ReturnNoOutEdge

DRAFT

Correction

Here is what's right.

In a control-flow graph we want to have a single entry node (where the execution must start) and a single exit node (where the execution must end).

Given that one can have multiple return statements in a method, we need to have an artificial exit node (usually a round node labeled “x”), and all return statements (or other statements that end execution of the method) have an outgoing edge to that exit node.

Note: This may not be an essential aspect for understanding how to represent control-flow with a CFG. The reason we want a single exit node is that we can run program analysis algorithms (the kind of algorithms a compiler or bug-finding tool would run over your program code) that perform their analysis in a backwards-fashion (from exit to entry node, in the opposite direction that a program execution would proceed). Given that in this course we don’t study program analysis algorithms, this “misconception” is not really that important.