diff options
| author | Vsevolod Stakhov <vsevolod@rambler-co.ru> | 2009-06-02 19:32:34 +0400 |
|---|---|---|
| committer | Vsevolod Stakhov <vsevolod@rambler-co.ru> | 2009-06-02 19:32:34 +0400 |
| commit | 7bae787900fea17ca82393886217c6287d7e8cea (patch) | |
| tree | 4f358b3624d7b2ba6c86a25057d4ba7db10965ae /src/tokenizers | |
| parent | 4eb2985d1ef3631fca82cbf18cc2e8d7aab9b096 (diff) | |
| download | rspamd-7bae787900fea17ca82393886217c6287d7e8cea.tar.gz rspamd-7bae787900fea17ca82393886217c6287d7e8cea.zip | |
* Rework url parsing algorithms
* Adopt all parts of rspamd for new url parser
* Improve url-extracter utility by avoiding cut&paste of mime parsing
* Small fixes to rspamc client
* Bump version to 0.1.3
Diffstat (limited to 'src/tokenizers')
| -rw-r--r-- | src/tokenizers/tokenizers.c | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/src/tokenizers/tokenizers.c b/src/tokenizers/tokenizers.c index 1b47289a2..4527e699c 100644 --- a/src/tokenizers/tokenizers.c +++ b/src/tokenizers/tokenizers.c @@ -122,6 +122,7 @@ tokenize_urls (memory_pool_t *pool, struct worker_task *task, GTree **tree) token_node_t *new = NULL; f_str_t url_domain; struct uri *url; + GList *cur; uint32_t h; if (*tree == NULL) { @@ -129,7 +130,9 @@ tokenize_urls (memory_pool_t *pool, struct worker_task *task, GTree **tree) memory_pool_add_destructor (pool, (pool_destruct_func)g_tree_destroy, *tree); } - TAILQ_FOREACH (url, &task->urls, next) { + cur = task->urls; + while (cur) { + url = cur->data; url_domain.begin = url->host; url_domain.len = url->hostlen; new = memory_pool_alloc (pool, sizeof (token_node_t)); @@ -139,6 +142,7 @@ tokenize_urls (memory_pool_t *pool, struct worker_task *task, GTree **tree) if (g_tree_lookup (*tree, new) == NULL) { g_tree_insert (*tree, new, new); } + cur = g_list_next (cur); } return TRUE; |