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 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.
17 * @author Christian Grothoff
18 * @author Krista Bennett
19 * @author Gerd Knorr <kraxel@bytesex.org>
20 * @author Ioana Patrascu
21 * @author Tzvetan Horozov
24 * Strings and string handling functions
26 * @defgroup strings Strings library
27 * Strings and string handling functions, including malloc and string tokenizing.
31 #ifndef GNUNET_STRINGS_LIB_H
32 #define GNUNET_STRINGS_LIB_H
34 /* we need size_t, and since it can be both unsigned int
35 or unsigned long long, this IS platform dependent;
36 but "stdlib.h" should be portable 'enough' to be
37 unconditionally available... */
43 #if 0 /* keep Emacsens' auto-indent happy */
48 #include "gnunet_time_lib.h"
52 * Convert a given fancy human-readable size to bytes.
54 * @param fancy_size human readable string (i.e. 1 MB)
55 * @param size set to the size in bytes
56 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
59 GNUNET_STRINGS_fancy_size_to_bytes (const char *fancy_size,
60 unsigned long long *size);
64 * Convert a given fancy human-readable time to our internal
67 * @param fancy_time human readable string (i.e. 1 minute)
68 * @param rtime set to the relative time
69 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
72 GNUNET_STRINGS_fancy_time_to_relative (const char *fancy_time,
73 struct GNUNET_TIME_Relative *rtime);
78 * Convert a given fancy human-readable time to our internal
79 * representation. The human-readable time is expected to be
80 * in local time, whereas the returned value will be in UTC.
82 * @param fancy_time human readable string (i.e. %Y-%m-%d %H:%M:%S)
83 * @param atime set to the absolute time
84 * @return #GNUNET_OK on success, #GNUNET_SYSERR on error
87 GNUNET_STRINGS_fancy_time_to_absolute (const char *fancy_time,
88 struct GNUNET_TIME_Absolute *atime);
92 * Convert a given filesize into a fancy human-readable format.
94 * @param size number of bytes
95 * @return fancy representation of the size (possibly rounded) for humans
98 GNUNET_STRINGS_byte_size_fancy (unsigned long long size);
102 * Convert the len characters long character sequence
103 * given in input that is in the given input charset
104 * to a string in given output charset.
106 * @param input input string
107 * @param len number of bytes in @a input
108 * @param input_charset character set used for @a input
109 * @param output_charset desired character set for the return value
110 * @return the converted string (0-terminated),
111 * if conversion fails, a copy of the orignal
112 * string is returned.
115 GNUNET_STRINGS_conv (const char *input, size_t len,
116 const char *input_charset,
117 const char *output_charset);
121 * Convert the len characters long character sequence
122 * given in input that is in the given charset
125 * @param input the input string (not necessarily 0-terminated)
126 * @param len the number of bytes in the @a input
127 * @param charset character set to convert from
128 * @return the converted string (0-terminated)
131 GNUNET_STRINGS_to_utf8 (const char *input,
133 const char *charset);
137 * Convert the len bytes-long UTF-8 string
138 * given in input to the given charset.
140 * @param input the input string (not necessarily 0-terminated)
141 * @param len the number of bytes in the @a input
142 * @param charset character set to convert to
143 * @return the converted string (0-terminated),
144 * if conversion fails, a copy of the orignal
145 * string is returned.
148 GNUNET_STRINGS_from_utf8 (const char *input,
150 const char *charset);
154 * Convert the utf-8 input string to lower case.
155 * Output needs to be allocated appropriately.
157 * @param input input string
158 * @param output output buffer
161 GNUNET_STRINGS_utf8_tolower (const char *input,
166 * Convert the utf-8 input string to upper case.
167 * Output needs to be allocated appropriately.
169 * @param input input string
170 * @param output output buffer
173 GNUNET_STRINGS_utf8_toupper (const char *input,
178 * Complete filename (a la shell) from abbrevition.
180 * @param fil the name of the file, may contain ~/ or
181 * be relative to the current directory
182 * @return the full file name,
183 * NULL is returned on error
186 GNUNET_STRINGS_filename_expand (const char *fil);
190 * Fill a buffer of the given size with count 0-terminated strings
191 * (given as varargs). If "buffer" is NULL, only compute the amount
192 * of space required (sum of "strlen(arg)+1").
194 * Unlike using "snprintf" with "%s", this function will add
195 * 0-terminators after each string. The
196 * "GNUNET_string_buffer_tokenize" function can be used to parse the
197 * buffer back into individual strings.
199 * @param buffer the buffer to fill with strings, can
200 * be NULL in which case only the necessary
201 * amount of space will be calculated
202 * @param size number of bytes available in buffer
203 * @param count number of strings that follow
204 * @param ... count 0-terminated strings to copy to buffer
205 * @return number of bytes written to the buffer
206 * (or number of bytes that would have been written)
209 GNUNET_STRINGS_buffer_fill (char *buffer,
216 * Given a buffer of a given size, find "count" 0-terminated strings
217 * in the buffer and assign the count (varargs) of type "const char**"
218 * to the locations of the respective strings in the buffer.
220 * @param buffer the buffer to parse FIXME: not 'const', is it?
221 * @param size size of the @a buffer
222 * @param count number of strings to locate
223 * @param ... pointers to where to store the strings
224 * @return offset of the character after the last 0-termination
225 * in the buffer, or 0 on error.
228 GNUNET_STRINGS_buffer_tokenize (const char *buffer,
230 unsigned int count, ...);
236 * Like `asctime`, except for GNUnet time. Converts a GNUnet internal
237 * absolute time (which is in UTC) to a string in local time.
238 * Note that the returned value will be overwritten if this function
241 * @param t the absolute time to convert
242 * @return timestamp in human-readable form in local time
245 GNUNET_STRINGS_absolute_time_to_string (struct GNUNET_TIME_Absolute t);
250 * Give relative time in human-readable fancy format.
251 * This is one of the very few calls in the entire API that is
254 * @param delta time in milli seconds
255 * @param do_round are we allowed to round a bit?
256 * @return string in human-readable form
259 GNUNET_STRINGS_relative_time_to_string (struct GNUNET_TIME_Relative delta,
265 * Returns a pointer to a part of filename (allocates nothing)!
267 * @param filename filename to extract basename from
268 * @return short (base) name of the file (that is, everything following the
269 * last directory separator in filename. If filename ends with a
270 * directory separator, the result will be a zero-length string.
271 * If filename has no directory separators, the result is filename
275 GNUNET_STRINGS_get_short_name (const char *filename);
279 * Convert binary data to ASCII encoding using CrockfordBase32.
280 * Does not append 0-terminator, but returns a pointer to the place where
281 * it should be placed, if needed.
283 * @param data data to encode
284 * @param size size of data (in bytes)
285 * @param out buffer to fill
286 * @param out_size size of the buffer. Must be large enough to hold
287 * ((size*8) + (((size*8) % 5) > 0 ? 5 - ((size*8) % 5) : 0)) / 5
288 * @return pointer to the next byte in 'out' or NULL on error.
291 GNUNET_STRINGS_data_to_string (const void *data,
298 * Return the base32crockford encoding of the given buffer.
300 * The returned string will be freshly allocated, and must be free'd
301 * with #GNUNET_free().
303 * @param buf buffer with data
304 * @param size size of the buffer @a buf
305 * @return freshly allocated, null-terminated string
308 GNUNET_STRINGS_data_to_string_alloc (const void *buf,
313 * Convert CrockfordBase32 encoding back to data.
314 * @a out_size must match exactly the size of the data before it was encoded.
316 * @param enc the encoding
317 * @param enclen number of characters in @a enc (without 0-terminator, which can be missing)
318 * @param out location where to store the decoded data
319 * @param out_size size of the output buffer @a out
320 * @return #GNUNET_OK on success, #GNUNET_SYSERR if result has the wrong encoding
323 GNUNET_STRINGS_string_to_data (const char *enc,
330 * Encode into Base64.
332 * @param data the data to encode
333 * @param len the length of the input
334 * @param output where to write the output (*output should be NULL,
336 * @return the size of the output
339 GNUNET_STRINGS_base64_encode (const char *data, size_t len, char **output);
343 * Decode from Base64.
345 * @param data the data to encode
346 * @param len the length of the input
347 * @param[out] output where to write the output (*output should be NULL,
349 * @return the size of the output
352 GNUNET_STRINGS_base64_decode (const char *data,
358 * Convert a peer path to a human-readable string.
360 * @param pids array of PIDs to convert to a string
361 * @param num_pids length of the @a pids array
362 * @return string representing the array of @a pids
365 GNUNET_STRINGS_pp2s (const struct GNUNET_PeerIdentity *pids,
366 unsigned int num_pids);
370 * Parse a path that might be an URI.
372 * @param path path to parse. Must be NULL-terminated.
373 * @param[out] scheme_part pointer to a string that
374 * represents the URI scheme will be stored. Can be NULL. The string is
375 * allocated by the function, and should be freed by GNUNET_free() when
376 * it is no longer needed.
377 * @param path_part a pointer to 'const char *' where a pointer to the path
378 * part of the URI will be stored. Can be NULL. Points to the same block
379 * of memory as @a path, and thus must not be freed. Might point to '\0',
380 * if path part is zero-length.
381 * @return #GNUNET_YES if it's an URI, #GNUNET_NO otherwise. If 'path' is not
382 * an URI, '* scheme_part' and '*path_part' will remain unchanged
383 * (if they weren't NULL).
386 GNUNET_STRINGS_parse_uri (const char *path,
388 const char **path_part);
392 * Check whether filename is absolute or not, and if it's an URI
394 * @param filename filename to check
395 * @param can_be_uri #GNUNET_YES to check for being URI, #GNUNET_NO - to
396 * assume it's not URI
397 * @param r_is_uri a pointer to an int that is set to #GNUNET_YES if 'filename'
398 * is URI and to GNUNET_NO otherwise. Can be NULL. If 'can_be_uri' is
399 * not #GNUNET_YES, *r_is_uri is set to #GNUNET_NO.
400 * @param r_uri_scheme a pointer to a char * that is set to a pointer to URI scheme.
401 * The string is allocated by the function, and should be freed with
402 * GNUNET_free (). Can be NULL.
403 * @return #GNUNET_YES if 'filename' is absolute, #GNUNET_NO otherwise.
406 GNUNET_STRINGS_path_is_absolute (const char *filename,
409 char **r_uri_scheme);
413 * Flags for what we should check a file for.
415 enum GNUNET_STRINGS_FilenameCheck
418 * Check that it exists.
420 GNUNET_STRINGS_CHECK_EXISTS = 0x00000001,
423 * Check that it is a directory.
425 GNUNET_STRINGS_CHECK_IS_DIRECTORY = 0x00000002,
428 * Check that it is a link.
430 GNUNET_STRINGS_CHECK_IS_LINK = 0x00000004,
433 * Check that the path is an absolute path.
435 GNUNET_STRINGS_CHECK_IS_ABSOLUTE = 0x00000008
440 * Perform checks on @a filename. FIXME: some duplication with
441 * "GNUNET_DISK_"-APIs. We should unify those.
443 * @param filename file to check
444 * @param checks checks to perform
445 * @return #GNUNET_YES if all checks pass, #GNUNET_NO if at least one of them
446 * fails, #GNUNET_SYSERR when a check can't be performed
449 GNUNET_STRINGS_check_filename (const char *filename,
450 enum GNUNET_STRINGS_FilenameCheck checks);
454 * Tries to convert @a zt_addr string to an IPv6 address.
455 * The string is expected to have the format "[ABCD::01]:80".
457 * @param zt_addr 0-terminated string. May be mangled by the function.
458 * @param addrlen length of zt_addr (not counting 0-terminator).
459 * @param r_buf a buffer to fill. Initially gets filled with zeroes,
460 * then its sin6_port, sin6_family and sin6_addr are set appropriately.
461 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
462 * case the contents of r_buf are undefined.
465 GNUNET_STRINGS_to_address_ipv6 (const char *zt_addr,
467 struct sockaddr_in6 *r_buf);
471 * Tries to convert @a zt_addr string to an IPv4 address.
472 * The string is expected to have the format "1.2.3.4:80".
474 * @param zt_addr 0-terminated string. May be mangled by the function.
475 * @param addrlen length of zt_addr (not counting 0-terminator).
476 * @param r_buf a buffer to fill.
477 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which case
478 * the contents of r_buf are undefined.
481 GNUNET_STRINGS_to_address_ipv4 (const char *zt_addr,
483 struct sockaddr_in *r_buf);
487 * Parse an address given as a string into a
490 * @param addr the address
491 * @param[out] af set to the parsed address family (i.e. AF_INET)
492 * @param[out] sa set to the parsed address
493 * @return 0 on error, otherwise number of bytes in @a sa
496 GNUNET_STRINGS_parse_socket_addr (const char *addr,
498 struct sockaddr **sa);
502 * Tries to convert @a addr string to an IP (v4 or v6) address.
503 * Will automatically decide whether to treat 'addr' as v4 or v6 address.
505 * @param addr a string, may not be 0-terminated.
506 * @param addrlen number of bytes in @a addr (if addr is 0-terminated,
507 * 0-terminator should not be counted towards addrlen).
508 * @param r_buf a buffer to fill.
509 * @return #GNUNET_OK if conversion succeded. #GNUNET_SYSERR otherwise, in which
510 * case the contents of r_buf are undefined.
513 GNUNET_STRINGS_to_address_ip (const char *addr,
515 struct sockaddr_storage *r_buf);
519 * Returns utf-8 encoded arguments. Does nothing (returns a copy of
520 * @a argc and @a argv) on any platform other than W32. Returned @a
521 * argv has `u8argv[u8argc] == NULL`. Returned @a argv is a single
522 * memory block, and can be freed with a single GNUNET_free() call.
524 * @param argc argc (as given by main())
525 * @param argv argv (as given by main())
526 * @param u8argc a location to store new argc in (though it's th same as argc)
527 * @param u8argv a location to store new argv in
528 * @return #GNUNET_OK on success, #GNUNET_SYSERR on failure
531 GNUNET_STRINGS_get_utf8_args (int argc,
534 char *const **u8argv);
537 /* ***************** IPv4/IPv6 parsing ****************** */
539 struct GNUNET_STRINGS_PortPolicy
543 * Starting port range (0 if none given).
548 * End of port range (0 if none given).
553 * #GNUNET_YES if the port range should be negated
556 int negate_portrange;
562 * @brief IPV4 network in CIDR notation.
564 struct GNUNET_STRINGS_IPv4NetworkPolicy
569 struct in_addr network;
574 struct in_addr netmask;
577 * Policy for port access.
579 struct GNUNET_STRINGS_PortPolicy pp;
585 * @brief network in CIDR notation for IPV6.
587 struct GNUNET_STRINGS_IPv6NetworkPolicy
592 struct in6_addr network;
597 struct in6_addr netmask;
600 * Policy for port access.
602 struct GNUNET_STRINGS_PortPolicy pp;
608 * Parse an IPv4 network policy. The argument specifies a list of
609 * subnets. The format is <tt>(network[/netmask][:[!]SPORT-DPORT];)*</tt>
610 * (no whitespace, must be terminated with a semicolon). The network
611 * must be given in dotted-decimal notation. The netmask can be given
612 * in CIDR notation (/16) or in dotted-decimal (/255.255.0.0).
614 * @param routeListX a string specifying the IPv4 subnets
615 * @return the converted list, terminated with all zeros;
616 * NULL if the synatx is flawed
618 struct GNUNET_STRINGS_IPv4NetworkPolicy *
619 GNUNET_STRINGS_parse_ipv4_policy (const char *routeListX);
623 * Parse an IPv6 network policy. The argument specifies a list of
624 * subnets. The format is <tt>(network[/netmask[:[!]SPORT[-DPORT]]];)*</tt>
625 * (no whitespace, must be terminated with a semicolon). The network
626 * must be given in colon-hex notation. The netmask must be given in
627 * CIDR notation (/16) or can be omitted to specify a single host.
628 * Note that the netmask is mandatory if ports are specified.
630 * @param routeListX a string specifying the policy
631 * @return the converted list, 0-terminated, NULL if the synatx is flawed
633 struct GNUNET_STRINGS_IPv6NetworkPolicy *
634 GNUNET_STRINGS_parse_ipv6_policy (const char *routeListX);
638 #if 0 /* keep Emacsens' auto-indent happy */
645 /* ifndef GNUNET_UTIL_STRING_H */
648 /** @} */ /* end of group */
650 /* end of gnunet_util_string.h */