> Hello Carlos,
>
> It is nothing new.
>
> A) Unicode Version
> I have written mostly the same to Jason about a year ago.
> I was no big deal to make a first unicode version, about 2 days
> (only for d2009/d2010).
> (This version is based on 4.8.7 on more than a year old).
> Jason has this version of mine.
> But he decided to go in another direction.

Are you still using that version you produced?

When I evaluated what I needed to do it looked unavoidable to me to make a
version of IBO in D2009/D2010 that explicitly maintained legacy backwards
compatibility to the extent possible. Just going through and changing
"string" to "ansistring" (among many other things like char, pchar, etc.) so
that my routines would all work again was a major undertaking. Strings are
the heart and soul of middleware.

It appeared to me that the other route would involve re-writing many
routines but they would no longer be backwards compatible either. Thus, the
old routines would need to be brought to a state where they functioned for
backwards compatibility anyway. So, the logical step was to get backwards
compatibility completed first in a separate branch of code and then in a
subsequent version to radically rewrite everything to take full advantage of
the new capabilities.

So, if I missed something that you offered I would be delighted to revisit
it. I would cry a river of tears if somehow there was actually a way I could
have made IBO function in D2009/D2010 in a stable manner with only a few
days of work.

> B) Test Suite
> You can found a first Test Suite for IBO on Sourceforge.
> Wenn I found some bugs I first write a test for it, then fix it.
> But the Test Suite contains only a few cases.
> If somebody have written testcode he/she is invited to join the project.

To my knowledge I have addressed all issues on Sourceforge. Would you please
validate that with the latest version of IBO and let me know if I have
missed anything?

Thanks,
Jason Wharton