View Single Post
Old 2016-08-21, 17:07   #1107
Serpentine Vermin Jar
Madpoo's Avatar
Jul 2014

1100110011012 Posts

Originally Posted by GP2 View Post
Here's an old thread from 2014 which says that the SQL Server database field is varchar rather than nvarchar, so it can't store Unicode. Unfortunate.

It would still be possible to fix a subset of the names, including "João" to "João".

Or all the accented names could be converted to use HTML &-style entity numbers, like webpages do.

Edit: hmm, that wouldn't work, see M50685343, where the name displays literally as "Vladan Vidaković" rather than "Vladan Vidaković"
There are indeed some oddities when characters fall outside the extended ASCII range. The SQL table for usernames is unfortunately varchar instead of nvarchar.

A fix for that would be to alter the table and convert to nvarchar... I imagine for the most part it would be an easy enough transition since SQL can typically do implicit casts between them and won't complain, but there would be issues in boolean operations where varchar <> nvarchar so we'd have to dig into it more to make sure we're good.
Madpoo is offline   Reply With Quote