[mythtv] [mythtv-commits] Ticket #6265: Mythvideo removed all (remote)files with umlauts in filename and don't readd them

[Janne Grunau]

On Fri, Aug 14, 2009 at 09:46:08AM +0200, Markus Schulz wrote:
> Am Wednesday 12 August 2009 schrieb MythTV:
> > #6265: Mythvideo removed all (remote)files with umlauts in filename
> > and don't readd them
> >
> > Comment:
> >
> >  This appears to be a non-bug setup (locale or mySQL) issue.  Closing
> > this for now, but I encourage discussion on the dev list so that we
> > can determine what the cause is so that we can troubleshoot it for
> > people. For reference, this (including all above examples) works
> > properly here.

With an UTF-8 locale?

> Sorry, but still not working:
> 
> 
> $ echo $LANG
> de_DE.UTF-8
> 
> $ mv 'Haende Wie Samt.avi' 'Hände Wie Samt.avi'
> 
> 2009-08-14 09:28:36.748 buildFileList directory = /data/dvd
> 2009-08-14 09:28:36.750 MythVideo::ScanVideoDirectory Scanning 
> (/data/dvd)
> 2009-08-14 09:28:36.751 buildFileList directory = /data/upload
> 2009-08-14 09:28:36.751 MythVideo::ScanVideoDirectory Scanning 
> (/data/upload)
> 2009-08-14 09:28:36.766 buildFileList directory = 
> myth://Videos@mythtv/data/media/
> 2009-08-14 09:28:36.766 MythVideo::ScanVideoDirectory Scanning Group 
> (myth://Videos@mythtv/data/media/)
> 2009-08-14 09:28:37.158 Removing file SG(mythtv) :filme/Adriano 
> Celentano/Haende Wie Samt.avi:
> 
> 
> The old file was removed but the new not added.

I suspect filename handling issue, shouldn't have anything to do with DB
charsets.


> $ awk '/^CREATE TABLE/ { NAME=$3 } /ENGINE/ { print NAME"->"$0 }'  
> mythconvergdump.sql                     
> `mythweb_sessions`->) ENGINE=MyISAM DEFAULT CHARSET=latin1;

mythweb table, missed from mythtv and plugins conversions

> `newssites`->) ENGINE=MyISAM DEFAULT CHARSET=latin1;

mythnews, apparently missed

> `schemalock`->) ENGINE=MyISAM DEFAULT CHARSET=latin1;

can't be converted from dbcheck.cpp since the table is locked during the
update. doesn't matter since it never would contain user or other UTF-8
data.

> `streams`->) ENGINE=MyISAM DEFAULT CHARSET=latin1;

doesn't seem to be used in mythtv and plugins

> Additionally there are some columns with latin1 encoding set:
> 
> $ awk '/^CREATE TABLE/ { NAME=$3 } /latin1/ { print NAME"->"$0 }'  
> mythconvergdump.sql | grep -v ENGINE
> 
> `archiveitems`->  `type` set('Recording','Video','File') character set 
> latin1 default NULL,

all the sets where probably missed since I didn't thought of them as
string type. It's not really a problem since they don't contain user
data and all existing contain only 7bit ASCII chars.

Janne