Re: utf-16 and glib (was: g_malloc overhead)

[Maciej Piechotka <uzytkownik2 gmail com>]

On Mon, 2009-01-26 at 22:49 +0100, Martin (OPENGeoMap) wrote:
> Maciej Piechotka escribiÃ:
> > On Mon, 2009-01-26 at 22:30 +0100, Martin (OPENGeoMap) wrote:
> >   
> > > hi:
> > >     
> > > > >     
> > > > >         
> > > > Well - what do you mean? Having 2 functions - one reciving utf-16 and
> > > > one utf-8? To be honest - it doesn't make any sense to me (it would
> > > > create much mess, double the code, make programming errors easier...).
> > > >
> > > > Converting? What's wrong with g_utf16_to_utf8?
> > > >   
> > > >       
> > > I was talking about a full utf16 and utf8 api in glib and use a macro to 
> > > work work intermediate string:
> > >
> > > For example in windows they have this types:
> > > LPSTR =char *
> > >     
> >
> > char * is used for utf-8 AFAIR
> >
> >   
> > > LPWSTR= utf16windowschar *
> > >
> > >     
> >
> > gunichar2
> >
> >   
> > > perhaps in glib we could have utf16 and utf8 in that way or am i wrong?
> > >
> > >     
> >
> > I'm not glib developer. As far as the module of operating on utf-16
> > strings is proposed I'm not against. However I would prefere to not have
> > 2 entries to each function.
> >   
>
> Hi:
>
> What is wrong with:
> gchar*  g_utf8_strncpy  (gchar *dest,const gchar *src,gsize n);

That's one not needed as strncpy should work.

> gunichar2 *  g_utf16_strncpy  (gunichar2*dest,const gunichar2*src,gsize n);

That's kind of support I'm not against.

> and the macro:
> gtext*  g_text_strncpy  (gtext*dest,const gtext*src,gsize n);
>
>
> regards.

With the entries - nothing. With macro - it may be just me but I percive
it shooting into foot. Just imagine that some header will assume gtext
to be utf-8. Other will turn on the macro (or user code) and change it
to utf-16. IMHO - having magic switch which might change the ABI is not
good.

Regards