[Freeciv-Dev] Re: (PR#7228) freeciv-client-gtk: City name problems (Unicode problem?)

["Vasco Alexandre da Silva Costa" <vasc@xxxxxxxxxxxxxx>]

<URL: http://rt.freeciv.org/Ticket/Display.html?id=7228 >

On Thu, 15 Jan 2004, Jason Short wrote:

> <URL: http://rt.freeciv.org/Ticket/Display.html?id=7228 >
>
> Raimar Falke wrote:
>
> >>The data is stored in the rulesets in latin1, as before.  On loading, it
> >>is converted into the local character set.
> >
> >
> > Where does this happens?
> >
> > $ grep -Irn iconv_open .|fgrep .c
> > ./client/gui-sdl/gui_iconv.c:104:  iconv_t cd = iconv_open(pTocode, 
> > pFromcode);
> > ./client/gui-sdl/gui_iconv.c:193:  cd = iconv_open(pTocode, pFromcode);
> > ./intl/loadmsgcat.c:300:          domain->conv = iconv_open (outcharset, 
> > charset);
>
> S1_14 only.  See PR#1824.  It is only a partial fix designed to avoid
> hideous bugs on single-player systems.  In multi-player it won't work
> very well.

They can connect if they use UTF-8 as well. This is also the proper
UNIX(TM) way of doing things. Programs are supposed to store strings
internally in the locale charset.

This way you can run the server in any locale you want by specifying it
via 'export LANG=' before running the server. Previously the charset was
always borked if you were not using a ISO-8859-1 charset.

Ideally, the network string format and the save game format should use
the same meta-charset (preferably Unicode based, for e.g. UTF-8) for
portability across systems, using the locale charset for presenting the
UI (GUI or text based) to the user. However, this is not trivial to
implement.

---
Vasco Alexandre da Silva Costa @ Instituto Superior Tecnico, Lisboa