Re: msgmerge problem

["Fernando Apesteguía" <fernando apesteguia gmail com>]

Sorry, but just to be clear:

If one of my translators send me a file specifying that he used an  ISO-XXX-X encoding, what should I do? Convert to UTF-8 and set charset=UTF-8 or keep the file intact and set charset=ISO-XXX-X?

What is the preferred behaviour for a GNOME application?

And another question so I don't send two mails this night :)
How does gtk and glib functions behave with enconding changes? Will be string sizes affected? Should I resize my buffers? I ask this because I make a extensive use of strings and buffers to perform some parsing operations.

Thanks in advance

On 6/19/06, Danilo Segan <danilo gnome org> wrote:

Hi Fernando,

У пон, 19. 06 2006. у 18:30 +0200, Fernando Apesteguía пише:

> You have made me really happy! :)
> That worked. Just a follow up question: The enconding conversion
> changed some characters ( e.g. the last vocal now appers to be a "c"
> instead of a extrange "e") is this normal?

Well, Croatian uses only č, ć, đ, and ž (these are not accented c, d or
z, but different letters), so it was easy for me to notice the problem
(I am Serbian, and Serbian is quite alike, if not the same as Croatian).

And I know that Croatian on free software systems would probably be
using ISO-8859-2 (other than UTF-8; this is also known as Latin-2).  On
Windows, it would use CP1250.

> And what to do if I face the same problem again? Should I change the
> encoding in the file's header or should I reencode with iconv? What is
> the difference?

You *must* (and I repeat, *must*) ask the translator who submitted a
translation in which encoding was it in — file will likely pass check if
you convert from any 8-bit ISO-8859 encoding (either using iconv of
header field), since they all cover almost same 8-bit space.

I felt safe in recommending ISO-8859-2 only because I am very well
familiar with what would Croatian need. :)

Cheers,
Danilo