| 90 |
90 |
$excerpt_array = array();
|
| 91 |
91 |
$word = '('.implode('|', $search_words).')';
|
| 92 |
92 |
// Build the regex-string
|
| 93 |
|
// start-sign: .!?; + INVERTED EXCLAMATION MARK - INVERTED QUESTION MARK - DOUBLE EXCLAMATION MARK - INTERROBANG - EXCLAMATION QUESTION MARK - QUESTION EXCLAMATION MARK - DOUBLE QUESTION MARK - HALFWIDTH IDEOGRAPHIC FULL STOP - IDEOGRAPHIC FULL STOP - IDEOGRAPHIC COMMA
|
| 94 |
|
$str1=".!?;"."\xC2\xA1"."\xC2\xBF"."\xE2\x80\xBC"."\xE2\x80\xBD"."\xE2\x81\x89"."\xE2\x81\x88"."\xE2\x81\x87"."\xEF\xBD\xA1"."\xE3\x80\x82"."\xE3\x80\x81";
|
| 95 |
|
// stop-sign: .!?; + DOUBLE EXCLAMATION MARK - INTERROBANG - EXCLAMATION QUESTION MARK - QUESTION EXCLAMATION MARK - DOUBLE QUESTION MARK - HALFWIDTH IDEOGRAPHIC FULL STOP - IDEOGRAPHIC FULL STOP - IDEOGRAPHIC COMMA
|
| 96 |
|
$str2=".!?;"."\xE2\x80\xBC"."\xE2\x80\xBD"."\xE2\x81\x89"."\xE2\x81\x88"."\xE2\x81\x87"."\xEF\xBD\xA1"."\xE3\x80\x82"."\xE3\x80\x81";
|
| 97 |
|
$regex='/(?:^|\b|['.$str1.'])([^'.$str1.']{0,200}?'.$word.'[^'.$str2.']{0,200}(?:['.$str2.']|\b|$))/Sisu';
|
| 98 |
|
if(version_compare(PHP_VERSION, '4.3.3', '>=')) {
|
|
93 |
if(strpos(strtoupper(PHP_OS), 'WIN')===0) { // windows -> see below
|
|
94 |
$str1=".!?;";
|
|
95 |
$str2=".!?;";
|
|
96 |
} else { // linux & Co.
|
|
97 |
// start-sign: .!?; + INVERTED EXCLAMATION MARK - INVERTED QUESTION MARK - DOUBLE EXCLAMATION MARK - INTERROBANG - EXCLAMATION QUESTION MARK - QUESTION EXCLAMATION MARK - DOUBLE QUESTION MARK - HALFWIDTH IDEOGRAPHIC FULL STOP - IDEOGRAPHIC FULL STOP - IDEOGRAPHIC COMMA
|
|
98 |
$str1=".!?;"."\xC2\xA1"."\xC2\xBF"."\xE2\x80\xBC"."\xE2\x80\xBD"."\xE2\x81\x89"."\xE2\x81\x88"."\xE2\x81\x87"."\xEF\xBD\xA1"."\xE3\x80\x82"."\xE3\x80\x81";
|
|
99 |
// stop-sign: .!?; + DOUBLE EXCLAMATION MARK - INTERROBANG - EXCLAMATION QUESTION MARK - QUESTION EXCLAMATION MARK - DOUBLE QUESTION MARK - HALFWIDTH IDEOGRAPHIC FULL STOP - IDEOGRAPHIC FULL STOP - IDEOGRAPHIC COMMA
|
|
100 |
$str2=".!?;"."\xE2\x80\xBC"."\xE2\x80\xBD"."\xE2\x81\x89"."\xE2\x81\x88"."\xE2\x81\x87"."\xEF\xBD\xA1"."\xE3\x80\x82"."\xE3\x80\x81";
|
|
101 |
}
|
|
102 |
$regex='/(?:^|\b|['.$str1.'])([^'.$str1.']{0,200}?'.$word.'[^'.$str2.']{0,200}(?:['.$str2.']|\b|$))/isu';
|
|
103 |
if(version_compare(PHP_VERSION, '4.3.3', '>=') &&
|
|
104 |
strpos(strtoupper(PHP_OS), 'WIN')!==0) { // this may crash windows server, so skip if on windows
|
| 99 |
105 |
// jump from match to match, get excerpt, stop if $max_excerpt_num is reached
|
| 100 |
106 |
$last_end = 0; $offset = 0;
|
| 101 |
107 |
while(preg_match('/'.$word.'/Sisu', $text, $match_array, PREG_OFFSET_CAPTURE, $last_end)) {
|
| 102 |
108 |
$offset = ($match_array[0][1]-206 < $last_end)?$last_end:$match_array[0][1]-206;
|
| 103 |
109 |
if(preg_match($regex, $text, $matches, PREG_OFFSET_CAPTURE, $offset)) {
|
| 104 |
110 |
$last_end = $matches[1][1]+strlen($matches[1][0])-1;
|
| 105 |
|
if(!preg_match('/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\./S', $matches[1][0])) // skip excerpts with email-addresses
|
|
111 |
if(!preg_match('/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\./', $matches[1][0])) // skip excerpts with email-addresses
|
| 106 |
112 |
$excerpt_array[] = trim($matches[1][0]);
|
| 107 |
113 |
if(count($excerpt_array)>=$max_excerpt_num) {
|
| 108 |
114 |
$excerpt_array = array_unique($excerpt_array);
|
| ... | ... | |
| 116 |
122 |
} else { // compatile, but may be very slow with many large pages
|
| 117 |
123 |
if(preg_match_all($regex, $text, $match_array)) {
|
| 118 |
124 |
foreach($match_array[1] AS $string) {
|
| 119 |
|
if(!preg_match('/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\./S', $string)) // skip excerpts with email-addresses
|
|
125 |
if(!preg_match('/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\./', $string)) // skip excerpts with email-addresses
|
| 120 |
126 |
$excerpt_array[] = trim($string);
|
| 121 |
127 |
}
|
| 122 |
128 |
}
|
| ... | ... | |
| 233 |
239 |
if($mod_no_highlight) // no highlighting
|
| 234 |
240 |
{ $mod_page_link_target = "&nohighlight=1".$mod_page_link_target; }
|
| 235 |
241 |
// clean the text:
|
| 236 |
|
$mod_text = preg_replace('/\s+/', ' ', $mod_text);
|
| 237 |
242 |
$mod_text = preg_replace('#<(!--.*--|style.*</style|script.*</script)>#SiU', ' ', $mod_text);
|
| 238 |
243 |
$mod_text = preg_replace('#<(br( /)?|dt|/dd|/?(h[1-6]|tr|table|p|li|ul|pre|code|div|hr))[^>]*>#Si', '.', $mod_text);
|
|
244 |
$mod_text = preg_replace('/\s+/', ' ', $mod_text);
|
|
245 |
$mod_text = preg_replace('/ \./', '.', $mod_text);
|
| 239 |
246 |
$mod_text = entities_to_umlauts($mod_text, 'UTF-8');
|
| 240 |
247 |
$anchor_text = $mod_text; // make an copy containing html-tags
|
| 241 |
248 |
$mod_text = strip_tags($mod_text);
|
search: fixed windows-related regex issue