Thanks Tom,
could you possibly send an example demonstrating this? I thought that 
atoms are always created using Expression::setAtom which sets the type 
right. However, even if they are, one certainly should be prevented from 
directly constructing an EXF_ATOM expression (e.g. by an assert). It's 
not too robust the way it is.

Tom Kaiser

Tom Moog wrote:

> In Sablotron 0.51 the ctor for Expression (expr.cpp line 1309) the
> switch statement fails to initialize Expression::type when the
> functor is EXF_ATOM.  This later causes problems in at  expr.cpp:1394
> where garbage is used in a switch(type).
> 
> 
> 
> 

Reply via email to