Re: gcc backend that can be interpreted by a java program?

[Adam Megacz <gcj at lists dot megacz dot com>]

Ranjit Mathew <rmathew@hotmail.com> writes:
> As far as I know, TrueType itself needs to be interpreted
> to draw the glyphs - so Adam, you'd end up interpreting
> the interpreter that interprets TrueType instructions. :-)

Right, but the bytecode language (and all interpreters of it) are
covered by Apple patents.  Freetype is unique (and very cool) because
it can *infer* the hints automatically -- so it doesn't use the
bytecodes at all (anymore), and can even hint fonts that lack hinting
bytecodes (like SVG fonts and PostScript fonts).


> I don't know why you can't use JNI/CNI, but IMHO porting

We're committed to using pure Java:

  - XWT runs (in the JVM) on a lot of architectures that I wouldn't be
    likely to produce freetype.so binaries for (Sparc comes to mind).

  - XWT runs on the Danger Hiptop, which only uses pure java

  - XWT will soon run on PalmOS via Jump2, which only takes bytecode as
    input.

  - I want XWT to be portable to architectures I can't even imagine
    today.  The JVM is a wonderfully simple architecture, and I'm
    confident that I can write a simulator (or even binary translater)
    for any future architecture.

  - a