Just as an FYI,
wiki.freeculture.org has mis-encoded UTF-8 for the better
part of the past four years. This is because we used the old Latin 1
schemas.
Now we don't have these problems anymore. I wrote up my notes at
http://wiki.freeculture.org/Fixing_text_encoding_corruption , but here
they are for y'all's convenience:
1. freeze writes to the main wiki
2. Dump freecult_wikidb to dump.sql
3. Create a fresh MW install (just for the table schemas) in
freecult_wikidb2
4. Create a temporary empty DB, and import dump.sql to it
5. In the temporary DB, ALTER TABLE on the text table so it has the
same columns as freecult_wikidb2's text table
6. Dump wikidb3 and have certainty that the column names will line up
(but don't copy the sucky old schema)
* mysqldump --no-create-info --add-locks --complete-insert
freecult_wikidb3 > sql
7. Import that into freecult_wikidb2, skipping the tables that are
missing
* mysql -f freecult_wikidb2 < sql
* WATCH for errors other than "skipping missing table"
8. php maintenance/rebuildall.php
* If this fails with key errors, just drop the recentchanges
table and recreate it with the wikidb2 schema
I've poured enough of my life into this issue, but if someone else wants
to take this up and document it better, by all means go ahead!
-- Asheesh.
--
It is often the case that the man who can't tell a lie thinks he is the best
judge of one.
-- Mark Twain, "Pudd'nhead Wilson's Calendar"