> I am mobin(Tuxtype Indic team). Actually we added SDL_Pango support in tuxtype for handling Indic language's rendering issues. So, if we > replace it with fribidi, all Indic languages + some of the south Asian Languages will face rendering issues ( shaping problem). So the expected output will not be same. Now in kerala( a state in India has malayalam as local language) schools we have been using tuxtype as the typing tutor. And also the work to add all other 21 official language's support is already going on. So it will make serious issues if we remove SDL_Pango.

> Since we have already fribidi support in SDL_Pango, I think it's very efficiently solving both the RTL and shaping problems

Hi mobin, actually my native language is arabic which RTL and has a shaping problems also, fribidi2 solve the shaping in arabic, but it is  very important issue is fribidi handles shaping in other language, I don't think so, as i understand fribidi do adhoc implementation for shaping arabic character unlike Pango which  support OpenType (I least that what I expect).

But when proposing fribidi, i mean SDL_Pango is relatively heavier and make porting harder, another issue raised by Tim and Caroline regarding BeOS porting, fribidi may be useful in this case fribidi has no dependencies, unlike SDL_Pango, so having partial RTL support is better than nothing, specially for languages using RTL system and bidi system without need a shaping like Hebrew, I did not means remove SDL_Pango but adding fribidi and using configuration flags to choose the suitable library depending on the depedenecies availability.

Finally, Thank you very much for you interesting point, This will force me to do more investigation opentype