Content Application Developer Manual / Version 2010
Table Of ContentsThe Content Application Engine comes with a source code generator that reads your content type definition file and generates beans for each content type. The bean interfaces mirror the content properties, and the class hierarchy mirrors the content type hierarchy. After generating the code, you can modify and adapt it to your application’s needs.
Since you need the code generator only once (or even not at all, if you start over with CoreMedia Blueprint), there is no need for a dedicated application. Just add a temporary dependency
<dependency> <groupId>com.coremedia.cms</groupId> <artifactId>bean-generator</artifactId> <scope>runtime</scope> </dependency>
to the POM file of your
CoreMedia CAE
module and
build it. Then the code generator can be invoked as follows from within the
WEB-INF
directory of the
CoreMedia CAE
:
java -jar lib/bean-generator-<version>.jar
This gives you all available options of the generator as shown in
Table 4.1, “Parameters of the Beangenerator”. You need to pass the location of your
content type definition file(s), a package for the generated classes and a target
directory for the source code. By default, the generator creates code with generics for link
lists as comments in order to be compatible with legacy Java versions. When you have
successfully created your content beans, you should delete the temporary dependency on
bean-generator
again so that your actual web application does not get
unnecessarily bloated.
The meaning of the parameters of the Beangenerator is described in the next table.
Parameter | Variable | Description |
---|---|---|
-d --document-types
| Name of the document types file to use | The URL of the document types file to use. You need at least one file. |
-baseonly
| Create only the base classes, but no implementations and no interfaces. | |
-beanmapping
| Name of a file | The file into which the bean mapping will be written (see Section 4.1.1.2, “Spring Configuration”). |
-c
| Name of a properties file that contains the package information. |
Define specific packages for specific document types in order to organize your bean files properly. There a three possible configuration types: One package for all files of a document type Example:
Different packages for the different files of a document type Example:
Settings for all document types Example:
|
-document
| Document type |
Generate the code for the defined content type. It must be included in one of the document type files
defined with the
-d
parameter.
|
-e, --exclude
| Comma separated list of document types | Exclude these types from generation process. |
-enc, --encoding
| One of the defined Java encodings | Encode the output files with the defined encoding (default is Cp1252) |
-f
| Overwrite existing files | |
-generics
| Use Java 1.5 Generics. | |
-gensetter
| Deprecated parameter. Generates code with public setter methods. | |
-avoidlist
| Generate methods with simple (non-list) parameters and return values for linklist types that have a maximum cardinality of '1'. | |
-i, --include
| Comma separated list of document types | Document types that should be included in the generation process (default is all) |
-noimpl
| Only create base classes and interfaces but no implementations | |
-o, -out
| Directory name | The directory where to generate the source code into. If it does not exist, it will be created. |
-p, --package
| Package name |
The package name used for the generated beans. A setting defined with
-c
overwrites
the package defined with -p . To avoid confusion only use
-p
or -c .
|
Table 4.1. Parameters of the Beangenerator
Example
If you use the following call from the lib directory of your content application web application,
java -jar beangenerator.jar -d /contentserver/config/contentserver/doctypes/menu-doctypes.xml -o /output/classes -c /properties/package.properties -beanmapping /properties/contentbeans.xml -e Picture -generics
with the following
package.properties
file,
doctype.Dish.package=com.menusite.dish package=com.menusite
this will generate interfaces, base classes and implementation classes for all document
types of the menu site example except for the
Picture
document type. The
document types are read from the
/contentserver/config/contentserver/doctypes/menu-doctypes.xml
file and the
generated classes are written to the
/output/classes
directory.
The Dish document type belongs to the
com.menusite.dish
package while the other
types are in the
com.menusite
package, this information is taken from the
/properties/package.properties
file. The bean mapping is written to the
/properties/contentbeans.xml
file. Generics are used which results in code like
the following:
public List<? extends Dish> getContent() { List<Content> contents = getContent().getLinks("content"); return createBeansFor(contents, Dish.class); }
Structure of the Generated Code
When you inspect the generated classes, you will find that three files per document type are generated:
An interface with the same name as the document type
An abstract class ending with “Base” and
A concrete class ending with “Impl"
The interface is what you should use in other classes, “*Base” contains the repository access code and “*Impl” is the actual class that is instantiated. This class is the place for you to modify. When a document type inherits from another type, its "*Base" class inherits the "*Impl" class of its parent. This way, it inherits the custom extensions made for the supertype. For content types that do not have a parent, the "*Base" class inherits from a framework class AbstractContentBean that defines the underlying content bean, factory, equality and hash code as well as a few convenience methods.
Note
The generated code may not compile under some circumstances, due to naming conflicts, for
example. A content property named ‘content’ will clash with the method
#getContent
inherited from
ContentBean.
In this case you should rename the generated getters in the interface and the
*Base
class.
The "*Base" class contains property getters for every user-defined property in the corresponding content type. Getters are not generated for metadata such as name or creation date. The property types are mapped to Java as follows:
Property Type | Java Type | Conversion |
---|---|---|
IntProperty
| int | Simply the value from the underlying content object |
StringProperty
| String | Simply the value from the underlying content object |
DateProperty
| Calendar | Simply the value from the underlying content object |
XmlProperty
(with grammar "coremedia-struct-2008")
| Struct |
The parsed
Struct
value from the underlying content object
|
XmlProperty
| Markup | The markup is transformed. Every internal xlink to a document or blob is transformed into the corresponding content bean id or blob id. |
BlobProperty
|
CapBlobRef
|
This is the result of
#getBlobRef
of the underlying content object
|
LinkListProperty
| List | Every content object in the link list is converted to a bean through the content bean factory |
Table 4.2. Document property to Java property mappings
You are free to modify the generated code. For example, you can do the following:
Add a new method to both the public interface and the "*Impl" class
Remove a method from the public interface of a bean. It will still be available for the implementation classes but not for clients.
Combine both to implement a derived property based on the now hidden content property, possibly with a different result type
Implement additional interfaces that may crosscut the document type hierarchy
When extending, you should not declare any fields in a content bean except for read-only, immutable service references.