[Adonthell-devel] Dlgedit i18n

[Kai Sterker]

After spending the morning with the gettext docs, I'm pretty confident
that i18n isn't a big problem. It certainly won't require huge changes.
Here are the details:

[*] The text will remain in the dialogue file, as it does now. So far I
thought it would be cleaner to seperate it, so it could be saved more
efficiently and loaded as neccessary, but that wouldn't work with
gettext. Of course you could argue that you wouldn't need gettext at
all, if the text was seperated from the code. You could easily create
different language versions of that text file. But it's better to stick
to some standart instead of doing your own things. Besides, other python
modules containing text (like the schedules) would need to be translated
as well, and here we need gettext.

[*] In the dialogue source, all text will be marked translateble. That
happens automatically and requires no special action from the person
creating the dialogue. With the 'pygettext' script, the strings can be
extracted into a .pot file that in turn serves as a template for the
localized text catalogues.

[*] The actual translation will happen on C++ side. As a dialogue will
contain plenty of strings, it would be too time consuming to translate
all of them when the script is loaded. (Especially since not all of the
strings will be needed.) Instead, the translation occurs when a string
is about to be displayed.

One question that remains is whether to put the text of all dialogues
into a single catalog (.po file), or use one per dialogue. I guess for a
start, using one .po file for everything is okay.

So after lunch I'll start coding that stuff, which will certainly take a
day or two. Feel free to post any comments or improvements regarding the
above. I guess once I actually dive into the code, some more issues will
turn up, but for now that should cover the most important things.

Kai