Re: [gts-devel] disabling GTS's memory management

[Rob M. <ro...@as...>]

> OK, I'll have a look at the "Gerris tutorial" you pointed me to and see
> if I can do this.  Will it be possible to make gts_vertex derive from
> say gts_notmemmanaged_object, or will i have to create a whole
> mirror-image inheritance tree?

Stephane is certainly the expert on this one, but from what I remember of
extending the GTS class structure, there is no way to slip a new layer 'in
the middle'.

If I recall correctly, faces are extended from triangles, which are extended
from objects.  And so on with the other entities.

When a surface is created, you pass parameters telling it what classes
should be used to provide faces, edges, and vertices.  That way, one may
extend an entity (say a face) and then use it in the surface.

There may be a way to extend this approach to replacing the fundamental
gts_object that everything else inherits from.  This would involve a
secondary constructor for each of the objects.  This constructor would take
a parameter specifying the gts_object--based class to build from.  (A
replacement default constructor would be written to maintain backward
compatibility).

I'm not sure how feasible this is, but it is the first approach I would take
a close look at.

Let me know if this doesn't make sense,

Also, the object-oriented in ANSI-C programming paradigm used by GTS is
borrowed heavily from glib & related projects.  It may serve you well to
poke around there for examples on how to accomplish this sort of thing.

                 Rob