[GRASS-translations] once again file encoding

Glynn Clements glynn at gclements.plus.com
Tue Jul 22 15:58:52 EDT 2008

Robert Nuske wrote:

> there was a discussion on the german grass list about weird display of german 
> umlaute in some strings. 
> We found out that the two "old" po-files (grassmods & grasslibs) are using the 
> iso-encoding and the two "newer" ones (grasstcl & grasswxpy) use utf-8. The 
> new po-files seem to produce "good looking" umlaute and the old ones mixed up 
> stuff.
> I guess we should do something, probably changing the encoding of two files. 
> But which encoding shall we keep?

Please use ISO-8859-* where possible. Unless you actually need the
Euro sign, you should probably use ISO-8859-1.

On systems using a modern version of GNU gettext, the encoding won't
matter; gettext will perform any recoding which is necessary. Systems
which don't perform recoding probably won't be using UTF-8 for the
locale's encoding (or for anything, really).

AFAICT, the specific problem here is likely to be that
locale/po/grassmods_de.po claims to be in ISO-8859-15, but it's
actually in UTF-8.

This change occurred in:

	r24806 | carlos | 2007-09-21 23:48:53 +0100 (Fri, 21 Sep 2007) | 2 lines
	Changed paths:
	   M /grass/trunk/locale/po/grassmods_de.po

	Robert Nuske translations and updated

Except, it's not quite UTF-8. There are some encoding errors, all
relating to the a-umlaut character. Some of them have a line break in
the middle of the two bytes which make up the character, while a few
others are encoded in ISO-8859-1/-15 rather than in UTF-8.

I have fixed the encoding errors (r32214 in SVN trunk), and converted
it back to ISO-8859-15 (r32215 in SVN trunk).

Glynn Clements <glynn at gclements.plus.com>

More information about the grass-translations mailing list