I'm trying to get into unit tests for a BizTalk application I'm working on, following the example in Michael Stephensons blog post and seemed to be getting somewhere
Then I got an error down the line, which I tracked back to a "invalid" XML test file I was using, but this was passing my validation against schema unit test ...
- reason being incorrect namespace
My puzzlement is why does the XmlReader think the XML is valid vs. the schema, but if I use the BizTalk IDE "Validate Instance" option I get the errors ...
... error BEC2004: Validate Instance failed for schema FromFrontOffice.xsd, file: ...
XmlSchema schema = XmlSchema.Read(schemaStream, null);
XmlReaderSettings xmlReaderSettings = new XmlReaderSettings();
xmlReaderSettings.Schemas.Add(schema);
xmlReaderSettings.ValidationType = ValidationType.Schema;
xmlReaderSettings.ValidationEventHandler += ValidationEventHandler;
XmlReader xmlReader = XmlReader.Create(xmlStream, xmlReaderSettings);
while (xmlReader.Read())
private void ValidationEventHandler(object sender, ValidationEventArgs args)
{
if (args.Exception == null) return;
_IsValid = false;
}
Think I've sorted it ... trick seems to be using ValidationFlags
xmlReaderSettings.ValidationFlags =
XmlSchemaValidationFlags.ReportValidationWarnings |
XmlSchemaValidationFlags.ProcessIdentityConstraints |
XmlSchemaValidationFlags.ProcessInlineSchema |
XmlSchemaValidationFlags.ProcessSchemaLocation;
Related
I have written the code using camel-sql which is working fine. Now I have to write test cases for the same. I have used in-memory database H2. I have initialized the database and assigned the datasource to sqlComponent.
// Setup code
#Override
protected JndiRegistry createRegistry() throws Exception {
JndiRegistry jndi = super.createRegistry();
// this is the database we create with some initial data for our unit test
database = new EmbeddedDatabaseBuilder()
.setType(EmbeddedDatabaseType.H2).addScript("createTableAndInsert.sql").build();
jndi.bind("myDataSource", database);
return jndi;
}
// Testcase code
#SuppressWarnings("unchecked")
#Test
public void testRoute() throws Exception {
Exchange receivedExchange = template.send("direct:myRoute", ExchangePattern.InOut ,exchange -> {
exchange.getIn().setHeader("ID", new Integer(1));
});
camelContext.start();
MyClass updatedEntity = (MyClass)jdbcTemplate.queryForObject("select * from MY_TABLE where id=?", new Long[] { 1l } ,
new RouteTest.CustomerRowMapper() );
// Here I can get the updatedEntity from jdbcTemplate
assertNotNull(receivedExchange);
assertNotNull(updatedEntity);
}
// Main code
from("direct:myRoute")
.routeId("pollDbRoute")
.transacted()
.to("sql:select * from MY_TABLE msg where msg.id = :#"+ID+"?dataSource=#myDataSource&outputType=SelectOne&outputClass=com.entity.MyClass")
.log(LoggingLevel.INFO,"Polled message from DB");
The problem is, as soon as the test case starts, it is saying
No bean could be found in the registry for: myDataSource of type: javax.sql.DataSource
I looked into camel-SQL component test cases and doing the same thing but the code is not able to find dataSource. Please help. Thanks in advance.
After spending a lot of time on this issue, I identified that H2 database was using JDBCUtils to fetch records and It was throwing ClassNotFoundException. I was getting it nowhere in Camel exception hierarchy because this exception was being suppressed and all I was getting a generic exception message. Here is the exception:
ClassNotFoundException: com.vividsolutions.jts.geom.Geometry
After searching for the issue I found out that It requires one more dependency. So I added it and it resolved the issue.
Issue URL: https://github.com/elastic/elasticsearch/issues/9891
Dependency: https://mvnrepository.com/artifact/com.vividsolutions/jts-core/1.14.0
Lets say i have a domain A which has a custom validator for property P
static constraints = {
P validator: { val, obj ->
(A.executeQuery("SELECT COUNT(*) FROM A cei WHERE cei.event.id = ?", [val.id])[0] <= 1)
}
In the unit test how can i mock the P property of domain A so that i don't get the error i am getting when running unit test. The error i get when running unit test is shown below. The setup code instantiates domain A as shown below.
void setUp(){
inv = new A(P: rg).save(flush: true)
Error is
java.lang.UnsupportedOperationException: String-based queries like [executeQuery] are currently not supported in this implementation of GORM. Use criteria instead.
Note: These are fake code.
The best option for me, is encapsulate the inner code of the validator inside a service. And then, mock that service in your unit test.
But the error thrown is that executeQuery is not available, but criteria is.
So change your code for using a criteria instead.
P validator: { val, obj ->
A.withCriteria{
eq('event', Event.findById(val.id))
projections{
count('id')
}
}[0] < = 1
}
I have been trying to figure this out for 2 days now and I am really stuck and frustrated. I have a domain object with a service which is being used for custom validation. The domain looks like this:
class Llama {
String name
transient myFetcherService
static transients = [
'myFetcherService'
]
static constraints = {
name validator: { val, obj ->
if (obj.nameExists(val) == true) {
//return some error here.
}
}
}
protected boolean nameExists(String name) {
List<Llama> llamasList = myFetcherService.fetchExistingLlamasByName(name)
if (llamasList.isEmpty()) {
return false
}
return true
}
}
Now, I have another Service, which simply saves a list of Llama objects. It looks like this:
class LlamaFactoryService {
public void createLlamas(List<String> llamaNames) {
llamaNames.each { name ->
new Llama(name: name).save()
}
}
}
In my test. I keep getting this error:
Failure: createLlamas should create Llammas (com.myLlamaProject.LlamaFactoryServiceSpec)
| java.lang.NullPointerException: Cannot invoke method myFetcherService on null object
I don't understand. In my tests, added a metaClass for the service in the "given" section. When it tries to save, it's telling that the service is null. This is what my test looks like:
given:
def myFetcherService = mockFor(MyFetcherService)
myFetcherService.demand.fetchExistingLlamasByName {def name -> return []}
Llama.metaClass.myFetcherService = myFetcherService.createMock()
when:
service.createLlamas(['Pablo','Juan','Carlos'])
then:
//some validations here....
I also tried using metaClass on the method nameExists() like:
Llama.metaClass.myFetcherService = { def name -> false }
, but it gives me the same nullPointerException as the one above. Could someone point me to the right direction? I'm a bit stuck. :(
Thanks in advance for reading and helping.
You're using a unit test and the general rule for unit tests is that beans generally aren't created for you, so you'll need to inject them yourself.
(Code edited to reflect the fact I misread the question)
I think you want a testing pattern something like:
given:
def mockMyFetcherService = Mock(MyFetcherService) // create the mock
Llama.metaClass.getMyFetcherService = { mockMyFetcherService } // inject the dependency
def returnList = [] // let's just define this here and you can re-use this pattern in other tests more easily
when:
service.createLlamas(['Pablo','Juan','Carlos'])
then:
// tell Spock what you expect to have happen with your Mock - return the List you defined above
3 * mockFetcherService.fetchExistingLlamasByName(_) >> returnList
If the injection of the service into the metaClass doesn't work (suggested here), you could always try using the defineBeans{} closure within the unit test itself (http://www.block-consult.com/blog/2011/08/17/inject-spring-security-service-into-domain-class-for-controller-unit-testing/).
Thus you could try:
defineBeans {
myFetcherService(MockMyFetcherService)
}
where MockMyFetcherService is defined in the same file that defines the test. This is the approach followed here:
See here for examples of more Spock interactions.
If you're using Grails 2.4.3 or below you'll need to put CGLIB in BuildConfig.groovy but I see here that it's already done for you in 2.4.4, so you should be ok just to use Mock(classname).
I'm trying to utilize the build-test-data plugin in my Grails (v2.4.3) app to assist with test data creation for unit testing, but while running my unit tests the plugin cannot find TestDataConfig.groovy to load my specified values (for unique constraint tests, etc).
I've installed the plugin via including it in BuildConfig.groovy:
plugins {
...
test ":build-test-data:2.2.0"
...
}
I've ran the following command to create the TestDataConfig.groovy template, which places the file at \grails-app\conf\:
grails install-build-test-data-config-template
I've followed the general instructions on the plugin wiki to come up with a properly formatted file:
testDataConfig {
sampleData {
'<path>.User' {
def a = 1
username = { -> "username${a++}" }
}
}
}
(Where path is the fully-qualified class name.)
In my tests, I am using the following general format:
import grails.buildtestdata.TestDataConfigurationHolder
import grails.buildtestdata.mixin.Build
import grails.test.mixin.TestFor
import spock.lang.Specification
import spock.lang.Unroll
#TestFor(User)
#Build(User)
class UserSpec extends Specification {
def setup() {
mockForConstraintsTests(User)
TestDataConfigurationHolder.reset()
user = User.buildWithoutSave()
}
#Unroll
void "test #field must be unique"() {
given: 'a User exists'
user.save(flush: true)
when: 'another User is created with a non-unique field value'
def nonUniqueUser = User.buildWithoutSave()
nonUniqueUser."$field" = user."$field"
then: 'validation fails and the field has a unique constraint error'
!nonUniqueUser.validate()
nonUniqueUser.errors.errorCount == 1
nonUniqueUser.errors.getFieldError("$field").code == 'unique'
where:
field << ['username', '<some other field>']
}
}
But, when the test is run (using IntelliJ IDEA) TestDataConfig.groovy cannot be found via the following method in the plugin:
static Class getDefaultTestDataConfigClass() {
GroovyClassLoader classLoader = new GroovyClassLoader(TestDataConfigurationHolder.classLoader)
String testDataConfig = Holders.config?.grails?.buildtestdata?.testDataConfig ?: 'TestDataConfig'
try {
return classLoader.loadClass(testDataConfig)
} catch (ClassNotFoundException ignored) {
log.warn "${testDataConfig}.groovy not found, build-test-data plugin proceeding without config file"
return null
}
}
So the test continues on without a config file and I do not get uniquely generated data sets.
I've even tried explicitly including the file in Config.groovy:
grails.buildtestdata.testDataConfig = "TestDataConfig"
But, the same method in the plugin shows that Holders.config? is null.
I've looked at a few solutions to a similar problem here on StackOverflow with nothing working in my case; I cannot figure out how to get my app to detect the presence of the TestDataConfig.groovy file.
Any ideas? Thanks so much!
I have a unit test that I am trying to get the validation to succeed on.
Basically, when I call badMockedSecureFile.validate(), it is not doing what I expect, and that is failing validation for the two fields, encryptedFileName and encryptedFileData.
When I break in the debugger, I simply get a null value for badMockedSecureFile.errors in the subsequent assert. Here are my two files:
Any input would be greatly appreciated. I couldn't find an exact similar question. I'm using grails 2.2.4 with Oracle JDK 1.7.0_25 if that matters any.
EDIT: I just wanted to note that I removed the mockForConstraintTests call and it seems to be working now. I get this feeling I didn't RTFM somewhere and this behaviour changed in unit testing, or is something else going on?
SecureFile.groovy
class SecureFile implements Serializable {
/**
* An unencrypted version of the file name. This file name is unencrypted
* when the appropriate password and key combo is used and it is never
* persisted to the database for security (see transients below).
*/
String fileName
/**
* Unencrypted version of the file data. Never persisted to the
* database for security (see transients below).
*/
byte[] fileData
String encryptedFileName
byte[] encryptedFileData
Date dateAdded
Date dateUpdated
Date dateDeleted
static constraints = {
encryptedFileName(nullable: false, blank: false)
encryptedFileData(nullable: false)
}
static transients = ["fileName", "fileData"]
static belongsTo = [user: User]
}
SecureFileTests.groovy
import static org.junit.Assert.*
import grails.test.mixin.*
import grails.test.mixin.support.*
import org.junit.*
/**
* See the API for {#link grails.test.mixin.support.GrailsUnitTestMixin} for usage instructions
*/
#TestFor(SecureFile)
class SecureFileTests {
static final String SAMPLE_PDF_FILE = "fileEncryptionTestSample.pdf"
void testConstraints() {
def samplePdfFile = new FileInputStream(SAMPLE_PDF_FILE)
// Not really encrypted for this mock.
def mockedSecureFile = new SecureFile(
encryptedFileName: "--Not-Really-Encrypted--",
encryptedFileData: samplePdfFile.getBytes()
)
mockForConstraintsTests(SecureFile, [mockedSecureFile])
// Validation should fail if both properties are null.
def badMockedSecureFile = new SecureFile()
assert !badMockedSecureFile.validate()
assert "nullable" == badMockedSecureFile.errors["encryptedFileName"].code
assert "nullable" == badMockedSecureFile.errors["encryptedFileData"].code
}
}
Remove code from badMockedSecureFile.errors["encryptedFileName"].code .
You will get as you expected.
mockForConstraintsTests(SecureFile)
// Validation should fail if both properties are null.
def badMockedSecureFile = new SecureFile()
assert !badMockedSecureFile.validate()
assert "nullable" == badMockedSecureFile.errors["encryptedFileName"]
assert "nullable" == badMockedSecureFile.errors["encryptedFileData"]