＋设为首页 ＋添加到收藏夹
second, we need to configure the code-generation options. these options are stored in a model called the genmodel. the genmodel decorates the ecore model with information specific to the solution domain (in our case java and eclipse). the third step is to generate the code. this step uses the ecore model and the genmodel to generate the implementation for the editor plugins.
the generated code can now be tested. it is feature complete, but a rather naive implementation of an editor. in most cases, a developer has to take over the implementation of sections of the generated code. emf has a feature that allows the developer to flag the code he/she wants to enhance.
after the code has been generated, a developer typically starts an iterative process where various discrepancies are fixed. the discrepancies can be separated into three categories: errors in the business model: requires update to the ecore model suboptimal configuration of the code generation options: requires changes to the genmodel requirements for a more sophisticated implementation than that generated: requires that the developer takes control of sections of the generated code and modifies the code. emf has excellent support for this iterative development process, allowing the developers to fix and regenerate.
creating the models
lets illustrate the various stages and artifacts in emf by creating a simpl... 下一页