Re: PO files on status pages transmitted with incorrect or nocharset information



On Tue, 2003-10-28 at 13:14, Carlos Perelló Marín wrote:
> El mar, 28-10-2003 a las 14:05, Åsmund Skjæveland escribió:
[...]
> > I Am Not An Expert©, but isn't it possible to tell Apache to send po
> > files with a HTTP header specifying UTF-8 as the encoding?
> 
> All .po files are not always UTF-8 files, that's the problem.

This is sort of related to Åsmund's original question. 

There are some UTF-8 .po files in CVS which have become mangled by being
interpreted as latin-1.

For example, if I take the name "Åsmund", save it in a file, then run
'recode latin1..utf8 asmund-test', the result is "Ã…smund". This can be
done accidentally by opening and saving a UTF-8 encoded file while using
a non-UTF-8 locale. There are (unfortunately) lots of examples of this
problem in the Irish translations.

Can anyone think of a good way to avoid this or fix it for all files?
Part of the problem is that the resulting mangled files are technically
valid in latin-1. Perhaps something could be done based on the
"Content-Type" and "Content-Transfer-Encoding" headers of the .po files.

Regards,

David

¹ See gnome-desktop.HEAD.ga.po in the 2.4 status pages, for example.

This is a digitally signed message part



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]