Field Detail

• public static final java.lang.String ENTITY_CLASS_RESOURCE_PATTERN

• public static final java.lang.String OSIV_CLASS_NAME

• public static final java.lang.String TRANSACTION_MANAGER_BEAN

• public static final java.lang.String WEB_APPLICATION_CONTEXT_CLASS_NAME

• protected java.lang.ClassLoader classLoader

• protected boolean secondaryDatastore

Property Detail

• PropertyResolver configuration

• java.lang.Object originalConfiguration

• java.util.Collection<java.lang.String> packages

• java.util.Collection<java.lang.Class> persistentClasses

• boolean registerApplicationIfNotPresent

• PathMatchingResourcePatternResolver resourcePatternResolver

Constructor Detail

• AbstractDatastoreInitializer()

• AbstractDatastoreInitializer(java.lang.ClassLoader classLoader, java.lang.String[] packages)

• AbstractDatastoreInitializer(java.lang.String[] packages)

• AbstractDatastoreInitializer(java.util.Collection<java.lang.Class> persistentClasses)

• AbstractDatastoreInitializer(java.lang.Class[] persistentClasses)

• AbstractDatastoreInitializer(PropertyResolver configuration, java.util.Collection<java.lang.Class> persistentClasses)

• AbstractDatastoreInitializer(PropertyResolver configuration, java.lang.Class[] persistentClasses)

• AbstractDatastoreInitializer(PropertyResolver configuration, java.lang.String[] packages)

• AbstractDatastoreInitializer(java.util.Map configuration, java.util.Collection<java.lang.Class> persistentClasses)

• AbstractDatastoreInitializer(java.util.Map configuration, java.lang.Class[] persistentClasses)

Method Detail

• protected java.util.Collection<java.lang.Class> collectMappedClasses(java.lang.String datastoreType)

  The classes this datastore maps.

  Parameters:

  datastoreType - the datastore the classes are being collected for

  Returns:

  the classes, which a datastore's definition takes through mappedClasses

• protected java.lang.Object configurationReference(BeanDefinitionRegistry registry)

  What a bean definition should hold in place of the configuration itself.

  A definition holding the resolver is a definition holding everything the resolver can reach, and generating code for such a definition writes those values out: an environment carries the machine's own variables, so the generated source ends up containing the environment of whatever machine built it, along with whatever was in it -- credentials among them -- and the application then reads its settings from there rather than from where it runs.

  So while code is being generated, and only then, the context's environment is named instead, which leaves the lookup to be made where the application runs. Every other time the resolver is held as it always was, which matters for a datastore brought up on its own: its configuration is whatever the caller passed and there is no environment holding it.

• @groovy.transform.CompileStatic
  ApplicationContext configure()

  Configures for an existing Mongo instance

  Parameters:

  mongo - The instance of Mongo

  Returns:

  The configured ApplicationContext

• @groovy.transform.CompileStatic
  void configureForBeanDefinitionRegistry(BeanDefinitionRegistry beanDefinitionRegistry)

  Configures an existing BeanDefinitionRegistry

  Parameters:

  beanDefinitionRegistry - The BeanDefinitionRegistry to configure

• @groovy.transform.CompileDynamic
  protected boolean containsRegisteredBean(java.lang.Object builder, BeanDefinitionRegistry registry, java.lang.String beanName)

• ApplicationEventPublisher findEventPublisher(BeanDefinitionRegistry beanDefinitionRegistry)

  Finds the event publisher to use

  Parameters:

  beanDefinitionRegistry - The event publisher

  Returns:

  The event publisher

• protected java.lang.Class<? extends ApplicationEventPublisher> findEventPublisherClass(BeanDefinitionRegistry beanDefinitionRegistry)

  Finds the publisher class a datastore definition should name, in place of a publisher itself.

  The same choice findEventPublisher makes -- one that publishes through the surrounding context, or the one that publishes nowhere -- but left to the container to build. A definition holding an already-constructed publisher is a definition that cannot be generated ahead of time: generating one means writing out what it holds, and a publisher bound to a running context is not something there is any way to write down.

  Parameters:

  beanDefinitionRegistry - the registry the definitions are being contributed to

  Returns:

  the class to register the publisher bean under

• MessageSource findMessageSource(BeanDefinitionRegistry beanDefinitionRegistry)

  Finds the message source to use

  Parameters:

  beanDefinitionRegistry - The registry

  Returns:

  The message source

• @groovy.transform.CompileDynamic
  groovy.lang.Closure getAdditionalBeansConfiguration(BeanDefinitionRegistry registry, java.lang.String type)

  Internal method aiding in datastore configuration.

  Parameters:

  registry - The BeanDefinitionRegistry

  type - The type of the datastore

  Returns:

  A closure containing bean definitions

• abstract groovy.lang.Closure getBeanDefinitions(BeanDefinitionRegistry beanDefinitionRegistry)

• @groovy.transform.CompileDynamic
  groovy.lang.Closure getCommonConfiguration(BeanDefinitionRegistry registry, java.lang.String type)

• @groovy.transform.CompileStatic
  protected java.lang.Class getGrailsApplicationClass()

• @groovy.transform.CompileStatic
  protected java.lang.Class getGrailsValidatorClass()

• protected abstract java.lang.Class<AbstractDatastorePersistenceContextInterceptor> getPersistenceInterceptorClass()

  Returns:

  The class used to define the persistence interceptor

• protected boolean isGrailsPresent()

• protected boolean isMappedClass(java.lang.String datastoreType, java.lang.Class cls)

• protected boolean isWebApplicationRegistry(BeanDefinitionRegistry registry)

  Determines whether the given registry belongs to a web application. The presence of the dispatcherServlet bean definition is only a reliable signal after Spring Boot auto-configuration has been processed; Grails plugin bean definitions register before that, so the type of the registry (the web application context itself) is checked as well.

  Parameters:

  registry - The bean definition registry

  Returns:

  true if the registry belongs to a web application

• @groovy.transform.CompileDynamic
  protected java.util.Map<java.lang.String, java.lang.Class<?>> loadDataServices(java.lang.String secondaryDatastore)

• protected java.lang.Class[] mappedClasses(java.lang.String datastoreType)

  The classes this datastore maps, as the array its constructor takes.

  The array type is what makes this survive ahead-of-time processing. A datastore takes its classes as a variable-argument parameter, so the constructor argument the generator writes out is looked up by that parameter's type when the bean is rebuilt. A collection does not answer to Class[], so the lookup misses it and the argument is resolved as a dependency instead, which for an array type means every Class bean in the context: none. The datastore is then built mapping nothing, and the first call on a domain class reports that it is not one.

  Kept apart from collectMappedClasses rather than folded into it, because that one is the seam a datastore outside this repository overrides to say which classes are its own. Changing what it returns would leave such an override no longer overriding anything, and silently unasked.

  Parameters:

  datastoreType - the datastore the classes are being collected for

  Returns:

  the classes, as the array a datastore's constructor takes

• @groovy.transform.CompileStatic
  protected void scanForPersistentClasses()

• @java.lang.Override
  void setResourceLoader(ResourceLoader resourceLoader)

• void setSecondaryDatastore(boolean secondaryDatastore)

  Whether this datastore is secondary to another primary datastore (example the SQL DB)

  Parameters:

  secondaryDatastore