[Mediawiki-l] Character encoding problem (when moving to another mysql db)

Kai Schlamp schlamp at gmx.de
Sat Sep 8 02:00:27 UTC 2007


Hello.

I have tried for several hours to move my mediawiki to another machine
resp. the mediawiki data to another MySQL DB.

server_old:
	5.0.32-Debian_7etch1-log
	Mediawiki 1.6.10
	MySQL charset:  UTF-8 Unicode (utf8)
	MySQL connection collation: utf8_general_ci

server_new:
	MySQL 5.0.38-Ubuntu_0ubuntu1-log
	rest same as above

When I look at the data directly in MySQL my german special chars look a
bit cryptic (like Ãœ for Ü).
But this seems to be ok, cause Mediawiki stores
everything in UTF8 format in the db not caring what MySQL uses (as I
read here somewhere).
Also Mediawiki on server_old has no problem with this at all.

But the problems begin, when I get a dump through phpmyadmin export
(can't do a mysqldumb on console on server_old, my old provider does not
permit this).
When I look at the dumped SQL file, everythings still seems to be ok.
Same cryptic chars for my german special characters as in the server_old db.
But when I import the SQL dump to the new database, the cryptic chars
change (for example the Ãœ for Ü is now Ö) and the special chars are
now corrupt when using the new Mediawiki.

file -i dump.sql gives me:
text/x-c; charset=utf-8

The only thing I can think of, is that by importing the file back to the
new db, the content of the dump file is encoded again.
And that's why those two cryptic chars, are now four.
Ü (1) -> Ãœ (2) -> Ö (4)

Any thoughts?

Best regards,
Kai






More information about the MediaWiki-l mailing list