2 This file is part of GNUnet.
3 (C) 2003, 2004, 2005, 2006, 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.
22 * @file util/container_meta_data.c
23 * @brief Storing of meta data
24 * @author Christian Grothoff
28 #include "gnunet_common.h"
29 #include "gnunet_container_lib.h"
30 #include "gnunet_strings_lib.h"
31 #include "gnunet_time_lib.h"
32 #include <extractor.h>
35 #define EXTRA_CHECKS ALLOW_EXTRA_CHECKS
39 EXTRACTOR_KeywordType type;
44 * Meta data to associate with a file, directory or namespace.
46 struct GNUNET_CONTAINER_MetaData
53 * Create a fresh struct CONTAINER_MetaData token.
55 struct GNUNET_CONTAINER_MetaData *
56 GNUNET_CONTAINER_meta_data_create ()
58 struct GNUNET_CONTAINER_MetaData *ret;
59 ret = GNUNET_malloc (sizeof (struct GNUNET_CONTAINER_MetaData));
69 GNUNET_CONTAINER_meta_data_destroy (struct GNUNET_CONTAINER_MetaData *md)
75 for (i = 0; i < md->itemCount; i++)
76 GNUNET_free (md->items[i].data);
77 GNUNET_array_grow (md->items, md->itemCount, 0);
82 * Add the current time as the publication date
86 GNUNET_CONTAINER_meta_data_add_publication_date (struct
87 GNUNET_CONTAINER_MetaData
91 struct GNUNET_TIME_Absolute t;
93 t = GNUNET_TIME_absolute_get ();
94 GNUNET_CONTAINER_meta_data_delete (md, EXTRACTOR_PUBLICATION_DATE, NULL);
95 dat = GNUNET_STRINGS_absolute_time_to_string (t);
96 GNUNET_CONTAINER_meta_data_insert (md, EXTRACTOR_PUBLICATION_DATE, dat);
102 * @return GNUNET_OK on success, GNUNET_SYSERR if this entry already exists
105 GNUNET_CONTAINER_meta_data_insert (struct GNUNET_CONTAINER_MetaData *md,
106 EXTRACTOR_KeywordType type,
112 GNUNET_assert (data != NULL);
113 for (idx = 0; idx < md->itemCount; idx++)
115 if ((md->items[idx].type == type) &&
116 (0 == strcmp (md->items[idx].data, data)))
117 return GNUNET_SYSERR;
120 GNUNET_array_grow (md->items, md->itemCount, md->itemCount + 1);
121 md->items[idx].type = type;
122 md->items[idx].data = p = GNUNET_strdup (data);
124 /* change OS native dir separators to unix '/' and others to '_' */
125 if (type == EXTRACTOR_FILENAME)
129 if (*p == DIR_SEPARATOR)
142 * @return GNUNET_OK on success, GNUNET_SYSERR if the item does not exist in md
145 GNUNET_CONTAINER_meta_data_delete (struct GNUNET_CONTAINER_MetaData *md,
146 EXTRACTOR_KeywordType type,
150 int ret = GNUNET_SYSERR;
151 for (idx = 0; idx < md->itemCount; idx++)
153 if ((md->items[idx].type == type) &&
154 ((data == NULL) || (0 == strcmp (md->items[idx].data, data))))
156 GNUNET_free (md->items[idx].data);
157 md->items[idx] = md->items[md->itemCount - 1];
158 GNUNET_array_grow (md->items, md->itemCount, md->itemCount - 1);
171 * Iterate over MD entries, excluding thumbnails.
173 * @return number of entries
176 GNUNET_CONTAINER_meta_data_get_contents (const struct
177 GNUNET_CONTAINER_MetaData *md,
178 GNUNET_CONTAINER_MetaDataProcessor
179 iterator, void *closure)
185 for (i = 0; i < md->itemCount; i++)
187 if (!EXTRACTOR_isBinaryType (md->items[i].type))
189 if ((iterator != NULL) &&
190 (GNUNET_OK != iterator (md->items[i].type,
191 md->items[i].data, closure)))
192 return GNUNET_SYSERR;
197 return (int) (md->itemCount - sub);
201 * Iterate over MD entries
203 * @return number of entries
206 GNUNET_CONTAINER_meta_data_get_by_type (const struct GNUNET_CONTAINER_MetaData
207 *md, EXTRACTOR_KeywordType type)
211 for (i = 0; i < md->itemCount; i++)
212 if (type == md->items[i].type)
213 return GNUNET_strdup (md->items[i].data);
218 * Iterate over MD entries
220 * @return number of entries
223 GNUNET_CONTAINER_meta_data_get_first_by_types (const struct
224 GNUNET_CONTAINER_MetaData *md,
229 EXTRACTOR_KeywordType type;
235 type = va_arg (args, EXTRACTOR_KeywordType);
238 ret = GNUNET_CONTAINER_meta_data_get_by_type (md, type);
247 * Get a thumbnail from the meta-data (if present).
249 * @param thumb will be set to the thumbnail data. Must be
250 * freed by the caller!
251 * @return number of bytes in thumbnail, 0 if not available
254 GNUNET_CONTAINER_meta_data_get_thumbnail (const struct
255 GNUNET_CONTAINER_MetaData * md,
256 unsigned char **thumb)
263 GNUNET_CONTAINER_meta_data_get_by_type (md, EXTRACTOR_THUMBNAIL_DATA);
266 if (strlen (encoded) == 0)
268 GNUNET_free (encoded);
269 return 0; /* invalid */
272 ret = EXTRACTOR_binaryDecode (encoded, thumb, &size);
273 GNUNET_free (encoded);
280 * Duplicate struct GNUNET_CONTAINER_MetaData.
282 struct GNUNET_CONTAINER_MetaData *
283 GNUNET_CONTAINER_meta_data_duplicate (const struct GNUNET_CONTAINER_MetaData
287 struct GNUNET_CONTAINER_MetaData *ret;
291 ret = GNUNET_CONTAINER_meta_data_create ();
292 for (i = 0; i < md->itemCount; i++)
293 GNUNET_CONTAINER_meta_data_insert (ret, md->items[i].type,
299 * Extract meta-data from a file.
301 * @return GNUNET_SYSERR on error, otherwise the number
302 * of meta-data items obtained
305 GNUNET_CONTAINER_meta_data_extract_from_file (struct GNUNET_CONTAINER_MetaData
306 *md, const char *filename,
307 EXTRACTOR_ExtractorList *
310 EXTRACTOR_KeywordList *head;
311 EXTRACTOR_KeywordList *pos;
314 if (filename == NULL)
315 return GNUNET_SYSERR;
316 if (extractors == NULL)
318 head = EXTRACTOR_getKeywords (extractors, filename);
319 head = EXTRACTOR_removeDuplicateKeywords (head,
320 EXTRACTOR_DUPLICATES_REMOVE_UNKNOWN);
326 GNUNET_CONTAINER_meta_data_insert (md, pos->keywordType,
331 EXTRACTOR_freeKeywords (head);
336 tryCompression (char *data, unsigned int oldSize)
342 dlen = compressBound (oldSize);
344 dlen = oldSize + (oldSize / 100) + 20;
345 /* documentation says 100.1% oldSize + 12 bytes, but we
346 should be able to overshoot by more to be safe */
348 tmp = GNUNET_malloc (dlen);
349 if (Z_OK == compress2 ((Bytef *) tmp,
350 &dlen, (const Bytef *) data, oldSize, 9))
354 memcpy (data, tmp, dlen);
364 * Decompress input, return the decompressed data
365 * as output, set outputSize to the number of bytes
368 * @return NULL on error
371 decompress (const char *input,
372 unsigned int inputSize, unsigned int outputSize)
378 output = GNUNET_malloc (olen);
379 if (Z_OK == uncompress ((Bytef *) output,
380 &olen, (const Bytef *) input, inputSize))
386 GNUNET_free (output);
392 * Flag in 'version' that indicates compressed meta-data.
394 #define HEADER_COMPRESSED 0x80000000
397 * Bits in 'version' that give the version number.
399 #define HEADER_VERSION_MASK 0x7FFFFFFF
401 struct MetaDataHeader
404 * The version of the MD serialization.
405 * The highest bit is used to indicate
408 * Version 0 is the current version;
409 * Version is 1 for a NULL pointer.
410 * Other version numbers are not yet defined.
415 * How many MD entries are there?
420 * Size of the MD (decompressed)
425 * This is followed by 'entries' values of type 'unsigned int' that
426 * correspond to EXTRACTOR_KeywordTypes. After that, the meta-data
427 * keywords follow (0-terminated). The MD block always ends with
428 * 0-termination, padding with 0 until a multiple of 8 bytes.
434 * Serialize meta-data to target.
436 * @param size maximum number of bytes available
437 * @param part is it ok to just write SOME of the
438 * meta-data to match the size constraint,
439 * possibly discarding some data?
440 * @return number of bytes written on success,
441 * GNUNET_SYSERR on error (typically: not enough
445 GNUNET_CONTAINER_meta_data_serialize (const struct GNUNET_CONTAINER_MetaData
446 *md, char *target, unsigned int max,
448 GNUNET_CONTAINER_MetaDataSerializationOptions
451 struct MetaDataHeader *hdr;
458 if (max < sizeof (struct MetaDataHeader))
459 return GNUNET_SYSERR; /* far too small */
460 ic = md ? md->itemCount : 0;
464 size = sizeof (struct MetaDataHeader);
465 size += sizeof (unsigned int) * ic;
466 for (i = 0; i < ic; i++)
467 size += 1 + strlen (md->items[i].data);
468 while (size % 8 != 0)
470 hdr = GNUNET_malloc (size);
471 hdr->version = htonl (md == NULL ? 1 : 0);
472 hdr->entries = htonl (ic);
473 for (i = 0; i < ic; i++)
474 ((unsigned int *) &hdr[1])[i] =
475 htonl ((unsigned int) md->items[i].type);
476 pos = sizeof (struct MetaDataHeader);
477 pos += sizeof (unsigned int) * ic;
478 for (i = 0; i < ic; i++)
480 len = strlen (md->items[i].data) + 1;
481 memcpy (&((char *) hdr)[pos], md->items[i].data, len);
485 hdr->size = htonl (size);
486 if ((part & GNUNET_CONTAINER_META_DATA_SERIALIZE_NO_COMPRESS) == 0)
488 pos = tryCompression ((char *) &hdr[1],
489 size - sizeof (struct MetaDataHeader));
493 pos = size - sizeof (struct MetaDataHeader);
495 if (pos < size - sizeof (struct MetaDataHeader))
497 hdr->version = htonl (HEADER_COMPRESSED);
498 size = pos + sizeof (struct MetaDataHeader);
505 if ((part & GNUNET_CONTAINER_META_DATA_SERIALIZE_PART) == 0)
507 return GNUNET_SYSERR; /* does not fit! */
509 /* partial serialization ok, try again with less meta-data */
511 ic = ic * 2 / 3; /* still far too big, make big reductions */
513 ic--; /* small steps, we're close */
515 GNUNET_assert (size <= max);
516 memcpy (target, hdr, size);
518 /* extra check: deserialize! */
521 struct GNUNET_CONTAINER_MetaData *mdx;
522 mdx = GNUNET_CONTAINER_meta_data_deserialize (target, size);
523 GNUNET_assert (NULL != mdx);
524 GNUNET_CONTAINER_meta_data_destroy (mdx);
531 * Estimate (!) the size of the meta-data in
532 * serialized form. The estimate MAY be higher
533 * than what is strictly needed.
536 GNUNET_CONTAINER_meta_data_get_serialized_size (const struct
537 GNUNET_CONTAINER_MetaData *md,
539 GNUNET_CONTAINER_MetaDataSerializationOptions
542 struct MetaDataHeader *hdr;
549 ic = md ? md->itemCount : 0;
550 size = sizeof (struct MetaDataHeader);
551 size += sizeof (unsigned int) * ic;
552 for (i = 0; i < ic; i++)
553 size += 1 + strlen (md->items[i].data);
554 while (size % 8 != 0)
556 hdr = GNUNET_malloc (size);
557 hdr->version = htonl (md == NULL ? 1 : 0);
558 hdr->entries = htonl (ic);
559 for (i = 0; i < ic; i++)
560 ((unsigned int *) &hdr[1])[i] = htonl ((unsigned int) md->items[i].type);
561 pos = sizeof (struct MetaDataHeader);
562 pos += sizeof (unsigned int) * ic;
563 for (i = 0; i < ic; i++)
565 len = strlen (md->items[i].data) + 1;
566 memcpy (&((char *) hdr)[pos], md->items[i].data, len);
569 if ((part & GNUNET_CONTAINER_META_DATA_SERIALIZE_NO_COMPRESS) == 0)
572 tryCompression ((char *) &hdr[1],
573 size - sizeof (struct MetaDataHeader));
577 pos = size - sizeof (struct MetaDataHeader);
579 if (pos < size - sizeof (struct MetaDataHeader))
580 size = pos + sizeof (struct MetaDataHeader);
586 * Deserialize meta-data. Initializes md.
587 * @param size number of bytes available
588 * @return MD on success, NULL on error (i.e.
591 struct GNUNET_CONTAINER_MetaData *
592 GNUNET_CONTAINER_meta_data_deserialize (const char *input, unsigned int size)
594 struct GNUNET_CONTAINER_MetaData *md;
595 const struct MetaDataHeader *hdr;
606 if (size < sizeof (struct MetaDataHeader))
608 hdr = (const struct MetaDataHeader *) input;
609 version = ntohl (MAKE_UNALIGNED (hdr->version)) & HEADER_VERSION_MASK;
611 return NULL; /* null pointer */
614 GNUNET_break_op (0); /* unsupported version */
617 ic = ntohl (MAKE_UNALIGNED (hdr->entries));
619 (ntohl (MAKE_UNALIGNED (hdr->version)) & HEADER_COMPRESSED) != 0;
623 ntohl (MAKE_UNALIGNED (hdr->size)) - sizeof (struct MetaDataHeader);
624 if (dataSize > 2 * 1042 * 1024)
627 return NULL; /* only 2 MB allowed [to make sure we don't blow
628 our memory limit because of a mal-formed
632 decompress ((const char *) &input[sizeof (struct MetaDataHeader)],
633 size - sizeof (struct MetaDataHeader), dataSize);
644 cdata = (const char *) &hdr[1];
645 dataSize = size - sizeof (struct MetaDataHeader);
646 if (size != ntohl (MAKE_UNALIGNED (hdr->size)))
653 if ((sizeof (unsigned int) * ic + ic) > dataSize)
658 if ((ic > 0) && (cdata[dataSize - 1] != '\0'))
664 md = GNUNET_CONTAINER_meta_data_create ();
665 GNUNET_array_grow (md->items, md->itemCount, ic);
667 pos = sizeof (unsigned int) * ic;
668 while ((pos < dataSize) && (i < ic))
670 len = strlen (&cdata[pos]) + 1;
671 md->items[i].type = (EXTRACTOR_KeywordType)
672 ntohl (MAKE_UNALIGNED (((const unsigned int *) cdata)[i]));
673 md->items[i].data = GNUNET_strdup (&cdata[pos]);
679 GNUNET_CONTAINER_meta_data_destroy (md);
682 GNUNET_free_non_null (data);
685 GNUNET_free_non_null (data);
686 return NULL; /* size too small */
690 * Test if two MDs are equal.
693 GNUNET_CONTAINER_meta_data_test_equal (const struct GNUNET_CONTAINER_MetaData
695 const struct GNUNET_CONTAINER_MetaData
702 if (md1->itemCount != md2->itemCount)
704 for (i = 0; i < md1->itemCount; i++)
707 for (j = 0; j < md2->itemCount; j++)
708 if ((md1->items[i].type == md2->items[j].type) &&
709 (0 == strcmp (md1->items[i].data, md2->items[j].data)))
714 if (found == GNUNET_NO)
721 /* end of container_meta_data.c */