[Digikam-devel] 0.8 branch

[Renchi Raju]

On Thu, 23 Jun 2005, Renchi Raju wrote:

> On Thu, 23 Jun 2005, Mikolaj Machowski wrote:
>
>>> Which is pretty weird because we depend on the locale for that afaik.
>> 
>> Sorry, I was too quick to conclusions. Locale seems to be detected
>> properly, in ~/.kde/share/config/digikamrc:
>> [General Settings]
>> Locale=pl_PL    <- which on *nix should automatically means iso-8859-2
>> 
>> Still, entry was broken. I understand currently descriptions are stored
>> in utf-8 (previously in iso-8859-2). It looked like entry was passed to
>> display as utf-8 while screen font is generally in iso-8859-2 encoding.
>
> the locale is no longer stored in digikamrc, but in the database. you can get 
> the current locale using:
> sqlite3 /pathto/digikam3.db "SELECT * FROM Settings WHERE keyword='Locale';"
>
> the utf-8 usage is internal to QString and that shouldn't affect the display. 
> if your font has the glyph for the character, it will render correctly.
>
> i'm trying to figure out where the breakage occured. can you try entering a 
> new comment with non-ascii characters and see if it is rendered correctly? if 
> it doesn't happen there, then the upgrade from the old to new database needs 
> reworking.

i did a few quick tests:
* changed my locale to pl_PL (the iso-8859-2 one, not the utf-8 one)
* added some comments with non-ascii characters using digikam trunk
* wiped out any digikam3.db, so that the old database is upgraded to a new 
sqlite3 db.
* started digikam branch: db gets upgraded to new sqlite3 format.
* the previously added comments look fine
* any newly added comments with non-ascii characters look fine too.

renchi