Mercurial > dnsbl

comparison src/scanner.cpp @ 52:a84752107aca

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
host names cannot have consecutive periods.
author carl
date Thu, 15 Jul 2004 23:01:41 -0700
parents 05957b18b2e6
children a39c813e8f7b
comparison
equal deleted inserted replaced
51:9f2971c692d0 52:a84752107aca
955 pending[--count] = '\0'; // null terminate host name by overwriting the terminator 955 pending[--count] = '\0'; // null terminate host name by overwriting the terminator
956 if (!strchr((const char *)pending, '@')) { 956 if (!strchr((const char *)pending, '@')) {
957 // not an email address or message id 957 // not an email address or message id
958 char *p1 = strchr((const char *)pending, '.'); 958 char *p1 = strchr((const char *)pending, '.');
959 char *p2 = strrchr((const char *)pending, '.'); 959 char *p2 = strrchr((const char *)pending, '.');
960 if (p1 && (p1 != p2)) { 960 char *p3 = strstr((const char *)pending, "..");
961 // have two periods, so three components 961 if (p1 && (p1 != p2) & !p3) {
962 // have two periods, so at least three components, and no empty components
962 for (int i=0; i<count; i++) pending[i] = tolower(pending[i]); 963 for (int i=0; i<count; i++) pending[i] = tolower(pending[i]);
963 // is last component a tld? 964 // is last component a tld?
964 string_set::iterator i = memory->tlds->find(p2+1); 965 string_set::iterator i = memory->tlds->find(p2+1);
965 if (i != memory->tlds->end()) memory->new_url((char*)pending); 966 if (i != memory->tlds->end()) memory->new_url((char*)pending);
966 } 967 }
1008 // url recognizer 1009 // url recognizer
1009 case u_reco: { 1010 case u_reco: {
1010 if (count > 13) { // need some minimal length host name after the protocol 1011 if (count > 13) { // need some minimal length host name after the protocol
1011 pending[--count] = '\0'; // null terminate host name by overwriting the terminator 1012 pending[--count] = '\0'; // null terminate host name by overwriting the terminator
1012 char *p = strrchr((const char *)pending, '/'); 1013 char *p = strrchr((const char *)pending, '/');
1013 if (p && // have a leading / 1014 if (p && // have a leading /
1014 strchr(p, '.') && // require at least one . in a dns name 1015 strchr(p, '.') && // require at least one . in a dns name
1016 !strstr(p, "..") && // no empty components in the dns name
1015 (strncasecmp((const char *)pending, "http", 4) == 0)) { // must start with protocol 1017 (strncasecmp((const char *)pending, "http", 4) == 0)) { // must start with protocol
1016 // we seem to have a host name 1018 // we seem to have a host name
1017 p++; // skip the last / 1019 p++; // skip the last /
1018 int c = strlen(p); 1020 int c = strlen(p);
1019 for (int i=0; i<c; i++) p[i] = tolower(p[i]); 1021 for (int i=0; i<c; i++) p[i] = tolower(p[i]);