2 This file is part of GNUnet.
3 (C) 2001-2013 Christian Grothoff (and other contributing authors)
5 GNUnet is free software; you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published
7 by the Free Software Foundation; either version 3, or (at your
8 option) any later version.
10 GNUnet is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with GNUnet; see the file COPYING. If not, write to the
17 Free Software Foundation, Inc., 59 Temple Place - Suite 330,
18 Boston, MA 02111-1307, USA.
22 * @file include/gnunet_strings_lib.h
23 * @brief strings and string handling functions (including malloc
24 * and string tokenizing)
26 * @author Christian Grothoff
27 * @author Krista Bennett
28 * @author Gerd Knorr <kraxel@bytesex.org>
29 * @author Ioana Patrascu
30 * @author Tzvetan Horozov
33 #ifndef GNUNET_STRINGS_LIB_H
34 #define GNUNET_STRINGS_LIB_H
36 /* we need size_t, and since it can be both unsigned int
37 or unsigned long long, this IS platform dependent;
38 but "stdlib.h" should be portable 'enough' to be
39 unconditionally available... */
45 #if 0 /* keep Emacsens' auto-indent happy */
50 #include "gnunet_time_lib.h"
54 * Convert a given fancy human-readable size to bytes.
56 * @param fancy_size human readable string (i.e. 1 MB)
57 * @param size set to the size in bytes
58 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
61 GNUNET_STRINGS_fancy_size_to_bytes (const char *fancy_size,
62 unsigned long long *size);
66 * Convert a given fancy human-readable time to our internal
69 * @param fancy_time human readable string (i.e. 1 minute)
70 * @param rtime set to the relative time
71 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
74 GNUNET_STRINGS_fancy_time_to_relative (const char *fancy_time,
75 struct GNUNET_TIME_Relative *rtime);
80 * Convert a given fancy human-readable time to our internal
81 * representation. The human-readable time is expected to be
82 * in local time, whereas the returned value will be in UTC.
84 * @param fancy_time human readable string (i.e. %Y-%m-%d %H:%M:%S)
85 * @param atime set to the absolute time
86 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
89 GNUNET_STRINGS_fancy_time_to_absolute (const char *fancy_time,
90 struct GNUNET_TIME_Absolute *atime);
94 * Convert a given filesize into a fancy human-readable format.
96 * @param size number of bytes
97 * @return fancy representation of the size (possibly rounded) for humans
100 GNUNET_STRINGS_byte_size_fancy (unsigned long long size);
104 * Convert the len characters long character sequence
105 * given in input that is in the given input charset
106 * to a string in given output charset.
108 * @param input input string
109 * @param len number of bytes in @a input
110 * @param input_charset character set used for @a input
111 * @param output_charset desired character set for the return value
112 * @return the converted string (0-terminated),
113 * if conversion fails, a copy of the orignal
114 * string is returned.
117 GNUNET_STRINGS_conv (const char *input, size_t len,
118 const char *input_charset,
119 const char *output_charset);
123 * Convert the len characters long character sequence
124 * given in input that is in the given charset
127 * @param input the input string (not necessarily 0-terminated)
128 * @param len the number of bytes in the @a input
129 * @param charset character set to convert from
130 * @return the converted string (0-terminated)
133 GNUNET_STRINGS_to_utf8 (const char *input,
135 const char *charset);
139 * Convert the len bytes-long UTF-8 string
140 * given in input to the given charset.
142 * @param input the input string (not necessarily 0-terminated)
143 * @param len the number of bytes in the @a input
144 * @param charset character set to convert to
145 * @return the converted string (0-terminated),
146 * if conversion fails, a copy of the orignal
147 * string is returned.
150 GNUNET_STRINGS_from_utf8 (const char *input,
152 const char *charset);
156 * Convert the utf-8 input string to lower case.
157 * Output needs to be allocated appropriately.
159 * @param input input string
160 * @param output output buffer
163 GNUNET_STRINGS_utf8_tolower (const char *input,
168 * Convert the utf-8 input string to upper case.
169 * Output needs to be allocated appropriately.
171 * @param input input string
172 * @param output output buffer
175 GNUNET_STRINGS_utf8_toupper (const char *input,
180 * Complete filename (a la shell) from abbrevition.
182 * @param fil the name of the file, may contain ~/ or
183 * be relative to the current directory
184 * @return the full file name,
185 * NULL is returned on error
188 GNUNET_STRINGS_filename_expand (const char *fil);
192 * Fill a buffer of the given size with count 0-terminated strings
193 * (given as varargs). If "buffer" is NULL, only compute the amount
194 * of space required (sum of "strlen(arg)+1").
196 * Unlike using "snprintf" with "%s", this function will add
197 * 0-terminators after each string. The
198 * "GNUNET_string_buffer_tokenize" function can be used to parse the
199 * buffer back into individual strings.
201 * @param buffer the buffer to fill with strings, can
202 * be NULL in which case only the necessary
203 * amount of space will be calculated
204 * @param size number of bytes available in buffer
205 * @param count number of strings that follow
206 * @param ... count 0-terminated strings to copy to buffer
207 * @return number of bytes written to the buffer
208 * (or number of bytes that would have been written)
211 GNUNET_STRINGS_buffer_fill (char *buffer,
218 * Given a buffer of a given size, find "count" 0-terminated strings
219 * in the buffer and assign the count (varargs) of type "const char**"
220 * to the locations of the respective strings in the buffer.
222 * @param buffer the buffer to parse
223 * @param size size of the @a buffer
224 * @param count number of strings to locate
225 * @param ... pointers to where to store the strings
226 * @return offset of the character after the last 0-termination
227 * in the buffer, or 0 on error.
230 GNUNET_STRINGS_buffer_tokenize (const char *buffer,
232 unsigned int count, ...);
238 * Like `asctime`, except for GNUnet time. Converts a GNUnet internal
239 * absolute time (which is in UTC) to a string in local time.
240 * Note that the returned value will be overwritten if this function
243 * @param t the absolute time to convert
244 * @return timestamp in human-readable form in local time
247 GNUNET_STRINGS_absolute_time_to_string (struct GNUNET_TIME_Absolute t);
252 * Give relative time in human-readable fancy format.
253 * This is one of the very few calls in the entire API that is
256 * @param delta time in milli seconds
257 * @param do_round are we allowed to round a bit?
258 * @return string in human-readable form
261 GNUNET_STRINGS_relative_time_to_string (struct GNUNET_TIME_Relative delta,
267 * Returns a pointer to a part of filename (allocates nothing)!
269 * @param filename filename to extract basename from
270 * @return short (base) name of the file (that is, everything following the
271 * last directory separator in filename. If filename ends with a
272 * directory separator, the result will be a zero-length string.
273 * If filename has no directory separators, the result is filename
277 GNUNET_STRINGS_get_short_name (const char *filename);
281 * Convert binary data to ASCII encoding. The ASCII encoding is rather
282 * GNUnet specific. It was chosen such that it only uses characters
283 * in [0-9A-V], can be produced without complex arithmetics and uses a
284 * small number of characters. The GNUnet encoding uses 103 characters.
285 * Does not append 0-terminator, but returns a pointer to the place where
286 * it should be placed, if needed.
288 * @param data data to encode
289 * @param size size of data (in bytes)
290 * @param out buffer to fill
291 * @param out_size size of the buffer. Must be large enough to hold
292 * ((size*8) + (((size*8) % 5) > 0 ? 5 - ((size*8) % 5) : 0)) / 5
293 * @return pointer to the next byte in 'out' or NULL on error.
296 GNUNET_STRINGS_data_to_string (const void *data,
303 * Convert ASCII encoding back to data
304 * out_size must match exactly the size of the data before it was encoded.
306 * @param enc the encoding
307 * @param enclen number of characters in 'enc' (without 0-terminator, which can be missing)
308 * @param out location where to store the decoded data
309 * @param out_size size of the output buffer @a out
310 * @return #GNUNET_OK on success, #GNUNET_SYSERR if result has the wrong encoding
313 GNUNET_STRINGS_string_to_data (const char *enc,
320 * Encode into Base64.
322 * @param data the data to encode
323 * @param len the length of the input
324 * @param output where to write the output (*output should be NULL,
326 * @return the size of the output
329 GNUNET_STRINGS_base64_encode (const char *data, size_t len, char **output);
332 * Decode from Base64.
334 * @param data the data to encode
335 * @param len the length of the input
336 * @param output where to write the output (*output should be NULL,
338 * @return the size of the output
341 GNUNET_STRINGS_base64_decode (const char *data, size_t len, char **output);
345 * Parse a path that might be an URI.
347 * @param path path to parse. Must be NULL-terminated.
348 * @param scheme_part a pointer to 'char *' where a pointer to a string that
349 * represents the URI scheme will be stored. Can be NULL. The string is
350 * allocated by the function, and should be freed by GNUNET_free() when
351 * it is no longer needed.
352 * @param path_part a pointer to 'const char *' where a pointer to the path
353 * part of the URI will be stored. Can be NULL. Points to the same block
354 * of memory as 'path', and thus must not be freed. Might point to '\0',
355 * if path part is zero-length.
356 * @return #GNUNET_YES if it's an URI, #GNUNET_NO otherwise. If 'path' is not
357 * an URI, '* scheme_part' and '*path_part' will remain unchanged
358 * (if they weren't NULL).
361 GNUNET_STRINGS_parse_uri (const char *path,
363 const char **path_part);
367 * Check whether filename is absolute or not, and if it's an URI
369 * @param filename filename to check
370 * @param can_be_uri #GNUNET_YES to check for being URI, #GNUNET_NO - to
371 * assume it's not URI
372 * @param r_is_uri a pointer to an int that is set to #GNUNET_YES if 'filename'
373 * is URI and to GNUNET_NO otherwise. Can be NULL. If 'can_be_uri' is
374 * not #GNUNET_YES, *r_is_uri is set to #GNUNET_NO.
375 * @param r_uri_scheme a pointer to a char * that is set to a pointer to URI scheme.
376 * The string is allocated by the function, and should be freed with
377 * GNUNET_free (). Can be NULL.
378 * @return #GNUNET_YES if 'filename' is absolute, #GNUNET_NO otherwise.
381 GNUNET_STRINGS_path_is_absolute (const char *filename,
384 char **r_uri_scheme);
388 * Flags for what we should check a file for.
390 enum GNUNET_STRINGS_FilenameCheck
393 * Check that it exists.
395 GNUNET_STRINGS_CHECK_EXISTS = 0x00000001,
398 * Check that it is a directory.
400 GNUNET_STRINGS_CHECK_IS_DIRECTORY = 0x00000002,
403 * Check that it is a link.
405 GNUNET_STRINGS_CHECK_IS_LINK = 0x00000004,
408 * Check that the path is an absolute path.
410 GNUNET_STRINGS_CHECK_IS_ABSOLUTE = 0x00000008
415 * Perform checks on @a filename. FIXME: some duplication with
416 * "GNUNET_DISK_"-APIs. We should unify those.
418 * @param filename file to check
419 * @param checks checks to perform
420 * @return #GNUNET_YES if all checks pass, #GNUNET_NO if at least one of them
421 * fails, #GNUNET_SYSERR when a check can't be performed
424 GNUNET_STRINGS_check_filename (const char *filename,
425 enum GNUNET_STRINGS_FilenameCheck checks);
429 * Tries to convert @a zt_addr string to an IPv6 address.
430 * The string is expected to have the format "[ABCD::01]:80".
432 * @param zt_addr 0-terminated string. May be mangled by the function.
433 * @param addrlen length of zt_addr (not counting 0-terminator).
434 * @param r_buf a buffer to fill. Initially gets filled with zeroes,
435 * then its sin6_port, sin6_family and sin6_addr are set appropriately.
436 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
437 * case the contents of r_buf are undefined.
440 GNUNET_STRINGS_to_address_ipv6 (const char *zt_addr,
442 struct sockaddr_in6 *r_buf);
446 * Tries to convert @a zt_addr string to an IPv4 address.
447 * The string is expected to have the format "1.2.3.4:80".
449 * @param zt_addr 0-terminated string. May be mangled by the function.
450 * @param addrlen length of zt_addr (not counting 0-terminator).
451 * @param r_buf a buffer to fill.
452 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which case
453 * the contents of r_buf are undefined.
456 GNUNET_STRINGS_to_address_ipv4 (const char *zt_addr,
458 struct sockaddr_in *r_buf);
462 * Tries to convert @a addr string to an IP (v4 or v6) address.
463 * Will automatically decide whether to treat 'addr' as v4 or v6 address.
465 * @param addr a string, may not be 0-terminated.
466 * @param addrlen number of bytes in @a addr (if addr is 0-terminated,
467 * 0-terminator should not be counted towards addrlen).
468 * @param r_buf a buffer to fill.
469 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
470 * case the contents of r_buf are undefined.
473 GNUNET_STRINGS_to_address_ip (const char *addr,
475 struct sockaddr_storage *r_buf);
479 * Returns utf-8 encoded arguments. Does nothing (returns a copy of
480 * @a argc and @a argv) on any platform other than W32. Returned @a
481 * argv has `u8argv[u8argc] == NULL`. Returned @a argv is a single
482 * memory block, and can be freed with a single GNUNET_free() call.
484 * @param argc argc (as given by main())
485 * @param argv argv (as given by main())
486 * @param u8argc a location to store new argc in (though it's th same as argc)
487 * @param u8argv a location to store new argv in
488 * @return #GNUNET_OK on success, #GNUNET_SYSERR on failure
491 GNUNET_STRINGS_get_utf8_args (int argc,
494 char *const **u8argv);
497 /* ***************** IPv4/IPv6 parsing ****************** */
499 struct GNUNET_STRINGS_PortPolicy
503 * Starting port range (0 if none given).
508 * End of port range (0 if none given).
513 * #GNUNET_YES if the port range should be negated
516 int negate_portrange;
522 * @brief IPV4 network in CIDR notation.
524 struct GNUNET_STRINGS_IPv4NetworkPolicy
529 struct in_addr network;
534 struct in_addr netmask;
537 * Policy for port access.
539 struct GNUNET_STRINGS_PortPolicy pp;
545 * @brief network in CIDR notation for IPV6.
547 struct GNUNET_STRINGS_IPv6NetworkPolicy
552 struct in6_addr network;
557 struct in6_addr netmask;
560 * Policy for port access.
562 struct GNUNET_STRINGS_PortPolicy pp;
568 * Parse an IPv4 network policy. The argument specifies a list of
569 * subnets. The format is <tt>(network[/netmask][:[!]SPORT-DPORT];)*</tt>
570 * (no whitespace, must be terminated with a semicolon). The network
571 * must be given in dotted-decimal notation. The netmask can be given
572 * in CIDR notation (/16) or in dotted-decimal (/255.255.0.0).
574 * @param routeListX a string specifying the IPv4 subnets
575 * @return the converted list, terminated with all zeros;
576 * NULL if the synatx is flawed
578 struct GNUNET_STRINGS_IPv4NetworkPolicy *
579 GNUNET_STRINGS_parse_ipv4_policy (const char *routeListX);
583 * Parse an IPv6 network policy. The argument specifies a list of
584 * subnets. The format is <tt>(network[/netmask[:[!]SPORT[-DPORT]]];)*</tt>
585 * (no whitespace, must be terminated with a semicolon). The network
586 * must be given in colon-hex notation. The netmask must be given in
587 * CIDR notation (/16) or can be omitted to specify a single host.
588 * Note that the netmask is mandatory if ports are specified.
590 * @param routeListX a string specifying the policy
591 * @return the converted list, 0-terminated, NULL if the synatx is flawed
593 struct GNUNET_STRINGS_IPv6NetworkPolicy *
594 GNUNET_STRINGS_parse_ipv6_policy (const char *routeListX);
598 #if 0 /* keep Emacsens' auto-indent happy */
606 /* ifndef GNUNET_UTIL_STRING_H */
608 /* end of gnunet_util_string.h */