Re: multilanguage environment

[Björn Gohla]

On Sunday 21 April 2002 20:31, Alexander Malmberg wrote:
> > i am having big trouble working with non latin scripts in gnustep, is
> > there nobody out there working with languages that use non-latin scripts?
>
> I don't really work much with non-latin scripts, but I do a lot of
> testing with them, and it does work.
>
> > does
> > gnustep not use unicode to encode strings internally?
>
> It does. (Sortof; strings might be stored differently for efficiency,
> but externally everything will behave like unicode strings.)
>
> > i do have iso-10646-1
> > fonts installed on my system, why can gnustep not use them?
>
> In what way does it not work? Do you get '*' instead of non-latin
> characters, or garbage, or nothing at all?

sometimes it shows asterisks sometimes garbage (at least when i tried 
japanese text) depending on what GNUSTEP_STRING_ENCODING i set.


> Problems are probably because of the xgps backend. You might want to add
> debug logging to setupAttributes in XGFont.m:232 and XftFontInfo.m:310
> to check the value of mostCompatibleEncoding and work from there (X
> might not be providing the correct attributes, or xgps might be using
> eg. *-foo-bar-iso8859-1 instead *-foo-bar-iso10646-1). Using the Xft
> anti-aliased fonts stuff might help.

anti-aliasing does not work with my X server, i can not check the version 
right now, but it is probably too old and i would rather not have to mess 
with it.


> Plug :) : my freetype/libart-based backend handles unicode properly, so
> if your iso-10646-1 fonts are truetype fonts that would be an option
> (type1 might work too, though I haven't tested that; possible freetype
> needs to be told that it needs to load the .pfm/.afm files as well).