Framework for testing concurrent algorithms on the JVM. Enables declarative test specification, automatically generates random scenarios, and verifies correctness properties using stress-testing or model checking.
Lincheck is a practical and user-friendly framework for writing deterministic and robust concurrent tests on JVM. When detecting an error, Lincheck provides a reproducible execution trace an ability to debug it step-by-step in IntelliJ IDEA.
⚠️ Lincheck 3.0 Changes️:
- Lincheck now focuses on testing arbitrary concurrent code still providing declarative API for testing concurrent data structures
- The API has been moved from
org.jetbrains.kotlinx.linchecktoorg.jetbrains.lincheck- The artifact is now published to Maven Central under the
org.jetbrains.lincheckgroup ID
To use Lincheck in your project, you first need to add it as a dependency. If you use Gradle, add the following lines to build.gradle.kts:
repositories {
mavenCentral()
}
dependencies {
// Lincheck dependency
testImplementation("org.jetbrains.lincheck:lincheck:3.6")
}
To write a Lincheck test,
you need to wrap your concurrent logic with the Lincheck.runConcurrentTest { ... } function.
Lincheck will automatically study different thread interleavings and report an error
if one leads to a test failure.
As an example, see the Lincheck test for a counter:
@Test // JUnit test
fun test() = Lincheck.runConcurrentTest {
var counter = 0
// Increment the counter concurrently
val t1 = thread { counter++ }
val t2 = thread { counter++ }
// Wait for the threads to finish
t1.join()
t2.join()
// Check both increments have been applied
assertEquals(2, counter)
}
If you run the counter test above, it finishes with an error, with Lincheck providing a step-by-step execution trace to reproduce it.
AssertionFailedError: expected:<2> but was:<1>
The following interleaving leads to the error:
| ------------------------------------------------------------------------------- |
| Main Thread | Thread 1 | Thread 2 |
| ------------------------------------------------------------------------------- |
| thread(block = Lambda#1): Thread#1 | | |
| thread( = Lambda
| switch (reason: waiting for Thread 1 to finish) | | |
| | | run() |
| | | counter ➜ 0 |
| | | switch |
| | run() | |
| | counter ➜ 0 | |
| | = | |
| | | = |
| Thread
| Thread
| ------------------------------------------------------------------------------- |
Lincheck provides a special API to simplify testing concurrent data structures. Instead of describing how to perform tests, you can specify what to test by just declaring all the data structure operations to examine. After that, Lincheck automatically generates a set of random concurrent scenarios, examines them using either stress-testing or bounded model checking, and verifies that the results of each invocation satisfy the required correctness property (linearizability by default).
The following Lincheck test easily finds a bug in the standard Java's ConcurrentLinkedDeque:
When running modelCheckingTest(), Lincheck provides a short concurrent scenario
that discovers the bug with a detailed execution trace to reproduce it.
= Invalid execution results =
| ------------------------------ |
| Thread 1 | Thread 2 |
| ------------------------------ |
| addLast(1) | |
| ------------------------------ |
| pollFirst(): 1 | addFirst(0): |
| | peekLast(): 1 |
| ------------------------------ |
---
All operations above the horizontal line | ----- | happen before those below the line
---
The following interleaving leads to the error:
| ------------------------------------------------------- |
| Thread 1 | Thread 2 |
| ------------------------------------------------------- |
| addLast(1) | |
| ------------------------------------------------------- |
| pollFirst(): 1 | |
| deque.pollFirst(): 1 | |
| first(): Node#2 | |
| p.item ➜ null | |
| p.next ➜ Node#1 | |
| p.item ➜ 1 | |
| first.prev ➜ null | |
| switch | |
| | addFirst(0) |
| | peekLast(): 1 |
| p.item.compareAndSet(1, null): true | |
| unlink(Node#1) | |
| result: 1 | |
| ------------------------------------------------------- |
Please see the official tutorial that showcases Lincheck features through examples.
You may also be interested in the following resources:
This is a fork of the Lin-Check framework by Devexperts.
class ConcurrentLinkedDequeTest {
private val deque = ConcurrentLinkedDeque<Int>()
fun addFirst(e: Int) = deque.addFirst(e)
fun addLast(e: Int) = deque.addLast(e)
fun pollFirst() = deque.pollFirst()
fun pollLast() = deque.pollLast()
fun peekFirst() = deque.peekFirst()
fun peekLast() = deque.peekLast()
// Run Lincheck in the stress testing mode
fun stressTest() = StressOptions().check(this::class)
// Run Lincheck in the model checking testing mode
fun modelCheckingTest() = ModelCheckingOptions().check(this::class)
}
Surfaced from shared tags and platforms — no rankings paid for.