Re: [CIL users] computeCFGInfo

[Benjamin Y. <ben...@fa...>]

On Sep 6, 2005, at 8:14 AM, Nathan Cooprider wrote:

> Before calling computeCFGInfo (in cil.ml), CIL has to call  
> prepareCFG.  From what I understand, this transforms switch, break  
> and continue statements into other statements with more-or-less  
> equivalent control flow.
> So my question is: why is this necessary?  I see in succpred_stmt  
> that the cases for those three types of statements are not written.   
> I have not tried to implement this myself, but I might in the near  
> future.  I assume the effort has already been tried.  Why did it  
> fail or not remain in the code?
> Nathan


Has there been any follow-up on this issue?  I recently started using  
the CFG information that Cil computes, which is very convenient.   
However, I would really like to see switches as specified in the  
source.  As Nathan said, it's not obvious why CFG information couldn't  
be computed in the presence of switches.  I suppose the chain of if- 
then-elses that prepareCFG creates could somehow be marked as all  
"coming from" a switch statement.  Any thoughts?

Thanks,
Ben