[gcompris-devel] GCompris still not ready for 1.0

[Bruno Coudoin]

le lun 25-02-2002 à 18:14, Mantas K. a écrit :
> Hi,
> 
> I think Gcompris still not ready for 1.0 version. There is why:
> 
> 1. I want to ask (for the second time) why the lithuanian words (for 
> wordsgame), I've sent Bruno aren't in CVS ? I'm attaching them to this 
> letter again.

Sorry for that. It's my fault. Your files are commited now.
There is now 9 languages supported in wordsgame.

> 
> 2. There are no instructions of how to compile CVS version of gcompris. 
> There are no config.sub, config.guess and other files. I tried to run 
> ./automake.sh and later make, but got error message:
> [root at mantas gcompris]# make
> Makefile:505: *** missing separator.  Stop.
> 
> In 505 line of Makefile I saw:
> 
> @INTLTOOL_DESKTOP_RULE@
> 
This is because gcompris now relies on intltool. You have to DL this
package from ximian.
I changed the INSTALL files to report that it is needed to compile.

Please tell us if this is enough to fix your problem.
I don't believe the file you mention are needed.

> 3. Is gcompris really internationalized now ? I mean other (non 
> iso-8859-1) encodings. Pre2 and pre3 still display 2 strange symbols 
> instead 1 non ascii lithuanian letter. Same is with russian and others 
> languages, that use non ISO-8859-1 encoding, I think.
> If gcompris uses utf8 in translated xml files, then fonts must be used 
> in utf8 too. I don't know if gnome 1.4 can use utf8 fonts.
> About the other possibility I talked in my earlied letters:
> We must have to change the last key '-u' to '-p' in 'aclocal.m4' and/or 
> 'configure' files string $(INTLTOOL_MERGE) $(top_srcdir)/po $< > After 
> doing it the strings won't be converted to UTF8 and we can use fonts 
> native encodings (ex. ISO-8859-13 for lithuanian, KOIR-8 for russian, etc.)
> Bruno, you haven't understood me very good, when saying "I can't 
> hardcode specific fonts depending on the language."
> I didn't mean harcoding, this was only an example. I think it would be 
> good if the change of locale setting in configuration board could change 
> font encoding too. This isn't hard to program: simplest way is to use 
> global variable for encoding and change this variable, when changing 
> locale setting in configuration board. Then in gdk_font_load function we 
> have to use encoding variable:
> gdk_font = gdk_font_load 
> (strcpy("-adobe-times-medium-r-normal--*-180-*-*-*-*-",encoding_variable))
> or maybe
> gdk_font = gdk_font_load 
> (strcpy("-adobe-times-medium-r-normal--*-180-*-*-*-*-",(char 
> *)encoding_variable))
> 
> I don't remember C string functions exactly :(, but think you understand 
> me. This works for me, when I tested. If gnome 1.4 can't use utf8 fonts, 
> then we have to do in this way.
> 
> Bruno Coudoin wrote:
>  >I propose to release a Pre4 monday. Hopefully, it will be the final 1.0 ;)
> ----
> 

Mantas, please understand that we are willing to support any language on
earth we can support. Now that said, we worked hard with Pascal to
implement and make gcompris works the way it is related to fonts locale
display.
The way it works now is that xml are full UTF-8 but when loaded in
gcompris they are converted to isolatin1 (see gameutil.c).
Now I believe that Russian and Lituanian are not isolatin1 so this won't
work. Doing it the way you propose is a lot of work and probably will
probably become useless in new gnome version with Pango.
I personnaly think it is better to wait than to hack something I don't
clearly understand. I would like to avoid to hardcode which locale is
which iso font to load.

What you could do is investigate and tell us how it is managed in Gnome2
and see what are the implications in our case.
Having a good vision in this area will let us take the good decision.

Bruno.