[M3devel] Never ending love with alignments, data sizes et al :)

[Daniel Alejandro Benavides D.]

Hi all:
I think this is just a failed check (an over check) of what it is suppose to think, is better to assert type errors than false execution ill-behavior (since this is not the case) one's assumption may be over checked, and thus you get the type error (otherwise you would like need some to think like this, I'm not against it just that, but this is pure dreams in one big head):
http://www.dorje.com/netstuff/jokes/prog.smurf

http://venus.is.s.u-tokyo.ac.jp/amo/AMO-Berlin/index.html

If you see it's akind of what you would want, type safe checking of C C++ superset of it with Obliq-like language. It's written in Gnu C. It's JIT

Thanks in advance

--- El sáb, 20/8/11, Dragiša Durić <dragisha at m3w.org> escribió:

> De: Dragiša Durić <dragisha at m3w.org>
> Asunto: [M3devel] Never ending love with alignments, data sizes et al :)
> Para: "m3devel" <m3devel at elegosoft.com>
> Fecha: sábado, 20 de agosto, 2011 13:40
> "../src/ArcGrid.m3", line 28:
> LOOPHOLE: expression's size differs from type's
> 
> PROCEDURE SwapFloat(f: REAL): REAL =
>   BEGIN
>     RETURN LOOPHOLE(NBOFrom(LOOPHOLE(f,
> Ctypes.unsigned)), REAL); <- line 28, of course :)
>   END SwapFloat;
> 
> BITSIZE is 32 for both.