Re: [PATCH] Set character set for libc.pot (preparing for 2.16 updatefor translation team).

[Carlos O'Donell <carlos at systemhalted dot org>]

On 6/13/2012 5:51 PM, Petr Baudis wrote:
> On Wed, Jun 13, 2012 at 04:53:35PM -0400, Carlos O'Donell wrote:
>> In preparing for the 2.16 release I need to regenerate the libc.pot file.
>>
>> The first bug I'd like to tackle is that we don't set the character set 
>> for the pot file. This results in a warning:
>>
>> ~~~
>> po/libc.pot.new: warning: Charset "CHARSET" is not a portable encoding name.
>>                           Message conversion to user's charset might not work.
>> ~~~
>>
>> We can either set it in the po/header.pot file or in the Makefile.
>>
>> I would prefer to set it in the Makefile so we don't loose the setting
>> if we regenerate the header from scratch.
>>
>> Is UTF-8 the correct value here?
>>
>> It's probably overkill, but that's fine.
>>
>> Comments?
> 
> AIUI, it just serves as default charset for new translations?
> In that case, I think UTF-8 is definitely the best idea.

My reading of the gettext documentation is that it's going to
serve as the charset for the template, and as the default
charset for the new translation. We should get it right, and I
think that UTF-8 is sufficiently right that it will work with
whatever strings we decide to localize.

I've checked this patch in to remove the warning and indicate
the charset for the pot file I'm going to send to the
translation team.

Cheers,
Carlos.