On 03/05/07, johan.boye@latecoere.fr johan.boye@latecoere.fr wrote:
# The search doesn't work on partial (incomplete) word. Got an article with the word "foobar", then I have a result when I type "foobar" but nothing when I type "foob". I have some exception, when I type "est"; it founds some pages with the word "test"... Is there a way to fix it?
The default search relies upon the MyISAM full text indexing engine, and may not be entirely useful, depending on your needs. You might have more luck with a custom search backend, such as the LuceneSearch extension, hooked up to the MWSearch daemon in Subversion, or something else entirely.
# Links (usually nasty file:// links) does not work with URL containing spaces
You need to URL-encode spaces in URLs. Use %20.
# Is there a plug-in to scan documents like .PDF or .DOC on the fly during the upload to index them in the search engine?
Yes and no; see http://lists.wikimedia.org/pipermail/mediawiki-l/2007-April/019490.html.
Is there a way to make a full backup of the website to make in available offline or export all pages to HTML in a save place on an automatic way?
You can write a shell script to do it. The procedure is documented at http://www.mediawiki.org/wiki/Manual:Backing_up_a_wiki.
Rob Church