[Mediawiki-l] Special Characters from database.

[Rob Church]

On 24/07/07, Adam Meyer <meyer7 at mindspring.com> wrote:
> I have an external program for my wiki that grabs mwt.old_text and
> puts it in a text box. Any time there are special characters (E.G. (r)
> (c) º (registered, copyright, degrees)) a capital A with an accent mark
> shows up before it. Im sure that this is just the way that it is
> store in the database, or an unicode issue, but how does mediawiki
> pull it with out this?

Some quick background; MediaWiki will, under the defaults, stick two
fingers up at the collation of the database and insert UTF-8 data into
a latin1 table, performing cleanup and normalisations and whatever
other Unicode goodies are required at a later stage. This is for
historical reasons, due to MySQL 4.0.x's lack of UTF-8 support, and
the poor UTF-8 support in later versions. Brion can probably give you
some more waffle about this.

I'd advise not accessing the table directly, since the old_text column
may not contain text; depending upon your configuration, it could
contain a pointer to an external store, or a pointer to another row,
or some compressed object, or compressed text. If at all possible, try
to use the MediaWiki Revision class to get at this text.


Rob Church