Re: [Adonthell-devel] First 0.3 Windows binaries

[Alexandre Courbot]

> Actually, I think the changes you made to nls.cc weren't neccessary.
> When ENABLE_NLS isn't defined, all the gettext related functions are
> defined to a noop in gettext.h. That should already suffice.

I fear it will be necessary when I'll compile with nls support. mingw32
doesn't have definitions for ONE language, which is not that bad :)

> > I had to make some changes however in order for the Python scripts to
> > work:
> 
> Uh, can it be that you forgot to commit some changes to the data file?
> Right now, dialogues don't work at all:
> 
>     *** Error loading dialogue script dialogues/oliver_start
>     AttributeError: 'dialogues/oliver_start' module has no attribute
> 'dialogues/oliver_start'

Uh? No, it's fine for me, and my CVS is up-to-date. This problem came
from character.data, which container dialogues/oliver_start like
entries, and I changed them to dialogues.oliver_start. Just copy the
file and gunzip it to check. Be carefull too, that saved games store
this information too. So if you loaded a game, you'll end up with the
old '/' style.

> And 'make install' seems to be broken as well.

Ah, maybe. I haven't tested, but I updated Makefile.ams and configure.in
carefully. I'll have a look.

> So do I. I made a quick test with win2k, but there were problems, as I
> said in the other mail. I also tried running with the -g option, just in
> case, but the result was the same. The window popped up for a sec,
> nothing else. stderr.txt contains the following

Hum, I've never tried with -g, and I fear that setting the base
directory to '.' doesn't work under windows (have tested that way
first). Don't know why. I just hope we don't have to stuck to a
hard-coded base directory.

>     ImportError: No module named schedules.control
>     ImportError: No module named init
> 
> but the files are definately there.

Yeah, that's what I've noticed too.

Alex.
-- 
http://www.gnurou.org