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

[Bug#223024: po2debconf spits up all over my ca.po file, blames iconv]



      Bon ditardanit!

      Com veeu, sembla que po2debconf troba alguns problemes en
  convertir els caràcters ela geminada d'UTF-8 (0x13F i 0x140) a «l·» en
  ISO-8859-1 i ISO-8859-15.  Què en fem al respecte?  Sembla que ara per
  ara no hi ha més remei que usar «l·» també en UTF-8...  Algú coneix
  alguna forma de remeiar-ho?  Jo crec que és un error que el
  funcionament d'«iconv» depenga del locale carregat?  N'enviem un
  informe?

----- Forwarded message from Denis Barbier <barbier@linuxfr.org> -----

X-Sieve: CMU Sieve 2.2
X-Resolved-to: degauss@fastmail.fm
X-Delivered-to: degauss@fastmail.fm
X-Mail-from: barbier@linuxfr.org
To: Ivan Vilata i Balaguer <ivan@selidor.net>,
	Branden Robinson <branden@debian.org>, 223024@bugs.debian.org
Subject: Re: Bug#223024: po-debconf: po2debconf spits up all over my ca.po file, blames iconv
From: barbier@linuxfr.org (Denis Barbier)

On Sat, Dec 06, 2003 at 01:27:04PM +0100, Ivan Vilata i Balaguer wrote:
> On Fri, Dec 05, 2003 at 11:07:22PM +0100, Denis Barbier wrote:
> > [...]
> > debian/po/ca.po is UTF-8 encoded. po2debconf tries to convert it to
> > iso-8859-15, but fails because l.48 contains a character Unicode 0x140:
> > LATIN SMALL LETTER L WITH MIDDLE DOT
> > which does not appear in iso-8859-15.
> > [...]
> 
>       In fact, I think the right thing would be that iconv translated
>   automatically the 0x140 Unicode character into iso-8859-{1,15} «l·»
>   when recoding.  The same would go for 0x13F, which is the capital
>   letter, into «L·».  This patch looks like a temporary hack to
>   circumvent iconv's limitation.  Should a bug be filed against iconv?
> 
>   (Unicode characters 0x13F and 0x140 are both _valid_ Catalan
>   characters, and the right ones for the letters they represent, IMO.)

Thanks for your explanations.
po2debconf does not run iconv with adding //TRANSLIT to encoding
output, which is why no transliteration is performed.
The reason is that
  LANG=ca_ES@euro iconv -f utf-8 -t iso-8859-15//TRANSLIT ca.po
does recode 0x140 into «l·», but
  LANG=C iconv -f utf-8 -t iso-8859-15//TRANSLIT ca.po
replaces it by a question mark, which is very bad.
So transliteration only works when the right locale is installed.

After sarge, templates will be UTF-8 encoded, so this problem
won't exist anymore.  for sarge, I suggest that you do not use
Unicode characters 0x13F and 0x140.  Could you discuss this
issue on debian-l10n-catalan?  

Denis

----- End forwarded message -----

-- 
Ivan Vilata i Balaguer  	@      FIGHT AGAINST SOFTWARE PATENTS!     @
http://www.selidor.net/ 	@ http://EuropeSwPatentFree.hispalinux.es/ @

Attachment: pgpxoSeAVxfA_.pgp
Description: PGP signature


Reply to: