FS#342 UTF8 and search issues

UTF-8/Unicode

• 

  2005-05-22

  I previously posted this to the discussion page but received no responses:
  
  I installed the lastest version of Dokuwiki (2005-05-07), PHP version 4.3.3) and am having problems with searching. The first problem was an error of PHP not finding utf8_decode(). I changed the call to @utf8_decode() and that fixed that problem.
  
  The second problem is in search.php() as it relates to utf8. The code there reads:
  
  $p = utf8_strpos($lctext,$q);
  
  This always results in $p being set to zero in my setup. If I change the call to get rid of the utf8 portion, namely:
  
  $p = strpos($lctext,$q);
  
  the index is properly set.
  
  Note that all wiki contents were generated with the latest version so conversion probably doesn’t come into play here.
  
  Any ideas why utf8 isn’t working properly for me?
  
  Wow: If I don’t replace the line
  
  $l = utf8_strlen($q) + 200;
  
  with the non utf8 version, no further next is output to the page. (For example, if I put a print statement right after the utf8_strlen() line, it doesn’t show up. If I replace the utf8_strlen() line, it does show up.)
  
  
• 

  2005-05-22 andi

  Well adding an @ in front of utf8_decode just supresses the error message. If the function is not found you will get no return value and thus string will contain nothing.
  
  So your problem is the missing utf8_decode. According to the manual at http://www.php.net/manual/en/function.utf8-decode.php this function should be available in all PHP4 and PHP5 versions. So it looks like your PHP installation is broken.
  
  According to a post at http://www.faqts.com/knowledge_base/view.phtml/aid/357/fid/17 you maybe need to compile PHP with --with-xml but this seems to be PHP3 related. An blog entry at http://www.letitshine.org/?p=2 suggests to install a php-xml package (if available for your distribution)
  
  So to sum it up: you need that utf8_decode function!
  
• 

  2005-05-23 skipole

  It looks like you are right on the money. With Mandrake, it appears that you need to install php-xml. My search results still aren't perfect (there are no links to the pages matching the results), but this is serious progress. andi++