[gecode-users] new IntArgs fails

[Denys Duchier]

"Christian Schulte" <cschulte at kth.se> writes:

> Argument arrays are really meant to just pass arguments, they are bad as
> general data structures. Why not use vector instead?

I was playing with generating gecode bindings for python (to better
understand the issue, the options, and where the problems are) and
e.g. the constructor IntSet(IntArgs&) requires IntArgs; vector<int> is
not an option.  The generated wrapper for IntArgs does a new and that
didn't compile; I was simply curious if that was intended.

I, of course, don't particularly want IntArgs in python.  That was just
a quick try to see how some things worked out (or not).

Cheers,

--Denys