[Bf-committers] Transform Refactoring

[Martin Poirier]

--- Jeremy Wall <zaphar at gmail.com> wrote:

> Currently constraints actually edit the object data
> themselves to
> enforce the constraint If I remember correctly. Are
> you planning to
> change to "test" method where the constraint just
> returns a pass/fail
> for a given transform? Or do you plan to  continue
> with a similar
> method to the currrent one for enforcing
> constraint's?

I think you're mixing two different things here. The
transform function has very little to do with Blender
Constraints. The constraining system used by the
transform function is more like a projection system,
mapping the input to the constraint space.

> Personally I would  prefer to see the constraint
> system return
> pass/fail values for a given transform.

How could a transform constraint fail?

> I think it would make
> implementing constraints easier and simpler.  It
> would also make the
> constraint data  types independent from the
> transform code.

I must be missing something here, but constraint have
to be transformation dependant. You don't constraint
to a single axis in a translation like you'd do in a
rotation.

> Just for clarification is your generic constraint
> data type a
> container for other constraint specific structs or 
> is it supposed to
> hold the actuall constraint data? I'm assuming its a
> container similar
> to the current implementation.

It holds the actual constraint data (which is much
simpler than Constraint datablocks).

Martin


		
_______________________________
Do you Yahoo!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com