pkellum Wrote:I thought XBMC supported Asian character sets? I'll try it tonight and see what happens. If that is the issue though, it needs to be corrected in XBMC's database code. I didn't think anyone actually used latin1 anymore.
XBMC does support utf8, but I remember I had a problem involving indexes creation when I was trying to migrate to MySQL, and it was caused by the collation of the database being utf8, I switched to latin1 and the problem was solved.
Here are the technical details of the problem: MySQL has a limitation that allows a max byte size of 767 for any fields that are part of an index, MySQL uses 3 byte per char if the collation of the db is set to utf8, and only 1 byte per char if the collation is set to something different than utf8. The table
files has a field
strFilename which type is varchar(512) and there is an index
ix_files that uses that field. If you define the db collation as utf8, that index creation will return an error, because 512x3= 1536 bytes which exceeds MySQL size restriction for indexes, but if you choose a different collation for your db, the size in bytes of the field would be 512x1= 512 bytes which doesn't exceed MySQL size restriction and that's why I suggested you to change the db collation to try and avoid the error.
When I discovered this issue some time ago, I posted the problem in the
thread opened by the developer of the MySQL for XBMC feature
firnsy and I suggested to change the db creation code in XBMC to solve the problem or as an alternative solution to stop creating the problematic index, but as the XBMC db is going through a redesign at present time the decission was to wait for the redesign first and then evaluate which way to go regarding this issue.
I hope that the explanation is clear enough.