110 likes | 232 Vues
Sakai Tool Naming Tips. Aaron Zeckoski azeckoski@gmail.com. Many tools exist in a typical Sakai installation As a developer, you have to be careful when working in a large scale environment The primary issue is with naming collisions
E N D
Sakai Tool Naming Tips Aaron Zeckoski azeckoski@gmail.com
Many tools exist in a typical Sakai installation As a developer, you have to be careful when working in a large scale environment The primary issue is with naming collisions Some names must be unique in the Sakai instance (not used by any other installed tools) Tools have many interconnections and use a custom request cycle so some programming practices must be followed Sakai tool environment
Most namespace collisions will cause an error to occur on tomcat startup Typically you will see that part of the spring tree died and therefore all spring beans are destroyed Sometimes Sakai will still load but the offending tool will not Tip: Watch tomcat logs when starting up new tools Namespace collisions URL: http://bugs.sakaiproject.org/confluence/display/BOOT/Sakai+app+and+tool+naming+tips
The tool id (e.g. sakai.tasklist) in the tool xml file must be unique Typical id is sakai.toolname Tool xml file located in tool/src/webapp/tools/ Typical filename is sakai.toolname.xml Sakai tool xml file naming <?xml version="1.0" encoding="UTF-8"?> <registration> <tool id="sakai.tasklist" title="Programmer's Cafe - Task List" description="Programmer's Cafe - Task List"> <category name="course" /> <category name="project" /> </tool> </registration>
The id (e.g. sakai-tasklist-tool) in each project.xml file must be unique Typical id is sakai-toolname-location Maven project.xml files are located throughout a typical Sakai app Sakai project.xml naming <?xml version="1.0" encoding="UTF-8"?> <project> <pomVersion>3</pomVersion> <extend>../../master/project.xml</extend> <name>Programmer's Cafe - Task List</name> <groupId>sakaiproject</groupId> <id>sakai-tasklist-tool</id> <currentVersion>${sakai.version}</currentVersion> ...
The id (e.g. org.sakaiproject.logic.CrudPlusLogic) of each bean in components.xml must be unique Convention for the id is to use the fully qualified classpath of the interface for the class component.xml must be located in impl/pack/src/webapp/WEB-INF/ Spring bean naming <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE beans PUBLIC "-//SPRING//DTD BEAN//EN" "http://www.springframework.org/dtd/spring-beans.dtd"> <beans> <bean id="org.sakaiproject.crudplus.logic.CrudPlusLogic" class="org.sakaiproject.crudplus.logic.impl.CrudPlusLogicImpl" init-method="init"> </beans>
Database table names have to be unique since Sakai shares a common schema (or database) Convention is to prefix the table name with the tool name (e.g. TOOLNAME_TABLE) Good names: SAM_ANSWER, EVAL_ANSWER, GB_GRADES Bad names: ANSWER, GRADE, CONFIG Database table naming
All Hibernate HBM files have to have unique names since Sakai uses one common Hibernate SessionFactory Convention is to prefix the hbm filename with the tool name (e.g. ToolnameItem.hbm.xml) Good names: EvalAnswer.hbm.xml, TasklistTask.hbm.xml Bad names: Answer.hbm.xml, Task.hbm.xml, Item.hbm.xml Hibernate HBM files
All Hibernate persistent classes must have unique classnames (this does not include the fully qualified classpath) Convention is to prefix the class name with the tool name (e.g. ToolnameItem.java) Good: EvalAnswer.java, TasklistTask.java Bad: Answer.java, Task.java Hibernate Persistent Classnames
The servlet-name in the web.xml file must match the tool id from the tool xml file web.xml is located in tool/src/webapp/WEB-INF If this does not match you will get an uninformative NPE (Null Pointer Exception) in the tomcat log and the tool will fail to load web.xml servlet name <servlet> <servlet-name>sakai.tasklist</servlet-name> <servlet-class>org.sakaiproject.tool.tasklist.TasklistTool</servlet-class> <load-on-startup>1</load-on-startup> </servlet>