | @end itemize | |
3366 | ||
19c3f208 | 3367 | The exact options that are permitted and how they are interpreted vary |
28f540f4 RM |
3368 | between the different conversion specifiers. See the descriptions of the |
3369 | individual conversions for information about the particular options that | |
3370 | they allow. | |
3371 | ||
3372 | With the @samp{-Wformat} option, the GNU C compiler checks calls to | |
3373 | @code{scanf} and related functions. It examines the format string and | |
3374 | verifies that the correct number and types of arguments are supplied. | |
3375 | There is also a GNU C syntax to tell the compiler that a function you | |
19c3f208 | 3376 | write uses a @code{scanf}-style format string. |
28f540f4 RM |
3377 | @xref{Function Attributes, , Declaring Attributes of Functions, |
3378 | gcc.info, Using GNU CC}, for more information. | |
3379 | ||
3380 | @node Table of Input Conversions | |
3381 | @subsection Table of Input Conversions | |
3382 | @cindex input conversions, for @code{scanf} | |
3383 | ||
3384 | Here is a table that summarizes the various conversion specifications: | |
3385 | ||
3386 | @table @asis | |
3387 | @item @samp{%d} | |
3388 | Matches an optionally signed integer written in decimal. @xref{Numeric | |
3389 | Input Conversions}. | |
3390 | ||
3391 | @item @samp{%i} | |
3392 | Matches an optionally signed integer in any of the formats that the C | |
3393 | language defines for specifying an integer constant. @xref{Numeric | |
3394 | Input Conversions}. | |
3395 | ||
3396 | @item @samp{%o} | |
3397 | Matches an unsigned integer written in octal radix. | |
3398 | @xref{Numeric Input Conversions}. | |
3399 | ||
3400 | @item @samp{%u} | |
3401 | Matches an unsigned integer written in decimal radix. | |
3402 | @xref{Numeric Input Conversions}. | |
3403 | ||
3404 | @item @samp{%x}, @samp{%X} | |
3405 | Matches an unsigned integer written in hexadecimal radix. | |
3406 | @xref{Numeric Input Conversions}. | |
3407 | ||
3408 | @item @samp{%e}, @samp{%f}, @samp{%g}, @samp{%E}, @samp{%G} | |
3409 | Matches an optionally signed floating-point number. @xref{Numeric Input | |
3410 | Conversions}. | |
3411 | ||
3412 | @item @samp{%s} | |
b5e73f56 | 3413 | |
28f540f4 | 3414 | Matches a string containing only non-whitespace characters. |
b5e73f56 UD |
3415 | @xref{String Input Conversions}. The presence of the @samp{l} modifier |
3416 | determines whether the output is stored as a wide character string or a | |
3417 | multibyte string. If @samp{%s} is used in a wide character function the | |
3418 | string is converted as with multiple calls to @code{wcrtomb} into a | |
3419 | multibyte string. This means that the buffer must provide room for | |
3420 | @code{MB_CUR_MAX} bytes for each wide character read. In case | |
3421 | @samp{%ls} is used in a multibyte function the result is converted into | |
3422 | wide characters as with multiple calls of @code{mbrtowc} before being | |
3423 | stored in the user provided buffer. | |
3424 | ||
3425 | @item @samp{%S} | |
3426 | This is an alias for @samp{%ls} which is supported for compatibility | |
3427 | with the Unix standard. | |
28f540f4 RM |
3428 | |
3429 | @item @samp{%[} | |
3430 | Matches a string of characters that belong to a specified set. | |
b5e73f56 UD |
3431 | @xref{String Input Conversions}. The presence of the @samp{l} modifier |
3432 | determines whether the output is stored as a wide character string or a | |
3433 | multibyte string. If @samp{%[} is used in a wide character function the | |
3434 | string is converted as with multiple calls to @code{wcrtomb} into a | |
3435 | multibyte string. This means that the buffer must provide room for | |
3436 | @code{MB_CUR_MAX} bytes for each wide character read. In case | |
3437 | @samp{%l[} is used in a multibyte function the result is converted into | |
3438 | wide characters as with multiple calls of @code{mbrtowc} before being | |
3439 | stored in the user provided buffer. | |
28f540f4 RM |
3440 | |
3441 | @item @samp{%c} | |
3442 | Matches a string of one or more characters; the number of characters | |
3443 | read is controlled by the maximum field width given for the conversion. | |
3444 | @xref{String Input Conversions}. | |
3445 | ||
b5e73f56 UD |
3446 | If the @samp{%c} is used in a wide stream function the read value is |
3447 | converted from a wide character to the corresponding multibyte character | |
3448 | before storing it. Note that this conversion can produce more than one | |
3449 | byte of output and therefore the provided buffer be large enough for up | |
3450 | to @code{MB_CUR_MAX} bytes for each character. If @samp{%lc} is used in | |
3451 | a multibyte function the input is treated as a multibyte sequence (and | |
3452 | not bytes) and the result is converted as with calls to @code{mbrtowc}. | |
3453 | ||
3454 | @item @samp{%C} | |
3455 | This is an alias for @samp{%lc} which is supported for compatibility | |
3456 | with the Unix standard. | |
3457 | ||
28f540f4 RM |
3458 | @item @samp{%p} |
3459 | Matches a pointer value in the same implementation-defined format used | |
3460 | by the @samp{%p} output conversion for @code{printf}. @xref{Other Input | |
3461 | Conversions}. | |
3462 | ||
3463 | @item @samp{%n} | |
3464 | This conversion doesn't read any characters; it records the number of | |
3465 | characters read so far by this call. @xref{Other Input Conversions}. | |
3466 | ||
3467 | @item @samp{%%} | |
3468 | This matches a literal @samp{%} character in the input stream. No | |
3469 | corresponding argument is used. @xref{Other Input Conversions}. | |
3470 | @end table | |
3471 | ||
3472 | If the syntax of a conversion specification is invalid, the behavior is | |
3473 | undefined. If there aren't enough function arguments provided to supply | |
3474 | addresses for all the conversion specifications in the template strings | |
3475 | that perform assignments, or if the arguments are not of the correct | |
3476 | types, the behavior is also undefined. On the other hand, extra | |
3477 | arguments are simply ignored. | |
3478 | ||
3479 | @node Numeric Input Conversions | |
3480 | @subsection Numeric Input Conversions | |
3481 | ||
3482 | This section describes the @code{scanf} conversions for reading numeric | |
3483 | values. | |
3484 | ||
3485 | The @samp{%d} conversion matches an optionally signed integer in decimal | |
3486 | radix. The syntax that is recognized is the same as that for the | |
3487 | @code{strtol} function (@pxref{Parsing of Integers}) with the value | |
3488 | @code{10} for the @var{base} argument. | |
3489 | ||
3490 | The @samp{%i} conversion matches an optionally signed integer in any of | |
3491 | the formats that the C language defines for specifying an integer | |
3492 | constant. The syntax that is recognized is the same as that for the | |
3493 | @code{strtol} function (@pxref{Parsing of Integers}) with the value | |
3494 | @code{0} for the @var{base} argument. (You can print integers in this | |
3495 | syntax with @code{printf} by using the @samp{#} flag character with the | |
3496 | @samp{%x}, @samp{%o}, or @samp{%d} conversion. @xref{Integer Conversions}.) | |
3497 | ||
3498 | For example, any of the strings @samp{10}, @samp{0xa}, or @samp{012} | |
3499 | could be read in as integers under the @samp{%i} conversion. Each of | |
3500 | these specifies a number with decimal value @code{10}. | |
3501 | ||
3502 | The @samp{%o}, @samp{%u}, and @samp{%x} conversions match unsigned | |
3503 | integers in octal, decimal, and hexadecimal radices, respectively. The | |
3504 | syntax that is recognized is the same as that for the @code{strtoul} | |
3505 | function (@pxref{Parsing of Integers}) with the appropriate value | |
3506 | (@code{8}, @code{10}, or @code{16}) for the @var{base} argument. | |
3507 | ||
3508 | The @samp{%X} conversion is identical to the @samp{%x} conversion. They | |
3509 | both permit either uppercase or lowercase letters to be used as digits. | |
3510 | ||
3511 | The default type of the corresponding argument for the @code{%d} and | |
3512 | @code{%i} conversions is @code{int *}, and @code{unsigned int *} for the | |
3513 | other integer conversions. You can use the following type modifiers to | |
3514 | specify other sizes of integer: | |
3515 | ||
3516 | @table @samp | |
cc3fa755 UD |
3517 | @item hh |
3518 | Specifies that the argument is a @code{signed char *} or @code{unsigned | |
3519 | char *}. | |
3520 | ||
ec751a23 | 3521 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 3522 | |
28f540f4 RM |
3523 | @item h |
3524 | Specifies that the argument is a @code{short int *} or @code{unsigned | |
3525 | short int *}. | |
3526 | ||
e852e889 UD |
3527 | @item j |
3528 | Specifies that the argument is a @code{intmax_t *} or @code{uintmax_t *}. | |
3529 | ||
ec751a23 | 3530 | This modifier was introduced in @w{ISO C99}. |
e852e889 | 3531 | |
28f540f4 RM |
3532 | @item l |
3533 | Specifies that the argument is a @code{long int *} or @code{unsigned | |
3534 | long int *}. Two @samp{l} characters is like the @samp{L} modifier, below. | |
3535 | ||
b5e73f56 UD |
3536 | If used with @samp{%c} or @samp{%s} the corresponding parameter is |
3537 | considered as a pointer to a wide character or wide character string | |
3538 | respectively. This use of @samp{l} was introduced in @w{Amendment 1} to | |
3539 | @w{ISO C90}. | |
3540 | ||
28f540f4 RM |
3541 | @need 100 |
3542 | @item ll | |
3543 | @itemx L | |
3544 | @itemx q | |
3545 | 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 | |
3546 | GNU C compiler. For systems that don't provide extra-long integers, this | |
3547 | is the same as @code{long int}.) | |
3548 | ||
3549 | The @samp{q} modifier is another name for the same thing, which comes | |
3550 | from 4.4 BSD; a @w{@code{long long int}} is sometimes called a ``quad'' | |
3551 | @code{int}. | |
e852e889 UD |
3552 | |
3553 | @item t | |
3554 | Specifies that the argument is a @code{ptrdiff_t *}. | |
3555 | ||
ec751a23 | 3556 | This modifier was introduced in @w{ISO C99}. |
e852e889 UD |
3557 | |
3558 | @item z | |
3559 | Specifies that the argument is a @code{size_t *}. | |
3560 | ||
ec751a23 | 3561 | This modifier was introduced in @w{ISO C99}. |
28f540f4 RM |
3562 | @end table |
3563 | ||
3564 | All of the @samp{%e}, @samp{%f}, @samp{%g}, @samp{%E}, and @samp{%G} | |
3565 | input conversions are interchangeable. They all match an optionally | |
3566 | signed floating point number, in the same syntax as for the | |
3567 | @code{strtod} function (@pxref{Parsing of Floats}). | |
3568 | ||
3569 | For the floating-point input conversions, the default argument type is | |
3570 | @code{float *}. (This is different from the corresponding output | |
3571 | conversions, where the default type is @code{double}; remember that | |
3572 | @code{float} arguments to @code{printf} are converted to @code{double} | |
3573 | by the default argument promotions, but @code{float *} arguments are | |
3574 | not promoted to @code{double *}.) You can specify other sizes of float | |
3575 | using these type modifiers: | |
3576 | ||
3577 | @table @samp | |
3578 | @item l | |
3579 | Specifies that the argument is of type @code{double *}. | |
3580 | ||
3581 | @item L | |
3582 | Specifies that the argument is of type @code{long double *}. | |
3583 | @end table | |
3584 | ||
2c6fe0bd UD |
3585 | For all the above number parsing formats there is an additional optional |
3586 | flag @samp{'}. When this flag is given the @code{scanf} function | |
3587 | expects the number represented in the input string to be formatted | |
3588 | according to the grouping rules of the currently selected locale | |
3589 | (@pxref{General Numeric}). | |
3590 | ||
3591 | If the @code{"C"} or @code{"POSIX"} locale is selected there is no | |
3592 | difference. But for a locale which specifies values for the appropriate | |
3593 | fields in the locale the input must have the correct form in the input. | |
3594 | Otherwise the longest prefix with a correct form is processed. | |
3595 | ||
28f540f4 RM |
3596 | @node String Input Conversions |
3597 | @subsection String Input Conversions | |
3598 | ||
3599 | This section describes the @code{scanf} input conversions for reading | |
b5e73f56 UD |
3600 | string and character values: @samp{%s}, @samp{%S}, @samp{%[}, @samp{%c}, |
3601 | and @samp{%C}. | |
28f540f4 RM |
3602 | |
3603 | You have two options for how to receive the input from these | |
3604 | conversions: | |
3605 | ||
3606 | @itemize @bullet | |
3607 | @item | |
b5e73f56 UD |
3608 | Provide a buffer to store it in. This is the default. You should |
3609 | provide an argument of type @code{char *} or @code{wchar_t *} (the | |
3610 | latter of the @samp{l} modifier is present). | |
28f540f4 RM |
3611 | |
3612 | @strong{Warning:} To make a robust program, you must make sure that the | |
3613 | input (plus its terminating null) cannot possibly exceed the size of the | |
3614 | buffer you provide. In general, the only way to do this is to specify a | |
3615 | maximum field width one less than the buffer size. @strong{If you | |
3616 | provide the buffer, always specify a maximum field width to prevent | |
3617 | overflow.} | |
3618 | ||
3619 | @item | |
3620 | Ask @code{scanf} to allocate a big enough buffer, by specifying the | |
3621 | @samp{a} flag character. This is a GNU extension. You should provide | |
3622 | an argument of type @code{char **} for the buffer address to be stored | |
3623 | in. @xref{Dynamic String Input}. | |
3624 | @end itemize | |
3625 | ||
3626 | The @samp{%c} conversion is the simplest: it matches a fixed number of | |
04b9968b | 3627 | characters, always. The maximum field width says how many characters to |
28f540f4 RM |
3628 | read; if you don't specify the maximum, the default is 1. This |
3629 | conversion doesn't append a null character to the end of the text it | |
3630 | reads. It also does not skip over initial whitespace characters. It | |
3631 | reads precisely the next @var{n} characters, and fails if it cannot get | |
3632 | that many. Since there is always a maximum field width with @samp{%c} | |
3633 | (whether specified, or 1 by default), you can always prevent overflow by | |
3634 | making the buffer long enough. | |
b5e73f56 UD |
3635 | @comment Is character == byte here??? --drepper |
3636 | ||
3637 | If the format is @samp{%lc} or @samp{%C} the function stores wide | |
3638 | characters which are converted using the conversion determined at the | |
3639 | time the stream was opened from the external byte stream. The number of | |
3640 | bytes read from the medium is limited by @code{MB_CUR_LEN * @var{n}} but | |
3641 | at most @var{n} wide character get stored in the output string. | |
28f540f4 RM |
3642 | |
3643 | The @samp{%s} conversion matches a string of non-whitespace characters. | |
3644 | It skips and discards initial whitespace, but stops when it encounters | |
3645 | more whitespace after having read something. It stores a null character | |
3646 | at the end of the text that it reads. | |
3647 | ||
3648 | For example, reading the input: | |
3649 | ||
3650 | @smallexample | |
3651 | hello, world | |
3652 | @end smallexample | |
3653 | ||
3654 | @noindent | |
3655 | with the conversion @samp{%10c} produces @code{" hello, wo"}, but | |
3656 | reading the same input with the conversion @samp{%10s} produces | |
3657 | @code{"hello,"}. | |
3658 | ||
3659 | @strong{Warning:} If you do not specify a field width for @samp{%s}, | |
3660 | then the number of characters read is limited only by where the next | |
3661 | whitespace character appears. This almost certainly means that invalid | |
3662 | input can make your program crash---which is a bug. | |
3663 | ||
b5e73f56 UD |
3664 | The @samp{%ls} and @samp{%S} format are handled just like @samp{%s} |
3665 | except that the external byte sequence is converted using the conversion | |
3666 | associated with the stream to wide characters with their own encoding. | |
3667 | A width or precision specified with the format do not directly determine | |
3668 | how many bytes are read from the stream since they measure wide | |
3669 | characters. But an upper limit can be computed by multiplying the value | |
3670 | of the width or precision by @code{MB_CUR_MAX}. | |
3671 | ||
28f540f4 RM |
3672 | To read in characters that belong to an arbitrary set of your choice, |
3673 | use the @samp{%[} conversion. You specify the set between the @samp{[} | |
3674 | character and a following @samp{]} character, using the same syntax used | |
9878ad46 | 3675 | in regular expressions for explicit sets of characters. As special cases: |
28f540f4 RM |
3676 | |
3677 | @itemize @bullet | |
19c3f208 | 3678 | @item |
28f540f4 RM |
3679 | A literal @samp{]} character can be specified as the first character |
3680 | of the set. | |
3681 | ||
19c3f208 | 3682 | @item |
28f540f4 RM |
3683 | An embedded @samp{-} character (that is, one that is not the first or |
3684 | last character of the set) is used to specify a range of characters. | |
3685 | ||
19c3f208 | 3686 | @item |
28f540f4 RM |
3687 | If a caret character @samp{^} immediately follows the initial @samp{[}, |
3688 | then the set of allowed input characters is the everything @emph{except} | |
3689 | the characters listed. | |
3690 | @end itemize | |
3691 | ||
3692 | The @samp{%[} conversion does not skip over initial whitespace | |
3693 | characters. | |
3694 | ||
9878ad46 OB |
3695 | Note that the @dfn{character class} syntax available in character sets |
3696 | that appear inside regular expressions (such as @samp{[:alpha:]}) is | |
3697 | @emph{not} available in the @samp{%[} conversion. | |
3698 | ||
28f540f4 RM |
3699 | Here are some examples of @samp{%[} conversions and what they mean: |
3700 | ||
3701 | @table @samp | |
3702 | @item %25[1234567890] | |
3703 | Matches a string of up to 25 digits. | |
3704 | ||
3705 | @item %25[][] | |
3706 | Matches a string of up to 25 square brackets. | |
3707 | ||
3708 | @item %25[^ \f\n\r\t\v] | |
3709 | Matches a string up to 25 characters long that doesn't contain any of | |
3710 | the standard whitespace characters. This is slightly different from | |
3711 | @samp{%s}, because if the input begins with a whitespace character, | |
3712 | @samp{%[} reports a matching failure while @samp{%s} simply discards the | |
3713 | initial whitespace. | |
3714 | ||
19c3f208 | 3715 | @item %25[a-z] |
28f540f4 RM |
3716 | Matches up to 25 lowercase characters. |
3717 | @end table | |
3718 | ||
b5e73f56 UD |
3719 | As for @samp{%c} and @samp{%s} the @samp{%[} format is also modified to |
3720 | produce wide characters if the @samp{l} modifier is present. All what | |
3721 | is said about @samp{%ls} above is true for @samp{%l[}. | |
3722 | ||
28f540f4 RM |
3723 | One more reminder: the @samp{%s} and @samp{%[} conversions are |
3724 | @strong{dangerous} if you don't specify a maximum width or use the | |
3725 | @samp{a} flag, because input too long would overflow whatever buffer you | |
3726 | have provided for it. No matter how long your buffer is, a user could | |
3727 | supply input that is longer. A well-written program reports invalid | |
3728 | input with a comprehensible error message, not with a crash. | |
3729 | ||
3730 | @node Dynamic String Input | |
3731 | @subsection Dynamically Allocating String Conversions | |
3732 | ||
3733 | A GNU extension to formatted input lets you safely read a string with no | |
3734 | maximum size. Using this feature, you don't supply a buffer; instead, | |
3735 | @code{scanf} allocates a buffer big enough to hold the data and gives | |
3736 | you its address. To use this feature, write @samp{a} as a flag | |
3737 | character, as in @samp{%as} or @samp{%a[0-9a-z]}. | |
3738 | ||
3739 | The pointer argument you supply for where to store the input should have | |
3740 | type @code{char **}. The @code{scanf} function allocates a buffer and | |
3741 | stores its address in the word that the argument points to. You should | |
3742 | free the buffer with @code{free} when you no longer need it. | |
3743 | ||
3744 | Here is an example of using the @samp{a} flag with the @samp{%[@dots{}]} | |
3745 | conversion specification to read a ``variable assignment'' of the form | |
3746 | @samp{@var{variable} = @var{value}}. | |
3747 | ||
3748 | @smallexample | |
3749 | @{ | |
3750 | char *variable, *value; | |
3751 | ||
3752 | if (2 > scanf ("%a[a-zA-Z0-9] = %a[^\n]\n", | |
162ba701 | 3753 | &variable, &value)) |
28f540f4 RM |
3754 | @{ |
3755 | invalid_input_error (); | |
3756 | return 0; | |
3757 | @} | |
3758 | ||
3759 | @dots{} | |
3760 | @} | |
3761 | @end smallexample | |
3762 | ||
3763 | @node Other Input Conversions | |
3764 | @subsection Other Input Conversions | |
3765 | ||
3766 | This section describes the miscellaneous input conversions. | |
3767 | ||
3768 | The @samp{%p} conversion is used to read a pointer value. It recognizes | |
04b9968b | 3769 | the same syntax used by the @samp{%p} output conversion for |
28f540f4 RM |
3770 | @code{printf} (@pxref{Other Output Conversions}); that is, a hexadecimal |
3771 | number just as the @samp{%x} conversion accepts. The corresponding | |
3772 | argument should be of type @code{void **}; that is, the address of a | |
3773 | place to store a pointer. | |
3774 | ||
3775 | The resulting pointer value is not guaranteed to be valid if it was not | |
3776 | originally written during the same program execution that reads it in. | |
3777 | ||
3778 | The @samp{%n} conversion produces the number of characters read so far | |
3779 | by this call. The corresponding argument should be of type @code{int *}. | |
3780 | This conversion works in the same way as the @samp{%n} conversion for | |
3781 | @code{printf}; see @ref{Other Output Conversions}, for an example. | |
3782 | ||
3783 | The @samp{%n} conversion is the only mechanism for determining the | |
3784 | success of literal matches or conversions with suppressed assignments. | |
3785 | If the @samp{%n} follows the locus of a matching failure, then no value | |
3786 | is stored for it since @code{scanf} returns before processing the | |
3787 | @samp{%n}. If you store @code{-1} in that argument slot before calling | |
3788 | @code{scanf}, the presence of @code{-1} after @code{scanf} indicates an | |
3789 | error occurred before the @samp{%n} was reached. | |
3790 | ||
3791 | Finally, the @samp{%%} conversion matches a literal @samp{%} character | |
3792 | in the input stream, without using an argument. This conversion does | |
3793 | not permit any flags, field width, or type modifier to be specified. | |
3794 | ||
3795 | @node Formatted Input Functions | |
3796 | @subsection Formatted Input Functions | |
3797 | ||
3798 | Here are the descriptions of the functions for performing formatted | |
3799 | input. | |
3800 | Prototypes for these functions are in the header file @file{stdio.h}. | |
3801 | @pindex stdio.h | |
3802 | ||
3803 | @comment stdio.h | |
f65fd747 | 3804 | @comment ISO |
28f540f4 RM |
3805 | @deftypefun int scanf (const char *@var{template}, @dots{}) |
3806 | The @code{scanf} function reads formatted input from the stream | |
3807 | @code{stdin} under the control of the template string @var{template}. | |
3808 | The optional arguments are pointers to the places which receive the | |
3809 | resulting values. | |
3810 | ||
3811 | The return value is normally the number of successful assignments. If | |
04b9968b UD |
3812 | an end-of-file condition is detected before any matches are performed, |
3813 | including matches against whitespace and literal characters in the | |
3814 | template, then @code{EOF} is returned. | |
28f540f4 RM |
3815 | @end deftypefun |
3816 | ||
b5e73f56 UD |
3817 | @comment wchar.h |
3818 | @comment ISO | |
3819 | @deftypefun int wscanf (const wchar_t *@var{template}, @dots{}) | |
3820 | The @code{wscanf} function reads formatted input from the stream | |
3821 | @code{stdin} under the control of the template string @var{template}. | |
3822 | The optional arguments are pointers to the places which receive the | |
3823 | resulting values. | |
3824 | ||
3825 | The return value is normally the number of successful assignments. If | |
3826 | an end-of-file condition is detected before any matches are performed, | |
3827 | including matches against whitespace and literal characters in the | |
3828 | template, then @code{WEOF} is returned. | |
3829 | @end deftypefun | |
3830 | ||
28f540f4 | 3831 | @comment stdio.h |
f65fd747 | 3832 | @comment ISO |
28f540f4 RM |
3833 | @deftypefun int fscanf (FILE *@var{stream}, const char *@var{template}, @dots{}) |
3834 | This function is just like @code{scanf}, except that the input is read | |
3835 | from the stream @var{stream} instead of @code{stdin}. | |
3836 | @end deftypefun | |
3837 | ||
b5e73f56 UD |
3838 | @comment wchar.h |
3839 | @comment ISO | |
3840 | @deftypefun int fwscanf (FILE *@var{stream}, const wchar_t *@var{template}, @dots{}) | |
3841 | This function is just like @code{wscanf}, except that the input is read | |
3842 | from the stream @var{stream} instead of @code{stdin}. | |
3843 | @end deftypefun | |
3844 | ||
28f540f4 | 3845 | @comment stdio.h |
f65fd747 | 3846 | @comment ISO |
28f540f4 RM |
3847 | @deftypefun int sscanf (const char *@var{s}, const char *@var{template}, @dots{}) |
3848 | This is like @code{scanf}, except that the characters are taken from the | |
3849 | null-terminated string @var{s} instead of from a stream. Reaching the | |
3850 | end of the string is treated as an end-of-file condition. | |
3851 | ||
3852 | The behavior of this function is undefined if copying takes place | |
3853 | between objects that overlap---for example, if @var{s} is also given | |
b5e73f56 UD |
3854 | as an argument to receive a string read under control of the @samp{%s}, |
3855 | @samp{%S}, or @samp{%[} conversion. | |
3856 | @end deftypefun | |
3857 | ||
3858 | @comment wchar.h | |
3859 | @comment ISO | |
8ded91fb | 3860 | @deftypefun int swscanf (const wchar_t *@var{ws}, const wchar_t *@var{template}, @dots{}) |
b5e73f56 UD |
3861 | This is like @code{wscanf}, except that the characters are taken from the |
3862 | null-terminated string @var{ws} instead of from a stream. Reaching the | |
3863 | end of the string is treated as an end-of-file condition. | |
3864 | ||
3865 | The behavior of this function is undefined if copying takes place | |
3866 | between objects that overlap---for example, if @var{ws} is also given as | |
3867 | an argument to receive a string read under control of the @samp{%s}, | |
3868 | @samp{%S}, or @samp{%[} conversion. | |
28f540f4 RM |
3869 | @end deftypefun |
3870 | ||
3871 | @node Variable Arguments Input | |
3872 | @subsection Variable Arguments Input Functions | |
3873 | ||
3874 | The functions @code{vscanf} and friends are provided so that you can | |
3875 | define your own variadic @code{scanf}-like functions that make use of | |
3876 | the same internals as the built-in formatted output functions. | |
3877 | These functions are analogous to the @code{vprintf} series of output | |
3878 | functions. @xref{Variable Arguments Output}, for important | |
3879 | information on how to use them. | |
3880 | ||
b5e73f56 UD |
3881 | @strong{Portability Note:} The functions listed in this section were |
3882 | introduced in @w{ISO C99} and were before available as GNU extensions. | |
28f540f4 RM |
3883 | |
3884 | @comment stdio.h | |
b5e73f56 | 3885 | @comment ISO |
28f540f4 | 3886 | @deftypefun int vscanf (const char *@var{template}, va_list @var{ap}) |
04b9968b | 3887 | This function is similar to @code{scanf}, but instead of taking |
28f540f4 RM |
3888 | a variable number of arguments directly, it takes an argument list |
3889 | pointer @var{ap} of type @code{va_list} (@pxref{Variadic Functions}). | |
3890 | @end deftypefun | |
3891 | ||
b5e73f56 UD |
3892 | @comment wchar.h |
3893 | @comment ISO | |
3894 | @deftypefun int vwscanf (const wchar_t *@var{template}, va_list @var{ap}) | |
3895 | This function is similar to @code{wscanf}, but instead of taking | |
3896 | a variable number of arguments directly, it takes an argument list | |
3897 | pointer @var{ap} of type @code{va_list} (@pxref{Variadic Functions}). | |
3898 | @end deftypefun | |
3899 | ||
28f540f4 | 3900 | @comment stdio.h |
b5e73f56 | 3901 | @comment ISO |
28f540f4 RM |
3902 | @deftypefun int vfscanf (FILE *@var{stream}, const char *@var{template}, va_list @var{ap}) |
3903 | This is the equivalent of @code{fscanf} with the variable argument list | |
3904 | specified directly as for @code{vscanf}. | |
3905 | @end deftypefun | |
3906 | ||
b5e73f56 UD |
3907 | @comment wchar.h |
3908 | @comment ISO | |
3909 | @deftypefun int vfwscanf (FILE *@var{stream}, const wchar_t *@var{template}, va_list @var{ap}) | |
3910 | This is the equivalent of @code{fwscanf} with the variable argument list | |
3911 | specified directly as for @code{vwscanf}. | |
3912 | @end deftypefun | |
3913 | ||
28f540f4 | 3914 | @comment stdio.h |
b5e73f56 | 3915 | @comment ISO |
28f540f4 RM |
3916 | @deftypefun int vsscanf (const char *@var{s}, const char *@var{template}, va_list @var{ap}) |
3917 | This is the equivalent of @code{sscanf} with the variable argument list | |
3918 | specified directly as for @code{vscanf}. | |
3919 | @end deftypefun | |
3920 | ||
b5e73f56 UD |
3921 | @comment wchar.h |
3922 | @comment ISO | |
3923 | @deftypefun int vswscanf (const wchar_t *@var{s}, const wchar_t *@var{template}, va_list @var{ap}) | |
3924 | This is the equivalent of @code{swscanf} with the variable argument list | |
3925 | specified directly as for @code{vwscanf}. | |
3926 | @end deftypefun | |
3927 | ||
28f540f4 RM |
3928 | In GNU C, there is a special construct you can use to let the compiler |
3929 | know that a function uses a @code{scanf}-style format string. Then it | |
3930 | can check the number and types of arguments in each call to the | |
3931 | function, and warn you when they do not match the format string. | |
88197030 | 3932 | For details, see @ref{Function Attributes, , Declaring Attributes of Functions, |
04b9968b | 3933 | gcc.info, Using GNU CC}. |
28f540f4 RM |
3934 | |
3935 | @node EOF and Errors | |
3936 | @section End-Of-File and Errors | |
3937 | ||
3938 | @cindex end of file, on a stream | |
3939 | Many of the functions described in this chapter return the value of the | |
3940 | macro @code{EOF} to indicate unsuccessful completion of the operation. | |
3941 | Since @code{EOF} is used to report both end of file and random errors, | |
3942 | it's often better to use the @code{feof} function to check explicitly | |
3943 | for end of file and @code{ferror} to check for errors. These functions | |
3944 | check indicators that are part of the internal state of the stream | |
3945 | object, indicators set if the appropriate condition was detected by a | |
3946 | previous I/O operation on that stream. | |
3947 | ||
28f540f4 | 3948 | @comment stdio.h |
f65fd747 | 3949 | @comment ISO |
28f540f4 | 3950 | @deftypevr Macro int EOF |
b5e73f56 UD |
3951 | This macro is an integer value that is returned by a number of narrow |
3952 | stream functions to indicate an end-of-file condition, or some other | |
1f77f049 | 3953 | error situation. With @theglibc{}, @code{EOF} is @code{-1}. In |
b5e73f56 UD |
3954 | other libraries, its value may be some other negative number. |
3955 | ||
3956 | This symbol is declared in @file{stdio.h}. | |
3957 | @end deftypevr | |
3958 | ||
3959 | @comment wchar.h | |
3960 | @comment ISO | |
3961 | @deftypevr Macro int WEOF | |
3962 | This macro is an integer value that is returned by a number of wide | |
3963 | stream functions to indicate an end-of-file condition, or some other | |
1f77f049 | 3964 | error situation. With @theglibc{}, @code{WEOF} is @code{-1}. In |
b5e73f56 UD |
3965 | other libraries, its value may be some other negative number. |
3966 | ||
3967 | This symbol is declared in @file{wchar.h}. | |
28f540f4 RM |
3968 | @end deftypevr |
3969 | ||
28f540f4 | 3970 | @comment stdio.h |
f65fd747 | 3971 | @comment ISO |
28f540f4 RM |
3972 | @deftypefun int feof (FILE *@var{stream}) |
3973 | The @code{feof} function returns nonzero if and only if the end-of-file | |
3974 | indicator for the stream @var{stream} is set. | |
b5e73f56 UD |
3975 | |
3976 | This symbol is declared in @file{stdio.h}. | |
28f540f4 RM |
3977 | @end deftypefun |
3978 | ||
7b4161bb UD |
3979 | @comment stdio.h |
3980 | @comment GNU | |
3981 | @deftypefun int feof_unlocked (FILE *@var{stream}) | |
3982 | The @code{feof_unlocked} function is equivalent to the @code{feof} | |
1dc843f7 | 3983 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
3984 | |
3985 | This function is a GNU extension. | |
b5e73f56 UD |
3986 | |
3987 | This symbol is declared in @file{stdio.h}. | |
7b4161bb UD |
3988 | @end deftypefun |
3989 | ||
28f540f4 | 3990 | @comment stdio.h |
f65fd747 | 3991 | @comment ISO |
28f540f4 RM |
3992 | @deftypefun int ferror (FILE *@var{stream}) |
3993 | The @code{ferror} function returns nonzero if and only if the error | |
3994 | indicator for the stream @var{stream} is set, indicating that an error | |
3995 | has occurred on a previous operation on the stream. | |
b5e73f56 UD |
3996 | |
3997 | This symbol is declared in @file{stdio.h}. | |
28f540f4 RM |
3998 | @end deftypefun |
3999 | ||
7b4161bb UD |
4000 | @comment stdio.h |
4001 | @comment GNU | |
4002 | @deftypefun int ferror_unlocked (FILE *@var{stream}) | |
4003 | The @code{ferror_unlocked} function is equivalent to the @code{ferror} | |
1dc843f7 | 4004 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
4005 | |
4006 | This function is a GNU extension. | |
b5e73f56 UD |
4007 | |
4008 | This symbol is declared in @file{stdio.h}. | |
7b4161bb UD |
4009 | @end deftypefun |
4010 | ||
28f540f4 RM |
4011 | In addition to setting the error indicator associated with the stream, |
4012 | the functions that operate on streams also set @code{errno} in the same | |
4013 | way as the corresponding low-level functions that operate on file | |
4014 | descriptors. For example, all of the functions that perform output to a | |
4015 | stream---such as @code{fputc}, @code{printf}, and @code{fflush}---are | |
4016 | implemented in terms of @code{write}, and all of the @code{errno} error | |
4017 | conditions defined for @code{write} are meaningful for these functions. | |
4018 | For more information about the descriptor-level I/O functions, see | |
4019 | @ref{Low-Level I/O}. | |
4020 | ||
460e040a UD |
4021 | @node Error Recovery |
4022 | @section Recovering from errors | |
4023 | ||
7ba73c63 | 4024 | You may explicitly clear the error and EOF flags with the @code{clearerr} |
460e040a UD |
4025 | function. |
4026 | ||
4027 | @comment stdio.h | |
4028 | @comment ISO | |
4029 | @deftypefun void clearerr (FILE *@var{stream}) | |
4030 | This function clears the end-of-file and error indicators for the | |
4031 | stream @var{stream}. | |
4032 | ||
4033 | The file positioning functions (@pxref{File Positioning}) also clear the | |
4034 | end-of-file indicator for the stream. | |
4035 | @end deftypefun | |
4036 | ||
7b4161bb UD |
4037 | @comment stdio.h |
4038 | @comment GNU | |
4039 | @deftypefun void clearerr_unlocked (FILE *@var{stream}) | |
4040 | The @code{clearerr_unlocked} function is equivalent to the @code{clearerr} | |
1dc843f7 | 4041 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
4042 | |
4043 | This function is a GNU extension. | |
4044 | @end deftypefun | |
4045 | ||
460e040a UD |
4046 | Note that it is @emph{not} correct to just clear the error flag and retry |
4047 | a failed stream operation. After a failed write, any number of | |
4048 | characters since the last buffer flush may have been committed to the | |
4049 | file, while some buffered data may have been discarded. Merely retrying | |
4050 | can thus cause lost or repeated data. | |
4051 | ||
4052 | A failed read may leave the file pointer in an inappropriate position for | |
4053 | a second try. In both cases, you should seek to a known position before | |
4054 | retrying. | |
4055 | ||
4056 | Most errors that can happen are not recoverable --- a second try will | |
4057 | always fail again in the same way. So usually it is best to give up and | |
4058 | report the error to the user, rather than install complicated recovery | |
4059 | logic. | |
4060 | ||
7ba73c63 | 4061 | One important exception is @code{EINTR} (@pxref{Interrupted Primitives}). |
460e040a UD |
4062 | Many stream I/O implementations will treat it as an ordinary error, which |
4063 | can be quite inconvenient. You can avoid this hassle by installing all | |
4064 | signals with the @code{SA_RESTART} flag. | |
4065 | ||
4066 | For similar reasons, setting nonblocking I/O on a stream's file | |
4067 | descriptor is not usually advisable. | |
4068 | ||
28f540f4 RM |
4069 | @node Binary Streams |
4070 | @section Text and Binary Streams | |
4071 | ||
a7a93d50 | 4072 | @gnusystems{} and other POSIX-compatible operating systems organize all |
28f540f4 RM |
4073 | files as uniform sequences of characters. However, some other systems |
4074 | make a distinction between files containing text and files containing | |
f65fd747 | 4075 | binary data, and the input and output facilities of @w{ISO C} provide for |
28f540f4 RM |
4076 | this distinction. This section tells you how to write programs portable |
4077 | to such systems. | |
4078 | ||
4079 | @cindex text stream | |
4080 | @cindex binary stream | |
4081 | When you open a stream, you can specify either a @dfn{text stream} or a | |
4082 | @dfn{binary stream}. You indicate that you want a binary stream by | |
4083 | specifying the @samp{b} modifier in the @var{opentype} argument to | |
4084 | @code{fopen}; see @ref{Opening Streams}. Without this | |
4085 | option, @code{fopen} opens the file as a text stream. | |
4086 | ||
4087 | Text and binary streams differ in several ways: | |
4088 | ||
4089 | @itemize @bullet | |
4090 | @item | |
4091 | The data read from a text stream is divided into @dfn{lines} which are | |
4092 | terminated by newline (@code{'\n'}) characters, while a binary stream is | |
4093 | simply a long series of characters. A text stream might on some systems | |
4094 | fail to handle lines more than 254 characters long (including the | |
4095 | terminating newline character). | |
4096 | @cindex lines (in a text file) | |
4097 | ||
4098 | @item | |
4099 | On some systems, text files can contain only printing characters, | |
4100 | horizontal tab characters, and newlines, and so text streams may not | |
4101 | support other characters. However, binary streams can handle any | |
4102 | character value. | |
4103 | ||
4104 | @item | |
4105 | Space characters that are written immediately preceding a newline | |
4106 | character in a text stream may disappear when the file is read in again. | |
4107 | ||
4108 | @item | |
4109 | More generally, there need not be a one-to-one mapping between | |
4110 | characters that are read from or written to a text stream, and the | |
4111 | characters in the actual file. | |
4112 | @end itemize | |
4113 | ||
4114 | Since a binary stream is always more capable and more predictable than a | |
4115 | text stream, you might wonder what purpose text streams serve. Why not | |
4116 | simply always use binary streams? The answer is that on these operating | |
4117 | systems, text and binary streams use different file formats, and the | |
4118 | only way to read or write ``an ordinary file of text'' that can work | |
4119 | with other text-oriented programs is through a text stream. | |
4120 | ||
1f77f049 | 4121 | In @theglibc{}, and on all POSIX systems, there is no difference |
28f540f4 RM |
4122 | between text streams and binary streams. When you open a stream, you |
4123 | get the same kind of stream regardless of whether you ask for binary. | |
4124 | This stream can handle any file content, and has none of the | |
4125 | restrictions that text streams sometimes have. | |
4126 | ||
4127 | @node File Positioning | |
4128 | @section File Positioning | |
4129 | @cindex file positioning on a stream | |
4130 | @cindex positioning a stream | |
4131 | @cindex seeking on a stream | |
4132 | ||
4133 | The @dfn{file position} of a stream describes where in the file the | |
4134 | stream is currently reading or writing. I/O on the stream advances the | |
a7a93d50 | 4135 | file position through the file. On @gnusystems{}, the file position is |
28f540f4 RM |
4136 | represented as an integer, which counts the number of bytes from the |
4137 | beginning of the file. @xref{File Position}. | |
4138 | ||
4139 | During I/O to an ordinary disk file, you can change the file position | |
4140 | whenever you wish, so as to read or write any portion of the file. Some | |
4141 | other kinds of files may also permit this. Files which support changing | |
4142 | the file position are sometimes referred to as @dfn{random-access} | |
4143 | files. | |
4144 | ||
4145 | You can use the functions in this section to examine or modify the file | |
4146 | position indicator associated with a stream. The symbols listed below | |
4147 | are declared in the header file @file{stdio.h}. | |
4148 | @pindex stdio.h | |
4149 | ||
4150 | @comment stdio.h | |
f65fd747 | 4151 | @comment ISO |
28f540f4 RM |
4152 | @deftypefun {long int} ftell (FILE *@var{stream}) |
4153 | This function returns the current file position of the stream | |
4154 | @var{stream}. | |
4155 | ||
4156 | This function can fail if the stream doesn't support file positioning, | |
4157 | or if the file position can't be represented in a @code{long int}, and | |
4158 | possibly for other reasons as well. If a failure occurs, a value of | |
4159 | @code{-1} is returned. | |
4160 | @end deftypefun | |
4161 | ||
a5a0310d UD |
4162 | @comment stdio.h |
4163 | @comment Unix98 | |
4164 | @deftypefun off_t ftello (FILE *@var{stream}) | |
04b9968b UD |
4165 | The @code{ftello} function is similar to @code{ftell}, except that it |
4166 | returns a value of type @code{off_t}. Systems which support this type | |
4167 | use it to describe all file positions, unlike the POSIX specification | |
4168 | which uses a long int. The two are not necessarily the same size. | |
4169 | Therefore, using ftell can lead to problems if the implementation is | |
4170 | written on top of a POSIX compliant low-level I/O implementation, and using | |
4171 | @code{ftello} is preferable whenever it is available. | |
4172 | ||
4173 | If this function fails it returns @code{(off_t) -1}. This can happen due | |
a5a0310d UD |
4174 | to missing support for file positioning or internal errors. Otherwise |
4175 | the return value is the current file position. | |
4176 | ||
4177 | The function is an extension defined in the Unix Single Specification | |
4178 | version 2. | |
a3a4a74e UD |
4179 | |
4180 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 4181 | 32 bit system this function is in fact @code{ftello64}. I.e., the |
a3a4a74e UD |
4182 | LFS interface transparently replaces the old interface. |
4183 | @end deftypefun | |
4184 | ||
4185 | @comment stdio.h | |
4186 | @comment Unix98 | |
4187 | @deftypefun off64_t ftello64 (FILE *@var{stream}) | |
4188 | This function is similar to @code{ftello} with the only difference that | |
4189 | the return value is of type @code{off64_t}. This also requires that the | |
4190 | stream @var{stream} was opened using either @code{fopen64}, | |
4191 | @code{freopen64}, or @code{tmpfile64} since otherwise the underlying | |
4192 | file operations to position the file pointer beyond the @math{2^31} | |
4193 | bytes limit might fail. | |
4194 | ||
4195 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
4196 | bits machine this function is available under the name @code{ftello} | |
4197 | and so transparently replaces the old interface. | |
a5a0310d UD |
4198 | @end deftypefun |
4199 | ||
28f540f4 | 4200 | @comment stdio.h |
f65fd747 | 4201 | @comment ISO |
28f540f4 RM |
4202 | @deftypefun int fseek (FILE *@var{stream}, long int @var{offset}, int @var{whence}) |
4203 | The @code{fseek} function is used to change the file position of the | |
4204 | stream @var{stream}. The value of @var{whence} must be one of the | |
4205 | constants @code{SEEK_SET}, @code{SEEK_CUR}, or @code{SEEK_END}, to | |
4206 | indicate whether the @var{offset} is relative to the beginning of the | |
4207 | file, the current file position, or the end of the file, respectively. | |
4208 | ||
4209 | This function returns a value of zero if the operation was successful, | |
4210 | and a nonzero value to indicate failure. A successful call also clears | |
4211 | the end-of-file indicator of @var{stream} and discards any characters | |
4212 | that were ``pushed back'' by the use of @code{ungetc}. | |
4213 | ||
4214 | @code{fseek} either flushes any buffered output before setting the file | |
4215 | position or else remembers it so it will be written later in its proper | |
4216 | place in the file. | |
4217 | @end deftypefun | |
4218 | ||
a5a0310d UD |
4219 | @comment stdio.h |
4220 | @comment Unix98 | |
4221 | @deftypefun int fseeko (FILE *@var{stream}, off_t @var{offset}, int @var{whence}) | |
4222 | This function is similar to @code{fseek} but it corrects a problem with | |
4223 | @code{fseek} in a system with POSIX types. Using a value of type | |
4224 | @code{long int} for the offset is not compatible with POSIX. | |
4225 | @code{fseeko} uses the correct type @code{off_t} for the @var{offset} | |
4226 | parameter. | |
4227 | ||
f2ea0f5b | 4228 | For this reason it is a good idea to prefer @code{ftello} whenever it is |
a5a0310d UD |
4229 | available since its functionality is (if different at all) closer the |
4230 | underlying definition. | |
4231 | ||
4232 | The functionality and return value is the same as for @code{fseek}. | |
4233 | ||
4234 | The function is an extension defined in the Unix Single Specification | |
4235 | version 2. | |
a3a4a74e UD |
4236 | |
4237 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 4238 | 32 bit system this function is in fact @code{fseeko64}. I.e., the |
a3a4a74e UD |
4239 | LFS interface transparently replaces the old interface. |
4240 | @end deftypefun | |
4241 | ||
4242 | @comment stdio.h | |
4243 | @comment Unix98 | |
4244 | @deftypefun int fseeko64 (FILE *@var{stream}, off64_t @var{offset}, int @var{whence}) | |
4245 | This function is similar to @code{fseeko} with the only difference that | |
4246 | the @var{offset} parameter is of type @code{off64_t}. This also | |
4247 | requires that the stream @var{stream} was opened using either | |
4248 | @code{fopen64}, @code{freopen64}, or @code{tmpfile64} since otherwise | |
4249 | the underlying file operations to position the file pointer beyond the | |
4250 | @math{2^31} bytes limit might fail. | |
4251 | ||
4252 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
4253 | bits machine this function is available under the name @code{fseeko} | |
4254 | and so transparently replaces the old interface. | |
a5a0310d UD |
4255 | @end deftypefun |
4256 | ||
4257 | @strong{Portability Note:} In non-POSIX systems, @code{ftell}, | |
4258 | @code{ftello}, @code{fseek} and @code{fseeko} might work reliably only | |
4259 | on binary streams. @xref{Binary Streams}. | |
28f540f4 RM |
4260 | |
4261 | The following symbolic constants are defined for use as the @var{whence} | |
4262 | argument to @code{fseek}. They are also used with the @code{lseek} | |
4263 | function (@pxref{I/O Primitives}) and to specify offsets for file locks | |
4264 | (@pxref{Control Operations}). | |
4265 | ||
4266 | @comment stdio.h | |
f65fd747 | 4267 | @comment ISO |
28f540f4 RM |
4268 | @deftypevr Macro int SEEK_SET |
4269 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
4270 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
4271 | the offset provided is relative to the beginning of the file. | |
28f540f4 RM |
4272 | @end deftypevr |
4273 | ||
4274 | @comment stdio.h | |
f65fd747 | 4275 | @comment ISO |
28f540f4 RM |
4276 | @deftypevr Macro int SEEK_CUR |
4277 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
4278 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
4279 | the offset provided is relative to the current file position. | |
28f540f4 RM |
4280 | @end deftypevr |
4281 | ||
4282 | @comment stdio.h | |
f65fd747 | 4283 | @comment ISO |
28f540f4 RM |
4284 | @deftypevr Macro int SEEK_END |
4285 | This is an integer constant which, when used as the @var{whence} | |
a5a0310d UD |
4286 | argument to the @code{fseek} or @code{fseeko} function, specifies that |
4287 | the offset provided is relative to the end of the file. | |
28f540f4 RM |
4288 | @end deftypevr |
4289 | ||
4290 | @comment stdio.h | |
f65fd747 | 4291 | @comment ISO |
28f540f4 RM |
4292 | @deftypefun void rewind (FILE *@var{stream}) |
4293 | The @code{rewind} function positions the stream @var{stream} at the | |
f2ea0f5b | 4294 | beginning of the file. It is equivalent to calling @code{fseek} or |
a5a0310d UD |
4295 | @code{fseeko} on the @var{stream} with an @var{offset} argument of |
4296 | @code{0L} and a @var{whence} argument of @code{SEEK_SET}, except that | |
4297 | the return value is discarded and the error indicator for the stream is | |
4298 | reset. | |
28f540f4 RM |
4299 | @end deftypefun |
4300 | ||
4301 | These three aliases for the @samp{SEEK_@dots{}} constants exist for the | |
4302 | sake of compatibility with older BSD systems. They are defined in two | |
4303 | different header files: @file{fcntl.h} and @file{sys/file.h}. | |
4304 | ||
4305 | @table @code | |
4306 | @comment sys/file.h | |
4307 | @comment BSD | |
4308 | @item L_SET | |
4309 | @vindex L_SET | |
4310 | An alias for @code{SEEK_SET}. | |
4311 | ||
4312 | @comment sys/file.h | |
4313 | @comment BSD | |
4314 | @item L_INCR | |
4315 | @vindex L_INCR | |
4316 | An alias for @code{SEEK_CUR}. | |
4317 | ||
4318 | @comment sys/file.h | |
4319 | @comment BSD | |
4320 | @item L_XTND | |
4321 | @vindex L_XTND | |
4322 | An alias for @code{SEEK_END}. | |
4323 | @end table | |
4324 | ||
4325 | @node Portable Positioning | |
4326 | @section Portable File-Position Functions | |
4327 | ||
a7a93d50 | 4328 | On @gnusystems{}, the file position is truly a character count. You |
a5a0310d UD |
4329 | can specify any character count value as an argument to @code{fseek} or |
4330 | @code{fseeko} and get reliable results for any random access file. | |
4331 | However, some @w{ISO C} systems do not represent file positions in this | |
4332 | way. | |
28f540f4 RM |
4333 | |
4334 | On some systems where text streams truly differ from binary streams, it | |
4335 | is impossible to represent the file position of a text stream as a count | |
4336 | of characters from the beginning of the file. For example, the file | |
4337 | position on some systems must encode both a record offset within the | |
4338 | file, and a character offset within the record. | |
4339 | ||
4340 | As a consequence, if you want your programs to be portable to these | |
4341 | systems, you must observe certain rules: | |
4342 | ||
4343 | @itemize @bullet | |
4344 | @item | |
4345 | The value returned from @code{ftell} on a text stream has no predictable | |
4346 | relationship to the number of characters you have read so far. The only | |
4347 | thing you can rely on is that you can use it subsequently as the | |
a5a0310d UD |
4348 | @var{offset} argument to @code{fseek} or @code{fseeko} to move back to |
4349 | the same file position. | |
28f540f4 | 4350 | |
19c3f208 | 4351 | @item |
a5a0310d | 4352 | In a call to @code{fseek} or @code{fseeko} on a text stream, either the |
04b9968b | 4353 | @var{offset} must be zero, or @var{whence} must be @code{SEEK_SET} and |
ded5b9b7 | 4354 | the @var{offset} must be the result of an earlier call to @code{ftell} |
04b9968b | 4355 | on the same stream. |
28f540f4 RM |
4356 | |
4357 | @item | |
4358 | The value of the file position indicator of a text stream is undefined | |
4359 | while there are characters that have been pushed back with @code{ungetc} | |
4360 | that haven't been read or discarded. @xref{Unreading}. | |
4361 | @end itemize | |
4362 | ||
4363 | But even if you observe these rules, you may still have trouble for long | |
4364 | files, because @code{ftell} and @code{fseek} use a @code{long int} value | |
4365 | to represent the file position. This type may not have room to encode | |
a5a0310d UD |
4366 | all the file positions in a large file. Using the @code{ftello} and |
4367 | @code{fseeko} functions might help here since the @code{off_t} type is | |
4368 | expected to be able to hold all file position values but this still does | |
4369 | not help to handle additional information which must be associated with | |
4370 | a file position. | |
28f540f4 RM |
4371 | |
4372 | So if you do want to support systems with peculiar encodings for the | |
4373 | file positions, it is better to use the functions @code{fgetpos} and | |
4374 | @code{fsetpos} instead. These functions represent the file position | |
4375 | using the data type @code{fpos_t}, whose internal representation varies | |
4376 | from system to system. | |
4377 | ||
4378 | These symbols are declared in the header file @file{stdio.h}. | |
4379 | @pindex stdio.h | |
4380 | ||
4381 | @comment stdio.h | |
f65fd747 | 4382 | @comment ISO |
28f540f4 RM |
4383 | @deftp {Data Type} fpos_t |
4384 | This is the type of an object that can encode information about the | |
4385 | file position of a stream, for use by the functions @code{fgetpos} and | |
4386 | @code{fsetpos}. | |
4387 | ||
a7a93d50 | 4388 | In @theglibc{}, @code{fpos_t} is an opaque data structure that |
5988b69d AJ |
4389 | contains internal data to represent file offset and conversion state |
4390 | information. In other systems, it might have a different internal | |
28f540f4 | 4391 | representation. |
a3a4a74e | 4392 | |
0be8752b | 4393 | When compiling with @code{_FILE_OFFSET_BITS == 64} on a 32 bit machine |
5988b69d AJ |
4394 | this type is in fact equivalent to @code{fpos64_t} since the LFS |
4395 | interface transparently replaces the old interface. | |
a3a4a74e UD |
4396 | @end deftp |
4397 | ||
4398 | @comment stdio.h | |
4399 | @comment Unix98 | |
4400 | @deftp {Data Type} fpos64_t | |
4401 | This is the type of an object that can encode information about the | |
4402 | file position of a stream, for use by the functions @code{fgetpos64} and | |
4403 | @code{fsetpos64}. | |
4404 | ||
a7a93d50 | 4405 | In @theglibc{}, @code{fpos64_t} is an opaque data structure that |
5988b69d AJ |
4406 | contains internal data to represent file offset and conversion state |
4407 | information. In other systems, it might have a different internal | |
a3a4a74e | 4408 | representation. |
28f540f4 RM |
4409 | @end deftp |
4410 | ||
4411 | @comment stdio.h | |
f65fd747 | 4412 | @comment ISO |
28f540f4 RM |
4413 | @deftypefun int fgetpos (FILE *@var{stream}, fpos_t *@var{position}) |
4414 | This function stores the value of the file position indicator for the | |
4415 | stream @var{stream} in the @code{fpos_t} object pointed to by | |
4416 | @var{position}. If successful, @code{fgetpos} returns zero; otherwise | |
4417 | it returns a nonzero value and stores an implementation-defined positive | |
4418 | value in @code{errno}. | |
a3a4a74e UD |
4419 | |
4420 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 4421 | 32 bit system the function is in fact @code{fgetpos64}. I.e., the LFS |
5988b69d | 4422 | interface transparently replaces the old interface. |
a3a4a74e UD |
4423 | @end deftypefun |
4424 | ||
4425 | @comment stdio.h | |
4426 | @comment Unix98 | |
4427 | @deftypefun int fgetpos64 (FILE *@var{stream}, fpos64_t *@var{position}) | |
4428 | This function is similar to @code{fgetpos} but the file position is | |
4429 | returned in a variable of type @code{fpos64_t} to which @var{position} | |
4430 | points. | |
4431 | ||
4432 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
4433 | bits machine this function is available under the name @code{fgetpos} | |
4434 | and so transparently replaces the old interface. | |
28f540f4 RM |
4435 | @end deftypefun |
4436 | ||
4437 | @comment stdio.h | |
f65fd747 | 4438 | @comment ISO |
a3a4a74e | 4439 | @deftypefun int fsetpos (FILE *@var{stream}, const fpos_t *@var{position}) |
28f540f4 RM |
4440 | This function sets the file position indicator for the stream @var{stream} |
4441 | to the position @var{position}, which must have been set by a previous | |
4442 | call to @code{fgetpos} on the same stream. If successful, @code{fsetpos} | |
4443 | clears the end-of-file indicator on the stream, discards any characters | |
4444 | that were ``pushed back'' by the use of @code{ungetc}, and returns a value | |
4445 | of zero. Otherwise, @code{fsetpos} returns a nonzero value and stores | |
4446 | an implementation-defined positive value in @code{errno}. | |
a3a4a74e UD |
4447 | |
4448 | When the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a | |
0be8752b | 4449 | 32 bit system the function is in fact @code{fsetpos64}. I.e., the LFS |
5988b69d | 4450 | interface transparently replaces the old interface. |
a3a4a74e UD |
4451 | @end deftypefun |
4452 | ||
4453 | @comment stdio.h | |
4454 | @comment Unix98 | |
4455 | @deftypefun int fsetpos64 (FILE *@var{stream}, const fpos64_t *@var{position}) | |
4456 | This function is similar to @code{fsetpos} but the file position used | |
4457 | for positioning is provided in a variable of type @code{fpos64_t} to | |
4458 | which @var{position} points. | |
4459 | ||
4460 | If the sources are compiled with @code{_FILE_OFFSET_BITS == 64} on a 32 | |
4461 | bits machine this function is available under the name @code{fsetpos} | |
4462 | and so transparently replaces the old interface. | |
28f540f4 RM |
4463 | @end deftypefun |
4464 | ||
4465 | @node Stream Buffering | |
4466 | @section Stream Buffering | |
4467 | ||
4468 | @cindex buffering of streams | |
4469 | Characters that are written to a stream are normally accumulated and | |
4470 | transmitted asynchronously to the file in a block, instead of appearing | |
4471 | as soon as they are output by the application program. Similarly, | |
4472 | streams often retrieve input from the host environment in blocks rather | |
4473 | than on a character-by-character basis. This is called @dfn{buffering}. | |
4474 | ||
4475 | If you are writing programs that do interactive input and output using | |
4476 | streams, you need to understand how buffering works when you design the | |
4477 | user interface to your program. Otherwise, you might find that output | |
4478 | (such as progress or prompt messages) doesn't appear when you intended | |
04b9968b | 4479 | it to, or displays some other unexpected behavior. |
28f540f4 RM |
4480 | |
4481 | This section deals only with controlling when characters are transmitted | |
4482 | between the stream and the file or device, and @emph{not} with how | |
4483 | things like echoing, flow control, and the like are handled on specific | |
4484 | classes of devices. For information on common control operations on | |
4485 | terminal devices, see @ref{Low-Level Terminal Interface}. | |
4486 | ||
4487 | You can bypass the stream buffering facilities altogether by using the | |
4488 | low-level input and output functions that operate on file descriptors | |
4489 | instead. @xref{Low-Level I/O}. | |
4490 | ||
4491 | @menu | |
4492 | * Buffering Concepts:: Terminology is defined here. | |
4493 | * Flushing Buffers:: How to ensure that output buffers are flushed. | |
4494 | * Controlling Buffering:: How to specify what kind of buffering to use. | |
4495 | @end menu | |
4496 | ||
4497 | @node Buffering Concepts | |
4498 | @subsection Buffering Concepts | |
4499 | ||
4500 | There are three different kinds of buffering strategies: | |
4501 | ||
4502 | @itemize @bullet | |
4503 | @item | |
4504 | Characters written to or read from an @dfn{unbuffered} stream are | |
4505 | transmitted individually to or from the file as soon as possible. | |
4506 | @cindex unbuffered stream | |
4507 | ||
4508 | @item | |
4509 | Characters written to a @dfn{line buffered} stream are transmitted to | |
4510 | the file in blocks when a newline character is encountered. | |
4511 | @cindex line buffered stream | |
4512 | ||
4513 | @item | |
4514 | Characters written to or read from a @dfn{fully buffered} stream are | |
4515 | transmitted to or from the file in blocks of arbitrary size. | |
4516 | @cindex fully buffered stream | |
4517 | @end itemize | |
4518 | ||
4519 | Newly opened streams are normally fully buffered, with one exception: a | |
4520 | stream connected to an interactive device such as a terminal is | |
4521 | initially line buffered. @xref{Controlling Buffering}, for information | |
4522 | on how to select a different kind of buffering. Usually the automatic | |
4523 | selection gives you the most convenient kind of buffering for the file | |
4524 | or device you open. | |
4525 | ||
4526 | The use of line buffering for interactive devices implies that output | |
4527 | messages ending in a newline will appear immediately---which is usually | |
4528 | what you want. Output that doesn't end in a newline might or might not | |
4529 | show up immediately, so if you want them to appear immediately, you | |
4530 | should flush buffered output explicitly with @code{fflush}, as described | |
4531 | in @ref{Flushing Buffers}. | |
4532 | ||
4533 | @node Flushing Buffers | |
4534 | @subsection Flushing Buffers | |
4535 | ||
4536 | @cindex flushing a stream | |
4537 | @dfn{Flushing} output on a buffered stream means transmitting all | |
4538 | accumulated characters to the file. There are many circumstances when | |
4539 | buffered output on a stream is flushed automatically: | |
4540 | ||
4541 | @itemize @bullet | |
4542 | @item | |
4543 | When you try to do output and the output buffer is full. | |
4544 | ||
4545 | @item | |
4546 | When the stream is closed. @xref{Closing Streams}. | |
4547 | ||
19c3f208 | 4548 | @item |
28f540f4 RM |
4549 | When the program terminates by calling @code{exit}. |
4550 | @xref{Normal Termination}. | |
4551 | ||
4552 | @item | |
4553 | When a newline is written, if the stream is line buffered. | |
4554 | ||
4555 | @item | |
4556 | Whenever an input operation on @emph{any} stream actually reads data | |
4557 | from its file. | |
4558 | @end itemize | |
4559 | ||
4560 | If you want to flush the buffered output at another time, call | |
4561 | @code{fflush}, which is declared in the header file @file{stdio.h}. | |
4562 | @pindex stdio.h | |
4563 | ||
4564 | @comment stdio.h | |
f65fd747 | 4565 | @comment ISO |
28f540f4 RM |
4566 | @deftypefun int fflush (FILE *@var{stream}) |
4567 | This function causes any buffered output on @var{stream} to be delivered | |
4568 | to the file. If @var{stream} is a null pointer, then | |
4569 | @code{fflush} causes buffered output on @emph{all} open output streams | |
4570 | to be flushed. | |
4571 | ||
4572 | This function returns @code{EOF} if a write error occurs, or zero | |
4573 | otherwise. | |
4574 | @end deftypefun | |
4575 | ||
7b4161bb UD |
4576 | @comment stdio.h |
4577 | @comment POSIX | |
4578 | @deftypefun int fflush_unlocked (FILE *@var{stream}) | |
4579 | The @code{fflush_unlocked} function is equivalent to the @code{fflush} | |
1dc843f7 | 4580 | function except that it does not implicitly lock the stream. |
7b4161bb UD |
4581 | @end deftypefun |
4582 | ||
4583 | The @code{fflush} function can be used to flush all streams currently | |
4584 | opened. While this is useful in some situations it does often more than | |
4585 | necessary since it might be done in situations when terminal input is | |
4586 | required and the program wants to be sure that all output is visible on | |
4587 | the terminal. But this means that only line buffered streams have to be | |
4588 | flushed. Solaris introduced a function especially for this. It was | |
1f77f049 | 4589 | always available in @theglibc{} in some form but never officially |
7b4161bb UD |
4590 | exported. |
4591 | ||
9deee4bb | 4592 | @comment stdio_ext.h |
7b4161bb UD |
4593 | @comment GNU |
4594 | @deftypefun void _flushlbf (void) | |
4595 | The @code{_flushlbf} function flushes all line buffered streams | |
4596 | currently opened. | |
4597 | ||
4598 | This function is declared in the @file{stdio_ext.h} header. | |
4599 | @end deftypefun | |
4600 | ||
28f540f4 RM |
4601 | @strong{Compatibility Note:} Some brain-damaged operating systems have |
4602 | been known to be so thoroughly fixated on line-oriented input and output | |
4603 | that flushing a line buffered stream causes a newline to be written! | |
4604 | Fortunately, this ``feature'' seems to be becoming less common. You do | |
a7a93d50 | 4605 | not need to worry about this with @theglibc{}. |
28f540f4 | 4606 | |
7b4161bb UD |
4607 | In some situations it might be useful to not flush the output pending |
4608 | for a stream but instead simply forget it. If transmission is costly | |
4609 | and the output is not needed anymore this is valid reasoning. In this | |
4610 | situation a non-standard function introduced in Solaris and available in | |
1f77f049 | 4611 | @theglibc{} can be used. |
7b4161bb UD |
4612 | |
4613 | @comment stdio_ext.h | |
4614 | @comment GNU | |
4615 | @deftypefun void __fpurge (FILE *@var{stream}) | |
4616 | The @code{__fpurge} function causes the buffer of the stream | |
4617 | @var{stream} to be emptied. If the stream is currently in read mode all | |
4618 | input in the buffer is lost. If the stream is in output mode the | |
4619 | buffered output is not written to the device (or whatever other | |
4620 | underlying storage) and the buffer the cleared. | |
4621 | ||
4622 | This function is declared in @file{stdio_ext.h}. | |
4623 | @end deftypefun | |
28f540f4 RM |
4624 | |
4625 | @node Controlling Buffering | |
4626 | @subsection Controlling Which Kind of Buffering | |
4627 | ||
4628 | After opening a stream (but before any other operations have been | |
4629 | performed on it), you can explicitly specify what kind of buffering you | |
4630 | want it to have using the @code{setvbuf} function. | |
4631 | @cindex buffering, controlling | |
4632 | ||
4633 | The facilities listed in this section are declared in the header | |
4634 | file @file{stdio.h}. | |
4635 | @pindex stdio.h | |
4636 | ||
4637 | @comment stdio.h | |
f65fd747 | 4638 | @comment ISO |
28f540f4 RM |
4639 | @deftypefun int setvbuf (FILE *@var{stream}, char *@var{buf}, int @var{mode}, size_t @var{size}) |
4640 | This function is used to specify that the stream @var{stream} should | |
4641 | have the buffering mode @var{mode}, which can be either @code{_IOFBF} | |
4642 | (for full buffering), @code{_IOLBF} (for line buffering), or | |
4643 | @code{_IONBF} (for unbuffered input/output). | |
4644 | ||
4645 | If you specify a null pointer as the @var{buf} argument, then @code{setvbuf} | |
4646 | allocates a buffer itself using @code{malloc}. This buffer will be freed | |
4647 | when you close the stream. | |
4648 | ||
4649 | Otherwise, @var{buf} should be a character array that can hold at least | |
4650 | @var{size} characters. You should not free the space for this array as | |
4651 | long as the stream remains open and this array remains its buffer. You | |
4652 | should usually either allocate it statically, or @code{malloc} | |
4653 | (@pxref{Unconstrained Allocation}) the buffer. Using an automatic array | |
4654 | is not a good idea unless you close the file before exiting the block | |
4655 | that declares the array. | |
4656 | ||
4657 | While the array remains a stream buffer, the stream I/O functions will | |
4658 | use the buffer for their internal purposes. You shouldn't try to access | |
4659 | the values in the array directly while the stream is using it for | |
4660 | buffering. | |
4661 | ||
4662 | The @code{setvbuf} function returns zero on success, or a nonzero value | |
4663 | if the value of @var{mode} is not valid or if the request could not | |
4664 | be honored. | |
4665 | @end deftypefun | |
4666 | ||
4667 | @comment stdio.h | |
f65fd747 | 4668 | @comment ISO |
28f540f4 RM |
4669 | @deftypevr Macro int _IOFBF |
4670 | The value of this macro is an integer constant expression that can be | |
4671 | used as the @var{mode} argument to the @code{setvbuf} function to | |
4672 | specify that the stream should be fully buffered. | |
4673 | @end deftypevr | |
4674 | ||
4675 | @comment stdio.h | |
f65fd747 | 4676 | @comment ISO |
28f540f4 RM |
4677 | @deftypevr Macro int _IOLBF |
4678 | The value of this macro is an integer constant expression that can be | |
4679 | used as the @var{mode} argument to the @code{setvbuf} function to | |
4680 | specify that the stream should be line buffered. | |
4681 | @end deftypevr | |
4682 | ||
4683 | @comment stdio.h | |
f65fd747 | 4684 | @comment ISO |
28f540f4 RM |
4685 | @deftypevr Macro int _IONBF |
4686 | The value of this macro is an integer constant expression that can be | |
4687 | used as the @var{mode} argument to the @code{setvbuf} function to | |
4688 | specify that the stream should be unbuffered. | |
4689 | @end deftypevr | |
4690 | ||
4691 | @comment stdio.h | |
f65fd747 | 4692 | @comment ISO |
28f540f4 RM |
4693 | @deftypevr Macro int BUFSIZ |
4694 | The value of this macro is an integer constant expression that is good | |
4695 | to use for the @var{size} argument to @code{setvbuf}. This value is | |
4696 | guaranteed to be at least @code{256}. | |
4697 | ||
4698 | The value of @code{BUFSIZ} is chosen on each system so as to make stream | |
19c3f208 | 4699 | I/O efficient. So it is a good idea to use @code{BUFSIZ} as the size |
28f540f4 RM |
4700 | for the buffer when you call @code{setvbuf}. |
4701 | ||
4702 | Actually, you can get an even better value to use for the buffer size | |
4703 | by means of the @code{fstat} system call: it is found in the | |
4704 | @code{st_blksize} field of the file attributes. @xref{Attribute Meanings}. | |
4705 | ||
4706 | Sometimes people also use @code{BUFSIZ} as the allocation size of | |
4707 | buffers used for related purposes, such as strings used to receive a | |
4708 | line of input with @code{fgets} (@pxref{Character Input}). There is no | |
4709 | particular reason to use @code{BUFSIZ} for this instead of any other | |
4710 | integer, except that it might lead to doing I/O in chunks of an | |
4711 | efficient size. | |
4712 | @end deftypevr | |
4713 | ||
4714 | @comment stdio.h | |
f65fd747 | 4715 | @comment ISO |
28f540f4 RM |
4716 | @deftypefun void setbuf (FILE *@var{stream}, char *@var{buf}) |
4717 | If @var{buf} is a null pointer, the effect of this function is | |
4718 | equivalent to calling @code{setvbuf} with a @var{mode} argument of | |
4719 | @code{_IONBF}. Otherwise, it is equivalent to calling @code{setvbuf} | |
4720 | with @var{buf}, and a @var{mode} of @code{_IOFBF} and a @var{size} | |
4721 | argument of @code{BUFSIZ}. | |
4722 | ||
4723 | The @code{setbuf} function is provided for compatibility with old code; | |
4724 | use @code{setvbuf} in all new programs. | |
4725 | @end deftypefun | |
4726 | ||
4727 | @comment stdio.h | |
4728 | @comment BSD | |
4729 | @deftypefun void setbuffer (FILE *@var{stream}, char *@var{buf}, size_t @var{size}) | |
4730 | If @var{buf} is a null pointer, this function makes @var{stream} unbuffered. | |
4731 | Otherwise, it makes @var{stream} fully buffered using @var{buf} as the | |
4732 | buffer. The @var{size} argument specifies the length of @var{buf}. | |
4733 | ||
4734 | This function is provided for compatibility with old BSD code. Use | |
4735 | @code{setvbuf} instead. | |
4736 | @end deftypefun | |
4737 | ||
4738 | @comment stdio.h | |
4739 | @comment BSD | |
4740 | @deftypefun void setlinebuf (FILE *@var{stream}) | |
4741 | This function makes @var{stream} be line buffered, and allocates the | |
4742 | buffer for you. | |
4743 | ||
4744 | This function is provided for compatibility with old BSD code. Use | |
4745 | @code{setvbuf} instead. | |
4746 | @end deftypefun | |
4747 | ||
7b4161bb | 4748 | It is possible to query whether a given stream is line buffered or not |
1f77f049 JM |
4749 | using a non-standard function introduced in Solaris and available in |
4750 | @theglibc{}. | |
7b4161bb UD |
4751 | |
4752 | @comment stdio_ext.h | |
4753 | @comment GNU | |
4754 | @deftypefun int __flbf (FILE *@var{stream}) | |
4755 | The @code{__flbf} function will return a nonzero value in case the | |
4756 | stream @var{stream} is line buffered. Otherwise the return value is | |
4757 | zero. | |
4758 | ||
4759 | This function is declared in the @file{stdio_ext.h} header. | |
4760 | @end deftypefun | |
4761 | ||
4762 | Two more extensions allow to determine the size of the buffer and how | |
4763 | much of it is used. These functions were also introduced in Solaris. | |
4764 | ||
4765 | @comment stdio_ext.h | |
4766 | @comment GNU | |
4767 | @deftypefun size_t __fbufsize (FILE *@var{stream}) | |
4768 | The @code{__fbufsize} function return the size of the buffer in the | |
4769 | stream @var{stream}. This value can be used to optimize the use of the | |
4770 | stream. | |
4771 | ||
4772 | This function is declared in the @file{stdio_ext.h} header. | |
4773 | @end deftypefun | |
4774 | ||
4775 | @comment stdio_ext.h | |
4776 | @comment GNU | |
cc6e48bc NB |
4777 | @deftypefun size_t __fpending (FILE *@var{stream}) |
4778 | The @code{__fpending} | |
f126ef67 UD |
4779 | function returns the number of bytes currently in the output buffer. |
4780 | For wide-oriented stream the measuring unit is wide characters. This | |
4781 | function should not be used on buffers in read mode or opened read-only. | |
7b4161bb UD |
4782 | |
4783 | This function is declared in the @file{stdio_ext.h} header. | |
4784 | @end deftypefun | |
4785 | ||
28f540f4 RM |
4786 | @node Other Kinds of Streams |
4787 | @section Other Kinds of Streams | |
4788 | ||
1f77f049 | 4789 | @Theglibc{} provides ways for you to define additional kinds of |
28f540f4 RM |
4790 | streams that do not necessarily correspond to an open file. |
4791 | ||
4792 | One such type of stream takes input from or writes output to a string. | |
4793 | These kinds of streams are used internally to implement the | |
4794 | @code{sprintf} and @code{sscanf} functions. You can also create such a | |
4795 | stream explicitly, using the functions described in @ref{String Streams}. | |
4796 | ||
4797 | More generally, you can define streams that do input/output to arbitrary | |
4798 | objects using functions supplied by your program. This protocol is | |
4799 | discussed in @ref{Custom Streams}. | |
4800 | ||
4801 | @strong{Portability Note:} The facilities described in this section are | |
4802 | specific to GNU. Other systems or C implementations might or might not | |
4803 | provide equivalent functionality. | |
4804 | ||
4805 | @menu | |
19c3f208 | 4806 | * String Streams:: Streams that get data from or put data in |
162ba701 | 4807 | a string or memory buffer. |
28f540f4 | 4808 | * Custom Streams:: Defining your own streams with an arbitrary |
162ba701 | 4809 | input data source and/or output data sink. |
28f540f4 RM |
4810 | @end menu |
4811 | ||
4812 | @node String Streams | |
4813 | @subsection String Streams | |
4814 | ||
4815 | @cindex stream, for I/O to a string | |
4816 | @cindex string stream | |
4817 | The @code{fmemopen} and @code{open_memstream} functions allow you to do | |
4818 | I/O to a string or memory buffer. These facilities are declared in | |
4819 | @file{stdio.h}. | |
4820 | @pindex stdio.h | |
4821 | ||
4822 | @comment stdio.h | |
4823 | @comment GNU | |
4824 | @deftypefun {FILE *} fmemopen (void *@var{buf}, size_t @var{size}, const char *@var{opentype}) | |
4825 | This function opens a stream that allows the access specified by the | |
4826 | @var{opentype} argument, that reads from or writes to the buffer specified | |
4827 | by the argument @var{buf}. This array must be at least @var{size} bytes long. | |
4828 | ||
4829 | If you specify a null pointer as the @var{buf} argument, @code{fmemopen} | |
04b9968b UD |
4830 | dynamically allocates an array @var{size} bytes long (as with @code{malloc}; |
4831 | @pxref{Unconstrained Allocation}). This is really only useful | |
28f540f4 RM |
4832 | if you are going to write things to the buffer and then read them back |
4833 | in again, because you have no way of actually getting a pointer to the | |
4834 | buffer (for this, try @code{open_memstream}, below). The buffer is | |
417d6b34 | 4835 | freed when the stream is closed. |
28f540f4 RM |
4836 | |
4837 | The argument @var{opentype} is the same as in @code{fopen} | |
8b7fb588 | 4838 | (@pxref{Opening Streams}). If the @var{opentype} specifies |
28f540f4 RM |
4839 | append mode, then the initial file position is set to the first null |
4840 | character in the buffer. Otherwise the initial file position is at the | |
4841 | beginning of the buffer. | |
4842 | ||
4843 | When a stream open for writing is flushed or closed, a null character | |
4844 | (zero byte) is written at the end of the buffer if it fits. You | |
4845 | should add an extra byte to the @var{size} argument to account for this. | |
4846 | Attempts to write more than @var{size} bytes to the buffer result | |
4847 | in an error. | |
4848 | ||
4849 | For a stream open for reading, null characters (zero bytes) in the | |
4850 | buffer do not count as ``end of file''. Read operations indicate end of | |
4851 | file only when the file position advances past @var{size} bytes. So, if | |
4852 | you want to read characters from a null-terminated string, you should | |
4853 | supply the length of the string as the @var{size} argument. | |
4854 | @end deftypefun | |
4855 | ||
4856 | Here is an example of using @code{fmemopen} to create a stream for | |
4857 | reading from a string: | |
4858 | ||
4859 | @smallexample | |
4860 | @include memopen.c.texi | |
4861 | @end smallexample | |
4862 | ||
4863 | This program produces the following output: | |
4864 | ||
4865 | @smallexample | |
4866 | Got f | |
4867 | Got o | |
4868 | Got o | |
4869 | Got b | |
4870 | Got a | |
4871 | Got r | |
4872 | @end smallexample | |
4873 | ||
4874 | @comment stdio.h | |
4875 | @comment GNU | |
4876 | @deftypefun {FILE *} open_memstream (char **@var{ptr}, size_t *@var{sizeloc}) | |
4877 | This function opens a stream for writing to a buffer. The buffer is | |
29402b12 RM |
4878 | allocated dynamically and grown as necessary, using @code{malloc}. |
4879 | After you've closed the stream, this buffer is your responsibility to | |
4880 | clean up using @code{free} or @code{realloc}. @xref{Unconstrained Allocation}. | |
28f540f4 RM |
4881 | |
4882 | When the stream is closed with @code{fclose} or flushed with | |
4883 | @code{fflush}, the locations @var{ptr} and @var{sizeloc} are updated to | |
4884 | contain the pointer to the buffer and its size. The values thus stored | |
4885 | remain valid only as long as no further output on the stream takes | |
4886 | place. If you do more output, you must flush the stream again to store | |
4887 | new values before you use them again. | |
4888 | ||
4889 | A null character is written at the end of the buffer. This null character | |
4890 | is @emph{not} included in the size value stored at @var{sizeloc}. | |
4891 | ||
a5a0310d UD |
4892 | You can move the stream's file position with @code{fseek} or |
4893 | @code{fseeko} (@pxref{File Positioning}). Moving the file position past | |
4894 | the end of the data already written fills the intervening space with | |
4895 | zeroes. | |
28f540f4 RM |
4896 | @end deftypefun |
4897 | ||
4898 | Here is an example of using @code{open_memstream}: | |
4899 | ||
4900 | @smallexample | |
4901 | @include memstrm.c.texi | |
4902 | @end smallexample | |
4903 | ||
4904 | This program produces the following output: | |
4905 | ||
4906 | @smallexample | |
4907 | buf = `hello', size = 5 | |
4908 | buf = `hello, world', size = 12 | |
4909 | @end smallexample | |
4910 | ||
28f540f4 RM |
4911 | @node Custom Streams |
4912 | @subsection Programming Your Own Custom Streams | |
4913 | @cindex custom streams | |
4914 | @cindex programming your own streams | |
4915 | ||
4916 | This section describes how you can make a stream that gets input from an | |
4917 | arbitrary data source or writes output to an arbitrary data sink | |
52cea457 UD |
4918 | programmed by you. We call these @dfn{custom streams}. The functions |
4919 | and types described here are all GNU extensions. | |
28f540f4 RM |
4920 | |
4921 | @c !!! this does not talk at all about the higher-level hooks | |
4922 | ||
4923 | @menu | |
4924 | * Streams and Cookies:: The @dfn{cookie} records where to fetch or | |
162ba701 | 4925 | store data that is read or written. |
28f540f4 | 4926 | * Hook Functions:: How you should define the four @dfn{hook |
162ba701 | 4927 | functions} that a custom stream needs. |
28f540f4 RM |
4928 | @end menu |
4929 | ||
4930 | @node Streams and Cookies | |
4931 | @subsubsection Custom Streams and Cookies | |
4932 | @cindex cookie, for custom stream | |
4933 | ||
4934 | Inside every custom stream is a special object called the @dfn{cookie}. | |
4935 | This is an object supplied by you which records where to fetch or store | |
4936 | the data read or written. It is up to you to define a data type to use | |
4937 | for the cookie. The stream functions in the library never refer | |
4938 | directly to its contents, and they don't even know what the type is; | |
4939 | they record its address with type @code{void *}. | |
4940 | ||
4941 | To implement a custom stream, you must specify @emph{how} to fetch or | |
4942 | store the data in the specified place. You do this by defining | |
4943 | @dfn{hook functions} to read, write, change ``file position'', and close | |
4944 | the stream. All four of these functions will be passed the stream's | |
4945 | cookie so they can tell where to fetch or store the data. The library | |
4946 | functions don't know what's inside the cookie, but your functions will | |
4947 | know. | |
4948 | ||
4949 | When you create a custom stream, you must specify the cookie pointer, | |
19c3f208 | 4950 | and also the four hook functions stored in a structure of type |
28f540f4 RM |
4951 | @code{cookie_io_functions_t}. |
4952 | ||
4953 | These facilities are declared in @file{stdio.h}. | |
4954 | @pindex stdio.h | |
4955 | ||
4956 | @comment stdio.h | |
4957 | @comment GNU | |
4958 | @deftp {Data Type} {cookie_io_functions_t} | |
19c3f208 | 4959 | This is a structure type that holds the functions that define the |
28f540f4 RM |
4960 | communications protocol between the stream and its cookie. It has |
4961 | the following members: | |
4962 | ||
4963 | @table @code | |
4964 | @item cookie_read_function_t *read | |
4965 | This is the function that reads data from the cookie. If the value is a | |
f2ea0f5b | 4966 | null pointer instead of a function, then read operations on this stream |
28f540f4 RM |
4967 | always return @code{EOF}. |
4968 | ||
4969 | @item cookie_write_function_t *write | |
4970 | This is the function that writes data to the cookie. If the value is a | |
4971 | null pointer instead of a function, then data written to the stream is | |
4972 | discarded. | |
4973 | ||
4974 | @item cookie_seek_function_t *seek | |
4975 | This is the function that performs the equivalent of file positioning on | |
4976 | the cookie. If the value is a null pointer instead of a function, calls | |
a5a0310d UD |
4977 | to @code{fseek} or @code{fseeko} on this stream can only seek to |
4978 | locations within the buffer; any attempt to seek outside the buffer will | |
4979 | return an @code{ESPIPE} error. | |
28f540f4 RM |
4980 | |
4981 | @item cookie_close_function_t *close | |
4982 | This function performs any appropriate cleanup on the cookie when | |
4983 | closing the stream. If the value is a null pointer instead of a | |
4984 | function, nothing special is done to close the cookie when the stream is | |
4985 | closed. | |
4986 | @end table | |
4987 | @end deftp | |
4988 | ||
4989 | @comment stdio.h | |
4990 | @comment GNU | |
4991 | @deftypefun {FILE *} fopencookie (void *@var{cookie}, const char *@var{opentype}, cookie_io_functions_t @var{io-functions}) | |
4992 | This function actually creates the stream for communicating with the | |
4993 | @var{cookie} using the functions in the @var{io-functions} argument. | |
4994 | The @var{opentype} argument is interpreted as for @code{fopen}; | |
4995 | see @ref{Opening Streams}. (But note that the ``truncate on | |
4996 | open'' option is ignored.) The new stream is fully buffered. | |
4997 | ||
4998 | The @code{fopencookie} function returns the newly created stream, or a null | |
4999 | pointer in case of an error. | |
5000 | @end deftypefun | |
5001 | ||
5002 | @node Hook Functions | |
5003 | @subsubsection Custom Stream Hook Functions | |
5004 | @cindex hook functions (of custom streams) | |
5005 | ||
5006 | Here are more details on how you should define the four hook functions | |
5007 | that a custom stream needs. | |
5008 | ||
5009 | You should define the function to read data from the cookie as: | |
5010 | ||
5011 | @smallexample | |
62ecda3d | 5012 | ssize_t @var{reader} (void *@var{cookie}, char *@var{buffer}, size_t @var{size}) |
28f540f4 RM |
5013 | @end smallexample |
5014 | ||
5015 | This is very similar to the @code{read} function; see @ref{I/O | |
5016 | Primitives}. Your function should transfer up to @var{size} bytes into | |
5017 | the @var{buffer}, and return the number of bytes read, or zero to | |
5018 | indicate end-of-file. You can return a value of @code{-1} to indicate | |
5019 | an error. | |
5020 | ||
5021 | You should define the function to write data to the cookie as: | |
5022 | ||
5023 | @smallexample | |
62ecda3d | 5024 | ssize_t @var{writer} (void *@var{cookie}, const char *@var{buffer}, size_t @var{size}) |
28f540f4 RM |
5025 | @end smallexample |
5026 | ||
5027 | This is very similar to the @code{write} function; see @ref{I/O | |
5028 | Primitives}. Your function should transfer up to @var{size} bytes from | |
5029 | the buffer, and return the number of bytes written. You can return a | |
247c3ede PB |
5030 | value of @code{0} to indicate an error. You must not return any |
5031 | negative value. | |
28f540f4 RM |
5032 | |
5033 | You should define the function to perform seek operations on the cookie | |
5034 | as: | |
5035 | ||
5036 | @smallexample | |
c1e36819 | 5037 | int @var{seeker} (void *@var{cookie}, off64_t *@var{position}, int @var{whence}) |
28f540f4 RM |
5038 | @end smallexample |
5039 | ||
5040 | For this function, the @var{position} and @var{whence} arguments are | |
c1e36819 | 5041 | interpreted as for @code{fgetpos}; see @ref{Portable Positioning}. |
28f540f4 | 5042 | |
19c3f208 | 5043 | After doing the seek operation, your function should store the resulting |
28f540f4 RM |
5044 | file position relative to the beginning of the file in @var{position}. |
5045 | Your function should return a value of @code{0} on success and @code{-1} | |
5046 | to indicate an error. | |
5047 | ||
5048 | You should define the function to do cleanup operations on the cookie | |
5049 | appropriate for closing the stream as: | |
5050 | ||
5051 | @smallexample | |
5052 | int @var{cleaner} (void *@var{cookie}) | |
5053 | @end smallexample | |
5054 | ||
5055 | Your function should return @code{-1} to indicate an error, and @code{0} | |
5056 | otherwise. | |
5057 | ||
5058 | @comment stdio.h | |
5059 | @comment GNU | |
5060 | @deftp {Data Type} cookie_read_function | |
5061 | This is the data type that the read function for a custom stream should have. | |
5062 | If you declare the function as shown above, this is the type it will have. | |
5063 | @end deftp | |
5064 | ||
5065 | @comment stdio.h | |
5066 | @comment GNU | |
5067 | @deftp {Data Type} cookie_write_function | |
5068 | The data type of the write function for a custom stream. | |
5069 | @end deftp | |
5070 | ||
5071 | @comment stdio.h | |
5072 | @comment GNU | |
5073 | @deftp {Data Type} cookie_seek_function | |
5074 | The data type of the seek function for a custom stream. | |
5075 | @end deftp | |
5076 | ||
5077 | @comment stdio.h | |
5078 | @comment GNU | |
5079 | @deftp {Data Type} cookie_close_function | |
5080 | The data type of the close function for a custom stream. | |
5081 | @end deftp | |
5082 | ||
5083 | @ignore | |
5084 | Roland says: | |
5085 | ||
5086 | @quotation | |
5087 | There is another set of functions one can give a stream, the | |
5088 | input-room and output-room functions. These functions must | |
5089 | understand stdio internals. To describe how to use these | |
5090 | functions, you also need to document lots of how stdio works | |
5091 | internally (which isn't relevant for other uses of stdio). | |
5092 | Perhaps I can write an interface spec from which you can write | |
5093 | good documentation. But it's pretty complex and deals with lots | |
5094 | of nitty-gritty details. I think it might be better to let this | |
5095 | wait until the rest of the manual is more done and polished. | |
5096 | @end quotation | |
5097 | @end ignore | |
5098 | ||
5099 | @c ??? This section could use an example. | |
0501d603 UD |
5100 | |
5101 | ||
5102 | @node Formatted Messages | |
5103 | @section Formatted Messages | |
5104 | @cindex formatted messages | |
5105 | ||
5106 | On systems which are based on System V messages of programs (especially | |
5107 | the system tools) are printed in a strict form using the @code{fmtmsg} | |
5108 | function. The uniformity sometimes helps the user to interpret messages | |
cf29ffbe | 5109 | and the strictness tests of the @code{fmtmsg} function ensure that the |
0501d603 UD |
5110 | programmer follows some minimal requirements. |
5111 | ||
5112 | @menu | |
5113 | * Printing Formatted Messages:: The @code{fmtmsg} function. | |
5114 | * Adding Severity Classes:: Add more severity classes. | |
5115 | * Example:: How to use @code{fmtmsg} and @code{addseverity}. | |
5116 | @end menu | |
5117 | ||
5118 | ||
5119 | @node Printing Formatted Messages | |
5120 | @subsection Printing Formatted Messages | |
5121 | ||
5122 | Messages can be printed to standard error and/or to the console. To | |
cf29ffbe | 5123 | select the destination the programmer can use the following two values, |
0501d603 UD |
5124 | bitwise OR combined if wanted, for the @var{classification} parameter of |
5125 | @code{fmtmsg}: | |
5126 | ||
5127 | @vtable @code | |
5128 | @item MM_PRINT | |
5129 | Display the message in standard error. | |
5130 | @item MM_CONSOLE | |
5131 | Display the message on the system console. | |
5132 | @end vtable | |
5133 | ||
f2ea0f5b | 5134 | The erroneous piece of the system can be signalled by exactly one of the |
0501d603 UD |
5135 | following values which also is bitwise ORed with the |
5136 | @var{classification} parameter to @code{fmtmsg}: | |
5137 | ||
5138 | @vtable @code | |
5139 | @item MM_HARD | |
5140 | The source of the condition is some hardware. | |
5141 | @item MM_SOFT | |
5142 | The source of the condition is some software. | |
5143 | @item MM_FIRM | |
5144 | The source of the condition is some firmware. | |
5145 | @end vtable | |
5146 | ||
5147 | A third component of the @var{classification} parameter to @code{fmtmsg} | |
5148 | can describe the part of the system which detects the problem. This is | |
5149 | done by using exactly one of the following values: | |
5150 | ||
5151 | @vtable @code | |
5152 | @item MM_APPL | |
f2ea0f5b | 5153 | The erroneous condition is detected by the application. |
0501d603 | 5154 | @item MM_UTIL |
f2ea0f5b | 5155 | The erroneous condition is detected by a utility. |
0501d603 | 5156 | @item MM_OPSYS |
f2ea0f5b | 5157 | The erroneous condition is detected by the operating system. |
0501d603 UD |
5158 | @end vtable |
5159 | ||
5160 | A last component of @var{classification} can signal the results of this | |
5161 | message. Exactly one of the following values can be used: | |
5162 | ||
5163 | @vtable @code | |
5164 | @item MM_RECOVER | |
5165 | It is a recoverable error. | |
5166 | @item MM_NRECOV | |
5167 | It is a non-recoverable error. | |
5168 | @end vtable | |
5169 | ||
5170 | @comment fmtmsg.h | |
5171 | @comment XPG | |
5172 | @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}) | |
5173 | Display a message described by its parameters on the device(s) specified | |
5174 | in the @var{classification} parameter. The @var{label} parameter | |
5175 | identifies the source of the message. The string should consist of two | |
5176 | colon separated parts where the first part has not more than 10 and the | |
04b9968b | 5177 | second part not more than 14 characters. The @var{text} parameter |
f2ea0f5b | 5178 | describes the condition of the error, the @var{action} parameter possible |
0501d603 UD |
5179 | steps to recover from the error and the @var{tag} parameter is a |
5180 | reference to the online documentation where more information can be | |
5181 | found. It should contain the @var{label} value and a unique | |
5182 | identification number. | |
5183 | ||
cf29ffbe | 5184 | Each of the parameters can be a special value which means this value |
0501d603 UD |
5185 | is to be omitted. The symbolic names for these values are: |
5186 | ||
5187 | @vtable @code | |
5188 | @item MM_NULLLBL | |
5189 | Ignore @var{label} parameter. | |
5190 | @item MM_NULLSEV | |
5191 | Ignore @var{severity} parameter. | |
5192 | @item MM_NULLMC | |
5193 | Ignore @var{classification} parameter. This implies that nothing is | |
5194 | actually printed. | |
5195 | @item MM_NULLTXT | |
5196 | Ignore @var{text} parameter. | |
5197 | @item MM_NULLACT | |
5198 | Ignore @var{action} parameter. | |
5199 | @item MM_NULLTAG | |
5200 | Ignore @var{tag} parameter. | |
5201 | @end vtable | |
5202 | ||
cf29ffbe | 5203 | There is another way certain fields can be omitted from the output to |
0501d603 | 5204 | standard error. This is described below in the description of |
0bc93a2f | 5205 | environment variables influencing the behavior. |
0501d603 UD |
5206 | |
5207 | The @var{severity} parameter can have one of the values in the following | |
5208 | table: | |
5209 | @cindex severity class | |
5210 | ||
5211 | @vtable @code | |
5212 | @item MM_NOSEV | |
5213 | Nothing is printed, this value is the same as @code{MM_NULLSEV}. | |
5214 | @item MM_HALT | |
5215 | This value is printed as @code{HALT}. | |
5216 | @item MM_ERROR | |
5217 | This value is printed as @code{ERROR}. | |
5218 | @item MM_WARNING | |
5219 | This value is printed as @code{WARNING}. | |
5220 | @item MM_INFO | |
5221 | This value is printed as @code{INFO}. | |
5222 | @end vtable | |
5223 | ||
5224 | The numeric value of these five macros are between @code{0} and | |
5225 | @code{4}. Using the environment variable @code{SEV_LEVEL} or using the | |
5226 | @code{addseverity} function one can add more severity levels with their | |
5227 | corresponding string to print. This is described below | |
5228 | (@pxref{Adding Severity Classes}). | |
5229 | ||
5230 | @noindent | |
5231 | If no parameter is ignored the output looks like this: | |
5232 | ||
5233 | @smallexample | |
5234 | @var{label}: @var{severity-string}: @var{text} | |
5235 | TO FIX: @var{action} @var{tag} | |
5236 | @end smallexample | |
5237 | ||
5238 | The colons, new line characters and the @code{TO FIX} string are | |
5239 | inserted if necessary, i.e., if the corresponding parameter is not | |
5240 | ignored. | |
5241 | ||
5242 | This function is specified in the X/Open Portability Guide. It is also | |
04b9968b | 5243 | available on all systems derived from System V. |
0501d603 | 5244 | |
cf29ffbe | 5245 | The function returns the value @code{MM_OK} if no error occurred. If |
0501d603 UD |
5246 | only the printing to standard error failed, it returns @code{MM_NOMSG}. |
5247 | If printing to the console fails, it returns @code{MM_NOCON}. If | |
cf29ffbe | 5248 | nothing is printed @code{MM_NOTOK} is returned. Among situations where |
0501d603 UD |
5249 | all outputs fail this last value is also returned if a parameter value |
5250 | is incorrect. | |
5251 | @end deftypefun | |
5252 | ||
0bc93a2f | 5253 | There are two environment variables which influence the behavior of |
0501d603 UD |
5254 | @code{fmtmsg}. The first is @code{MSGVERB}. It is used to control the |
5255 | output actually happening on standard error (@emph{not} the console | |
04b9968b | 5256 | output). Each of the five fields can explicitly be enabled. To do |
0501d603 | 5257 | this the user has to put the @code{MSGVERB} variable with a format like |
cf29ffbe | 5258 | the following in the environment before calling the @code{fmtmsg} function |
0501d603 UD |
5259 | the first time: |
5260 | ||
5261 | @smallexample | |
95fdc6a0 | 5262 | MSGVERB=@var{keyword}[:@var{keyword}[:@dots{}]] |
0501d603 UD |
5263 | @end smallexample |
5264 | ||
5265 | Valid @var{keyword}s are @code{label}, @code{severity}, @code{text}, | |
5266 | @code{action}, and @code{tag}. If the environment variable is not given | |
5267 | or is the empty string, a not supported keyword is given or the value is | |
5268 | somehow else invalid, no part of the message is masked out. | |
5269 | ||
0bc93a2f | 5270 | The second environment variable which influences the behavior of |
0501d603 | 5271 | @code{fmtmsg} is @code{SEV_LEVEL}. This variable and the change in the |
0bc93a2f | 5272 | behavior of @code{fmtmsg} is not specified in the X/Open Portability |
0501d603 | 5273 | Guide. It is available in System V systems, though. It can be used to |
cf29ffbe | 5274 | introduce new severity levels. By default, only the five severity levels |
0501d603 UD |
5275 | described above are available. Any other numeric value would make |
5276 | @code{fmtmsg} print nothing. | |
5277 | ||
5278 | If the user puts @code{SEV_LEVEL} with a format like | |
5279 | ||
5280 | @smallexample | |
95fdc6a0 | 5281 | SEV_LEVEL=[@var{description}[:@var{description}[:@dots{}]]] |
0501d603 UD |
5282 | @end smallexample |
5283 | ||
5284 | @noindent | |
5285 | in the environment of the process before the first call to | |
5286 | @code{fmtmsg}, where @var{description} has a value of the form | |
5287 | ||
5288 | @smallexample | |
5289 | @var{severity-keyword},@var{level},@var{printstring} | |
5290 | @end smallexample | |
5291 | ||
5292 | The @var{severity-keyword} part is not used by @code{fmtmsg} but it has | |
5293 | to be present. The @var{level} part is a string representation of a | |
5294 | number. The numeric value must be a number greater than 4. This value | |
5295 | must be used in the @var{severity} parameter of @code{fmtmsg} to select | |
5296 | this class. It is not possible to overwrite any of the predefined | |
5297 | classes. The @var{printstring} is the string printed when a message of | |
5298 | this class is processed by @code{fmtmsg} (see above, @code{fmtsmg} does | |
5299 | not print the numeric value but instead the string representation). | |
5300 | ||
5301 | ||
5302 | @node Adding Severity Classes | |
5303 | @subsection Adding Severity Classes | |
5304 | @cindex severity class | |
5305 | ||
04b9968b | 5306 | There is another possibility to introduce severity classes besides using |
0501d603 UD |
5307 | the environment variable @code{SEV_LEVEL}. This simplifies the task of |
5308 | introducing new classes in a running program. One could use the | |
5309 | @code{setenv} or @code{putenv} function to set the environment variable, | |
cf29ffbe | 5310 | but this is toilsome. |
0501d603 UD |
5311 | |
5312 | @deftypefun int addseverity (int @var{severity}, const char *@var{string}) | |
04b9968b | 5313 | This function allows the introduction of new severity classes which can be |
0501d603 UD |
5314 | addressed by the @var{severity} parameter of the @code{fmtmsg} function. |
5315 | The @var{severity} parameter of @code{addseverity} must match the value | |
04b9968b | 5316 | for the parameter with the same name of @code{fmtmsg}, and @var{string} |
0501d603 UD |
5317 | is the string printed in the actual messages instead of the numeric |
5318 | value. | |
5319 | ||
5320 | If @var{string} is @code{NULL} the severity class with the numeric value | |
5321 | according to @var{severity} is removed. | |
5322 | ||
cf29ffbe UD |
5323 | It is not possible to overwrite or remove one of the default severity |
5324 | classes. All calls to @code{addseverity} with @var{severity} set to one | |
5325 | of the values for the default classes will fail. | |
5326 | ||
0501d603 UD |
5327 | The return value is @code{MM_OK} if the task was successfully performed. |
5328 | If the return value is @code{MM_NOTOK} something went wrong. This could | |
5329 | mean that no more memory is available or a class is not available when | |
5330 | it has to be removed. | |
5331 | ||
5332 | This function is not specified in the X/Open Portability Guide although | |
cf29ffbe | 5333 | the @code{fmtsmg} function is. It is available on System V systems. |
0501d603 UD |
5334 | @end deftypefun |
5335 | ||
5336 | ||
5337 | @node Example | |
5338 | @subsection How to use @code{fmtmsg} and @code{addseverity} | |
5339 | ||
5340 | Here is a simple example program to illustrate the use of the both | |
5341 | functions described in this section. | |
5342 | ||
5343 | @smallexample | |
5344 | @include fmtmsgexpl.c.texi | |
5345 | @end smallexample | |
5346 | ||
04b9968b UD |
5347 | The second call to @code{fmtmsg} illustrates a use of this function as |
5348 | it usually occurs on System V systems, which heavily use this function. | |
5349 | It seems worthwhile to give a short explanation here of how this system | |
5350 | works on System V. The value of the | |
0bc93a2f | 5351 | @var{label} field (@code{UX:cat}) says that the error occurred in the |
0501d603 UD |
5352 | Unix program @code{cat}. The explanation of the error follows and the |
5353 | value for the @var{action} parameter is @code{"refer to manual"}. One | |
04b9968b | 5354 | could be more specific here, if necessary. The @var{tag} field contains, |
0501d603 UD |
5355 | as proposed above, the value of the string given for the @var{label} |
5356 | parameter, and additionally a unique ID (@code{001} in this case). For | |
5357 | a GNU environment this string could contain a reference to the | |
5358 | corresponding node in the Info page for the program. | |
5359 | ||
5360 | @noindent | |
5361 | Running this program without specifying the @code{MSGVERB} and | |
5362 | @code{SEV_LEVEL} function produces the following output: | |
5363 | ||
5364 | @smallexample | |
5365 | UX:cat: NOTE2: invalid syntax | |
5366 | TO FIX: refer to manual UX:cat:001 | |
5367 | @end smallexample | |
5368 | ||
5369 | We see the different fields of the message and how the extra glue (the | |
5370 | colons and the @code{TO FIX} string) are printed. But only one of the | |
5371 | three calls to @code{fmtmsg} produced output. The first call does not | |
5372 | print anything because the @var{label} parameter is not in the correct | |
8b7fb588 UD |
5373 | form. The string must contain two fields, separated by a colon |
5374 | (@pxref{Printing Formatted Messages}). The third @code{fmtmsg} call | |
0501d603 UD |
5375 | produced no output since the class with the numeric value @code{6} is |
5376 | not defined. Although a class with numeric value @code{5} is also not | |
04b9968b | 5377 | defined by default, the call to @code{addseverity} introduces it and |
0ea554bf | 5378 | the second call to @code{fmtmsg} produces the above output. |
0501d603 UD |
5379 | |
5380 | When we change the environment of the program to contain | |
5381 | @code{SEV_LEVEL=XXX,6,NOTE} when running it we get a different result: | |
5382 | ||
5383 | @smallexample | |
5384 | UX:cat: NOTE2: invalid syntax | |
5385 | TO FIX: refer to manual UX:cat:001 | |
5386 | label:foo: NOTE: text | |
5387 | TO FIX: action tag | |
5388 | @end smallexample | |
5389 | ||
04b9968b | 5390 | Now the third call to @code{fmtmsg} produced some output and we see how |
0501d603 UD |
5391 | the string @code{NOTE} from the environment variable appears in the |
5392 | message. | |
5393 | ||
04b9968b | 5394 | Now we can reduce the output by specifying which fields we are |
0501d603 UD |
5395 | interested in. If we additionally set the environment variable |
5396 | @code{MSGVERB} to the value @code{severity:label:action} we get the | |
5397 | following output: | |
5398 | ||
5399 | @smallexample | |
5400 | UX:cat: NOTE2 | |
5401 | TO FIX: refer to manual | |
5402 | label:foo: NOTE | |
5403 | TO FIX: action | |
5404 | @end smallexample | |
5405 | ||
5406 | @noindent | |
5407 | I.e., the output produced by the @var{text} and the @var{tag} parameters | |
cf29ffbe | 5408 | to @code{fmtmsg} vanished. Please also note that now there is no colon |
0501d603 | 5409 | after the @code{NOTE} and @code{NOTE2} strings in the output. This is |
04b9968b | 5410 | not necessary since there is no more output on this line because the text |
0501d603 | 5411 | is missing. |