First of all many thanks to all who invested the time to read and
look thru all the things I supplied (and asked) and additionally
also took the time to investigate further. This was a big help.
I meanwhile got the error fixed which made the AS script go slow,
but I'm still struggling to set up a conversion from ISO 6937 to
Unicode/ASCII since I only got a few samples of how the text should
look like (the customer don't know either). I didn't find a rule
yet, so it's some kind of trial and error.
Again thanks for the tremendous input - any further ideas are
appreciate as well ;-)
The quick and dirty way to convert ISO 6937 to Unicode (based on a
quick look at the Wikipedia article) looks like it's to simply take
any byte in the 0xC1-0xCF range, turn it into the appropriate Unicode
combining mark (for example, 0xC1 would become U+0300, a grave mark)
and emit that *after* the following character. Otherwise, just copy
the value as is, since ISO 646 is isomorphic to Unicode. Presto,
canonically decomposed Unicode. (Details of which UTF to emit and
how are left as an exercise for the reader.)
