On Mon, 14 Jul 2003 16:17+0100, Gaute B Strokkenes wrote:
On 14 jul 2003, trond@ramstind.gtf.ol.no wrote:
On Sun, 13 Jul 2003 18:44-0400, Karl Berry wrote:
I've received a bug report (thanks, Gaute and Josip) that the Norwegian translation included in the Texinfo 4.6 release contains a spurious byte with hex value 0x81 before every non-ASCII character. Can you look into this please?
I have noticed this behaviour before. I have usually removed all occurences of this byte before submitting a new no.po file. Since this bug keeps returning, I don't think it's an error on my part, then again I could be wrong.
Do you use emacs? Having these odd bytes pop up in annoying places is a common bug with certain versions.
I use GNU Emacs 20.3.1 (I know my Emacs is old, but that's the one I have) and po-mode 2.01 from GNU gettext 0.12.1.
I should clarify that I have never had any problems with the translation file after editing it with my version of Emacs, i.e. no spurious bytes.
As far as I can tell, the updated no.po file I mailed Karl the other day, has no spurious bytes whatsoever. If the spurious bytes are introduced again, I suspect the problem is somewhere else along the chain.
However, almost every time I have downloaded a new (pretest) version of GNU Texinfo, there has been spurious bytes in the no.po file.
Could this be caused by some misbehaving conversion between ISO-8859-1 and UTF-8 and back again?
(God, how I hate UTF-8! It's a bad compromise between ISO-8859-X (or is it ISO-646?) and plain Unicode, but that's my view of the subject.)