2 This file is part of GNUnet.
3 (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009 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 2, 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.
23 * @brief Parses and produces uri strings.
24 * @author Igor Wronsky, Christian Grothoff
26 * GNUnet URIs are of the general form "gnunet://MODULE/IDENTIFIER".
27 * The specific structure of "IDENTIFIER" depends on the module and
28 * maybe differenciated into additional subcategories if applicable.
29 * This module only deals with fs identifiers (MODULE = "fs").
32 * This module only parses URIs for the AFS module. The FS URIs fall
33 * into four categories, "chk", "sks", "ksk" and "loc". The first three
34 * categories were named in analogy (!) to Freenet, but they do NOT
35 * work in exactly the same way. They are very similar from the user's
36 * point of view (unique file identifier, subspace, keyword), but the
37 * implementation is rather different in pretty much every detail.
38 * The concrete URI formats are:
42 * First, there are URIs that identify a file. They have the format
43 * "gnunet://fs/chk/HEX1.HEX2.SIZE". These URIs can be used to
44 * download the file. The description, filename, mime-type and other
45 * meta-data is NOT part of the file-URI since a URI uniquely
46 * identifies a resource (and the contents of the file would be the
47 * same even if it had a different description).
51 * The second category identifies entries in a namespace. The format
52 * is "gnunet://fs/sks/NAMESPACE/IDENTIFIER" where the namespace
53 * should be given in HEX. Applications may allow using a nickname
54 * for the namespace if the nickname is not ambiguous. The identifier
55 * can be either an ASCII sequence or a HEX-encoding. If the
56 * identifier is in ASCII but the format is ambiguous and could denote
57 * a HEX-string a "/" is appended to indicate ASCII encoding.
61 * The third category identifies ordinary searches. The format is
62 * "gnunet://fs/ksk/KEYWORD[+KEYWORD]*". Using the "+" syntax
63 * it is possible to encode searches with the boolean "AND" operator.
64 * "+" is used since it indicates a commutative 'and' operation and
65 * is unlikely to be used in a keyword by itself.
69 * The last category identifies a datum on a specific machine. The
70 * format is "gnunet://fs/loc/HEX1.HEX2.SIZE.PEER.SIG.EXPTIME". PEER is
71 * the BinName of the public key of the peer storing the datum. The
72 * signature (SIG) certifies that this peer has this content.
73 * HEX1, HEX2 and SIZE correspond to a 'chk' URI.
77 * The encoding for hexadecimal values is defined in the hashing.c
78 * module in the gnunetutil library and discussed there.
82 #include "gnunet_fs_service.h"
83 #include "gnunet_signatures.h"
88 * Get a unique key from a URI. This is for putting URIs
89 * into HashMaps. The key may change between FS implementations.
91 * @param uri uri to convert to a unique key
92 * @param key wherer to store the unique key
95 GNUNET_FS_uri_to_key (const struct GNUNET_FS_Uri *uri,
96 GNUNET_HashCode * key)
101 *key = uri->data.chk.chk.query;
104 GNUNET_CRYPTO_hash (uri->data.sks.identifier,
105 strlen (uri->data.sks.identifier), key);
108 if (uri->data.ksk.keywordCount > 0)
109 GNUNET_CRYPTO_hash (uri->data.ksk.keywords[0],
110 strlen (uri->data.ksk.keywords[0]), key);
113 GNUNET_CRYPTO_hash (&uri->data.loc.fi,
114 sizeof (struct FileIdentifier) +
115 sizeof (struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded), key);
118 memset (key, 0, sizeof (GNUNET_HashCode));
125 * Convert keyword URI to a human readable format
126 * (i.e. the search query that was used in the first place)
128 * @param uri ksk uri to convert to a string
129 * @return string with the keywords
132 GNUNET_FS_uri_ksk_to_string_fancy (const struct GNUNET_FS_Uri *uri)
139 unsigned int keywordCount;
141 if ((uri == NULL) || (uri->type != ksk))
146 keywords = uri->data.ksk.keywords;
147 keywordCount = uri->data.ksk.keywordCount;
148 n = keywordCount + 1;
149 for (i = 0; i < keywordCount; i++)
151 keyword = keywords[i];
152 n += strlen (keyword) - 1;
153 if (NULL != strstr (&keyword[1], " "))
155 if (keyword[0] == '+')
158 ret = GNUNET_malloc (n);
160 for (i = 0; i < keywordCount; i++)
162 keyword = keywords[i];
163 if (NULL != strstr (&keyword[1], " "))
166 if (keyword[0] == '+')
167 strcat (ret, keyword);
169 strcat (ret, &keyword[1]);
174 if (keyword[0] == '+')
175 strcat (ret, keyword);
177 strcat (ret, &keyword[1]);
186 * Given a keyword with %-encoding (and possibly quotes to protect
187 * spaces), return a copy of the keyword without %-encoding and
188 * without double-quotes (%22). Also, add a space at the beginning
189 * if there is not a '+'.
191 * @param in string with %-encoding
192 * @param emsg where to store the parser error message (if any)
193 * @return decodded string with leading space (or preserved plus)
196 percent_decode_keyword (const char *in, char **emsg)
204 out = GNUNET_strdup (in);
207 while (out[rpos] != '\0')
209 if (out[rpos] == '%')
211 if (1 != sscanf (&out[rpos + 1], "%2X", &hx))
214 *emsg = GNUNET_strdup (_("`%' must be followed by HEX number"));
219 continue; /* skip double quote */
220 out[wpos++] = (char) hx;
224 out[wpos++] = out[rpos++];
230 ret = GNUNET_strdup (out);
234 /* need to prefix with space */
235 ret = GNUNET_malloc (strlen (out) + 2);
247 * @param s an uri string
248 * @param emsg where to store the parser error message (if any)
249 * @return NULL on error, otherwise the KSK URI
251 static struct GNUNET_FS_Uri *
252 uri_ksk_parse (const char *s, char **emsg)
254 struct GNUNET_FS_Uri *ret;
264 GNUNET_assert (s != NULL);
266 pos = strlen (GNUNET_FS_URI_PREFIX GNUNET_FS_URI_KSK_INFIX);
267 if ( (slen <= pos) ||
268 (0 != strncmp (s, GNUNET_FS_URI_PREFIX GNUNET_FS_URI_KSK_INFIX,
270 return NULL; /* not KSK URI */
271 if ( (s[slen - 1] == '+') ||
274 *emsg = GNUNET_strdup (_("Malformed KSK URI (must not begin or end with `+')"));
279 for (i = pos; i < slen; i++)
281 if ((s[i] == '%') && (&s[i] == strstr (&s[i], "%22")))
283 saw_quote = (saw_quote + 1) % 2;
287 if ((s[i] == '+') && (saw_quote == 0))
292 *emsg = GNUNET_strdup (_("`++' not allowed in KSK URI"));
299 *emsg = GNUNET_strdup (_("Quotes not balanced in KSK URI"));
303 dup = GNUNET_strdup (s);
304 keywords = GNUNET_malloc (max * sizeof (char *));
305 for (i = slen - 1; i >= pos; i--)
307 if ((s[i] == '%') && (&s[i] == strstr (&s[i], "%22")))
309 saw_quote = (saw_quote + 1) % 2;
313 if ((dup[i] == '+') && (saw_quote == 0))
315 keywords[--max] = percent_decode_keyword (&dup[i + 1], emsg);
316 if (NULL == keywords[max])
321 keywords[--max] = percent_decode_keyword (&dup[pos], emsg);
322 if (NULL == keywords[max])
324 GNUNET_assert (max == 0);
326 ret = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
328 ret->data.ksk.keywordCount = iret;
329 ret->data.ksk.keywords = keywords;
332 for (i = 0; i < max; i++)
333 GNUNET_free_non_null (keywords[i]);
334 GNUNET_free (keywords);
343 * @param s an uri string
344 * @param emsg where to store the parser error message (if any)
345 * @return NULL on error, SKS URI otherwise
347 static struct GNUNET_FS_Uri *
348 uri_sks_parse (const char *s, char **emsg)
350 struct GNUNET_FS_Uri *ret;
351 GNUNET_HashCode namespace;
355 char enc[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)];
357 GNUNET_assert (s != NULL);
359 pos = strlen (GNUNET_FS_URI_PREFIX GNUNET_FS_URI_SKS_INFIX);
360 if ( (slen <= pos) ||
361 (0 != strncmp (s, GNUNET_FS_URI_PREFIX GNUNET_FS_URI_SKS_INFIX,
363 return NULL; /* not an SKS URI */
364 if ( (slen < pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded)) ||
365 (s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) - 1] != '/') )
367 *emsg = GNUNET_strdup (_("Malformed SKS URI"));
370 memcpy (enc, &s[pos], sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded));
371 enc[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)-1] = '\0';
372 if (GNUNET_OK != GNUNET_CRYPTO_hash_from_string (enc, &namespace))
374 *emsg = GNUNET_strdup (_("Malformed SKS URI"));
377 identifier = GNUNET_strdup (&s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded)]);
378 ret = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
380 ret->data.sks.namespace = namespace;
381 ret->data.sks.identifier = identifier;
389 * @param s an uri string
390 * @param emsg where to store the parser error message (if any)
391 * @return NULL on error, CHK URI otherwise
393 static struct GNUNET_FS_Uri *
394 uri_chk_parse (const char *s, char **emsg)
396 struct GNUNET_FS_Uri *ret;
397 struct FileIdentifier fi;
399 unsigned long long flen;
401 char h1[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)];
402 char h2[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)];
404 GNUNET_assert (s != NULL);
407 pos = strlen (GNUNET_FS_URI_PREFIX GNUNET_FS_URI_CHK_INFIX);
408 if ( (slen < pos + 2 * sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) + 1) ||
409 (0 != strncmp (s, GNUNET_FS_URI_PREFIX GNUNET_FS_URI_CHK_INFIX,
411 return NULL; /* not a CHK URI */
412 if ( (s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) - 1] != '.') ||
413 (s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) * 2 - 1] != '.') )
415 *emsg = GNUNET_strdup (_("Malformed CHK URI"));
420 sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded));
421 h1[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)-1] = '\0';
423 &s[pos + sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)],
424 sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded));
425 h2[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)-1] = '\0';
427 if ((GNUNET_OK != GNUNET_CRYPTO_hash_from_string (h1,
429 (GNUNET_OK != GNUNET_CRYPTO_hash_from_string (h2,
431 (1 != SSCANF (&s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) * 2],
435 *emsg = GNUNET_strdup (_("Malformed CHK URI"));
438 fi.file_length = GNUNET_htonll (flen);
439 ret = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
447 * Convert a character back to the binary value
448 * that it represents (given base64-encoding).
450 * @param a character to convert
451 * @return offset in the "tbl" array
454 c2v (unsigned char a)
456 if ((a >= '0') && (a <= '9'))
458 if ((a >= 'A') && (a <= 'Z'))
459 return (a - 'A' + 10);
460 if ((a >= 'a') && (a <= 'z'))
461 return (a - 'a' + 36);
471 * Convert string back to binary data.
473 * @param input '\\0'-terminated string
474 * @param data where to write binary data
475 * @param size how much data should be converted
476 * @return number of characters processed from input,
480 enc2bin (const char *input, void *data, size_t size)
488 if (((size * 8) % 6) != 0)
490 if (strlen (input) < len)
491 return -1; /* error! */
495 for (pos = 0; pos < size; pos++)
499 bits |= (c2v (input[len++]) << hbits);
502 (((unsigned char *) data)[pos]) = (unsigned char) bits;
511 * Structure that defines how the
512 * contents of a location URI must be
513 * assembled in memory to create or
514 * verify the signature of a location
517 struct LocUriAssembly
519 struct GNUNET_CRYPTO_RsaSignaturePurpose purpose;
521 struct GNUNET_TIME_AbsoluteNBO exptime;
523 struct FileIdentifier fi;
525 struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded peer;
532 * Also verifies validity of the location URI.
534 * @param s an uri string
535 * @param emsg where to store the parser error message (if any)
536 * @return NULL on error, valid LOC URI otherwise
538 static struct GNUNET_FS_Uri *
539 uri_loc_parse (const char *s, char **emsg)
541 struct GNUNET_FS_Uri *uri;
542 char h1[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)];
543 char h2[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)];
546 unsigned long long exptime;
547 unsigned long long flen;
548 struct GNUNET_TIME_Absolute et;
549 struct GNUNET_CRYPTO_RsaSignature sig;
550 struct LocUriAssembly ass;
554 GNUNET_assert (s != NULL);
556 pos = strlen (GNUNET_FS_URI_PREFIX GNUNET_FS_URI_LOC_INFIX);
557 if ( (slen < pos + 2 * sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) + 1) ||
558 (0 != strncmp (s, GNUNET_FS_URI_PREFIX GNUNET_FS_URI_LOC_INFIX,
560 return NULL; /* not an SKS URI */
561 if ( (s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) - 1] != '.') ||
562 (s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) * 2 - 1] != '.') )
564 *emsg = GNUNET_strdup (_("SKS URI malformed"));
569 sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded));
570 h1[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)-1] = '\0';
572 &s[pos + sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)],
573 sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded));
574 h2[sizeof(struct GNUNET_CRYPTO_HashAsciiEncoded)-1] = '\0';
576 if ((GNUNET_OK != GNUNET_CRYPTO_hash_from_string (h1,
578 (GNUNET_OK != GNUNET_CRYPTO_hash_from_string (h2,
579 &ass.fi.chk.query)) ||
580 (1 != SSCANF (&s[pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) * 2],
584 *emsg = GNUNET_strdup (_("SKS URI malformed"));
587 ass.fi.file_length = GNUNET_htonll (flen);
589 npos = pos + sizeof (struct GNUNET_CRYPTO_HashAsciiEncoded) * 2;
590 while ((s[npos] != '\0') && (s[npos] != '.'))
594 *emsg = GNUNET_strdup (_("SKS URI malformed"));
598 ret = enc2bin (&s[npos],
600 sizeof (struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded));
603 *emsg = GNUNET_strdup (_("SKS URI malformed (could not decode public key)"));
607 if (s[npos++] != '.')
609 *emsg = GNUNET_strdup (_("SKS URI malformed (could not find signature)"));
612 ret = enc2bin (&s[npos],
614 sizeof (struct GNUNET_CRYPTO_RsaSignature));
617 *emsg = GNUNET_strdup (_("SKS URI malformed (could not decode signature)"));
621 if (s[npos++] != '.')
623 *emsg = GNUNET_strdup (_("SKS URI malformed"));
626 if (1 != SSCANF (&s[npos], "%llu", &exptime))
628 *emsg = GNUNET_strdup (_("SKS URI malformed (could not parse expiration time)"));
631 ass.purpose.size = htonl(sizeof(struct LocUriAssembly));
632 ass.purpose.purpose = htonl(GNUNET_SIGNATURE_PURPOSE_PEER_PLACEMENT);
634 ass.exptime = GNUNET_TIME_absolute_hton (et);
636 GNUNET_CRYPTO_rsa_verify (GNUNET_SIGNATURE_PURPOSE_PEER_PLACEMENT,
641 *emsg = GNUNET_strdup (_("SKS URI malformed (signature failed validation)"));
644 uri = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
646 uri->data.loc.fi = ass.fi;
647 uri->data.loc.peer = ass.peer;
648 uri->data.loc.expirationTime = et;
649 uri->data.loc.contentSignature = sig;
658 * Convert a UTF-8 String to a URI.
660 * @param uri string to parse
661 * @param emsg where to store the parser error message (if any)
662 * @return NULL on error
664 struct GNUNET_FS_Uri *
665 GNUNET_FS_uri_parse (const char *uri,
668 struct GNUNET_FS_Uri *ret;
674 if ( (NULL != (ret = uri_chk_parse (uri, emsg))) ||
675 (NULL != (ret = uri_ksk_parse (uri, emsg))) ||
676 (NULL != (ret = uri_sks_parse (uri, emsg))) ||
677 (NULL != (ret = uri_loc_parse (uri, emsg))) )
680 *emsg = GNUNET_strdup (_("Unrecognized URI type"));
690 * @param uri uri to free
693 GNUNET_FS_uri_destroy (struct GNUNET_FS_Uri *uri)
697 GNUNET_assert (uri != NULL);
701 for (i = 0; i < uri->data.ksk.keywordCount; i++)
702 GNUNET_free (uri->data.ksk.keywords[i]);
703 GNUNET_array_grow (uri->data.ksk.keywords, uri->data.ksk.keywordCount,
707 GNUNET_free (uri->data.sks.identifier);
719 * How many keywords are ANDed in this keyword URI?
721 * @param uri ksk uri to get the number of keywords from
722 * @return 0 if this is not a keyword URI
725 GNUNET_FS_uri_ksk_get_keyword_count (const struct GNUNET_FS_Uri *uri)
727 if (uri->type != ksk)
729 return uri->data.ksk.keywordCount;
734 * Iterate over all keywords in this keyword URI.
736 * @param uri ksk uri to get the keywords from
737 * @param iterator function to call on each keyword
738 * @param iterator_cls closure for iterator
739 * @return -1 if this is not a keyword URI, otherwise number of
740 * keywords iterated over until iterator aborted
743 GNUNET_FS_uri_ksk_get_keywords (const struct GNUNET_FS_Uri *uri,
744 GNUNET_FS_KeywordIterator iterator,
750 if (uri->type != ksk)
752 if (iterator == NULL)
753 return uri->data.ksk.keywordCount;
754 for (i = 0; i < uri->data.ksk.keywordCount; i++)
756 keyword = uri->data.ksk.keywords[i];
757 /* first character of keyword indicates
758 if it is mandatory or not */
759 if (GNUNET_OK != iterator (iterator_cls,
769 * Obtain the identity of the peer offering the data
771 * @param uri the location URI to inspect
772 * @param peer where to store the identify of the peer (presumably) offering the content
773 * @return GNUNET_SYSERR if this is not a location URI, otherwise GNUNET_OK
776 GNUNET_FS_uri_loc_get_peer_identity (const struct GNUNET_FS_Uri *uri,
777 struct GNUNET_PeerIdentity * peer)
779 if (uri->type != loc)
780 return GNUNET_SYSERR;
781 GNUNET_CRYPTO_hash (&uri->data.loc.peer,
782 sizeof (struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded),
789 * Obtain the expiration of the LOC URI.
791 * @param uri location URI to get the expiration from
792 * @return expiration time of the URI
794 struct GNUNET_TIME_Absolute
795 GNUNET_FS_uri_loc_get_expiration (const struct GNUNET_FS_Uri *uri)
797 GNUNET_assert (uri->type == loc);
798 return uri->data.loc.expirationTime;
804 * Obtain the URI of the content itself.
806 * @param uri location URI to get the content URI from
807 * @return NULL if argument is not a location URI
809 struct GNUNET_FS_Uri *
810 GNUNET_FS_uri_loc_get_uri (const struct GNUNET_FS_Uri *uri)
812 struct GNUNET_FS_Uri *ret;
814 if (uri->type != loc)
816 ret = GNUNET_malloc (sizeof (struct GNUNET_FS_Uri));
818 ret->data.chk = uri->data.loc.fi;
824 * Construct a location URI (this peer will be used for the location).
826 * @param baseUri content offered by the sender
827 * @param cfg configuration information (used to find our hostkey)
828 * @param expiration_time how long will the content be offered?
829 * @return the location URI, NULL on error
831 struct GNUNET_FS_Uri *
832 GNUNET_FS_uri_loc_create (const struct GNUNET_FS_Uri *baseUri,
833 struct GNUNET_CONFIGURATION_Handle *cfg,
834 struct GNUNET_TIME_Absolute expiration_time)
836 struct GNUNET_FS_Uri *uri;
837 struct GNUNET_CRYPTO_RsaPrivateKey *my_private_key;
838 struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded my_public_key;
840 struct LocUriAssembly ass;
842 if (baseUri->type != chk)
845 GNUNET_CONFIGURATION_get_value_filename (cfg,
847 "HOSTKEY", &keyfile))
849 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
851 ("Lacking key configuration settings.\n"));
854 my_private_key = GNUNET_CRYPTO_rsa_key_create_from_file (keyfile);
855 if (my_private_key == NULL)
857 GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
858 _("Could not access hostkey file `%s'.\n"),
860 GNUNET_free (keyfile);
863 GNUNET_free (keyfile);
864 GNUNET_CRYPTO_rsa_key_get_public (my_private_key, &my_public_key);
865 ass.purpose.size = htonl(sizeof(struct LocUriAssembly));
866 ass.purpose.purpose = htonl(GNUNET_SIGNATURE_PURPOSE_PEER_PLACEMENT);
867 ass.exptime = GNUNET_TIME_absolute_hton (expiration_time);
868 ass.fi = baseUri->data.chk;
869 ass.peer = my_public_key;
870 uri = GNUNET_malloc (sizeof (struct GNUNET_FS_Uri));
872 uri->data.loc.fi = baseUri->data.chk;
873 uri->data.loc.expirationTime = expiration_time;
874 uri->data.loc.peer = my_public_key;
875 GNUNET_assert (GNUNET_OK ==
876 GNUNET_CRYPTO_rsa_sign (my_private_key,
878 &uri->data.loc.contentSignature));
879 GNUNET_CRYPTO_rsa_key_free (my_private_key);
885 * Canonicalize a keyword.
887 * @param in input string (the keyword)
888 * @return canonicalized keyword
891 canonicalize_keyword (const char *in)
897 ret = GNUNET_strdup (in);
900 while ('\0' != *rpos)
902 switch (tolower(*rpos))
913 /* skip characters listed above */
936 /* convert characters listed above to lower case */
937 *wpos = tolower(*rpos);
943 /* keep characters listed above without changes */
948 /* replace characters listed above with '_' */
959 * Canonicalize keyword URI. Performs operations such
960 * as decapitalization and removal of certain characters.
961 * (useful for search).
963 * @param uri the URI to canonicalize
964 * @return canonicalized version of the URI, NULL on error
966 struct GNUNET_FS_Uri *
967 GNUNET_FS_uri_ksk_canonicalize (const struct GNUNET_FS_Uri *uri)
969 struct GNUNET_FS_Uri *ret;
974 kc = uri->data.ksk.keywordCount;
975 kl = GNUNET_malloc (kc*sizeof(char*));
977 kl[i] = canonicalize_keyword (uri->data.ksk.keywords[i]);
978 ret = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
980 ret->data.ksk.keywordCount = kc;
981 ret->data.ksk.keywords = kl;
987 * Merge the sets of keywords from two KSK URIs.
988 * (useful for merging the canonicalized keywords with
989 * the original keywords for sharing).
991 * @param u1 first uri
992 * @param u2 second uri
993 * @return merged URI, NULL on error
995 struct GNUNET_FS_Uri *
996 GNUNET_FS_uri_ksk_merge (const struct GNUNET_FS_Uri *u1,
997 const struct GNUNET_FS_Uri *u2)
999 struct GNUNET_FS_Uri *ret;
1007 if ( (u1->type != ksk) ||
1013 kc = u1->data.ksk.keywordCount;
1014 kl = GNUNET_malloc ((kc+u2->data.ksk.keywordCount)*sizeof(char*));
1015 for (i=0;i<u1->data.ksk.keywordCount;i++)
1016 kl[i] = GNUNET_strdup (u1->data.ksk.keywords[i]);
1017 for (i=0;i<u2->data.ksk.keywordCount;i++)
1019 kp = u2->data.ksk.keywords[i];
1021 for (j=0;j<u1->data.ksk.keywordCount;j++)
1022 if (0 == strcmp(kp + 1,
1031 kl[kc++] = GNUNET_strdup (kp - 1);
1033 ret = GNUNET_malloc (sizeof(struct GNUNET_FS_Uri));
1035 ret->data.ksk.keywordCount = kc;
1036 ret->data.ksk.keywords = kl;
1044 * @param uri the URI to duplicate
1045 * @return copy of the URI
1047 struct GNUNET_FS_Uri *
1048 GNUNET_FS_uri_dup (const struct GNUNET_FS_Uri *uri)
1050 struct GNUNET_FS_Uri *ret;
1055 ret = GNUNET_malloc (sizeof (struct GNUNET_FS_Uri));
1056 memcpy (ret, uri, sizeof (struct GNUNET_FS_Uri));
1060 if (ret->data.ksk.keywordCount > 0)
1062 ret->data.ksk.keywords
1063 = GNUNET_malloc (ret->data.ksk.keywordCount * sizeof (char *));
1064 for (i = 0; i < ret->data.ksk.keywordCount; i++)
1065 ret->data.ksk.keywords[i] =
1066 GNUNET_strdup (uri->data.ksk.keywords[i]);
1069 ret->data.ksk.keywords = NULL; /* just to be sure */
1072 ret->data.sks.identifier = GNUNET_strdup (uri->data.sks.identifier);
1084 * Create an FS URI from a single user-supplied string of keywords.
1085 * The string is broken up at spaces into individual keywords.
1086 * Keywords that start with "+" are mandatory. Double-quotes can
1087 * be used to prevent breaking up strings at spaces (and also
1088 * to specify non-mandatory keywords starting with "+").
1090 * Keywords must contain a balanced number of double quotes and
1091 * double quotes can not be used in the actual keywords (for
1092 * example, the string '""foo bar""' will be turned into two
1093 * "OR"ed keywords 'foo' and 'bar', not into '"foo bar"'.
1095 * @param keywords the keyword string
1096 * @param emsg where to store an error message
1097 * @return an FS URI for the given keywords, NULL
1098 * if keywords is not legal (i.e. empty).
1100 struct GNUNET_FS_Uri *
1101 GNUNET_FS_uri_ksk_create (const char *keywords,
1105 unsigned int num_Words;
1108 struct GNUNET_FS_Uri *uri;
1112 if (keywords == NULL)
1117 searchString = GNUNET_strdup (keywords);
1122 while ('\0' != *pos)
1124 if ((saw_quote == 0) && (isspace (*pos)))
1128 else if (0 == inWord)
1134 saw_quote = (saw_quote + 1) % 2;
1139 GNUNET_free (searchString);
1140 *emsg = GNUNET_strdup (_("No keywords specified!\n"));
1145 GNUNET_free (searchString);
1146 *emsg = GNUNET_strdup (_("Number of double-quotes not balanced!\n"));
1149 keywordarr = GNUNET_malloc (num_Words * sizeof (char *));
1153 while ('\0' != *pos)
1155 if ((saw_quote == 0) && (isspace (*pos)))
1160 else if (0 == inWord)
1162 keywordarr[num_Words] = pos;
1167 saw_quote = (saw_quote + 1) % 2;
1171 GNUNET_FS_uri_ksk_create_from_args (num_Words,
1172 (const char **) keywordarr);
1173 GNUNET_free (keywordarr);
1174 GNUNET_free (searchString);
1180 * Create an FS URI from a user-supplied command line of keywords.
1181 * Arguments should start with "+" to indicate mandatory
1184 * @param argc number of keywords
1185 * @param argv keywords (double quotes are not required for
1186 * keywords containing spaces; however, double
1187 * quotes are required for keywords starting with
1188 * "+"); there is no mechanism for having double
1189 * quotes in the actual keywords (if the user
1190 * did specifically specify double quotes, the
1191 * caller should convert each double quote
1192 * into two single quotes).
1193 * @return an FS URI for the given keywords, NULL
1194 * if keywords is not legal (i.e. empty).
1196 struct GNUNET_FS_Uri *
1197 GNUNET_FS_uri_ksk_create_from_args (unsigned int argc,
1201 struct GNUNET_FS_Uri *uri;
1202 const char *keyword;
1210 /* allow URI to be given as one and only keyword and
1211 handle accordingly */
1214 (strlen(argv[0]) > strlen(GNUNET_FS_URI_PREFIX)) &&
1215 (0 == strncmp(argv[0], GNUNET_FS_URI_PREFIX, strlen(GNUNET_FS_URI_PREFIX)) ) &&
1216 (NULL != (uri = GNUNET_FS_uri_parse(argv[0], &emsg)) ) )
1218 GNUNET_free_non_null (emsg);
1219 uri = GNUNET_malloc (sizeof (struct GNUNET_FS_Uri));
1221 uri->data.ksk.keywordCount = argc;
1222 uri->data.ksk.keywords = GNUNET_malloc (argc * sizeof (char *));
1223 for (i = 0; i < argc; i++)
1226 if (keyword[0] == '+')
1227 val = GNUNET_strdup (keyword);
1229 GNUNET_asprintf (&val, " %s", keyword);
1240 uri->data.ksk.keywords[i] = val;
1247 * Test if two URIs are equal.
1249 * @param u1 one of the URIs
1250 * @param u2 the other URI
1251 * @return GNUNET_YES if the URIs are equal
1254 GNUNET_FS_uri_test_equal (const struct GNUNET_FS_Uri *u1,
1255 const struct GNUNET_FS_Uri *u2)
1261 GNUNET_assert (u1 != NULL);
1262 GNUNET_assert (u2 != NULL);
1263 if (u1->type != u2->type)
1268 if (0 == memcmp (&u1->data.chk,
1270 sizeof (struct FileIdentifier)))
1274 if ((0 == memcmp (&u1->data.sks.namespace,
1275 &u2->data.sks.namespace,
1276 sizeof (GNUNET_HashCode))) &&
1277 (0 == strcmp (u1->data.sks.identifier,
1278 u2->data.sks.identifier)))
1283 if (u1->data.ksk.keywordCount != u2->data.ksk.keywordCount)
1285 for (i = 0; i < u1->data.ksk.keywordCount; i++)
1288 for (j = 0; j < u2->data.ksk.keywordCount; j++)
1290 if (0 == strcmp (u1->data.ksk.keywords[i],
1291 u2->data.ksk.keywords[j]))
1297 if (ret == GNUNET_NO)
1302 if (memcmp (&u1->data.loc,
1304 sizeof (struct FileIdentifier) +
1305 sizeof (struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded) +
1306 sizeof (struct GNUNET_TIME_Absolute) +
1307 sizeof (unsigned short) + sizeof (unsigned short)) != 0)
1317 * Is this a namespace URI?
1319 * @param uri the uri to check
1320 * @return GNUNET_YES if this is an SKS uri
1323 GNUNET_FS_uri_test_sks (const struct GNUNET_FS_Uri *uri)
1325 return uri->type == sks;
1330 * Get the ID of a namespace from the given
1333 * @param uri the uri to get the namespace ID from
1334 * @param nsid where to store the ID of the namespace
1335 * @return GNUNET_OK on success
1338 GNUNET_FS_uri_sks_get_namespace (const struct GNUNET_FS_Uri *uri,
1339 GNUNET_HashCode * nsid)
1341 if (! GNUNET_FS_uri_test_sks (uri))
1344 return GNUNET_SYSERR;
1346 *nsid = uri->data.sks.namespace;
1352 * Get the content identifier of an SKS URI.
1354 * @param uri the sks uri
1355 * @return NULL on error (not a valid SKS URI)
1358 GNUNET_FS_uri_sks_get_content_id (const struct GNUNET_FS_Uri *uri)
1360 if (!GNUNET_FS_uri_test_sks (uri))
1365 return GNUNET_strdup (uri->data.sks.identifier);
1370 * Convert namespace URI to a human readable format
1371 * (using the namespace description, if available).
1373 * @param cfg configuration to use
1374 * @param uri SKS uri to convert
1375 * @return NULL on error (not an SKS URI)
1378 GNUNET_FS_uri_sks_to_string_fancy (struct GNUNET_CONFIGURATION_Handle *cfg,
1379 const struct GNUNET_FS_Uri *uri)
1384 if (uri->type != sks)
1386 name = GNUNET_PSEUDONYM_id_to_name (cfg, &uri->data.sks.namespace);
1388 return GNUNET_FS_uri_to_string (uri);
1389 GNUNET_asprintf (&ret,
1392 uri->data.sks.identifier);
1399 * Is this a keyword URI?
1401 * @param uri the uri
1402 * @return GNUNET_YES if this is a KSK uri
1405 GNUNET_FS_uri_test_ksk (const struct GNUNET_FS_Uri *uri)
1410 if (uri->type == ksk)
1412 for (i = uri->data.ksk.keywordCount - 1; i >= 0; i--)
1413 GNUNET_assert (uri->data.ksk.keywords[i] != NULL);
1416 return uri->type == ksk;
1421 * Is this a file (or directory) URI?
1423 * @param uri the uri to check
1424 * @return GNUNET_YES if this is a CHK uri
1427 GNUNET_FS_uri_test_chk (const struct GNUNET_FS_Uri *uri)
1429 return uri->type == chk;
1434 * What is the size of the file that this URI
1437 * @param uri the CHK URI to inspect
1438 * @return size of the file as specified in the CHK URI
1441 GNUNET_FS_uri_chk_get_file_size (const struct GNUNET_FS_Uri *uri)
1446 return GNUNET_ntohll (uri->data.chk.file_length);
1448 return GNUNET_ntohll (uri->data.loc.fi.file_length);
1452 return 0; /* unreachable */
1457 * Is this a location URI?
1459 * @param uri the uri to check
1460 * @return GNUNET_YES if this is a LOC uri
1463 GNUNET_FS_uri_test_loc (const struct GNUNET_FS_Uri *uri)
1465 return uri->type == loc;
1470 * Function called on each value in the meta data.
1471 * Adds it to the URI.
1473 * @param cls URI to update
1474 * @param plugin_name name of the plugin that produced this value;
1475 * special values can be used (i.e. '<zlib>' for zlib being
1476 * used in the main libextractor library and yielding
1478 * @param type libextractor-type describing the meta data
1479 * @param format basic format information about data
1480 * @param data_mime_type mime-type of data (not of the original file);
1481 * can be NULL (if mime-type is not known)
1482 * @param data actual meta-data found
1483 * @param data_len number of bytes in data
1484 * @return 0 (always)
1487 gather_uri_data (void *cls,
1488 const char *plugin_name,
1489 enum EXTRACTOR_MetaType type,
1490 enum EXTRACTOR_MetaFormat format,
1491 const char *data_mime_type,
1495 struct GNUNET_FS_Uri *uri = cls;
1499 if ( (format != EXTRACTOR_METAFORMAT_UTF8) &&
1500 (format != EXTRACTOR_METAFORMAT_C_STRING) )
1502 for (j = uri->data.ksk.keywordCount - 1; j >= 0; j--)
1503 if (0 == strcmp (&uri->data.ksk.keywords[j][1], data))
1505 nkword = GNUNET_malloc (strlen (data) + 2);
1506 strcpy (nkword, " "); /* not mandatory */
1507 strcat (nkword, data);
1508 uri->data.ksk.keywords[uri->data.ksk.keywordCount++] = nkword;
1514 * Construct a keyword-URI from meta-data (take all entries
1515 * in the meta-data and construct one large keyword URI
1516 * that lists all keywords that can be found in the meta-data).
1518 * @param md metadata to use
1519 * @return NULL on error, otherwise a KSK URI
1521 struct GNUNET_FS_Uri *
1522 GNUNET_FS_uri_ksk_create_from_meta_data (const struct GNUNET_CONTAINER_MetaData *md)
1524 struct GNUNET_FS_Uri *ret;
1528 ret = GNUNET_malloc (sizeof (struct GNUNET_FS_Uri));
1530 ret->data.ksk.keywordCount = 0;
1531 ret->data.ksk.keywords = NULL;
1532 ret->data.ksk.keywords
1533 = GNUNET_malloc (sizeof (char *) *
1534 GNUNET_CONTAINER_meta_data_iterate (md, NULL, NULL));
1535 GNUNET_CONTAINER_meta_data_iterate (md, &gather_uri_data, ret);
1542 * In URI-encoding, does the given character
1543 * need to be encoded using %-encoding?
1546 needs_percent (char c)
1548 return (!((isalnum (c)) ||
1549 (c == '-') || (c == '_') || (c == '.') || (c == '~')));
1554 * Convert a KSK URI to a string.
1556 * @param uri the URI to convert
1557 * @return NULL on error (i.e. keywordCount == 0)
1560 uri_ksk_to_string (const struct GNUNET_FS_Uri *uri)
1563 unsigned int keywordCount;
1570 const char *keyword;
1572 if (uri->type != ksk)
1574 keywords = uri->data.ksk.keywords;
1575 keywordCount = uri->data.ksk.keywordCount;
1577 keywordCount + strlen (GNUNET_FS_URI_PREFIX) +
1578 strlen (GNUNET_FS_URI_KSK_INFIX) + 1;
1579 for (i = 0; i < keywordCount; i++)
1581 keyword = keywords[i];
1582 slen = strlen (keyword);
1584 for (j = 0; j < slen; j++)
1586 if ((j == 0) && (keyword[j] == ' '))
1589 continue; /* skip leading space */
1591 if (needs_percent (keyword[j]))
1592 n += 2; /* will use %-encoding */
1595 ret = GNUNET_malloc (n);
1596 strcpy (ret, GNUNET_FS_URI_PREFIX);
1597 strcat (ret, GNUNET_FS_URI_KSK_INFIX);
1598 wpos = strlen (ret);
1599 for (i = 0; i < keywordCount; i++)
1601 keyword = keywords[i];
1602 slen = strlen (keyword);
1603 for (j = 0; j < slen; j++)
1605 if ((j == 0) && (keyword[j] == ' '))
1606 continue; /* skip leading space */
1607 if (needs_percent (keyword[j]))
1609 sprintf (&ret[wpos], "%%%02X", keyword[j]);
1614 ret[wpos++] = keyword[j];
1617 if (i != keywordCount - 1)
1625 * Convert SKS URI to a string.
1627 * @param uri sks uri to convert
1628 * @return NULL on error
1631 uri_sks_to_string (const struct GNUNET_FS_Uri *uri)
1633 const GNUNET_HashCode * namespace;
1634 const char *identifier;
1636 struct GNUNET_CRYPTO_HashAsciiEncoded ns;
1638 if (uri->type != sks)
1640 namespace = &uri->data.sks.namespace;
1641 identifier = uri->data.sks.identifier;
1642 GNUNET_CRYPTO_hash_to_enc (namespace, &ns);
1643 GNUNET_asprintf (&ret,
1645 GNUNET_FS_URI_PREFIX,
1646 GNUNET_FS_URI_SKS_INFIX,
1647 (const char *) &ns, identifier);
1653 * Convert a CHK URI to a string.
1655 * @param uri chk uri to convert
1656 * @return NULL on error
1659 uri_chk_to_string (const struct GNUNET_FS_Uri *uri)
1661 const struct FileIdentifier * fi;
1663 struct GNUNET_CRYPTO_HashAsciiEncoded keyhash;
1664 struct GNUNET_CRYPTO_HashAsciiEncoded queryhash;
1666 if (uri->type != chk)
1668 fi = &uri->data.chk;
1669 GNUNET_CRYPTO_hash_to_enc (&fi->chk.key, &keyhash);
1670 GNUNET_CRYPTO_hash_to_enc (&fi->chk.query, &queryhash);
1672 GNUNET_asprintf (&ret,
1674 GNUNET_FS_URI_PREFIX,
1675 GNUNET_FS_URI_CHK_INFIX,
1676 (const char *) &keyhash,
1677 (const char *) &queryhash,
1678 GNUNET_ntohll (fi->file_length));
1683 * Convert binary data to a string.
1685 * @param data binary data to convert
1686 * @param size number of bytes in data
1687 * @return converted data
1690 bin2enc (const void *data, size_t size)
1693 * 64 characters for encoding, 6 bits per character
1696 "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz_=";
1704 GNUNET_assert (strlen (tbl) == 64);
1706 if (((size * 8) % 6) != 0)
1708 ret = GNUNET_malloc (len + 1);
1713 for (pos = 0; pos < size; pos++)
1715 bits |= ((((const unsigned char *) data)[pos]) << hbits);
1719 ret[len++] = tbl[bits & 63];
1725 ret[len++] = tbl[bits & 63];
1731 * Convert a LOC URI to a string.
1733 * @param uri loc uri to convert
1734 * @return NULL on error
1737 uri_loc_to_string (const struct GNUNET_FS_Uri *uri)
1740 struct GNUNET_CRYPTO_HashAsciiEncoded keyhash;
1741 struct GNUNET_CRYPTO_HashAsciiEncoded queryhash;
1745 GNUNET_CRYPTO_hash_to_enc (&uri->data.loc.fi.chk.key, &keyhash);
1746 GNUNET_CRYPTO_hash_to_enc (&uri->data.loc.fi.chk.query, &queryhash);
1747 peerId = bin2enc (&uri->data.loc.peer,
1748 sizeof (struct GNUNET_CRYPTO_RsaPublicKeyBinaryEncoded));
1749 peerSig = bin2enc (&uri->data.loc.contentSignature,
1750 sizeof (struct GNUNET_CRYPTO_RsaSignature));
1751 GNUNET_asprintf (&ret,
1752 "%s%s%s.%s.%llu.%s.%s.%llu",
1753 GNUNET_FS_URI_PREFIX,
1754 GNUNET_FS_URI_LOC_INFIX,
1755 (const char *) &keyhash,
1756 (const char *) &queryhash,
1757 (unsigned long long) GNUNET_ntohll (uri->data.loc.fi.file_length),
1760 (unsigned long long) uri->data.loc.expirationTime.value);
1761 GNUNET_free (peerSig);
1762 GNUNET_free (peerId);
1768 * Convert a URI to a UTF-8 String.
1770 * @param uri uri to convert to a string
1771 * @return the UTF-8 string
1774 GNUNET_FS_uri_to_string (const struct GNUNET_FS_Uri *uri)
1784 return uri_ksk_to_string (uri);
1786 return uri_sks_to_string (uri);
1788 return uri_chk_to_string (uri);
1790 return uri_loc_to_string (uri);
1797 /* end of fs_uri.c */