I'm about to translate glob2 into Persian, and as you probably know, Persian uses Arabic alphabet; which its letters, unlike Hebrew (which is also right-to-left), sometimes (most of the times) stick together and form a unified shape. For example, the Persian translation for "Defense Tower" would be "برج مراقبت", but right now it will show up as "ب ر ج م ر ا ق ب ت" i.e. characters are not merged.
I'm a little frustrated by this problem. Because until it get solved, the Persian (and Arabic) translation would be useless.
I have the latest beta (0.9.4.1). I had this problem in 0.9.3 and 0.8.21 which I had before, and had installed both of them from my Debian repository. But I have compiled the new version myself (having libfribidi-dev (and lib-fribidi) installed) and the problem still resists.
Is it a bug? or something wrong in my system? because I searched for the problem and it seems that it have happened to no one else than me!
I even saw the Arabic translation contributor's weblog (http://the-very-end.blogspot.com/2007/0 ... rabic.html) and the letters in the screenshots were rendered correctly! He says one needs "to get latest fribidi from CVS in order to find arabic character shaped correctly". as the post dates back to June 2007, I don't think I still need to get the fribidi from CVS when the fribidi installed in my system is probably newer than the one in CVS in 2007 (which WORKED!). ...or do I need to do so?!
![Sad :(](./images/smilies/icon_sad.gif)
Thanks in advance.