From: Johan De Meersman Date: August 28 2012 8:49am Subject: Re: MySQL, UTF8 and collations List-Archive: http://lists.mysql.com/mysql/228092 Message-Id: <20868749.3521.1346143768826.JavaMail.root@tuxera.be> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable ----- Original Message ----- > From: "Rick James" >=20 > There is a section on German Sharp-s in http://mysql.rjweb.org/doc.php/ch= arcoll > I agree with "dirty". Yes, seen it, but thank you. I'm not having character set issues, everythin= g is nicely UTF8. I'm merely running into the schei=C3=9Fe bug, but while u= tf8_unicode_ci does properly handle that, it stil barfs on c=C5=93r (that's= oe in one character) and I'm wondering why. > As I understand it, about 5.1.24, they said "Oops, sharp-s is > collated wrong, let's fix it". The fix broke things, but they stuck > by the "correct" sorting. Eventually they said "Oh, let's add > another, compatible, collation". >=20 > They were caught between a rock and a hard place. The hard place being having to go "oops, sorry we broke things, we reverted= the change and will introduce a new default collation in the next version"= ? Yes, poor software vendors, why do people always expect them to admit the= y made a mistake in judgement. And, of course, "alter table foo collate utf8_general_mysql500_ci" doesn't = do what you hope, either. It changes the table's default collation nicely, = and then explicitly adds the *old* collation to all the rows you were hopin= g to have fixed. Guess I'll be fixing it manually (well, sed is my friend) in a mysqldump be= fore syncing up the second node after it's been upgraded. --=20 Linux Bier Wanderung 2012, now also available in Belgium! August, 12 to 19, Diksmuide, Belgium - http://lbw2012.tuxera.be