Prosper is an integration testing library for Adobe Experience Manager projects using Spock, a Groovy-based testing framework notable for it's expressive specification language. The library contains a base Spock specification that provides an in-memory JCR, Sling framework support, and a mock OSGi context for registering and testing services.
- Test AEM projects outside of an OSGi container in the standard Maven build lifecycle.
- Write test specifications in Groovy using Spock, a JUnit-based testing framework with an elegant syntax for writing tests quickly and efficiently.
- Sling
ResourceResolver
instance backed by an in-memory Jackrabbit Oak JCR instance supporting the complete set of repository operations. - Supplies mock OSGi bundle and Sling contexts for registering services, adapters, and Sling models.
- Utilizes Groovy builders from the AEM Groovy Extension to provide a simple DSL for creating test content.
- Provides additional builders for mock Sling requests and responses to simplify setup of test cases.
- Maven 3.x
- Familiarity with Groovy language and the Spock specification syntax (or see included tests for examples).
Prosper Version(s) | AEM Version(s) |
---|---|
14.x.x | 6.3, 6.4, 6.5 |
13.x.x, 12.x.x | 6.4 |
11.x.x | 6.3 |
10.x.x, 9.x.x, 8.x.x | 6.2 |
7.x.x, 6.x.x, 5.x.x, 4.x.x | 6.1 |
3.x.x, 2.x.x, 1.x.x | 6.0 |
<= 0.10.x | 5.6 (CQ) |
Add Maven dependency to project pom.xml
.
<dependency>
<groupId>com.icfolson.aem.prosper</groupId>
<artifactId>prosper</artifactId>
<version>15.0.0</version>
<scope>test</scope>
</dependency>
Create a src/test/groovy
directory in your project structure and add a Spock specification extending the base ProsperSpec
.
import com.icfolson.aem.prosper.specs.ProsperSpec
class ExampleSpec extends ProsperSpec {
def setupSpec() {
// use PageBuilder from base spec to create test content
pageBuilder.content {
home("Home") {
"jcr:content"("sling:resourceType": "foundation/components/page")
}
}
}
// basic content assertions provided
def "home page exists"() {
expect:
assertPageExists("/content/home")
}
// Node metaclass provided by AEM Groovy Extension simplifies JCR operations
def "home page has expected resource type"() {
setup:
def contentNode = session.getNode("/content/home/jcr:content")
expect:
contentNode.get("sling:resourceType") == "foundation/components/page"
}
}
Configure Groovy compiler and Surefire plugin in Maven pom.xml
. Additional configurations details for projects with mixed Java/Groovy sources can be found here.
<build>
<sourceDirectory>src/main/groovy</sourceDirectory>
<testSourceDirectory>src/test/groovy</testSourceDirectory>
<plugins>
<plugin>
<artifactId>maven-compiler-plugin</artifactId>
<version>3.8.0</version>
<configuration>
<compilerId>groovy-eclipse-compiler</compilerId>
<source>1.8</source>
<target>1.8</target>
<encoding>utf-8</encoding>
</configuration>
<dependencies>
<dependency>
<groupId>org.codehaus.groovy</groupId>
<artifactId>groovy-eclipse-compiler</artifactId>
<version>2.9.2-01</version>
</dependency>
<dependency>
<groupId>org.codehaus.groovy</groupId>
<artifactId>groovy-eclipse-batch</artifactId>
<version>2.4.3-01</version>
</dependency>
</dependencies>
</plugin>
<plugin>
<artifactId>maven-surefire-plugin</artifactId>
<version>2.19.1</version>
<configuration>
<includes>
<!-- Spock naming convention. Change as needed if your test classes have a different naming strategy. -->
<include>**/*Spec*</include>
</includes>
</configuration>
</plugin>
</plugins>
</build>
<dependencies>
<dependency>
<groupId>org.codehaus.groovy</groupId>
<artifactId>groovy-all</artifactId>
<version>2.4.15</version>
</dependency>
</dependencies>
Finally, run mvn test
from the command line to verify that specifications are found and execute successfully.
The Spock documentation outlines the features and methods that define a Spock specification (and their JUnit analogues, for those more familiar with Java-based testing), but the setupSpec
fixture method is of critical importance when testing AEM classes. This method, executed prior to the first feature method of the specification, is the conventional location for creating test content in the JCR. Likewise, the cleanup
and cleanupSpec
fixture methods are the appropriate place to remove test content following the execution of a test (or set of tests). However, the cleanupSpec
method will be implemented less frequently, as the base ProsperSpec
removes all test content from the JCR after every specification is executed to prevent cross-contamination of content between specifications.
class ExampleSpec extends ProsperSpec {
def setupSpec() {
// runs before first feature method, create test content and initialize shared resources here
}
def setup() {
// runs before each feature method, less commonly used
}
def "a feature method"() {
setup:
// create test-specific content, instances and/or mocks
expect:
// result
}
def "another feature method"() {
setup:
// create test content, etc.
when:
// stimulus
then:
// response
}
def cleanup() {
// optionally call the method below to remove all test content after each feature method
removeAllNodes()
}
def cleanupSpec() {
// runs after all feature methods, implemented in base spec but less commonly used
}
}
The base specification exposes a number of fields and methods for use in tests.
Field Name | Type | Description |
---|---|---|
session | javax.jcr.Session | Administrative JCR session |
resourceResolver | org.apache.sling.api.resource.ResourceResolver | Administrative Sling Resource Resolver |
pageManager | com.day.cq.wcm.api.PageManager | AEM Page Manager |
nodeBuilder | com.icfolson.aem.groovy.extension.builders.NodeBuilder | JCR Node Builder |
pageBuilder | com.icfolson.aem.groovy.extension.builders.PageBuilder | AEM Page Builder |
slingContext | com.icfolson.aem.prosper.context.SlingContextProvider | Prosper extension of Sling/OSGi Context |
See the ProsperSpec
GroovyDoc for details on available methods.
A test specification will often require content such as pages, components, or supporting node structures to facilitate the interactions of the class under test. Creating a large and/or complex content hierarchy using the JCR and Sling APIs can be tedious and time consuming. The base ProsperSpec
simplifies the content creation process by defining two Groovy builder instances, pageBuilder
and nodeBuilder
, that greatly reduce the amount of code needed to produce a working content structure in the JCR.
PageBuilder
creates nodes of type cq:Page
by default.
def setupSpec() {
pageBuilder.content {
beer { // page with no title
styles("Styles") { // create page with title "Styles"
"jcr:content"("jcr:lastModifiedBy": "mdaugherty", "jcr:lastModified": Calendar.instance) {
data("sling:Folder") // descendant of "jcr:content", argument sets node type instead of title
navigation("sling:resourceType: "components/navigation", "rootPath": "/content/beer")
}
dubbel("Dubbel")
tripel("Tripel")
saison("Saison")
}
// create a page with title "Breweries" and set properties on it's "jcr:content" node
breweries("Breweries", "jcr:lastModifiedBy": "mdaugherty", "jcr:lastModified": Calendar.instance)
}
}
}
Nodes named jcr:content
, however, are treated as unstructured nodes to allow the creation of descendant component/data nodes that are not of type cq:Page
. Descendants of jcr:content
nodes can specify a node type using a String
value as the first argument in the tree syntax (see /content/beer/styles/jcr:content/data
in the above example). As with page nodes, additional properties can be passed with a map argument.
This builder should be used when creating non-page content hierarchies, such as descendants of /etc
in the JCR. The syntax is similar to PageBuilder
, but the first argument is used to specify the node type rather than the jcr:title
property.
def setupSpec() {
nodeBuilder.etc { // unstructured node
designs("sling:Folder") { // node with type
site("sling:Folder", "jcr:title": "Site", "inceptionYear": 2014) // node with type and properties
}
}
}
The above example will create an nt:unstructured
(the default type) node at /etc
and sling:Folder
nodes at /etc/designs
and /etc/designs/site
. An additional map argument will set properties on the target node in the same manner as PageBuilder
.
Both builders automatically save the underlying JCR session after executing the provided closure.
In addition to the content builders, the session and pageManager instances provided by the base specification can be used directly to create test content in the JCR.
Another way to generate supporting content is to import a vault exported/packaged content structure. The content import is completely automatic and will run for all of your specs when it detects content to import. To take advantage of the content import, simply create a SLING-INF/content
directory within your project's test resources location (ex. src/test/resources/SLING-INF/content
). The content
directory must contain a child jcr_root
directory and META-INF
directory. The jcr_root
directory will contain all the vault exported/packaged content. The META-INF
directory will contain all the vault configuration XML files typically found within an AEM package.
You can specify an alternative filter.xml
file by using the class level com.icfolson.aem.prosper.annotations.ContentFilters
annotation. Simply provide the path to the filter.xml
file in the XML element and it will be used instead of the filter.xml
file within the META-INF/vault directory. The example below shows how you can provide a path to a non-default filter.xml
file.
@ContentFilters(
xml = "/SLING-INF/content/META-INF/vault/alt-filter.xml"
)
class MySpec extends ProsperSpec {
}
You can also dynamically generate spec-specific filters by using the various content filter annotations. This allows you to isolate the content you need for your individual specs. The example below shows how you can define a dynamic filter.
@ContentFilters(
filters = [
@ContentFilter(
root = "/etc",
mode = ImportMode.REPLACE,
rules = [
@ContentFilterRule(
type = ContentFilterRuleType.EXCLUDE,
pattern = "/etc/tags"
)
]
)
]
)
class MySpec extends ProsperSpec {
}
It is also possible to extend the provided filter.xml
file through dynamic filters. This allows you to provide common filters in the XML file and define specific filters for your spec with the annotations. The example below shows how you can extend an existing filter.xml
file.
@ContentFilters(
xml = "/SLING-INF/content/META-INF/vault/alt-filter.xml",
filters = [
@ContentFilter(
root = "/etc",
mode = ImportMode.REPLACE,
rules = [
@ContentFilterRule(
type = ContentFilterRuleType.EXCLUDE,
pattern = "/etc/tags"
)
]
)
]
)
class MySpec extends ProsperSpec {
}
In some cases, you may not want to import content for your spec. To skip the content import, annotate your spec class with @SkipContentImport
. The example below shows a spec that skips the content import.
@SkipContentImport
class MySpec extends ProsperSpec {
}
The AEM Groovy Extension decorates the com.day.cq.wcm.api.Page
, javax.jcr.Node
, and javax.jcr.Binary
classes with additional methods to simplify common operations. See the extension library Groovydocs for details of these additions. The metaclasses are registered automatically and available for use in all test methods.
Prosper's built-in assertion methods are used within Spock's then
and expect
blocks to verify the state of content in the transient repository following execution of a test. For example, a test that creates a node with property values (either directly or as a side effect of other operations) will want to confirm that the node was created and that the desired property name and values exist in the JCR.
Since expressions in these blocks are implicitly treated as boolean conditions by Spock, Prosper's assertion methods eliminate the need to logically combine expressions for the complex conditions required to assert JCR state. This is best illustrated with an example.
import com.day.cq.commons.jcr.JcrConstants
import com.day.cq.wcm.api.NameConstants
def "create content"() {
setup: "create a page with some properties"
def pageProperties = ["sling:resourceType": "foundation/components/page",
"jcr:description": "Prosper is an integration testing library for AEM."]
pageBuilder.content {
prosper("Prosper") {
"jcr:content"(pageProperties)
}
}
expect: "page is created and properties match expected values"
session.nodeExists("/content/prosper")
&& session.getNode("/content/prosper").primaryNodeType.name == NameConstants.NT_PAGE
&& session.getNode("/content/prosper").hasNode(JcrConstants.JCR_CONTENT)
&& pageProperties.every { name, value ->
session.getNode("/content/prosper").getNode(JcrConstants.JCR_CONTENT).get(name) == value
}
}
Thankfully, the expect
block can be simplified using an assertion method from the base ProsperSpec
.
expect: "page is created and properties match expected values"
assertPageExists("/content/prosper", pageProperties)
All available assert...
methods are detailed in the Prosper GroovyDocs.
Testing servlets and request-scoped supporting classes require mocking the SlingHttpServletRequest
and SlingHttpServletResponse
objects. The RequestBuilder
and ResponseBuilder
instances acquired through the base spec leverage Groovy closures to set the necessary properties and state on these mock objects with a minimal amount of initialization code.
Given a Sling servlet:
import groovy.json.JsonBuilder
import org.apache.sling.api.SlingHttpServletRequest
import org.apache.sling.api.SlingHttpServletResponse
import org.apache.sling.api.servlets.SlingAllMethodsServlet
import javax.jcr.Session
import javax.servlet.ServletException
import static com.google.common.base.Preconditions.checkNotNull
class TestServlet extends SlingAllMethodsServlet {
@Override
protected void doPost(SlingHttpServletRequest request,
SlingHttpServletResponse response) throws ServletException, IOException {
def path = checkNotNull(request.getParameter("path"))
def selector = request.requestPathInfo.selectors[1]
def session = request.resourceResolver.adaptTo(Session)
def node = session.getNode(path)
node.setProperty("testProperty", selector)
session.save()
new JsonBuilder([path: path, value: selector]).writeTo(response.writer)
}
}
A Prosper specification for this servlet will use the request and response builders to create the necessary method arguments to simulate a POST request to the servlet and verify both the JCR content updates and the JSON output resulting from the servlet execution.
class ServletSpec extends ProsperSpec {
def setupSpec() {
nodeBuilder.content {
prosper()
}
}
def "missing request parameter throws exception"() {
setup:
def servlet = new TestServlet()
// requestBuilder and responseBuilder are inherited from the base spec
def request = requestBuilder.build()
def response = responseBuilder.build()
when:
servlet.doPost(request, response)
then:
thrown(NullPointerException)
}
def "valid request parameter sets node property and returns JSON response"() {
setup:
def servlet = new TestServlet()
def request = requestBuilder.build {
parameterMap = [path: "/content/prosper"]
selectors = ["one", "two"]
contentType = "application/json"
}
def response = responseBuilder.build()
when:
servlet.doPost(request, response)
then:
assertNodeExists("/content/prosper", [testProperty: "two"])
and:
response.outputAsString == '{"path":"/content/prosper","value":"two"}'
}
}
The mock request and response objects delegate to the MockSlingHttpServletRequest and MockSlingHttpServletResponse objects from the Sling Mocks framework. The setter methods exposed by these classes are thus made available in the build
closures for the request and response builders to assist in setting appropriate mock values for the class under test.
See the Groovydoc for complete details of the available service registration and additional context methods; specific Sling adaptable examples are provided below.
Override the getAemContext()
method in a test specification to supply a custom Resource Resolver configuration or add context callbacks using the wcm.io AemContextBuilder.
class CustomContextSpec extends ProsperSpec {
@Override
AemContext getAemContext() {
new AemContextBuilder(ResourceResolverType.JCR_OAK)
.beforeSetUp(new AemContextCallback() {
@Override
void execute(AemContext context) throws Exception {
context.registerService(new CustomService())
}
})
.resourceResolverFactoryActivatorProps(["resource.resolver.mapping": ["/content/:/", "/-/"] as String[]])
.build()
}
def "test"() {
}
}
OSGi services can be mocked (fully or partially) using Spock's mocking API. Real service objects can also be used if all required dependency services are registered in the Sling context.
import com.day.cq.replication.ReplicationActionType
import com.day.cq.replication.ReplicationException
import com.day.cq.replication.Replicator
import groovy.util.logging.Slf4j
import org.apache.felix.scr.annotations.Reference
import org.apache.felix.scr.annotations.sling.SlingServlet
import org.apache.sling.api.SlingHttpServletRequest
import org.apache.sling.api.SlingHttpServletResponse
import javax.jcr.Session
import javax.servlet.ServletException
@SlingServlet(paths = "/bin/replicate/custom")
@Slf4j("LOG")
class CustomReplicationServlet extends SlingAllMethodsServlet {
@Reference
Replicator replicator
@Override
protected void doPost(SlingHttpServletRequest request, SlingHttpServletResponse response) throws
ServletException, IOException {
def path = request.getParameter("path")
def session = request.resourceResolver.adaptTo(Session)
try {
replicator.replicate(session, ReplicationActionType.ACTIVATE, path)
} catch (ReplicationException e) {
LOG.error("replication error", e)
}
}
}
The Prosper specification for this servlet can then set a mocked Replicator
instance and verify the expected interactions using the Spock syntax. Alternatively, services can inject other service references using the slingContext.registerInjectActivateService
methods.
def "servlet with mock service"() {
setup:
def servlet = new CustomReplicationServlet()
def replicator = Mock(Replicator)
slingContext.registerService(Replicator, replicator)
slingContext.registerInjectActivateService(servlet)
def request = requestBuilder.build {
parameters = [path: "/content"]
}
def response = responseBuilder.build()
when:
servlet.doPost(request, response)
then:
1 * replicator.replicate(_, _, "/content")
}
Specs can register adapters in the Sling context by adding AdapterFactory
instances or by providing mappings from adapter instances to closures that instantiate these instances from a Resource
, ResourceResolver
or SlingHttpRequestServlet
. Adapters will be registered with the mock BundleContext
and their adaptables and adapters properties will be respected when an adapter is chosen. Registered AdapterFactory
instances will pull these properties from the XML metadata files located in the classpath at /OSGI-INF, or the adaptable/adapter properites can be explicitly specified in the method arguments. Registered adapter closures will use the Resource
, ResourceResolver
or SlingHttpRequestServlet
as the adaptables property and the adapter instance class as the adapters property. The methods for registering adapters are illustrated in the examples below.
class ExampleAdapterFactory implements AdapterFactory {
@Override
<AdapterType> AdapterType getAdapter(Object adaptable, Class<AdapterType> type) {
AdapterType result = null
if (type == String) {
if (adaptable instanceof ResourceResolver) {
result = "Hello."
} else if (adaptable instanceof Resource) {
result = "Goodbye."
}
}
result
}
}
class ExampleSpec extends ProsperSpec {
def setupSpec() {
// example adapter factory
slingContext.registerAdapterFactory(new ExampleAdapterFactory(),
["org.apache.sling.api.resource.ResourceResolver", "org.apache.sling.api.resource.Resource"] as String[],
["java.lang.String"] as String[])
// resource adapters
slingContext.registerResourceAdapter(Integer, { Resource resource ->
resource.name.length()
})
slingContext.registerResourceAdapter(Map, { Resource resource ->
resource.metadata
})
// resource resolver adapters
slingContext.registerResourceResolverAdapter(Integer, { ResourceResolver resourceResolver ->
resourceResolver.searchPath.length
})
slingContext.registerResourceResolverAdapter(Node, { ResourceResolver resourceResolver ->
resourceResolver.getResource("/").adaptTo(Node)
})
// request adapters
slingContext.registerRequestAdapter(Integer, { SlingHttpServletRequest request ->
request.pathInfo.length()
})
}
def "resource is adaptable to multiple types"() {
setup:
def resource = resourceResolver.getResource("/")
expect:
resource.adaptTo(type) == result
where:
type | result
Integer | 0
Map | [:]
}
def "resource resolver is adaptable to multiple types"() {
expect:
resourceResolver.adaptTo(String) == "Hello."
resourceResolver.adaptTo(Integer) == 0
resourceResolver.adaptTo(Node).path == "/"
}
def "request is adaptable"() {
expect:
requestBuilder.build {
path = "/request/path.html"
}.adaptTo(Integer) == 18
}
}
Sling Model classes and injectors can also be registered in the Sling context for use in tests. The addModelsForPackage
method from the SlingContextProvider
scans the given package for classes annotated with @org.apache.sling.models.annotations.Model
and registers them in the mock bundle context, while the registerInjector
allows for registration of custom Sling injectors. A model instance can then be acquired by adapting from a Sling resource or request as shown below.
package com.icfolson.aem.prosper
import org.apache.sling.models.annotations.Model
import org.apache.sling.models.annotations.injectorspecific.Self
@Model(adaptables = [Resource, SlingHttpServletRequest])
class ProsperModel {
@Self
Resource resource
String getPath() {
resource.path
}
}
class ProsperModelSpec extends ProsperSpec {
def setupSpec() {
pageBuilder.content {
prosper()
}
}
def "adapt resource to model"() {
setup:
slingContext.addModelsForPackage("com.icfolson.aem.prosper")
def resource = getResource("/content/prosper")
def model = resource.adaptTo(ProsperModel)
expect:
model.path == "/content/prosper"
}
}
Many of the common AEM, JCR, and Sling namespaces and node types are registered when the Prosper test repository is created. Additional namespaces and node types may be added at runtime by annotating a test spec with the @NodeTypes
annotation and supplying an array containing paths to classpath .cnd file resources. For more information on the CND node type notation, see Node Type Notation in the Apache Jackrabbit documentation. An example of the annotation usage is presented below.
import com.icfolson.aem.prosper.annotations.NodeTypes
@NodeTypes("SLING-INF/nodetypes/spock.cnd")
class ExampleSpec extends ProsperSpec {
}
Traits are a Groovy language feature that can be utilized to "mix in" new functionality to test specs. Custom traits can be defined to support domain-specific features.
import com.icfolson.aem.prosper.builders.RequestBuilder
import org.apache.sling.api.SlingHttpServletRequest
trait MobileRequestTrait {
abstract RequestBuilder getRequestBuilder()
SlingHttpServletRequest buildMobileRequest(Map<String, Object> parameters) {
requestBuilder.setSelectors(["mobile"]).setParameterMap(parameters).build()
}
}
Specs can then implement the trait to add the new functionality. Note that the getRequestBuilder()
abstract method does not have to be implemented by the new spec, since this spec (as seen below) already inherits the required method from ProsperSpec
. Traits can thus "borrow" functionality from the base ProsperSpec
by defining abstract methods that match the corresponding method signatures in ProsperSpec
.
import spock.lang.Shared
class MobileRequestTraitSpec extends ProsperSpec implements MobileRequestTrait {
def "trait usage"() {
setup:
def request = buildMobileRequest([:])
expect:
request.requestPathInfo.selectors[0] == "mobile"
}
}
Follows Semantic Versioning guidelines.