Hi all!
Why does my Wiki do the following?
I have a page called Linux:Quickies, and somewhere in this page is the string "iptraf". But when I search for iptraf using the wiki search, nothing is found!
Other strings on this page are found very well, even shorter strings like "lib" (I set indexing to 3 letters search in my my.conf).
Any explanations?
- Moritz
I should add:
I already did a php < rebuildtextindex.php
My versions are MediaWiki: 1.4beta5 PHP: 4.3.10-2 (apache2handler) MySQL: 4.0.23_Debian-4-log
- m:o
On Wed, Feb 23, 2005 at 02:13:25PM +0100, Moritz Karbach wrote: # Hi all! # # Why does my Wiki do the following? # # I have a page called Linux:Quickies, and somewhere in this page is the string # "iptraf". But when I search for iptraf using the wiki search, nothing is # found! # # Other strings on this page are found very well, even shorter strings like # "lib" (I set indexing to 3 letters search in my my.conf). # # Any explanations?
These questions have been answered many times. Check the list archives.
Hi,
These questions have been answered many times. Check the list archives.
I'm sorry, I couldn't find anything similar to my problem in the archive, since - my query doesn't contain logical expressions - my query is longer than 3 letters - I've got no problems with fuzzy search
But maybe I've overseen something...
- Moritz
Hi again,
I just figured out some more about this behavior. My Linux page contained
--- snip --- ; source code patchen : patch < patch.diff : patch -p 0 < ../patch.diff : -p 0 ist wichtig, falls der patch nicht im selben Verzeichnis liegt wie die zu patchende Datei. Zur Not einfach andere Zahlen als 0 ausprobieren... --- snap ---
and many lines later
--- snip --- ; proftp aus dem Autostart nehmen <pre> uhland01:/etc# rm rc0.d/K50proftpd uhland01:/etc# rm rc1.d/K50proftpd uhland01:/etc# rm rc2.d/S50proftpd uhland01:/etc# rm rc3.d/S50proftpd uhland01:/etc# rm rc4.d/S50proftpd uhland01:/etc# rm rc5.d/S50proftpd uhland01:/etc# rm rc6.d/K50proftpd uhland01:/etc# init.d/proftpd stop </pre> --- snap ---
But in the searchindex table appeared only
--- snip --- source code patchen patch uhland01 etc# rm rc0 d k50proftpd uhland01 etc# rm rc1 d k50proftpd uhland01 etc# rm rc2 d s50proftpd uhland01 etc# rm rc3 d s50proftpd uhland01 etc# rm rc4 d s50proftpd uhland01 etc# rm rc5 d s50proftpd uhland01 etc# rm rc6 d k50proftpd uhland01 etc# init d proftpd stop ---- ; --- snap ---
That is, anything between ": patch <" and "<pre>uhland01" is missing! And in this missing area there was my "iptraf" search query...
I would consider this as a bug, but maybe it's a feature?
- Moritz
Moritz Karbach wrote: [snip]
That is, anything between ": patch <" and "<pre>uhland01" is missing! And in this missing area there was my "iptraf" search query...
I would consider this as a bug, but maybe it's a feature?
The search index strips HTML tags; apparently this isn't too careful and is removing a lot of material from your page. File a bug report at http://bugzilla.wikimedia.org, and if you want to get a head start take a look in SearchUpdate.php for the bit that does this.
-- brion vibber (brion @ pobox.com)
mediawiki-l@lists.wikimedia.org