Simon Marlow simonmar at
Thu Aug 28 06:34:59 EDT 2003

> 3) link against another library such as libunicode which provides its
> own classification routines (this could be done optionally at compile
> time...)

libunicode is no good, I tried it.  It only recognises Unicode
characters up to '\xffff'.

The only right way to do this, it seems, is to generate the tables from
UnicodeData.txt.  However, I'm prepared to live with the current
solution as long as we document its shortcomings.  After all, it does
the right thing on the majority of our installed base.


More information about the FFI mailing list