commit 4184776: [Fix] Fix tld compositon rules

[Vsevolod Stakhov]

Author: Vsevolod Stakhov
Date: 2020-02-25 17:01:00 +0000
URL: https://github.com/rspamd/rspamd/commit/4184776a1744bfa1e73eac80e8126573d8aae0f0 (HEAD -> master)

[Fix] Fix tld compositon rules
Issue: #3248, #3268

---
 src/libserver/url.c        |  7 +++----
 src/libutil/multipattern.c | 10 +++++-----
 2 files changed, 8 insertions(+), 9 deletions(-)

diff --git a/src/libserver/url.c b/src/libserver/url.c
index dcf13cfda..0407490d8 100644
--- a/src/libserver/url.c
+++ b/src/libserver/url.c
@@ -442,7 +442,6 @@ rspamd_url_parse_tld_file (const gchar *fname,
 
 		flags = URL_FLAG_NOHTML | URL_FLAG_TLD_MATCH;
 
-#ifndef WITH_HYPERSCAN
 		if (linebuf[0] == '*') {
 			flags |= URL_FLAG_STAR_MATCH;
 			p = strchr (linebuf, '.');
@@ -456,9 +455,6 @@ rspamd_url_parse_tld_file (const gchar *fname,
 		else {
 			p = linebuf;
 		}
-#else
-		p = linebuf;
-#endif
 
 		m.flags = flags;
 		rspamd_multipattern_add_pattern (url_scanner->search_trie, p,
@@ -1553,6 +1549,9 @@ rspamd_tld_trie_callback (struct rspamd_multipattern *mp,
 			ndots--;
 			pos = p + 1;
 		}
+		else {
+			pos = p;
+		}
 
 		p--;
 	}
diff --git a/src/libutil/multipattern.c b/src/libutil/multipattern.c
index 6eec009b3..f243c1d8a 100644
--- a/src/libutil/multipattern.c
+++ b/src/libutil/multipattern.c
@@ -108,7 +108,6 @@ rspamd_multipattern_escape_tld_hyperscan (const gchar *pattern, gsize slen,
 	 */
 
 	if (pattern[0] == '*') {
-		len = slen + 4;
 		p = strchr (pattern, '.');
 
 		if (p == NULL) {
@@ -119,12 +118,13 @@ rspamd_multipattern_escape_tld_hyperscan (const gchar *pattern, gsize slen,
 			p ++;
 		}
 
-		prefix = ".*.";
+		prefix = "\\.";
+		len = slen + strlen (prefix);
 	}
 	else {
-		len = slen + 1;
-		prefix = ".";
+		prefix = "\\.";
 		p = pattern;
+		len = slen + strlen (prefix);
 	}
 
 	res = g_malloc (len + 1);
@@ -208,7 +208,7 @@ rspamd_multipattern_pattern_filter (const gchar *pattern, gsize len,
 			tmp = rspamd_multipattern_escape_tld_hyperscan (pattern, len, &tlen);
 
 			ret = rspamd_str_regexp_escape (tmp, tlen, dst_len,
-					gl_flags|RSPAMD_REGEXP_ESCAPE_GLOB);
+					gl_flags|RSPAMD_REGEXP_ESCAPE_RE);
 			g_free (tmp);
 		}
 		else if (flags & RSPAMD_MULTIPATTERN_RE) {