The following three graphs show the control flow graph of a procedural program. The nodes contain the text of statements as labels. Not all edges have labels. The visualized control flow comes from the following nonsense program, which consists of a procedure test and a main routine in a pseudo imperative language.
PROCEDURE test( VAR b : INTEGER; c : INTEGER );
BEGIN
b := c + 5;
END
BEGIN /* main routine of a nonsense program */
x := 1;
WHILE (x = 1) DO
x := 2;
test ( x, 1 );
x := 3;
OD;
WHILE (x = 1) DO
x := 4;
x := 5;
test ( x, 2 );
OD;
WHILE (x = 1) DO
x := 6;
IF (x = 7) THEN x := 8; ELSE test ( x, 3 );
FI;
OD;
END.
Example 7: Control Flow Graph 1
This example shows a simple visualization of the control flow graph.
graph: {title: "Control Flow Graph" layoutalgorithm: dfs finetuning : no display_edge_labels: yes yspace : 55 edge.thickness : 3 node: { title:"18" label: "test_b := test_c+5"} node: { title:"17" label: "Exit" } node: { title:"16" label: "test (x,3)" } node: { title:"15" label: "x := 8" } node: { title:"14" label: "x=7" } node: { title:"13" label: "x := 6" } node: { title:"12" label: "x=1" } node: { title:"11" label: "test (x,2)" } node: { title:"10" label: "x := 5" } node: { title:"9" label: "x := 4" } node: { title:"8" label: "x=1" } node: { title:"7" label: "x := 3" } node: { title:"6" label: "test (x,1)" } node: { title:"5" label: "x := 2" } node: { title:"4" label: "x=1" } node: { title:"3" label: "x := 1" } node: { title:"2" label: "Start" } node: { title:"1" label: "Exit point\ntest" } node: { title:"0" label: "Entry point\ntest" } edge: { source:"18" target:"1" } edge: { source:"0" target:"18" } edge: { source:"12" target:"17" label: "false"} edge: { source:"8" target:"12" label: "false"} edge: { source:"16" target:"12" label: "back" } edge: { source:"15" target:"12" label: "back" } edge: { source:"13" target:"14" } edge: { source:"14" target:"16" label: "false"} edge: { source:"14" target:"15" label: "true" } edge: { source:"12" target:"13" label: "true" } edge: { source:"4" target:"8" label: "false"} edge: { source:"11" target:"8" label: "back" } edge: { source:"10" target:"11" } edge: { source:"9" target:"10" } edge: { source:"8" target:"9" label: "true" } edge: { source:"3" target:"4" } edge: { source:"7" target:"4" label: "back" } edge: { source:"6" target:"7" } edge: { source:"5" target:"6" } edge: { source:"4" target:"5" label: "true" } edge: { source:"2" target:"3" } } |
 |
Example 8: Control Flow Graph 2
This example shows an improved visualization of the control flow graph. The start, exit and branch nodes are drawn in different shapes so that they can be better recognized. The edges closing a cycle are specified as back edges in order to see the uniform flow of control in the other edges. The edges at the branch nodes are anchored at the left and right via the bent near edge specification.
graph: {title: "Control Flow Graph" layoutalgorithm: dfs finetuning : no display_edge_labels: yes yspace : 55 edge.thickness : 3 node: { title:"18" label: "test_b := test_c+5"} node: { title:"17" label: "Exit" shape: ellipse} node: { title:"16" label: "test (x,3)" } node: { title:"15" label: "x := 8" } node: { title:"14" label: "x=7" shape: rhomb } node: { title:"13" label: "x := 6" } node: { title:"12" label: "x=1" shape: rhomb } node: { title:"11" label: "test (x,2)" } node: { title:"10" label: "x := 5" } node: { title:"9" label: "x := 4" } node: { title:"8" label: "x=1" shape: rhomb } node: { title:"7" label: "x := 3" } node: { title:"6" label: "test (x,1)" } node: { title:"5" label: "x := 2" } node: { title:"4" label: "x=1" shape: rhomb } node: { title:"3" label: "x := 1" } node: { title:"2" label: "Start" shape: ellipse} node: { title:"1" label: "Exit point\ntest" shape: ellipse} node: { title:"0" label: "Entry point\ntest" shape: ellipse} edge: { source:"18" target:"1" } edge: { source:"0" target:"18" } bentnearedge: { source:"12" target:"17" label: "false"} bentnearedge: { source:"8" target:"12" label: "false"} backedge: { source:"16" target:"12" label: "back" } backedge: { source:"15" target:"12" label: "back" } edge: { source:"13" target:"14" } bentnearedge: { source:"14" target:"16" label: "false"} bentnearedge: { source:"14" target:"15" label: "true" } bentnearedge: { source:"12" target:"13" label: "true" } bentnearedge: { source:"4" target:"8" label: "false"} backedge: { source:"11" target:"8" label: "back" } edge: { source:"10" target:"11" } edge: { source:"9" target:"10" } bentnearedge: { source:"8" target:"9" label: "true" } edge: { source:"3" target:"4" } backedge: { source:"7" target:"4" label: "back" } edge: { source:"6" target:"7" } edge: { source:"5" target:"6" } bentnearedge: { source:"4" target:"5" label: "true" } edge: { source:"2" target:"3" } } |
 |
Example 9: Control Flow Graph 3
|
This example shows another improved visualization of the control flow graph of Example 7. Here an orthogonal layout is used so that the graph looks like a typical flowchart. For orthogonal layout, a large down factor (and a large near factor as well as an up factor of zero) is recommended. This improves the layout of long vertical edges. Add the following lines to the graph specification in Example 8: manhattan_edges: yes equalydist : yes layout_downfactor: 8 |
 Control flow graph 3: Manhattan edges |
Last modified on 22 November 2001 by webmaster. © 2000-2001 AbsInt