[luatex] OCP broken?

[Khaled Hosny]

On Mon, Jun 21, 2010 at 09:08:07AM +0200, Taco Hoekwater wrote:
> Khaled Hosny wrote:
> >On Sun, Jun 20, 2010 at 09:56:56AM +0200, Taco Hoekwater wrote:
> >>Khaled Hosny wrote:
> >>>I know such Omega extensions are rarely tested, if at all, but any idea
> >>>why the attached OCP is no longer working? It used to work while ago.
> >>Any idea when it was still ok? I have to disect through past revisions,
> >>I fear.
> >
> >Last time I checked was long ago, I'll try to see if I can find when it
> >broke.
> 
> This has been broken since ages. I tried some older versions which did
> actually run the otp but produced '987654321TLT ridtxet' because luatex
> does not interpret commands in the return values from the otp buffer
> and has not done that since a very,very long time (at least since
> 0.30).
> 
> The otp output buffer enters luatex at a different level than in Omega,
> past the 'mouth'. This is a side-effect of the changes to the main
> control loop that were made for the separation between glyphs and
> characters (getting rid of the ligature building in the mouth).
> 
> I could attempt to go back to the state where it runs the otp, but to
> be honest I don't see much point in it as it will not work properly
> anyway. In fact this may be a good moment to do away with otps altogether.

I don't have strong need for OTPs either, all what I need can be done in
lua, I was just a bit lazy and thought I'd reuse some old code. I'm fine
with no OTP, just make sure it is documented.


-- 
 Khaled Hosny
 Arabic localiser and member of Arabeyes.org team
 Free font developer