diff --git a/conf/dokuwiki.php b/conf/dokuwiki.php
index f3fd8286ad7b49aed7bc0b356445bcf9566a0459..0df72bce43aea171d36c1a5d4f28c7351bdf095c 100644
--- a/conf/dokuwiki.php
+++ b/conf/dokuwiki.php
@@ -130,3 +130,5 @@ $conf['ftp']['user'] = 'user';
$conf['ftp']['pass'] = 'password';
$conf['ftp']['root'] = '/home/user/htdocs';
+/* FIXME: delete when no longer needed */
+$conf['test_indexer'] = 0;
diff --git a/inc/indexer.php b/inc/indexer.php
index 3789098e0e06144e7f421b98c8c2040cd02fe0d2..09b43a1b5bc61442fffd84d30a2700e673051166 100644
--- a/inc/indexer.php
+++ b/inc/indexer.php
@@ -17,14 +17,47 @@
// Ranges taken from http://en.wikipedia.org/wiki/Unicode_block
// I'm no language expert. If you think some ranges are wrongly chosen or
// a range is missing, please contact me
-define('IDX_ASIAN','['.
- '\x{0E00}-\x{0E7F}'. // Thai
- '\x{2E80}-\x{D7AF}'. // CJK -> Hangul
+define('IDX_ASIAN1','[\x{0E00}-\x{0E7F}]'); // Thai
+define('IDX_ASIAN2','['.
+ '\x{2E80}-\x{3040}'. // CJK -> Hangul
+ '\x{309D}-\x{30A0}'.
+ '\x{30FB}-\x{31EF}\x{3200}-\x{D7AF}'.
'\x{F900}-\x{FAFF}'. // CJK Compatibility Ideographs
'\x{FE30}-\x{FE4F}'. // CJK Compatibility Forms
']');
+define('IDX_ASIAN3','['. // Hiragana/Katakana (can be two characters)
+ '\x{3042}\x{3044}\x{3046}\x{3048}'.
+ '\x{304A}-\x{3062}\x{3064}-\x{3082}'.
+ '\x{3084}\x{3086}\x{3088}-\x{308D}'.
+ '\x{308F}-\x{3094}'.
+ '\x{30A2}\x{30A4}\x{30A6}\x{30A8}'.
+ '\x{30AA}-\x{30C2}\x{30C4}-\x{30E2}'.
+ '\x{30E4}\x{30E6}\x{30E8}-\x{30ED}'.
+ '\x{30EF}-\x{30F4}\x{30F7}-\x{30FA}'.
+ ']['.
+ '\x{3041}\x{3043}\x{3045}\x{3047}\x{3049}'.
+ '\x{3063}\x{3083}\x{3085}\x{3087}\x{308E}\x{3095}-\x{309C}'.
+ '\x{30A1}\x{30A3}\x{30A5}\x{30A7}\x{30A9}'.
+ '\x{30C3}\x{30E3}\x{30E5}\x{30E7}\x{30EE}\x{30F5}\x{30F6}\x{30FB}\x{30FC}'.
+ '\x{31F0}-\x{31FF}'.
+ ']?');
+/**
+ * Measure the length of a string.
+ * Differs from strlen in handling of asian characters.
+ *
+ * @author Tom N Harris <tnharris@whoopdedo.org>
+ */
+function wordlen($w){
+ $l = strlen($w);
+ // If left alone, all chinese "words" will get put into w3.idx
+ // So the "length" of a "word" is faked
+ if(preg_match('/'.IDX_ASIAN2.'/u',$w))
+ $l += ord($w) - 0xE1; // Lead bytes from 0xE2-0xEF
+ return $l;
+}
+
/**
* Write a list of strings to an index file.
*
@@ -113,7 +146,7 @@ function idx_getPageWords($page){
$arr = idx_tokenizer($word,$stopwords);
$arr = array_count_values($arr);
foreach ($arr as $w => $c) {
- $l = strlen($w);
+ $l = wordlen($w);
if(isset($words[$l])){
$words[$l][$w] = $c * $count + (isset($words[$l][$w]) ? $words[$l][$w] : 0);
}else{
@@ -130,8 +163,8 @@ function idx_getPageWords($page){
foreach ($words[$wlen] as $word => $freq) {
$wid = array_search("$word\n",$word_idx);
if(!is_int($wid)){
+ $wid = count($word_idx);
$word_idx[] = "$word\n";
- $wid = count($word_idx)-1;
}
if(!isset($index[$wlen]))
$index[$wlen] = array();
@@ -169,8 +202,10 @@ function idx_addPage($page){
$page_idx[] = "$page\n";
$pid = count($page_idx)-1;
// page was new - write back
- if (!idx_saveIndex('page','',$page_idx))
+ if (!idx_saveIndex('page','',$page_idx)){
+ trigger_error("Failed to write page index", E_USER_ERROR);
return false;
+ }
}
// get word usage in page
@@ -179,47 +214,23 @@ function idx_addPage($page){
if(!count($words)) return true;
foreach(array_keys($words) as $wlen){
- // Open index and temp file
- $fn = $conf['indexdir']."/i$wlen";
- idx_touchIndex('i',$wlen);
- $idx = fopen($fn.'.idx','r');
- $tmp = fopen($fn.'.tmp','w');
- if(!$idx || !$tmp){
- trigger_error("Failed to open index files", E_USER_ERROR);
- return false;
- }
-
- // copy from index to temp file, modifying where needed
- $lno = 0;
- $line = '';
- while (!feof($idx)) {
- // read full line
- $line .= fgets($idx, 4096);
- if(substr($line,-1) != "\n") continue;
-
- // write a new Line to temp file
- idx_writeIndexLine($tmp,$line,$pid,$words[$wlen][$lno]);
-
- $line = ''; // reset line buffer
- $lno++; // increase linecounter
+ $index = idx_getIndex('i',$wlen);
+ foreach($words[$wlen] as $wid => $freq){
+ if($wid<count($index)){
+ $index[$wid] = idx_updateIndexLine($index[$wid],$pid,$freq);
+ }else{
+ // New words **should** have been added in increasing order
+ // starting with the first unassigned index.
+ // If someone can show how this isn't true, then I'll need to sort
+ // or do something special.
+ $index[$wid] = idx_updateIndexLine('',$pid,$freq);
+ }
}
- fclose($idx);
-
- // add missing lines (usually index and word should contain
- // the same number of lines, however if the page contained
- // new words the word file has some more lines which need to
- // be added here
- $word_idx = idx_getIndex('w',$wlen);
- $wcnt = count($word_idx);
- for($lno; $lno<$wcnt; $lno++){
- idx_writeIndexLine($tmp,'',$pid,$words[$wlen][$lno]);
+ // save back word index
+ if(!idx_saveIndex('i',$wlen,$index)){
+ trigger_error("Failed to write index", E_USER_ERROR);
+ return false;
}
-
- // close the temp file and move it over to be the new one
- fclose($tmp);
- if($conf['fperm']) chmod($fn.'.tmp', $conf['fperm']);
- // try rename first (fast) fallback to copy (slow)
- io_rename($fn.'.tmp', $fn.'.idx');
}
return true;
@@ -232,11 +243,26 @@ function idx_addPage($page){
* given filehandle. It removes the given document from
* the given line and readds it when $count is >0.
*
+ * @deprecated - see idx_updateIndexLine
* @author Andreas Gohr <andi@splitbrain.org>
*/
function idx_writeIndexLine($fh,$line,$pid,$count){
- $line = trim($line);
+ fwrite($fh,idx_updateIndexLine($line,$pid,$count));
+}
+/**
+ * Modify an index line with new information
+ *
+ * This returns a line of the index. It removes the
+ * given document from the line and readds it if
+ * $count is >0.
+ *
+ * @author Tom N Harris <tnharris@whoopdedo.org>
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+function idx_updateIndexLine($line,$pid,$count){
+ $line = trim($line);
+ $updated = array();
if($line != ''){
$parts = explode(':',$line);
// remove doc from given line
@@ -244,18 +270,17 @@ function idx_writeIndexLine($fh,$line,$pid,$count){
if($part == '') continue;
list($doc,$cnt) = explode('*',$part);
if($doc != $pid){
- fwrite($fh,"$doc*$cnt:");
+ $updated[] = $part;
}
}
}
// add doc
if ($count){
- fwrite($fh,"$pid*$count");
+ $updated[] = "$pid*$count";
}
- // add newline
- fwrite($fh,"\n");
+ return join(':',$updated)."\n";
}
/**
@@ -266,20 +291,30 @@ function idx_writeIndexLine($fh,$line,$pid,$count){
*
* @author Tom N Harris <tnharris@whoopdedo.org>
*/
-function idx_indexLengths($minlen){
+function idx_indexLengths(&$filter){
global $conf;
$dir = @opendir($conf['indexdir']);
if($dir===false)
return array();
$idx = array();
- // Exact match first.
- if(@file_exists($conf['indexdir']."/i$minlen.idx"))
- $idx[] = $minlen;
- while (($f = readdir($dir)) !== false) {
- if (substr($f,0,1) == 'i' && substr($f,-4) == '.idx'){
- $i = substr($f,1,-4);
- if (is_numeric($i) && $i > $minlen)
- $idx[] = $i;
+ if(is_array($filter)){
+ while (($f = readdir($dir)) !== false) {
+ if (substr($f,0,1) == 'i' && substr($f,-4) == '.idx'){
+ $i = substr($f,1,-4);
+ if (is_numeric($i) && isset($filter[(int)$i]))
+ $idx[] = (int)$i;
+ }
+ }
+ }else{
+ // Exact match first.
+ if(@file_exists($conf['indexdir']."/i$filter.idx"))
+ $idx[] = $filter;
+ while (($f = readdir($dir)) !== false) {
+ if (substr($f,0,1) == 'i' && substr($f,-4) == '.idx'){
+ $i = substr($f,1,-4);
+ if (is_numeric($i) && $i > $filter)
+ $idx[] = (int)$i;
+ }
}
}
closedir($dir);
@@ -287,118 +322,182 @@ function idx_indexLengths($minlen){
}
/**
- * Lookup words in index
+ * Find the the index number of each search term.
*
- * Takes an array of word and will return a list of matching
- * documents for each one.
+ * There are two variation: Simple and Sorted.
+ * The simple version just takes the words one at a time.
+ * The sorted version will group together words that appear in the same index.
+ * So it should perform better, because it only opens each index once.
+ * Actually, it's not that great. (in my experience) Probably because of the disk cache.
+ * And the sorted function does more work, making it slightly slower in some cases.
*
- * Important: No ACL checking is done here! All results are
- * returned, regardless of permissions
+ * For now, you can choose to use the sorted version by setting $conf['test_indexer'] = 1
+ * Eventually, the more worthy will be chosen and the loser cast into the deepest depths.
*
- * @author Andreas Gohr <andi@splitbrain.org>
+ * @param array $words The query terms. Words should only contain valid characters,
+ * with a '*' at either the beginning or end of the word (or both)
+ * @param arrayref $result Set to word => array("length*id" ...), use this to merge the
+ * index locations with the appropriate query term.
+ * @return array Set to length => array(id ...)
+ *
+ * @author Tom N Harris <tnharris@whoopdedo.org>
*/
-function idx_lookup($words){
- global $conf;
-
- $result = array();
-
- // load known words and documents
- $page_idx = idx_getIndex('page','');
-
+function idx_getIndexWordsSimple($words, &$result){
// get word IDs
$wids = array();
foreach($words as $word){
$result[$word] = array();
$wild = 0;
$xword = $word;
- $wlen = strlen($word);
+ $wlen = wordlen($word);
// check for wildcards
if(substr($xword,0,1) == '*'){
$xword = substr($xword,1);
- $wild = 1;
- $ptn = '/'.preg_quote($xword,'/').'$/';
+ $wild |= 1;
$wlen -= 1;
-# $l = -1*strlen($xword)-1;
}
if(substr($xword,-1,1) == '*'){
$xword = substr($xword,0,-1);
- $wild += 2;
+ $wild |= 2;
$wlen -= 1;
}
if ($wlen < 3 && $wild == 0 && !is_numeric($xword)) continue;
// look for the ID(s) for the given word
if($wild){ // handle wildcard search
+ $ptn = preg_quote($xword,'/');
+ if(($wild&1) == 0) $ptn = '^'.$ptn;
+ if(($wild&2) == 0) $ptn = $ptn.'$';
+ $ptn = '/'.$ptn.'/';
foreach (idx_indexLengths($wlen) as $ixlen){
$word_idx = idx_getIndex('w',$ixlen);
- $cnt = count($word_idx);
- for($wid=0; $wid<$cnt; $wid++){
- $iword = $word_idx[$wid];
- if( (($wild==3) && is_int(strpos($iword,$xword))) ||
-# (($wild==1) && ("$xword\n" == substr($iword,$l))) ||
- (($wild==1) && preg_match($ptn,$iword)) ||
-# (($wild==2) && ($xword == substr($iword,0,strlen($xword))))
- (($wild==2) && (0 === strpos($iword,$xword)))
-
- ){
- if(!isset($wids[$ixlen])) $wids[$ixlen] = array();
- $wids[$ixlen][] = $wid;
- $result[$word][] = "$ixlen*$wid";
- }
+ foreach(array_keys(preg_grep($ptn,$word_idx)) as $wid){
+ $wids[$ixlen][] = $wid;
+ $result[$word][] = "$ixlen*$wid";
}
}
}else{ // handle exact search
$word_idx = idx_getIndex('w',$wlen);
$wid = array_search("$word\n",$word_idx);
if(is_int($wid)){
- $wids[$wlen] = array($wid);
+ $wids[$wlen][] = $wid;
$result[$word][] = "$wlen*$wid";
}else{
$result[$word] = array();
}
}
}
+ return $wids;
+}
+function idx_getIndexWordsSorted($words,&$result){
+ // parse and sort tokens
+ $tokens = array();
+ $tokenlength = array();
+ $tokenwild = array();
+ foreach($words as $word){
+ $result[$word] = array();
+ $wild = 0;
+ $xword = $word;
+ $wlen = wordlen($word);
- $docs = array(); // hold docs found
- foreach(array_keys($wids) as $wlen){
- sort($wids[$wlen]);
- $wids[$wlen] = array_unique($wids[$wlen]);
-
- // Open index
- idx_touchIndex('i',$wlen);
- $idx = fopen($conf['indexdir']."/i$wlen.idx",'r');
- if(!$idx){
- msg("Failed to open index file",-1);
- return false;
+ // check for wildcards
+ if(substr($xword,0,1) == '*'){
+ $xword = substr($xword,1);
+ $wild |= 1;
+ $wlen -= 1;
}
-
- // Walk the index til the lines are found
- $lno = 0;
- $line = '';
- $ixids =& $wids[$wlen];
- $srch = array_shift($ixids); // which word do we look for?
- while (!feof($idx)) {
- // read full line
- $line .= fgets($idx, 4096);
- if(substr($line,-1) != "\n") continue;
- if($lno > $srch) break; // shouldn't happen
-
- // do we want this line?
- if($lno == $srch){
- // add docs to list
- $docs["$wlen*$srch"] = idx_parseIndexLine($page_idx,$line);
-
- $srch = array_shift($ixids); // next word to look up
- if($srch == null) break; // no more words
+ if(substr($xword,-1,1) == '*'){
+ $xword = substr($xword,0,-1);
+ $wild |= 2;
+ $wlen -= 1;
+ }
+ if ($wlen < 3 && $wild == 0 && !is_numeric($xword)) continue;
+ if(!isset($tokens[$xword])){
+ $tokenlength[$wlen][] = $xword;
+ }
+ if($wild){
+ $ptn = preg_quote($xword,'/');
+ if(($wild&1) == 0) $ptn = '^'.$ptn;
+ if(($wild&2) == 0) $ptn = $ptn.'$';
+ $tokens[$xword][] = array($word, '/'.$ptn.'/');
+ if(!isset($tokenwild[$xword])) $tokenwild[$xword] = $wlen;
+ }else
+ $tokens[$xword][] = array($word, null);
+ }
+ asort($tokenwild);
+ // $tokens = array( base word => array( [ query word , grep pattern ] ... ) ... )
+ // $tokenlength = array( base word length => base word ... )
+ // $tokenwild = array( base word => base word length ... )
+
+ $length_filter = empty($tokenwild) ? $tokenlength : min(array_keys($tokenlength));
+ $indexes_known = idx_indexLengths($length_filter);
+ if(!empty($tokenwild)) sort($indexes_known);
+ // get word IDs
+ $wids = array();
+ echo "\n";
+ foreach($indexes_known as $ixlen){
+ $word_idx = idx_getIndex('w',$ixlen);
+ // handle exact search
+ if(isset($tokenlength[$ixlen])){
+ foreach($tokenlength[$ixlen] as $xword){
+ $wid = array_search("$xword\n",$word_idx);
+ if(is_int($wid)){
+ $wids[$ixlen][] = $wid;
+ foreach($tokens[$xword] as $w)
+ $result[$w[0]][] = "$ixlen*$wid";
+ }
+ }
+ }
+ // handle wildcard search
+ foreach($tokenwild as $xword => $wlen){
+ if($wlen >= $ixlen) break;
+ foreach($tokens[$xword] as $w){
+ if(is_null($w[1])) continue;
+ foreach(array_keys(preg_grep($w[1],$word_idx)) as $wid){
+ $wids[$ixlen][] = $wid;
+ $result[$w[0]][] = "$ixlen*$wid";
+ }
}
-
- $line = ''; // reset line buffer
- $lno++; // increase linecounter
}
- fclose($idx);
}
+ return $wids;
+}
+/**
+ * Lookup words in index
+ *
+ * Takes an array of word and will return a list of matching
+ * documents for each one.
+ *
+ * Important: No ACL checking is done here! All results are
+ * returned, regardless of permissions
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+function idx_lookup($words){
+ global $conf;
+
+ $result = array();
+
+ if(isset($conf['test_indexer']) && ($conf['test_indexer']&1))
+ $wids = idx_getIndexWordsSorted($words, $result);
+ else
+ $wids = idx_getIndexWordsSimple($words, $result);
+ if(empty($wids)) return array();
+
+ // load known words and documents
+ $page_idx = idx_getIndex('page','');
+
+ $docs = array(); // hold docs found
+ foreach(array_keys($wids) as $wlen){
+ $wids[$wlen] = array_unique($wids[$wlen]);
+ $index = idx_getIndex('i',$wlen);
+ foreach($wids[$wlen] as $ixid){
+ if($ixid < count($index))
+ $docs["$wlen*$ixid"] = idx_parseIndexLine($page_idx,$index[$ixid]);
+ }
+ }
// merge found pages into final result array
$final = array();
@@ -453,8 +552,6 @@ function idx_parseIndexLine(&$page_idx,$line){
* @param string $string the query as given by the user
* @param arrayref $stopwords array of stopwords
* @param boolean $wc are wildcards allowed?
- *
- * @todo make combined function to use alone or in getPageWords
*/
function idx_tokenizer($string,&$stopwords,$wc=false){
$words = array();
@@ -462,7 +559,7 @@ function idx_tokenizer($string,&$stopwords,$wc=false){
if(preg_match('/[^0-9A-Za-z]/u', $string)){
// handle asian chars as single words (may fail on older PHP version)
- $asia = @preg_replace('/('.IDX_ASIAN.')/u','\1 ',$string);
+ $asia = @preg_replace('/('.IDX_ASIAN1.'|'.IDX_ASIAN2.'|'.IDX_ASIAN3.')/u',' \1 ',$string);
if(!is_null($asia)) $string = $asia; //recover from regexp failure
$arr = explode(' ', utf8_stripspecials($string,' ','\._\-:'.$wc));
diff --git a/inc/utf8.php b/inc/utf8.php
index e32b64b1744bfc5a9b5bbd22e785fd53535dc273..8da981fd9ccf82e4526c944bcb7eb03cf2f613e5 100644
--- a/inc/utf8.php
+++ b/inc/utf8.php
@@ -1027,11 +1027,20 @@ $UTF8_SPECIAL_CHARS = array(
0x279f, 0x27a0, 0x27a1, 0x27a2, 0x27a3, 0x27a4, 0x27a5, 0x27a6, 0x27a7, 0x27a8,
0x27a9, 0x27aa, 0x27ab, 0x27ac, 0x27ad, 0x27ae, 0x27af, 0x27b1, 0x27b2, 0x27b3,
0x27b4, 0x27b5, 0x27b6, 0x27b7, 0x27b8, 0x27b9, 0x27ba, 0x27bb, 0x27bc, 0x27bd,
- 0x27be, 0xf6d9, 0xf6da, 0xf6db, 0xf8d7, 0xf8d8, 0xf8d9, 0xf8da, 0xf8db, 0xf8dc,
+ 0x27be, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x300c,
+ 0x300d, 0x300e, 0x300f, 0x3010, 0x3011, 0x3012, 0x3014, 0x3015, 0x3016, 0x3017,
+ 0x3018, 0x3019, 0x301a, 0x301b, 0x3036,
+ 0xf6d9, 0xf6da, 0xf6db, 0xf8d7, 0xf8d8, 0xf8d9, 0xf8da, 0xf8db, 0xf8dc,
0xf8dd, 0xf8de, 0xf8df, 0xf8e0, 0xf8e1, 0xf8e2, 0xf8e3, 0xf8e4, 0xf8e5, 0xf8e6,
0xf8e7, 0xf8e8, 0xf8e9, 0xf8ea, 0xf8eb, 0xf8ec, 0xf8ed, 0xf8ee, 0xf8ef, 0xf8f0,
0xf8f1, 0xf8f2, 0xf8f3, 0xf8f4, 0xf8f5, 0xf8f6, 0xf8f7, 0xf8f8, 0xf8f9, 0xf8fa,
0xf8fb, 0xf8fc, 0xf8fd, 0xf8fe, 0xfe7c, 0xfe7d,
+ 0xff01, 0xff02, 0xff03, 0xff04, 0xff05, 0xff06, 0xff07, 0xff08, 0xff09,
+ 0xff09, 0xff0a, 0xff0b, 0xff0c, 0xff0d, 0xff0e, 0xff0f, 0xff1a, 0xff1b, 0xff1c,
+ 0xff1d, 0xff1e, 0xff1f, 0xff20, 0xff3b, 0xff3c, 0xff3d, 0xff3e, 0xff40, 0xff5b,
+ 0xff5c, 0xff5d, 0xff5e, 0xff5f, 0xff60, 0xff61, 0xff62, 0xff63, 0xff64, 0xff65,
+ 0xffe0, 0xffe1, 0xffe2, 0xffe3, 0xffe4, 0xffe5, 0xffe6, 0xffe8, 0xffe9, 0xffea,
+ 0xffeb, 0xffec, 0xffed, 0xffee,
);
// utf8 version of above data
@@ -1051,8 +1060,12 @@ $UTF8_SPECIAL_CHARS2 =
'✲✳✴✵✶✷✸✹✺✻✼✽✾✿â€ââ‚âƒâ„â…â†â‡âˆâ‰âŠâ‹ï¿½'.
'�âââ‘â’â–â˜â™âšâ›âœââžâ¡â¢â£â¤â¥â¦â§â¿âž‰âž“➔➘➙➚�'.
'��➜âžâžžâžŸâž ➡➢➣➤➥➦➧➨➩➪➫➬âžâž®âž¯âž±âž²âž³âž´âžµâž¶'.
- '➷➸➹➺➻➼➽➾ï£ï£žï£Ÿï£ �'.
- '�ï£ï£®ï£¯ï£°ï£±ï£²ï£³ï£´ï£µï£¶ï£·ï£¸ï£¹ï£ºï£»ï£¼ï£½ï£¾ï¹¼ï¹½';
+ '➷➸➹➺➻➼➽➾'.
+ '　ã€ã€‚〃〈〉《》「ã€ã€Žã€ã€ã€‘〒〔〕〖〗〘〙〚〛〶'.
+ 'ï£ï£žï£Ÿï£ �'.
+ '�ï£ï£®ï£¯ï£°ï£±ï£²ï£³ï£´ï£µï£¶ï£·ï£¸ï£¹ï£ºï£»ï£¼ï£½ï£¾ï¹¼ï¹½'.
+ 'ï¼ï¼‚＃＄％＆＇（）＊＋，ï¼ï¼Žï¼ï¼šï¼›ï¼œï¼ï¼žï¼Ÿï¼ ［＼］＾｀｛｜ï½ï½ž'.
+ 'ï½Ÿï½ ï½¡ï½¢ï½£ï½¤ï½¥ï¿ ï¿¡ï¿¢ï¿£ï¿¤ï¿¥ï¿¦ï¿¨ï¿©ï¿ªï¿«ï¿¬ï¿ï¿®';
/**
* Romanization lookup table
diff --git a/lib/exe/indexer.php b/lib/exe/indexer.php
index f5fc07bc3ce906600f84a662610ad5c639c248a7..057ab06f1bf2459334250b9c5da62b22b95f8898 100644
--- a/lib/exe/indexer.php
+++ b/lib/exe/indexer.php
@@ -13,6 +13,9 @@ require_once(DOKU_INC.'inc/events.php');
session_write_close(); //close session
if(!defined('NL')) define('NL',"\n");
+// Version tag used to force rebuild on upgrade
+define(INDEXER_VERSION, 1);
+
// keep running after browser closes connection
@ignore_user_abort(true);
@@ -136,10 +139,15 @@ function runIndexer(){
if(!$ID) return false;
// check if indexing needed
- $last = @filemtime(metaFN($ID,'.indexed'));
- if($last > @filemtime(wikiFN($ID))){
- print "runIndexer(): index for $ID up to date".NL;
- return false;
+ $idxtag = metaFN($ID,'.indexed');
+ if(@file_exists($idxtag)){
+ if(io_readFile($idxtag) >= INDEXER_VERSION){
+ $last = @filemtime($idxtag);
+ if($last > @filemtime(wikiFN($ID))){
+ print "runIndexer(): index for $ID up to date".NL;
+ return false;
+ }
+ }
}
// try to aquire a lock
@@ -163,7 +171,7 @@ function runIndexer(){
idx_addPage($ID);
// we're finished - save and free lock
- io_saveFile(metaFN($ID,'.indexed'),' ');
+ io_saveFile(metaFN($ID,'.indexed'),INDEXER_VERSION);
@rmdir($lock);
print "runIndexer(): finished".NL;
return true;
diff --git a/lib/plugins/config/lang/en/lang.php b/lib/plugins/config/lang/en/lang.php
index 526ac4a0e7b855076fcd29c0485f350dc4d270c3..24b8db147ecb3499902161ebfcc1522b4ac97940 100644
--- a/lib/plugins/config/lang/en/lang.php
+++ b/lib/plugins/config/lang/en/lang.php
@@ -180,3 +180,5 @@ $lang['compression_o_0'] = 'none';
$lang['compression_o_gz'] = 'gzip';
$lang['compression_o_bz2'] = 'bz2';
+/* FIXME: delete when no longer needed */
+$lang['test_indexer'] = 'New Indexer testing bitfield: 0x1 -> sorted searching';
diff --git a/lib/plugins/config/settings/config.metadata.php b/lib/plugins/config/settings/config.metadata.php
index 36b2be38baaa58a2fa976ffb553154e1241e2b7c..44b121dfa343c30a8af2548eb6e431f9d9be940e 100644
--- a/lib/plugins/config/settings/config.metadata.php
+++ b/lib/plugins/config/settings/config.metadata.php
@@ -163,6 +163,7 @@ $meta['rss_type'] = array('multichoice','_choices' => array('rss','rss1','rss
$meta['rss_linkto'] = array('multichoice','_choices' => array('diff','page','rev','current'));
$meta['rss_update'] = array('numeric');
$meta['recent_days'] = array('numeric');
+$meta['test_indexer'] = array('numeric'); // FIXME: delete when no longer needed
$meta['_network'] = array('fieldset');
$meta['proxy____host'] = array('string','_pattern' => '#^[a-z0-9\-\.+]+?#i');