[dokuwiki] Indexing of modified pages

  • From: "Todd Augsburger" <todd@xxxxxxxxxxxxxxxx>
  • To: <dokuwiki@xxxxxxxxxxxxx>
  • Date: Thu, 19 Jul 2007 08:21:12 -0400

I've been thinking about the indexing process.

Is there any attempt to "remove" items from the indexes when pages are 
modified? (or removed?) I find that after significant modification, searches 
return "phantom" page results where key words no longer exist within the 
page.

If not, how could it be done?
(Currently I use the Searchindex Manager plugin periodically, which I 
patched to simply delete all the index files in the current dokuwiki release 
before reindexing--not an acceptable method for real-time indexing.)

Would it make sense to use INDEXER_PAGE_ADD and a plugin to "remove" all the 
page entries from all the indexes prior to the add? This seems like it might 
be a bit slow, but I can't think of any better way. My first thought is to 
simply open each index file and use a string replacement to remove all 
occurrences of a page index. Would it be better (safer?) to read them into 
arrays and rewrite them? Does anyone foresee any problem with leaving blank 
lines within the indexes?

Or should this be "core" code? <grin>

Todd Augsburger
todd@xxxxxxxxxxxxxxxx
Roller Organs
www.rollerorgans.com
 

-- 
DokuWiki mailing list - more info at
http://wiki.splitbrain.org/wiki:mailinglist

Other related posts: