• {SF(aid=>788425)}{SF} priority 7, medium security flaw
• {SF(aid=>783605)}{SF}
• Fields not searched
  • page description
  • articles body using search module (works in the list articles page) Chealer9 20031025
  • blog posts title
• Files found have real filename as link title instead of Tiki name.
• FAQ : Searching for FAQs displays multiple results (as many as there are questions/answers for the found FAQ, try to test "Linux" on this site) Chealer9
• If by example tiki_p_view is not assigned to anonymous but it is on specific pages, search result (only when searching in Wiki pages) is "Permission denied you cannot view this section". This is a minor but possibly confusing bug. Chealer9

• {SF(aid=>813756,tag=>rfe)}{SF}
• {SF(aid=>665440,tag=>rfe)}{SF}
• {SF(aid=>721265,tag=>rfe)}{SF}
• {SF(aid=>791710,tag=>rfe)}{SF}
• {SF(aid=>802268,tag=>rfe)}{SF} (priority 5)
• {SF(aid=>814666,tag=>rfe)}{SF} (priority 2)
• {SF(aid=>739646,tag=>rfe)}{SF} (priority 5)
• {SF(aid=>792892,tag=>rfe)}{SF} Maybe later...
• a wiki page with a name "MultilingualDev" must be found by a query "multilingual" even if the full search feature is on (idem a query "rss" must find the pages named RSSFeedDev and RssFeedDev.)
• full-text search on the wiki page titles and descriptions only
  • The Search Wiki Page Name module is nice, but it only finds pages if multiple words are entered in the exact order as the page names with no missing words in between

It's obvious that we could use advanced search features, multiple filtering options. This will be hard to do with DB abstraction. Let's stil list some trackers, you can add your ideas.

• {SF(aid=>813345,tag=>rfe)}{SF}

http://php.warpedweb.net/sitesearch/
mwexler: Marc on the dev list mentioned mindmeld as a potential partner. While I like the idea, it does not solve the search problem. Mindmeld requires that the world be structured into thoughts and questions, and that's a bit more structured than tiki. It would be a great module to add to the left nav, however. Perhaps something like Wordindex at sourceforge would be more useful. 2003/09/12
Marc Laporte: interesting. It can index PDFs and other file types, it skips a list of "bad" words, etc But it uses Perl...

Are the comments part of the container or not? If I look for a word that occurs only in a page comment, do the search give back the page? Idem does a found word in a comment change the relevance of the page?
Marc Laporte thinks yes

http://www.hansanderson.com/php/search/
broubrou: I suggest using SEARCHpHp instead of writing the whole search thing from scratch. It's an already working solution and it's all php-based. From what I understand of the problem, SEARCpHp does pretty much everything you need.

• Index is completely portable, copy it to any machine using PHP and SEARCpHp, Unix or other
• You can index from file system, databases, the web, whatever. If you can access it in PHP, you can index it
• It's Free, and always will be, under the same license as PHP itself

http://phpmysearch.web4.hm/index.php
mwexler: I stumbled on this one while seeing what people thought of SEARCHpHp. Support is commercial, but code is open source. Neither of these had overwhelming crowd support, but either is probably better than starting from scratch.

http://www.mnogosearch.ru/
mwexler: Another one I stumbled across. Commcercial for Windows, but free for Perl and PHP versions. Same comments as above.

Deno: of the three above mentioned solutions, the mnoGoSearch is the only one I saw used in a "real life" situation. It's used by mysql.com, and it works quite well. Interesting part is that there isn't really a need to fully integrate mnoGoSearch into tiki - all that needs to be done is a good documentation page explaining how to fully disable the tiki built-in search, and how to get the most out of mnoGoSearch+tiki combination.