System and method for generating and reusing software application code with source definition files

The following related documents are hereby incorporated by reference in their entirety: BEAW EB L OGIC S ERVER ®: A SSEMBLING AND C ONFIGURING W EB A PPLICATIONS (Release 7.0, Jun. 28, 2002); BEAW EB L OGIC S ERVER ®: P ROGRAMMING W EB L OGIC E NTERPRISE J AVA B EANS ® (Release 7.0, Jun. 28, 2002); BEAW EB L OGIC S ERVER ®: D EVELOPING W EB L OGIC S ERVER A PPLICATIONS (Release 7.0, Jun. 28, 2002); BEAW EB L OGIC S ERVER ®: W EB L OGIC B UILDER O NLINE H ELP (Release 7.0, Jun. 28, 2002); BEAW EB L OGIC S ERVER ®: P ROGRAMMING WEBL OGIC XML (Release 7.0, Jun. 28, 2002); BEAW EB L OGIC S ERVER ®: P ROGRAMMING WEBL OGIC W EB S ERVICES (Release 7.0, Jun. 28, 2002); BEA WebLogic Server®: Programming WebLogic E NTERPRISE J AVA B EANS ® (Release 7.0, Jun. 28, 2002).

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

The present invention disclosure relates to the field of automatic code generation, in particular, automatic generation of Enterprise Java™ Bean source code files from an annotated class definition and the reverse.

Enterprise Java™ Beans are reusable software components written in the Java™ programming language. An Enterprise Java™ Bean (EJB) is typically declared in a number of separate source files which define various aspects of the EJB, such as its class, interfaces and deployment behavior. Each time a modification is made to one EJB source file, the other associated EJB source files must also be modified accordingly. This activity is cumbersome and error prone, since the changes must be manually propagated to different files. What is needed is a way to generate source files for EJB's with minimal editing and duplication of information.

FIG. 1 is a block diagram of an EJB code generation system in accordance to one embodiment of the invention.

FIG. 2 is a block diagram of a reverse EJB code generation system in accordance to one embodiment of the invention.

The invention is illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “an” or “one” embodiment in this disclosure are not necessarily to the same embodiment, and such references mean at least one.

In one embodiment, instead of having to edit and maintain several EJB files for each EJB (e.g., the bean class, remote and home classes, deployment descriptors, etc.), editing is limited to a single file that contains an annotated description of one or more EJB class definitions. EJB's are declared in the Java™ programming language. The Java™ programming language and run-time environment are available from Sun Microsystems, Inc., of Palo Alto, Calif. The annotated description file is processed in accordance to one embodiment and the EJB files required for deployment on a server are automatically generated. Annotations are directives that contain information needed to create an EJB's source files. In one embodiment, annotations are interspersed between Java™ source code statements. In another embodiment, annotations are embedded in Java™ comments. In yet another embodiment, an annotation is a Javadoc tag. Javadoc is the tool for generating source code documentation in hypertext markup language (“HTML”) format from tags embedded in Java™ source code comments. Javadoc is available from Sun Microsystems, Inc. In another embodiment, annotations are provided separately from the description of one or more EJB classes. For example, annotations can be provided in a graphical user interface, a separate file, or created dynamically based on run-time properties a user has associated with an EJB.

In one embodiment, by way of example, annotations can take the form: @ejbgen-name [attribute[=expression]] . . . attribute[=expression]], where items in square brackets are optional arguments. Name can be the type of annotation and the optional attribute=expression list can be comprised of one or more attribute-expression pairs. In one embodiment, attribute is the name of a required or optional parameter associated with an annotation and the optional expression can be, for example, an infix, postfix or prefix expression that evaluates to a constant (e.g., a number or an alpha-numeric string). In another embodiment, an expression's operators can be arithmetic, logical and string-based. For example, the following Java™ code segment has a comment block that contains three annotations:

Conventionally, an EJB's abstract accessor methods define container-managed persistence (“CMP”) data fields of the same name. For example, in the code segment above, assuming the accessor method getAccountId( ) was an EJB method, a corresponding data field called “accountId” would be defined upon deployment of the EJB. In one embodiment, the annotation “@ejbgen:cmp-field” specifies that the Accountld CMP field will be stored in column named “acct_id” of a database upon deployment of the EJB. The annotation “@ejbgen:primkey-field” indicates that Accountld will be a primary key in the database. Finally, the annotation “@ejbgen:remote-method” specifies that transaction isolation is required for the method getAccountId( ). All three annotations can be considered method annotations since they come before the declaration of a method.

FIG. 1 is a block diagram in accordance to one embodiment of the invention. An annotated description 1 is processed by analyzer 2 . Annotated description 1 can include one or more EJB definitions. In one embodiment, annotated description 1 can reside in a computer file. In another embodiment, annotated description 1 can reside in a computer memory. In yet a further embodiment, annotated description 1 can be transmitted via a computer network (not shown) to analyzer 2 . Analyzer 2 employs parser 4 to divide the Java™ source text in annotated description 1 into discrete units and extract annotations therefrom. Analyzer 2 may perform preprocessing of annotated description 1 before making use of parser 4 . In one embodiment, parser 4 can exploit the doclet application programming interface available in Javadoc. The Javadoc doclet interface provides programmatic access to a parsed representation of a Java™ source file. In one embodiment, parser 4 can process the entire annotated description 1 at once. In another embodiment, parser 4 can be employed repeatedly by analyzer 2 to incrementally process annotated description 1 . Each annotation is verified by comparing what was found in the annotated description 1 with a formal definition in annotation library 5 . In one embodiment, by way of example, annotation library 5 can be a database that includes the definition of each annotation (e.g., its required and optional parameters, permissible values, etc.). In another embodiment, annotation library 5 can be implemented as a cache, tree or any other suitable data structure without departing from the spirit or scope of the present disclosure.

If the annotation and its arguments are proper, analyzer 2 evaluates any attribute-expression arguments provided. The can be done during the validation of each annotation, or in another embodiment, after all annotations have been verified. Values for variable names that occur in expressions are retrieved from variable library 6 . In one embodiment, variable library 6 can be implemented in a similar fashion to annotation library 5 . In a further embodiment, variable names can be constructed dynamically from expressions using string concatenation operators. In one embodiment, each attribute-expression is evaluated, comparing the value of the expression to permissible values allowed for a given attribute as dictated by annotation library 5 . If an expression's value is out of range or erroneous, an error can be declared. In another embodiment, a suitable default value can be provided in place of the errant expression. The form of expressions can be, for example, infix, postfix and prefix. Means for parsing and evaluating such expressions are well known in the art. Based on the parsed annotations, code generator 3 emits the classes or files 7 necessary to deploy the EJB(s) described in annotated description 1 .

In one embodiment, annotations can be described in terms of their attributes, their scope (e.g., class or method) and the type of EJB they are applicable to (e.g., message, stateless session, stateful session, entity, etc.). Annotations having class scope come before a class declaration and annotations having method scope come before a method declaration. Annotations can also be adjacent to or associated with other Java™ elements such as fields or variables. In a further embodiment, and by way of illustration, annotations can be defined as in the following tables. The Annotation/Attribute column contains the annotation name in the first row of each table and any associated attribute names in the subsequent rows. The Description column contains a description of the annotation or attribute. The For column indicates whether the annotation is applicable to a class or a method. Finally, the EJB Type column indicates the type of EJB the annotation is applicable to (e.g., session, message, entity, etc.). It will be apparent to those skilled in the art that many such annotations are possible and therefore this illustration is illustrative and should not be construed to be limiting in any respect.

In one embodiment, by way of example, the following class annotations accompany a Java™ code fragment that declares a stateless session EJB “Trader” having a JNDI name of “ejb20-statelessSession-TraderHome”, and an environment variable “tradeLimit” with a default value of 500.

The type of session EJB (e.g., stateful or stateless) can be automatically determined. For example, if there is an ejbCreate( ) method having a non-empty parameter list, the EJB is stateful. Otherwise, the EJB is stateless. Two methods associated with this EJB are declared below with method annotations indicating that they are available through a remote interface (i.e., to clients residing outside of the Java™ application server in which this EJB is deployed). Notice that both methods make use of the environment variable “tradeLimit” which was declared in the class annotation.

The @ejbgen:remote-method annotation indicates that a method will be available in a remote interface. A remote interface declares an EJB's methods that are accessible to applications outside of an EJB's container. Two Java™ files and two deployment descriptors are automatically generated in accordance to an embodiment based on the code fragment above:

In one embodiment, a value associated with an attribute can be specified with a constant or a variable. In another embodiment, a variable definition can be separate from the EJB definition (e.g., in another file). For example, the variable “remote-jndi-name” can be defined as “Account” as follows:

remote-jndi-name=Account

Within the annotated description, variables can be accessed with the delimiters “${“and”}”. For example:

@ejbgen:jndi-name remote=${remote-jndi-name}

Thus, “remote” would be equal to “Account”.

In another embodiment, variables can be used anywhere after an annotation and can contain whole annotation definitions. Assuming the following variable definition:

jndi-name-tag=remote=RemoteAccount local=LocalAccount

The following annotation:

@ejbgen:jndi-name ${jndi-name-tag}

would expand to:

@ejbgen:jndi-name remote=RemoteAccount Local=LocalAccount

In another embodiment, predefined variables are recognized. Predefined variables have default values that can be overridden. In one embodiment, predefined variables can be defined as follows:

In another embodiment, predefined variables can be prefixed with a class name. For example, if the following predefined variables are overridden as follows:

Account.home-baseClass=BaseAccountHome home.baseClass=BaseHome value.package=value containerManaged.value.package=valuePackageForContainerManaged

Thus, all home interfaces generated will extend the class BaseHome except the home of EJB “Account”, which will extend BaseAccountHome.

In another embodiment, EJB annotations can be inherited according to Java™ class inheritance. For example, if an EJB named AccountEJB has a base class BaseAccount:

The JNDI remote name “BaseRemote” and max-beans-in-cache of 300 that are defined in the class BaseAccount are inherited by the subclass AccountEJB. The subclass can optionally override these values. For example, the value of max-beans-in-cache is overridden to be 400 instead of 300.

In another embodiment, EJB relations can be inherited if a value for the attribute ‘role-name’ is specified. All “role-name” attributes in the Java™ inheritance path can be merged. For example, the following parent class defines a partial relation:

The following subclass inherits the above relation and completes its definition:

It will be apparent to those skilled in the art that the process of generating EJB source files from an annotated description of an EJB can be performed in reverse. All of the files or classes needed to deploy an EJB can be reduced into a single annotated description of the EJB which can, in accordance to one embodiment, be used in turn to reproduce the EJB source files. Annotations can be generated automatically based information contained in the class definitions, deployment descriptors and relationships between classes.

FIG. 2 is a block diagram in accordance to one embodiment of the invention. A plurality of input files 7 containing EJB classes and deployment information is processed by analyzer 8 . Input files 7 can include one or more EJB class definitions. In one embodiment, input files 7 can reside in a single file, such as a Java™ Archive File (“JAR”). In another embodiment, input files 7 can reside in a computer memory. In yet a further embodiment, input files 7 can be transmitted via a computer network (not shown) to analyzer 8 . Analyzer 8 employs parser 10 to extract information related to class structure, class relationships and deployment from input files 7 . Analyzer 8 may perform preprocessing of input files 7 before making use of parser 10 . In one embodiment, parser 10 can exploit the doclet application programming interface available in Javadoc. In one embodiment, parser 10 can process all input files 7 at once. In another embodiment, parser 10 can be employed repeatedly by analyzer 8 to incrementally process input files 7 . During or after parsing, analyzer 8 builds an EJB database 11 containing all of the extracted information pertaining to each EJB represented by input files 7 . The EJB database 11 information is then examined to identify which annotation(s) from annotation library 5 adequately describe the EJB's. The these annotations (not shown) along with the EJB database 11 information are provided to code generator 9 . Code generator 9 then emits annotated Java™ code 1 to describe the EJB class(s) in input files 7 .

The foregoing description of the preferred embodiments of the present invention has been provided for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will be apparent to the practitioner skilled in the art. Embodiments were chosen and described in order to best describe the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention, the various embodiments and with various modifications that are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the following claims and their equivalents.