Bug 114735

can you attach a screenshot, please?

Created attachment 48574 [details]
screenshot of yast2

I marked the problematic Umlaut in red. 

it seems translation string was in some 8-bit encoding and its trying to  show
it as unicode

Created attachment 48600 [details]
fixed cups_de

fixed corruptions in /usr/share/locale/de/cups_de

Hi Karl, the only place I could find this was in /usr/share/locale/de/cups_de
which is not part of our file set. It looks like it may be generated
automatically but we can't find the strings in our resources. However, we fixed
the corruptions in the file and I've attached it above if this is of use to you.

Thanks for the report and fix!

This bug is not for me.
But I don't know whom to assign it.
/usr/share/locale/de/cups_de belongs to the package cups-client
but is it really a CUPS bug?
Karl, do you know more about what is going on here?

Strange.
I definitely set "ke@suse.de" as info provider
but somehow it was set to Klaus...

The system default encoding is UTF-8.  It might help to make use of UTF-8 for
/usr/share/locale/de/cups_de . 

Fine that our system default encoding is UTF-8, cups default encoding is not UTF-8. :-) 
 
/usr/share/locale/de/cups_de needs to be kept as is, otherwise the webinterface and other 
things will break IMHO. 
 
cups is able to detect which encoding the user has, converts to the appropriate encoding, 
and then responds with the correct charset. 
  
Nevertheless, if it really is latin-1 encoding, then I'm asking myself, why my snapshot is 
displaying broken two-byte characters for every digraph and not one-byte characters? 
Therefore I doubt that this encoding problem is generate by the latin-9 encoding cups. 
Sorry guys, but I strongly think, its the fault of YaST. Assigning back to ke@suse.de  

aj please to clarify what is the default encoding on SUSE Linux.

Default encoding is UTF-8.

If that causes a problem with cups, we need to convert the file somehow for
YaST display.

Klaus, please work with the YaST team on this.

ke: please tell me, at which point of the YaST code you were reading the CUPS   
internal locale settings:   
- which functions do you use for that?  
- how can internal data of CUPS accessed by an external program, like YaST?  
  
Additional question: why is this the only translated text in YaST, whereas  
YaST is using English texts for configuration everywhere else?  
You might notice that the screenshot shows "Media Type", which is not 
translated to "Medium". But the later is also present in the cups_de 
translation. This is correct behavior. But why does YaST use some other locale 
data and not none? Very confusing. 
   
Another question: how does it happen, that a 8bit digraph (in latin9) is   
displayed as a two-byte digraph in YaST?  

I'm not a yast hacker.

I assigned this bug to mzugec@suse.de

According to what I found out, it seems YaST tries to recode what it
got from CUPS to UTF8 in src/modules/Printerdb.ycp: getPpdFileOptions
but I don't understand the special handling of "PageSize" and "Custom"
in this code.

To test it, you must use a PPD file with a "CustomPageSize" option, e.g.:
/usr/share/cups/model/Epson/Stylus_C64-gimp-print.ppd.gz

this translation comes from cups package

Michal: You are the maintainer of the appropriate YaST module. Why don't you just answer  
the open questions regarding YaST, so we can work on it? Instead you deliver NO answer  
and especially no content, but reassign the bug to me. Thats not helpful. Please do. 

The problem is that /usr/share/locale/de_DE/cups_de_DE provided by cups-client
is encoded in ISO-8859-15 instead of UTF-8
When I convert it by iconv it works fine!
I will inform to Ines Pozzo on Monday (for RC2)

fixed