2 This file is part of GNUnet.
3 Copyright (C) 2001-2013 GNUnet e.V.
5 GNUnet is free software: you can redistribute it and/or modify it
6 under the terms of the GNU Affero General Public License as published
7 by the Free Software Foundation, either version 3 of the License,
8 or (at your 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 Affero General Public License for more details.
15 You should have received a copy of the GNU Affero General Public License
16 along with this program. If not, see <http://www.gnu.org/licenses/>.
20 * @author Christian Grothoff
21 * @author Krista Bennett
22 * @author Gerd Knorr <kraxel@bytesex.org>
23 * @author Ioana Patrascu
24 * @author Tzvetan Horozov
27 * Strings and string handling functions
29 * @defgroup strings Strings library
30 * Strings and string handling functions, including malloc and string tokenizing.
34 #ifndef GNUNET_STRINGS_LIB_H
35 #define GNUNET_STRINGS_LIB_H
37 /* we need size_t, and since it can be both unsigned int
38 or unsigned long long, this IS platform dependent;
39 but "stdlib.h" should be portable 'enough' to be
40 unconditionally available... */
46 #if 0 /* keep Emacsens' auto-indent happy */
51 #include "gnunet_time_lib.h"
55 * Convert a given fancy human-readable size to bytes.
57 * @param fancy_size human readable string (i.e. 1 MB)
58 * @param size set to the size in bytes
59 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
62 GNUNET_STRINGS_fancy_size_to_bytes (const char *fancy_size,
63 unsigned long long *size);
67 * Convert a given fancy human-readable time to our internal
70 * @param fancy_time human readable string (i.e. 1 minute)
71 * @param rtime set to the relative time
72 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
75 GNUNET_STRINGS_fancy_time_to_relative (const char *fancy_time,
76 struct GNUNET_TIME_Relative *rtime);
81 * Convert a given fancy human-readable time to our internal
82 * representation. The human-readable time is expected to be
83 * in local time, whereas the returned value will be in UTC.
85 * @param fancy_time human readable string (i.e. %Y-%m-%d %H:%M:%S)
86 * @param atime set to the absolute time
87 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
90 GNUNET_STRINGS_fancy_time_to_absolute (const char *fancy_time,
91 struct GNUNET_TIME_Absolute *atime);
95 * Convert a given filesize into a fancy human-readable format.
97 * @param size number of bytes
98 * @return fancy representation of the size (possibly rounded) for humans
101 GNUNET_STRINGS_byte_size_fancy (unsigned long long size);
105 * Convert the len characters long character sequence
106 * given in input that is in the given input charset
107 * to a string in given output charset.
109 * @param input input string
110 * @param len number of bytes in @a input
111 * @param input_charset character set used for @a input
112 * @param output_charset desired character set for the return value
113 * @return the converted string (0-terminated),
114 * if conversion fails, a copy of the orignal
115 * string is returned.
118 GNUNET_STRINGS_conv (const char *input, size_t len,
119 const char *input_charset,
120 const char *output_charset);
124 * Convert the len characters long character sequence
125 * given in input that is in the given charset
128 * @param input the input string (not necessarily 0-terminated)
129 * @param len the number of bytes in the @a input
130 * @param charset character set to convert from
131 * @return the converted string (0-terminated)
134 GNUNET_STRINGS_to_utf8 (const char *input,
136 const char *charset);
140 * Convert the len bytes-long UTF-8 string
141 * given in input to the given charset.
143 * @param input the input string (not necessarily 0-terminated)
144 * @param len the number of bytes in the @a input
145 * @param charset character set to convert to
146 * @return the converted string (0-terminated),
147 * if conversion fails, a copy of the orignal
148 * string is returned.
151 GNUNET_STRINGS_from_utf8 (const char *input,
153 const char *charset);
157 * Convert the utf-8 input string to lower case.
158 * Output needs to be allocated appropriately.
160 * @param input input string
161 * @param output output buffer
164 GNUNET_STRINGS_utf8_tolower (const char *input,
169 * Convert the utf-8 input string to upper case.
170 * Output needs to be allocated appropriately.
172 * @param input input string
173 * @param output output buffer
176 GNUNET_STRINGS_utf8_toupper (const char *input,
181 * Complete filename (a la shell) from abbrevition.
183 * @param fil the name of the file, may contain ~/ or
184 * be relative to the current directory
185 * @return the full file name,
186 * NULL is returned on error
189 GNUNET_STRINGS_filename_expand (const char *fil);
193 * Fill a buffer of the given size with count 0-terminated strings
194 * (given as varargs). If "buffer" is NULL, only compute the amount
195 * of space required (sum of "strlen(arg)+1").
197 * Unlike using "snprintf" with "%s", this function will add
198 * 0-terminators after each string. The
199 * "GNUNET_string_buffer_tokenize" function can be used to parse the
200 * buffer back into individual strings.
202 * @param buffer the buffer to fill with strings, can
203 * be NULL in which case only the necessary
204 * amount of space will be calculated
205 * @param size number of bytes available in buffer
206 * @param count number of strings that follow
207 * @param ... count 0-terminated strings to copy to buffer
208 * @return number of bytes written to the buffer
209 * (or number of bytes that would have been written)
212 GNUNET_STRINGS_buffer_fill (char *buffer,
219 * Given a buffer of a given size, find "count" 0-terminated strings
220 * in the buffer and assign the count (varargs) of type "const char**"
221 * to the locations of the respective strings in the buffer.
223 * @param buffer the buffer to parse FIXME: not 'const', is it?
224 * @param size size of the @a buffer
225 * @param count number of strings to locate
226 * @param ... pointers to where to store the strings
227 * @return offset of the character after the last 0-termination
228 * in the buffer, or 0 on error.
231 GNUNET_STRINGS_buffer_tokenize (const char *buffer,
233 unsigned int count, ...);
239 * Like `asctime`, except for GNUnet time. Converts a GNUnet internal
240 * absolute time (which is in UTC) to a string in local time.
241 * Note that the returned value will be overwritten if this function
244 * @param t the absolute time to convert
245 * @return timestamp in human-readable form in local time
248 GNUNET_STRINGS_absolute_time_to_string (struct GNUNET_TIME_Absolute t);
253 * Give relative time in human-readable fancy format.
254 * This is one of the very few calls in the entire API that is
257 * @param delta time in milli seconds
258 * @param do_round are we allowed to round a bit?
259 * @return string in human-readable form
262 GNUNET_STRINGS_relative_time_to_string (struct GNUNET_TIME_Relative delta,
268 * Returns a pointer to a part of filename (allocates nothing)!
270 * @param filename filename to extract basename from
271 * @return short (base) name of the file (that is, everything following the
272 * last directory separator in filename. If filename ends with a
273 * directory separator, the result will be a zero-length string.
274 * If filename has no directory separators, the result is filename
278 GNUNET_STRINGS_get_short_name (const char *filename);
282 * Convert binary data to ASCII encoding using CrockfordBase32.
283 * Does not append 0-terminator, but returns a pointer to the place where
284 * it should be placed, if needed.
286 * @param data data to encode
287 * @param size size of data (in bytes)
288 * @param out buffer to fill
289 * @param out_size size of the buffer. Must be large enough to hold
290 * ((size*8) + (((size*8) % 5) > 0 ? 5 - ((size*8) % 5) : 0)) / 5
291 * @return pointer to the next byte in 'out' or NULL on error.
294 GNUNET_STRINGS_data_to_string (const void *data,
301 * Return the base32crockford encoding of the given buffer.
303 * The returned string will be freshly allocated, and must be free'd
304 * with #GNUNET_free().
306 * @param buf buffer with data
307 * @param size size of the buffer @a buf
308 * @return freshly allocated, null-terminated string
311 GNUNET_STRINGS_data_to_string_alloc (const void *buf,
316 * Convert CrockfordBase32 encoding back to data.
317 * @a out_size must match exactly the size of the data before it was encoded.
319 * @param enc the encoding
320 * @param enclen number of characters in @a enc (without 0-terminator, which can be missing)
321 * @param out location where to store the decoded data
322 * @param out_size size of the output buffer @a out
323 * @return #GNUNET_OK on success, #GNUNET_SYSERR if result has the wrong encoding
326 GNUNET_STRINGS_string_to_data (const char *enc,
333 * Encode into Base64.
335 * @param data the data to encode
336 * @param len the length of the input
337 * @param output where to write the output (*output should be NULL,
339 * @return the size of the output
342 GNUNET_STRINGS_base64_encode (const void *in,
348 * Decode from Base64.
350 * @param data the data to encode
351 * @param len the length of the input
352 * @param[out] output where to write the output (*output should be NULL,
354 * @return the size of the output
357 GNUNET_STRINGS_base64_decode (const char *data,
363 * Convert a peer path to a human-readable string.
365 * @param pids array of PIDs to convert to a string
366 * @param num_pids length of the @a pids array
367 * @return string representing the array of @a pids
370 GNUNET_STRINGS_pp2s (const struct GNUNET_PeerIdentity *pids,
371 unsigned int num_pids);
375 * Parse a path that might be an URI.
377 * @param path path to parse. Must be NULL-terminated.
378 * @param[out] scheme_part pointer to a string that
379 * represents the URI scheme will be stored. Can be NULL. The string is
380 * allocated by the function, and should be freed by GNUNET_free() when
381 * it is no longer needed.
382 * @param path_part a pointer to 'const char *' where a pointer to the path
383 * part of the URI will be stored. Can be NULL. Points to the same block
384 * of memory as @a path, and thus must not be freed. Might point to '\0',
385 * if path part is zero-length.
386 * @return #GNUNET_YES if it's an URI, #GNUNET_NO otherwise. If 'path' is not
387 * an URI, '* scheme_part' and '*path_part' will remain unchanged
388 * (if they weren't NULL).
391 GNUNET_STRINGS_parse_uri (const char *path,
393 const char **path_part);
397 * Check whether filename is absolute or not, and if it's an URI
399 * @param filename filename to check
400 * @param can_be_uri #GNUNET_YES to check for being URI, #GNUNET_NO - to
401 * assume it's not URI
402 * @param r_is_uri a pointer to an int that is set to #GNUNET_YES if 'filename'
403 * is URI and to GNUNET_NO otherwise. Can be NULL. If 'can_be_uri' is
404 * not #GNUNET_YES, *r_is_uri is set to #GNUNET_NO.
405 * @param r_uri_scheme a pointer to a char * that is set to a pointer to URI scheme.
406 * The string is allocated by the function, and should be freed with
407 * GNUNET_free (). Can be NULL.
408 * @return #GNUNET_YES if 'filename' is absolute, #GNUNET_NO otherwise.
411 GNUNET_STRINGS_path_is_absolute (const char *filename,
414 char **r_uri_scheme);
418 * Flags for what we should check a file for.
420 enum GNUNET_STRINGS_FilenameCheck
423 * Check that it exists.
425 GNUNET_STRINGS_CHECK_EXISTS = 0x00000001,
428 * Check that it is a directory.
430 GNUNET_STRINGS_CHECK_IS_DIRECTORY = 0x00000002,
433 * Check that it is a link.
435 GNUNET_STRINGS_CHECK_IS_LINK = 0x00000004,
438 * Check that the path is an absolute path.
440 GNUNET_STRINGS_CHECK_IS_ABSOLUTE = 0x00000008
445 * Perform checks on @a filename. FIXME: some duplication with
446 * "GNUNET_DISK_"-APIs. We should unify those.
448 * @param filename file to check
449 * @param checks checks to perform
450 * @return #GNUNET_YES if all checks pass, #GNUNET_NO if at least one of them
451 * fails, #GNUNET_SYSERR when a check can't be performed
454 GNUNET_STRINGS_check_filename (const char *filename,
455 enum GNUNET_STRINGS_FilenameCheck checks);
459 * Tries to convert @a zt_addr string to an IPv6 address.
460 * The string is expected to have the format "[ABCD::01]:80".
462 * @param zt_addr 0-terminated string. May be mangled by the function.
463 * @param addrlen length of zt_addr (not counting 0-terminator).
464 * @param r_buf a buffer to fill. Initially gets filled with zeroes,
465 * then its sin6_port, sin6_family and sin6_addr are set appropriately.
466 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
467 * case the contents of r_buf are undefined.
470 GNUNET_STRINGS_to_address_ipv6 (const char *zt_addr,
472 struct sockaddr_in6 *r_buf);
476 * Tries to convert @a zt_addr string to an IPv4 address.
477 * The string is expected to have the format "1.2.3.4:80".
479 * @param zt_addr 0-terminated string. May be mangled by the function.
480 * @param addrlen length of zt_addr (not counting 0-terminator).
481 * @param r_buf a buffer to fill.
482 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which case
483 * the contents of r_buf are undefined.
486 GNUNET_STRINGS_to_address_ipv4 (const char *zt_addr,
488 struct sockaddr_in *r_buf);
492 * Parse an address given as a string into a
495 * @param addr the address
496 * @param[out] af set to the parsed address family (i.e. AF_INET)
497 * @param[out] sa set to the parsed address
498 * @return 0 on error, otherwise number of bytes in @a sa
501 GNUNET_STRINGS_parse_socket_addr (const char *addr,
503 struct sockaddr **sa);
507 * Tries to convert @a addr string to an IP (v4 or v6) address.
508 * Will automatically decide whether to treat 'addr' as v4 or v6 address.
510 * @param addr a string, may not be 0-terminated.
511 * @param addrlen number of bytes in @a addr (if addr is 0-terminated,
512 * 0-terminator should not be counted towards addrlen).
513 * @param r_buf a buffer to fill.
514 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
515 * case the contents of r_buf are undefined.
518 GNUNET_STRINGS_to_address_ip (const char *addr,
520 struct sockaddr_storage *r_buf);
524 * Returns utf-8 encoded arguments. Does nothing (returns a copy of
525 * @a argc and @a argv) on any platform other than W32. Returned @a
526 * argv has `u8argv[u8argc] == NULL`. Returned @a argv is a single
527 * memory block, and can be freed with a single GNUNET_free() call.
529 * @param argc argc (as given by main())
530 * @param argv argv (as given by main())
531 * @param u8argc a location to store new argc in (though it's th same as argc)
532 * @param u8argv a location to store new argv in
533 * @return #GNUNET_OK on success, #GNUNET_SYSERR on failure
536 GNUNET_STRINGS_get_utf8_args (int argc,
539 char *const **u8argv);
542 /* ***************** IPv4/IPv6 parsing ****************** */
544 struct GNUNET_STRINGS_PortPolicy
548 * Starting port range (0 if none given).
553 * End of port range (0 if none given).
558 * #GNUNET_YES if the port range should be negated
561 int negate_portrange;
567 * @brief IPV4 network in CIDR notation.
569 struct GNUNET_STRINGS_IPv4NetworkPolicy
574 struct in_addr network;
579 struct in_addr netmask;
582 * Policy for port access.
584 struct GNUNET_STRINGS_PortPolicy pp;
590 * @brief network in CIDR notation for IPV6.
592 struct GNUNET_STRINGS_IPv6NetworkPolicy
597 struct in6_addr network;
602 struct in6_addr netmask;
605 * Policy for port access.
607 struct GNUNET_STRINGS_PortPolicy pp;
613 * Parse an IPv4 network policy. The argument specifies a list of
614 * subnets. The format is <tt>(network[/netmask][:[!]SPORT-DPORT];)*</tt>
615 * (no whitespace, must be terminated with a semicolon). The network
616 * must be given in dotted-decimal notation. The netmask can be given
617 * in CIDR notation (/16) or in dotted-decimal (/255.255.0.0).
619 * @param routeListX a string specifying the IPv4 subnets
620 * @return the converted list, terminated with all zeros;
621 * NULL if the synatx is flawed
623 struct GNUNET_STRINGS_IPv4NetworkPolicy *
624 GNUNET_STRINGS_parse_ipv4_policy (const char *routeListX);
628 * Parse an IPv6 network policy. The argument specifies a list of
629 * subnets. The format is <tt>(network[/netmask[:[!]SPORT[-DPORT]]];)*</tt>
630 * (no whitespace, must be terminated with a semicolon). The network
631 * must be given in colon-hex notation. The netmask must be given in
632 * CIDR notation (/16) or can be omitted to specify a single host.
633 * Note that the netmask is mandatory if ports are specified.
635 * @param routeListX a string specifying the policy
636 * @return the converted list, 0-terminated, NULL if the synatx is flawed
638 struct GNUNET_STRINGS_IPv6NetworkPolicy *
639 GNUNET_STRINGS_parse_ipv6_policy (const char *routeListX);
643 #if 0 /* keep Emacsens' auto-indent happy */
650 /* ifndef GNUNET_UTIL_STRING_H */
653 /** @} */ /* end of group */
655 /* end of gnunet_util_string.h */