[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[GNUnet-SVN] r18853 - gnunet/src/fs
From: |
gnunet |
Subject: |
[GNUnet-SVN] r18853 - gnunet/src/fs |
Date: |
Fri, 30 Dec 2011 23:25:33 +0100 |
Author: grothoff
Date: 2011-12-30 23:25:33 +0100 (Fri, 30 Dec 2011)
New Revision: 18853
Modified:
gnunet/src/fs/Makefile.am
gnunet/src/fs/fs_uri.c
Log:
LRN: normalize keywords (decapitalize, split) using libunistring - #2052
Modified: gnunet/src/fs/Makefile.am
===================================================================
--- gnunet/src/fs/Makefile.am 2011-12-30 22:13:26 UTC (rev 18852)
+++ gnunet/src/fs/Makefile.am 2011-12-30 22:25:33 UTC (rev 18853)
@@ -43,7 +43,7 @@
$(top_builddir)/src/datastore/libgnunetdatastore.la \
$(top_builddir)/src/util/libgnunetutil.la \
-lextractor \
- $(GN_LIBINTL) $(XLIB)
+ $(GN_LIBINTL) $(XLIB) -lunistring
libgnunetfs_la_LDFLAGS = \
$(GN_LIB_LDFLAGS) $(WINFLAGS) \
Modified: gnunet/src/fs/fs_uri.c
===================================================================
--- gnunet/src/fs/fs_uri.c 2011-12-30 22:13:26 UTC (rev 18852)
+++ gnunet/src/fs/fs_uri.c 2011-12-30 22:25:33 UTC (rev 18853)
@@ -82,8 +82,13 @@
#include "gnunet_fs_service.h"
#include "gnunet_signatures.h"
#include "fs_api.h"
+#include <unicase.h>
+#include <unistr.h>
+#include <unistdio.h>
+#include <uniconv.h>
+
/**
* Get a unique key from a URI. This is for putting URIs
* into HashMaps. The key may change between FS implementations.
@@ -1493,7 +1498,49 @@
return GNUNET_NO;
}
+static char *
+normalize_metadata (enum EXTRACTOR_MetaFormat format, const char *data,
+ size_t data_len)
+{
+ uint8_t *free_str = NULL;
+ uint8_t *str_to_normalize = (uint8_t *) data;
+ uint8_t *normalized;
+ size_t r_len;
+ if (str_to_normalize == NULL)
+ return NULL;
+ /* Don't trust libextractor */
+ if (format == EXTRACTOR_METAFORMAT_UTF8)
+ {
+ free_str = (uint8_t *) u8_check ((const uint8_t *) data, data_len);
+ if (free_str == NULL)
+ free_str = NULL;
+ else
+ format = EXTRACTOR_METAFORMAT_C_STRING;
+ }
+ if (format == EXTRACTOR_METAFORMAT_C_STRING)
+ {
+ free_str = u8_strconv_from_encoding (data, locale_charset (),
iconveh_escape_sequence);
+ if (free_str == NULL)
+ return NULL;
+ }
+ normalized = u8_tolower (str_to_normalize, strlen ((char *)
str_to_normalize), NULL, UNINORM_NFD, NULL, &r_len);
+ /* free_str is allocated by libunistring internally, use free() */
+ if (free_str != NULL)
+ free (free_str);
+ if (normalized != NULL)
+ {
+ /* u8_tolower allocates a non-NULL-terminated string! */
+ free_str = GNUNET_malloc (r_len + 1);
+ memcpy (free_str, normalized, r_len);
+ free_str[r_len] = '\0';
+ free (normalized);
+ normalized = free_str;
+ }
+ return (char *) normalized;
+}
+
+
/**
* Break the filename up by matching [], () and {} pairs to make
* keywords. In case of nesting parentheses only the inner pair counts.
@@ -1551,14 +1598,29 @@
{
if (NULL != array)
{
+ char *normalized;
tmp = close_paren[0];
close_paren[0] = '\0';
- if (GNUNET_NO == find_duplicate ((const char *) &open_paren[1], (const
char **) array, index + count))
+ if (GNUNET_NO == find_duplicate ((const char *) &open_paren[1],
+ (const char **) array, index + count))
{
insert_non_mandatory_keyword ((const char *) &open_paren[1], array,
index + count);
count++;
}
+ normalized = normalize_metadata (EXTRACTOR_METAFORMAT_UTF8,
+ &open_paren[1], close_paren - &open_paren[1]);
+ if (normalized != NULL)
+ {
+ if (GNUNET_NO == find_duplicate ((const char *) normalized,
+ (const char **) array, index + count))
+ {
+ insert_non_mandatory_keyword ((const char *) normalized, array,
+ index + count);
+ count++;
+ }
+ GNUNET_free (normalized);
+ }
close_paren[0] = tmp;
}
else
@@ -1601,12 +1663,26 @@
{
if (NULL != array)
{
+ char *normalized;
if (GNUNET_NO == find_duplicate (p, (const char **) array, index + seps))
{
insert_non_mandatory_keyword (p, array,
index + seps);
seps++;
}
+ normalized = normalize_metadata (EXTRACTOR_METAFORMAT_UTF8,
+ p, strlen (p));
+ if (normalized != NULL)
+ {
+ if (GNUNET_NO == find_duplicate ((const char *) normalized,
+ (const char **) array, index + seps))
+ {
+ insert_non_mandatory_keyword ((const char *) normalized, array,
+ index + seps);
+ seps++;
+ }
+ GNUNET_free (normalized);
+ }
}
else
seps++;
@@ -1616,7 +1692,6 @@
}
#undef TOKENS
-
/**
* Function called on each value in the meta data.
* Adds it to the URI.
@@ -1640,15 +1715,28 @@
const char *data_mime_type, const char *data, size_t data_len)
{
struct GNUNET_FS_Uri *uri = cls;
+ char *normalized_data;
if ((format != EXTRACTOR_METAFORMAT_UTF8) &&
(format != EXTRACTOR_METAFORMAT_C_STRING))
return 0;
- if (find_duplicate (data, (const char **) uri->data.ksk.keywords,
uri->data.ksk.keywordCount))
- return GNUNET_OK;
- insert_non_mandatory_keyword (data,
- uri->data.ksk.keywords,
uri->data.ksk.keywordCount);
- uri->data.ksk.keywordCount++;
+ normalized_data = normalize_metadata (format, data, data_len);
+ if (!find_duplicate (data, (const char **) uri->data.ksk.keywords,
uri->data.ksk.keywordCount))
+ {
+ insert_non_mandatory_keyword (data,
+ uri->data.ksk.keywords,
uri->data.ksk.keywordCount);
+ uri->data.ksk.keywordCount++;
+ }
+ if (normalized_data != NULL)
+ {
+ if (!find_duplicate (normalized_data, (const char **)
uri->data.ksk.keywords, uri->data.ksk.keywordCount))
+ {
+ insert_non_mandatory_keyword (normalized_data,
+ uri->data.ksk.keywords,
uri->data.ksk.keywordCount);
+ uri->data.ksk.keywordCount++;
+ }
+ GNUNET_free (normalized_data);
+ }
return 0;
}
@@ -1690,8 +1778,9 @@
tok_keywords = get_keywords_from_tokens (filename, NULL, 0);
paren_keywords = get_keywords_from_parens (filename, NULL, 0);
}
+ /* x2 because there might be a normalized variant of every keyword */
ret->data.ksk.keywords = GNUNET_malloc (sizeof (char *) * (ent
- + tok_keywords + paren_keywords));
+ + tok_keywords + paren_keywords) * 2);
GNUNET_CONTAINER_meta_data_iterate (md, &gather_uri_data, ret);
}
if (tok_keywords > 0)
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [GNUnet-SVN] r18853 - gnunet/src/fs,
gnunet <=