[opendx-users] Positions and double

[Marco Morandini]

I know taht positions can't be defined with doubles.
However, this would allow
to use opendx data to restart a simulation.

I'd like to get rid of the problem and to understand were
OpenDx has to be modified in oredr to accept double positions.
Is this an affordable work (one or two subroutines) or will require
too much work (I'm a simple Dx user)?

I begun the work by modifiying box.c, from

if (type!=TYPE_FLOAT || category!=CATEGORY_REAL || (rank!=0 && rank!=1))

to 

if ((type!=TYPE_FLOAT && type!=TYPE_DOUBLE) || category!=CATEGORY_REAL
|| (rank!=0 && rank!=1))

After the patch Dx gives this error:

ERROR: Image: AutoCamera: Bad parameter: zero length direction vector

I think the error raised by
if (fabs(DXLength(DXSub(p->f, p->t))) < FUZZ) {
        if (p->isauto)
            DXSetError(ERROR_BAD_PARAMETER, "#11822", "direction");
.....

in camera.c, line 768 (dx 4.1.0)

Could somone explain me what Dx is doing 
(which routines are used to set-up the camera),
were I need to write/modify code and how
much work would this require?

Thanks in advance,

Marco Morandini