Paul Coghlan a écrit:
I am struggling somewhat with some Canadian place names that include extended characters.
I create a page through the standard MW UI and called it Saint-Étienne but it arrives into the table (collated utf8-bin) with Saint-Étienne as the page_title?!
This in turn is not accessible using the page_title to build the URL.
Can anyone tell me where I went wrong? I am hoping to find a way to prevent this conversion of the characters.
Many thanks, Paul
It's not wrong. The utf8 characters are stored utf8 encoded in the binary latin1 mysql table. So mysql prints 'Saint-Étienne' (latin1) but mediawiki knows that the field is utf8 ecoded and correctly uses 'Saint-Étienne'
The only problem you may have with this is at backup time, where mysqldump may "convert" it, breaking what it thought was a latin1 field (see http://www.mediawiki.org/wiki/Manual:Backing_up_a_wiki ).