The container bootstrap sequence begins with initializing extensions. The container performs the following actions:
The container fires the
BeforeBeanDiscovery event which allows extensions to register scopes and stereotypes.
The container combines the registered scopes and stereotypes with scopes associated with the built-in contexts and built-in
stereotypes and makes the resulting set available through
TypeDiscoveryConfiguration#getKnownBeanDefiningAnnotations()
Afterwards, an EE7-compatible integrator performs type discovery in the following locations:
- Library jars
- EJB jars or application client jars
- The WEB-INF/classes directory of a war
- Directories in the JVM classpath
These locations are referred to as available archives hereafter.
Next, the integrator builds the set of bean defining annotations. The set contains:
- The set of annotations returned from
TypeDiscoveryConfiguration#getKnownBeanDefiningAnnotations()
- Every Java annotation discovered by the integrator that is annotated with
NormalScope or
Stereotype
The resulting set is referred to as bean defining annotations hereafter.
Next, the integrator processes available archives according to these rules. The rules are exclusive.
If an archive:
- contains the
beans.xml
file and the file does not contain the bean-discovery-mode
attribute, or
- contains the
beans.xml
file and the value of the bean-discovery-mode
attribute is set to
all
this archive is an explicit bean archive.
If an archive:
- contains the
beans.xml
file and the bean-discovery-mode
attribute is set to
annotated
, or
- does not contain the
beans.xml
file nor any implementation of the
Extension interface but the
archive contains types annotated with a bean defining annotation or session beans
this archive is an implicit bean archive.
If an archive:
- contains the
beans.xml
file and the bean-discovery-mode
attribute is set to
none
, or
- does not contain the
beans.xml
file and contains an implementation of the
Extension interface,
or
- does not contain the
beans.xml
file and does not contain any types annotated with a bean defining
annotation nor session bean
this archive is not a bean archive.
For each explicit bean archive the integrator creates an instance of
BeanDeploymentArchive representing this
archive. The
BeanDeploymentArchive#getBeanClasses() or
BeanDeploymentArchive#getLoadedBeanClasses() method returns a collection of all types present within the
archive. Filtering rules defined in
BeansXml#getScanning() are not required to be applied by the integrator and
are applied later on by Weld. The
BeanDeploymentArchive#getEjbs() method returns a collection of EJB descriptors
for EJBs present in the archive.
For each implicit bean archive the integrator creates an instance of
BeanDeploymentArchive representing this
archive. The
BeanDeploymentArchive#getBeanClasses() or
BeanDeploymentArchive#getLoadedBeanClasses() method of the bean archive returns all the types found in the
archive which are annotated with a bean defining annotations or are Session bean definitions. Filtering rules (
BeansXml#getScanning()) are not required to be applied by the integrator. The
BeanDeploymentArchive#getEjbs() method returns a collection of EJB descriptors for Session beans present in the
archive.
Initially, the integrator does not need to create a
BeanDeploymentArchive instance for an archive which is not a
bean archive. Note that although these archives are ignored when building the initial collection of bean archives, the
integrator may be required to create a
BeanDeploymentArchive later upon a call to
Deployment#loadBeanDeploymentArchive(Class). For example, when the method is called for a extension class that is
deployed in an archive that is not a bean archive.