Nick Glencross wrote:
Hmmm. still don't think that's what I want. Obviously this is in the context of the NCI callback stuff that I'm looking into -- no surprises there.
Ah callbacks again. As these are called by C code, you really should not a function signature at compile-time.
In case 2), which I've just tried, I think that I'm right in saying that the *PMCArray will arrive as the first argument of _call_back, which will automagically convert to an int, being its length.
Sure. But that's the case, where I think we need an extension to signatures in *runops_fromc*: '*' := flatten array to callee parameters.
Nick
leo
