llvm dev - Dec 2002 - [LLVMdev] Shouldn't need a "Module Graph"

The way we are doing it, per Vikram's recommendation, is to provide a 
mapping between the callers nodes and the callees nodes at the time of the 
call.  Given that the only way they can communicate is through the 
parameters, globals, and return node, this is fairly easy.  You must take 
into account that the DSGraphs may be different shapes though.

Dave

Thanks for the tip, but what do you mean by "DSGraphs may be different
shapes?

On Sun, 1 Dec 2002, David Crowe wrote:
> The way we are doing it, per Vikram's recommendation, is to provide a
> mapping between the callers nodes and the callees nodes at the time of the
> call.  Given that the only way they can communicate is through the
> parameters, globals, and return node, this is fairly easy.  You must take
> into account that the DSGraphs may be different shapes though.

Ok, here is a trivial example:

In the Callee: A->(itself)

In the caller: B->C->itself

Your mapping should provide:

B->A
C->A
A->B  (not sure about this one, is there a better way?)

I don't know how this example could be manufactured in code, but I'm
sure
there are more complex cases which need the same principle...

Dave

On Sun, 1 Dec 2002, Juan Nicolas Ruiz wrote:
> Thanks for the tip, but what do you mean by "DSGraphs may be different
> shapes?
> 
> On Sun, 1 Dec 2002, David Crowe wrote:
> 
> > The way we are doing it, per Vikram's recommendation, is to
provide a
> > mapping between the callers nodes and the callees nodes at the time of
the
> > call.  Given that the only way they can communicate is through the
> > parameters, globals, and return node, this is fairly easy.  You must
take
> > into account that the DSGraphs may be different shapes though.
>