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 ------------------------------------------------------------------------------ _______________________________________________ CIL-users mailing list CIL-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cil-users
