[metapost] memory problem

[Stephan Hennig]

Denis.Roegel at loria.fr schrieb:

>> Again speaking about MetaObj: Is it true that speeding up tree or
>> matrix handling isn't possible by changing MetaPost's equation
>> management since all objects (and attached paths) create one large
>> equation system there?
> 
> In a given metaobj object, all equations make up only one system,
> so the solution is (for me) to slightly change the handling of
> some memory consuming paths (such as round corners).

Ok, these are changes internal to MetaObj -- not MetaPost.


> metapost won't be able to separate these equations, since they are
> not made up of different sets of independent equations.
> 
> What creates different sets of independent equations are equations
> produced by different (unlinked) objects.

After giving it another thought, MetaObj in the current implementation
/should/ benefit from a smarter MetaPost wrt equation handling. The
/creation/ of an object with an attached path should speed up since its
equation system is independent from others at that time.

Internal changes to MetaObj -- such as delaying calculations until a
draw operation -- are welcome, too, of course. :)

Regards,
Stephan Hennig