Flow graph

slicer allows you to explore the control flow of a method using a top-to-bottom flow diagram.

This view can be opened in the context menu of a project entry (Open as -> Flow graph) or in the menu bar for an already opened view of a class file (View -> Flow graph).

A flow graph is composed of two parts: nodes and edges.

A node is a block of instructions, which are executed sequentially. The node containing the zero-offset (first) instruction is called the entrypoint node and is highlighted with a special border (primary color of the UI - white, black, …).

Every node that does not naturally (without a jump) pass off execution to another node must end with a branch (if, goto, switch or subroutine jump) or terminal instruction (return, throw or subroutine return), otherwise it does not pass class file verification (execution fell off at end of code).

An edge is a line that connects the execution of two blocks in a specific direction. An edge can represent:

• a conditional/unconditional jump (a full line)
• a natural flow of execution between two consecutive nodes (a dashed line)
• a connection to an exception handler (a full red line)

Edges may also have hint labels, like:

• true/false (“branch if comparison succeeds” instructions)
• default (switch instruction default branch)
• an integer value (switch instruction branch case)
• java/lang/Throwable (or any other exception type; exception handler)

A node typically has 1 (unconditional jump), 2 (conditional jump) or more (switch conditional jump) edges, but it can also have no edges, which means it’s dead code - it will never be executed.

Flow graphs can be exported into SVG or PNG format in the context menu (right click -> Export).