In response to a request in the bugzilla I have implemented a basic glyph shape selection based on context for scripted languages such as Arabic that have initial/medial/final/stand-alone glyphs for a single character. The behaviour is extremely basic and will only work for languages with 'legacy' code pages in Unicode where each such a shape is given a separate code point; it will not work for non- legacy code sets such as Syriac. It should allow us to expand the bidi-user base for the time being, until such a time when we get to use 'a proper' layout engine, such as Pango.
At the moment the data contained in the new class is for Hebrew only, but I will try to get support for Arabic as soon as possible. This feature is enabled only in the bidi build. Tomas
