Presented at ESEC/FSE 2005.

1. DynaMine
Finding Common Error Patterns by
Mining Software Revision Histories
Benjamin Livshits Thomas Zimmermann
Stanford University Saarland University

2. Error Pattern Iceberg
null dereferences,
buffer overruns, ...
The usual suspects
J2EE servlets
Application
specific
Device drivers Linux code

3. Co-changed items = patterns

4. Co-added Method Calls
public void createPartControl(Composite parent) {
...
// add listener for editor page activation
getSite().getPage().addPartListener(partListener);
}
public void dispose() {
...
getSite().getPage().removePartListener(partListener);
}

5. Co-added Method Calls
public void createPartControl(Composite parent) {
...
// add listener for editor page activation
getSite().getPage().addPartListener(partListener);
}
public void dispose() {
...
getSite().getPage().removePartListener(partListener);
}

6. Co-added Method Calls
public void createPartControl(Composite parent) {
...
// add listener for editor page activation
getSite().getPage().addPartListener(partListener);
}
co-added
public void dispose() {
...
getSite().getPage().removePartListener(partListener);
}

7. How DynaMine Works
mine CVS rank and
revision
patterns
history mining histories filter
instrument relevant
method calls
run the application
dynamic
analysis
post-process
usage error unlikely
patterns patterns patterns
report report
reporting
patterns bugs

8. Mining Patterns
mine CVS rank and
revision
patterns
history mining histories filter
instrument relevant
method calls
run the application
dynamic
analysis
post-process
usage error unlikely
patterns patterns patterns
report report
reporting
patterns bugs

9. Mining Method Calls
o1.addListener()
Foo.java
1.12
o1.removeListener()
Bar.java o2.addListener()
1.47
o2.removeListener()
System.out.println()
Baz.java o3.addListener()
1.23
o3.removeListener()
list.iterator()
iter.hasNext()
iter.next()
o4.addListener()
Qux.java
1.41
System.out.println()
o4.removeListener()
1.42

10. Finding Pairs
1 Pair
o1.addListener()
Foo.java
1.12
o1.removeListener()
Bar.java o2.addListener()
1 Pair
1.47
o2.removeListener()
System.out.println()
Baz.java o3.addListener()
1.23
o3.removeListener()
2 Pairs
list.iterator()
iter.hasNext()
iter.next()
0 Pairs
o4.addListener()
Qux.java
1.41
System.out.println()
0 Pairs
o4.removeListener()
1.42

11. Finding Patterns
Find “frequent itemsets” (with Apriori)
o.enterAlignment()
o.enterAlignment()
o.enterAlignment()
o.enterAlignment()
o.exitAlignment()
o.exitAlignment()
o.exitAlignment()
o.exitAlignment()
o.redoAlignment()
o.redoAlignment()
o.redoAlignment()
o.redoAlignment()
iter.hasNext()
iter.hasNext()
iter.hasNext()
iter.hasNext()
iter.next()
iter.next()
iter.next()
iter.next()
{enterAlignment(), exitAlignment(),
redoAlignment()}

12. Ranking Patterns
o1.addListener()
Foo.java
1.12
o1.removeListener()
Bar.java o2.addListener()
1.47
o2.removeListener()
Support count =
System.out.println()
#occurrences of a pattern
Baz.java o3.addListener()
1.23
o3.removeListener()
Confidence =
list.iterator()
strength of a pattern, P(A|B)
iter.hasNext()
iter.next()
o4.addListener()
Qux.java
1.41
System.out.println()
o4.removeListener()
1.42

13. Ranking Patterns
o1.addListener()
Foo.java
1.12
o1.removeListener()
Bar.java o2.addListener()
1.47
o2.removeListener()
System.out.println()
Baz.java o3.addListener()
1.23
o3.removeListener()
list.iterator()
iter.hasNext()
iter.next()
This is a fix!
o4.addListener()
Qux.java
1.41
System.out.println() Rank removeListener()
patterns higher
o4.removeListener()
1.42

14. Investigated Projects
JEDIT ECLIPSE
since 2000 2001
developers 92 112
lines of code 700,000 2,900,000
revisions 40,000 400,000

15. Simple Method Pairs
GUIs & Listener
Usage pattern
o.addWidget() o.removeWidget()
o.addPropertyChangeListener() o.removePropertyChangeListener()
Error pattern

16. Simple Method Pairs
Locking of Resources
o.HLock() o.HUnlock()
Not hit at runtime

17. State Machines in Eclipse
Pretty-printing Usage pattern
o.redoAlignment()
o.enterAlignment() o.exitAlignment()

18. State Machines in Eclipse
Memory context manupilation
Not hit at runtime
OS.PmMemFlush()
OS.PmMemStart() OS.PmMemStop()
OS.PmMemCreateMC() OS.PmMemReleaseMC()

19. State Machines in JEdit
Compound edits (for undo/redo)
Usage pattern
o.insert()
o.remove()
o.beginCompoundEdit() o.endCompoundEdit()

20. Complex Patterns
try {
monitor.beginTask(null, Policy.totalWork);
int depth = -1;
try {
workspace.prepareOperation(null, monitor);
workspace.beginOperation(true);
depth = workspace.getWorkManager().beginUnprotected();
! return runInWorkspace
(Policy.subMonitorFor(monitor, Policy.opWork,
SubProgressMonitor.PREPEND_MAIN_LABEL_TO_SUBTASK));
} catch (OperationCanceledException e) {
workspace.getWorkManager().operationCanceled();
return Status.CANCEL_STATUS;
} finally {
if (depth >= 0)
workspace.getWorkManager().endUnprotected(depth);
workspace.endOperation(null, false,
Policy.subMonitorFor(monitor, Policy.endOpWork));
}
} catch (CoreException e) {
return e.getStatus();
} finally {
monitor.done();
}

21. Complex Patterns
try {
monitor.beginTask(null, Policy.totalWork);
int depth = -1;
try {
workspace.prepareOperation(null, monitor);
workspace.beginOperation(true);
depth = workspace.getWorkManager().beginUnprotected();
! return runInWorkspace
(Policy.subMonitorFor(monitor, Policy.opWork,
SubProgressMonitor.PREPEND_MAIN_LABEL_TO_SUBTASK));
} catch (OperationCanceledException e) {
workspace.getWorkManager().operationCanceled();
return Status.CANCEL_STATUS;
} finally {
if (depth >= 0)
workspace.getWorkManager().endUnprotected(depth);
workspace.endOperation(null, false,
Policy.subMonitorFor(monitor, Policy.endOpWork));
}
} catch (CoreException e) {
return e.getStatus();
} finally {
monitor.done();
}

22. Complex Patterns
try {
monitor.beginTask(null, Policy.totalWork);
int depth = -1;
try {
workspace.prepareOperation(null, monitor);
workspace.beginOperation(true);
depth = workspace.getWorkManager().beginUnprotected();
! return runInWorkspace
(Policy.subMonitorFor(monitor, Policy.opWork,
SubProgressMonitor.PREPEND_MAIN_LABEL_TO_SUBTASK));
} catch (OperationCanceledException e) {
workspace.getWorkManager().operationCanceled();
return Status.CANCEL_STATUS;
} finally {
if (depth >= 0)
workspace.getWorkManager().endUnprotected(depth);
workspace.endOperation(null, false,
Policy.subMonitorFor(monitor, Policy.endOpWork));
}
} catch (CoreException e) {
return e.getStatus();
} finally {
monitor.done();
}

23. Workspace Transactions
Usage pattern
S → O*
O → w.prepareOperation()
w.beginOperation()
U*
w.endOperation()
U → w.getWorkManager().beginUnprotected()
S
[w.getWorkManager().operationCanceled()]
w.getWorkManager().beginUnprotected()

24. Dynamic Validation
mine CVS rank and
revision
patterns
history mining histories filter
instrument relevant
method calls
run the application
dynamic
analysis
post-process
usage error unlikely
patterns patterns patterns
report report
reporting
patterns bugs

25. Pattern classification
post-process
v validations, e violations
usage error unlikely
patterns patterns patterns
ev/10 v/10=e=2v otherwise

26. Experiments
usage pattern
15
not hit
24
error pattern
8
unlikely
11
total 56 patterns

27. Future Work
Automatically generate state machines
Additional patterns by textual matching
Programmer assist tools
Programmers who inserted a call to open()
inserted a call to close()
Aspect Mining

28. Contibutions
DynaMine learns usage patterns from
large version archives.
DynaMine ranks patterns effectively,
especially for finding error patterns.
DynaMine increases trust in patterns by
dynamic analysis