converting unicode ttf to cp1252 type1

[Philip Lehman <lehman@gmx.net>]

I'm trying to convert a bunch of true type fonts with unicode encoding
to type1 with microsoft-cp1252. I'm afraid it has to be cp1252 (or
adobe-fontspecific for that matter) since I need them for word
processing and standard latin encoding is not enough.

As I understand it, iso8859-1 does not define the codes 127 to 160
while cp1252 makes use of these codes and still sticks to a 8bit map.
So I figured it should be possible to build type1 fonts with cp1252
encoding. Has anybody tried to do this before or am I getting
something completely wrong here? Provided I'm not completely mistaken,
can anybody recommend some tools that convert ttf to type1 and support
custom *output* encodings?

I tried the latest ttf2pt1 package with the modifications by Sergey
Babkin. The ttf2pt1 scripts support custom encoding tables, but
although I wrote tables for iso10646-1 and ms-cp1252, re-mapping
doesn't work. Codes 127-160 remain blank and I even get funny results
for codes >160: the latin characters are replaced by cyrillic ones.

So I took a look at the ttf2pfb binary that comes with freetype. It
supports multiple encoding tables, but it looks like this works for
various *input* encodings only. The resulting type1 fonts I got were
always standard latin.

I spent the entire weekend trying to figure this out and it drives me
crazy... Can anybody help me out here?

-- 
Philip Lehman <lehman@gmx.net>