[...]
> Other than color, which, as everyone will be more that willing to point out,
> would be way slow, there was one real problem with the 6 Oct stuff, didn't
> handle large (ie full unicode) fonts, it seems to munge the CHARS (# chars
> in set), STARTCHAR, and ENCODING data. All other data seemed to come across,
> even data immediately following these, very strange. At this point I don't
> really have the time to look into it too much, it didn't seem to be any of
> the obvious problems, like bit/byte/... ordering, the data all seemed to
> be shifted down, maybe a structure alignment problem... Argh, I enjoy problem
> solving too much, but I've got too many other problems to solve.

The double-byte font SNF files are corrupted due to what I decided was
a compiler bug.  I have forgotten whether the R5 fonts files were
similarly corrupted, but you could just pick up the appropriate SNF
files from any known-good installation.  I gave a workaround patch to
Bernd, so if he does another distribution the files should be fixed then.