[mythtv-users] error when fixing database forced by mythtv upgrade

[Petr Stehlik]

Michael T. Dean píše v Ne 27. 09. 2009 v 15:48 -0400:
> > |   2871 | Samuel Ivaška (Slovenská republika 1983)                        |   
> 
> If that's what you're seeing from the mysql command-line client

yes that's it.

> queried against your 0.21-fixes database, it proves that every single 
> one of those lines is corrupt.  The "š", the "í", and the "á" characters 
> should not show up properly in the mysql client unless the column or 
> connection was actually set up as a UTF-8 column or connection at one 
> time (when the data was inserted).  They should show up as 2 
> characters.  ("š" = "Å¡", "í" = "í­­­" (which is 2 chars, but one isn't 

Yes, I understand the UTF-8 encoding.

> visible), and "á" = "á")  You'll note that in your query after 
> restoring the "uncorrupted" backup, they're showing the proper characters.

well you call it "proper" but it's the old Myth's utf-8 hidden in latin1
"encoding". Right now I got it "corrupted" in the way that it's actually
the right utf8 encoding (judging from the mysql client's SELECT output
above). I am wondering if I really need to go through "fixing" the
database by first re-encoding it back to latin1 and then letting Myth to
convert it back to utf8 (where I already am now, IIUIC).

> (You never followed the instructions that used to be on the wiki that 
> changed the database schema, converting some columns to utf8, to try to 
> "fix" the OSD data, etc., did you?

nope. Never had an issue with OSD data.

> > well, I have upgraded Etch to Lenny recently... Does that count as
> > moving between distros?

> All that matters is how the database server was configured--if it was 
> ever misconfigured at any time during your use of MythTV, data during 
> that time was corrupted.

I understand that but the data seems consistent at a first glance over
the SQL dump. I could run it through recode to see if it's all proper
utf8...

Petr