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

Re: Bug#412312: iconv fails when converting Catalan "ela geminada" character

Pierre Habouzit a écrit :
> On Tue, Feb 27, 2007 at 10:35:31PM +0100, Robert Millan wrote:
>>> No it does not, iconv is meant as a 1:1 converter. If you want more
>>> subtle and evolved conversions, please use recode.
>>>   ŀ and Ŀ do not exist in latin1/9 so it's valid not to be able to
>>> convert them.
>>>   Please do consider that echo $foo | iconv -f utf8 -t latin1 | iconv -f latin1 -t utf8 
>>> should be identity (if every character in $foo exists in utf8). With
>>> your request, it's no longer true, as converting e.g. ŀ back and forth
>>> would give you l· in utf-8 as well, which is incorrect.
>> Hi Pierre,
>> My real intention was gaining the ability to use ŀ and Ŀ in gettext
>> translations, with the certainty that gettext will manage to convert them for
>> people still using latin1.
>> I was under the impression that iconv relied on the same backend that gettext
>> does, and so that fixing iconv would automaticaly fix gettext.  But as you put
>> it, sounds like that's not the case.
>> Do you know what is needed to archieve what we want?
>   I think you can pass //TRANSLITERATE or sth like that to do what you
> want (I think). I'm not sure it's iconv that provides that though.

If I remember correctly, you have to add //TRANSLIT to the destination
code. But you should really use recode for that which has be designed
for that.

  .''`.  Aurelien Jarno	            | GPG: 1024D/F1BCDB73
 : :' :  Debian developer           | Electrical Engineer
 `. `'   aurel32@debian.org         | aurelien@aurel32.net
   `-    people.debian.org/~aurel32 | www.aurel32.net

Reply to: