Mercurial > dnsbl

comparison src/scanner.h @ 147:812c80305f26 stable-5-26

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
fix 5.23 bug and add fsa debug logging
author carl
date Mon, 04 Dec 2006 21:49:09 -0800
parents ecb40aa3eaa5
children c7fc218686f5
comparison
equal deleted inserted replaced
146:7278c9766e26 147:812c80305f26
17 class recorder 17 class recorder
18 { 18 {
19 mlfiPriv *priv; // needed for syslog 19 mlfiPriv *priv; // needed for syslog
20 string_set *html_tags; // valid tags 20 string_set *html_tags; // valid tags
21 string_set *tlds; // valid tlds 21 string_set *tlds; // valid tlds
22 string_set *cctlds; // valid cctlds 22 string_set *cctlds; // valid cctlds
23 string_set hosts; 23 string_set hosts;
24 int bad_html_tags; 24 int bad_html_tags;
25 int binary_tags; 25 int binary_tags;
26 26
27 public: 27 public:
29 ~recorder() { empty(); }; 29 ~recorder() { empty(); };
30 void empty(); 30 void empty();
31 void new_url(char *host); 31 void new_url(char *host);
32 void new_tag(char *tag); 32 void new_tag(char *tag);
33 void binary(); 33 void binary();
34 void syslog(char *buf) { my_syslog(priv, buf); };
34 mlfiPriv *get_priv() { return priv; }; 35 mlfiPriv *get_priv() { return priv; };
35 string_set *get_cctlds() { return cctlds; }; 36 string_set *get_cctlds() { return cctlds; };
36 string_set *get_tlds() { return tlds; }; 37 string_set *get_tlds() { return tlds; };
37 string_set &get_hosts() { return hosts; }; 38 string_set &get_hosts() { return hosts; };
38 bool excessive_bad_tags(int limit) { return (limit > 0) && (bad_html_tags > limit) && (bad_html_tags > 3*binary_tags); }; 39 bool excessive_bad_tags(int limit) { return (limit > 0) && (bad_html_tags > limit) && (bad_html_tags > 3*binary_tags); };
39 bool excessive_hosts(int limit) { return (limit > 0) && (hosts.size() > limit); }; 40 bool excessive_hosts(int limit) { return (limit > 0) && (hosts.size() > limit); };
41
40 }; 42 };
41 43
42 44
43 //////////////////////////////////////////////// 45 ////////////////////////////////////////////////
44 // the content scanner 46 // the content scanner