• “多团队大规模”开发模式


    应用SAP HANA “官方”开发模式的伙伴们在转到“多团队大规模”开发模式时会遇到各式各样的心理不适应的状况,各种纠结。比如GIT Repository和HANA Repository冲突什么的。

    这些问题主要是思路没有完全转换过来,两种开发模式在不断的“打架”。做一个简单的开发模式映射这样大家或许比较容易去接受“多团队大规模”的开发模式。这个映射不一定完全正确,主要是帮助大家理解。

    伙伴们想一想平常我们做JAVA或者.NET项目是怎么做的?一般情况是使用各种框架写JAVA/.NET的代码,然后用ORM映射数据库。写SQL Query时基本上都是在SQL里面去写,去测试然后拼接到自己的JAVA/.NET代码中。

     

    那么我们在SAP HANA平台上开发同样可以学习这样的模式,当然这种模式可能比“官方”模式厚重点,但是好处也是显而易见的,比如前文提到的那些好处。前后端的开发都可以在Eclipse中,HANA Studio的作用完全等同拼写检查SQL Query等SQL层面的工作。这样我们在开发过程中就可以应用很多手段,方便的传递参数。比如替换rootpackage,通过maven的插件就很容易的可以做到。这样HANA Instance就扮演了一个编译器的作用。前端的代码编译好之后,和后端代码整合到一起Activate到HANA Repository中,这时的Activate就像是compile一样。Maven的插件,或者Ant的脚本把Activate这些HANA特有的动作都封装起来,开发人员完全还是遵循着传统多团队大规模开发的模式——写代码,编译,调试等。

     

    由于Maven涉及到公司内部提供的插件,没有开源。所以就提供一个Ant的脚本以饷读者。

    <project name="regi" basedir=".">
        <description>Regi wrapper</description>
    
         <!-- Properties for using REGI -->
        <property name="userstore.key" value="BUILD"/>
        <property name="userstore.cmdline" value="${client.dir}/hdbuserstore"/>
        <property name="regi.cmdline" value="${client.dir}/regi"/>
        <property name="regi.connection.args" value="--key=BUILD" />
    
        <macrodef name="regi">
            <attribute name="workspacedir" />
            <attribute name="command" />
    
            <sequential>
                <!-- make sure the directory exists -->
                <mkdir dir="@{workspacedir}" />
    
                <!-- Create the workspace -->
                <exec executable="${regi.cmdline}" dir="@{workspacedir}" failonerror="true">
                  <arg line="@{command}" />
                </exec>
            </sequential>
        </macrodef>
    
        <macrodef name="regi_apply">
            <attribute name="workspacedir" />
            <attribute name="command" />
            <element name="files" />
    
            <sequential>
                <!-- make sure the directory exists -->
                <mkdir dir="@{workspacedir}" />
    
                <!-- Create the workspace -->
                <apply executable="${regi.cmdline}" dir="@{workspacedir}" failonerror="true" relative="true" parallel="true">
                  <arg line="@{command}" />
                  <files />
                </apply>
            </sequential>
        </macrodef>
    
        <macrodef name="regi_setupuserstore">
          <attribute name="workspacedir" />
          <sequential>
        <!-- Setup a key in the user store for use by REGI -->
        <exec executable="${userstore.cmdline}" dir="@{workspacedir}" failonerror="true">
          <arg line="SET ${userstore.key} ${hana.host} ${hana.user} ${hana.password}" />
          </exec>
        </sequential>
          </macrodef>
        
        <macrodef name="regi_createworkspace">
            <attribute name="workspacedir" />
    
            <sequential>
                <!-- make sure the directory exists (needed for keystore operation) -->
                <mkdir dir="@{workspacedir}" />
    
                <!-- Create the workspace -->
                <mkdir dir="@{workspacedir}" />
                <exec executable="${regi.cmdline}" dir="@{workspacedir}" failonerror="true">
                  <arg line="${regi.connection.args} create workspace --force" />
                </exec>
            </sequential>
        </macrodef>
    
        <macrodef name="regi_export">
            <attribute name="workspacedir" />
            <attribute name="deliveryUnit" default="" />
            <attribute name="vendor" default="" />
            <attribute name="fileName" default="" />
    
            <sequential>        
                <!-- Export the desired delivery unit to the specified file -->
                <regi workspacedir="@{workspacedir}" command="export @{deliveryUnit} @{vendor} @{fileName}" />
            </sequential>
        </macrodef>
    
        <macrodef name="regi_track">
            <attribute name="workspacedir" />
            <attribute name="package" default="" />
    
            <sequential>
                <echo>regi track @{package}</echo>
    
                <!-- Track the desired package -->
                <regi workspacedir="@{workspacedir}" command="track @{package}" />
            </sequential>
        </macrodef>
    
        <macrodef name="regi_commit">
            <attribute name="workspacedir" />
    
            <sequential>
                <regi workspacedir="@{workspacedir}" command="commit" />
                <regi workspacedir="@{workspacedir}" command="activate" />
            </sequential>
        </macrodef>
    
        <macrodef name="regi_import">
            <attribute name="workspacedir" />
            <attribute name="import" />
    
            <sequential>
                <regi workspacedir="@{workspacedir}" command="import @{import}" />
            </sequential>
        </macrodef>
    
        <macrodef name="regi_cleanpackage">
            <attribute name="package"/>
    
            <sequential>
                <echo message="Deleting regi package: @{package}" />
    
                <!-- Brutally crush any and all evidence of the repo objects for the passed package -->
                <hdbsqlsql failonerror="false" sql="
                    delete from &quot;_SYS_REPO&quot;.&quot;ACTIVE_OBJECT&quot;   where PACKAGE_ID = '@{package}' OR PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;OBJECT_HISTORY&quot;  where PACKAGE_ID = '@{package}' OR PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;RUNTIME_OBJECTS&quot; where PACKAGE_ID = '@{package}' OR PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;PACKAGE_CATALOG&quot; where PACKAGE_ID = '@{package}' OR PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;INACTIVE_OBJECT&quot; where PACKAGE_ID = '@{package}' OR PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;ACTIVE_OBJECTCROSSREF&quot; where FROM_PACKAGE_ID = '@{package}' OR FROM_PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;ACTIVE_OBJECTCROSSREF&quot; where TO_PACKAGE_ID = '@{package}' OR TO_PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;INACTIVE_OBJECTCROSSREF&quot; where FROM_PACKAGE_ID = '@{package}' OR FROM_PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                    delete from &quot;_SYS_REPO&quot;.&quot;INACTIVE_OBJECTCROSSREF&quot; where TO_PACKAGE_ID = '@{package}' OR TO_PACKAGE_ID LIKE '@{package}.%' ${line.separator}
                "/>
            </sequential>
        </macrodef>
    
        <macrodef name="regi_clean">
            <attribute name="workspacedir" />
    
            <sequential>
                <delete dir="@{workspacedir}/._SYS_REGI_settings"/>
            </sequential>
        </macrodef>
        
        <macrodef name="regi_checkout_trackedPackages">
            <attribute name="workspacedir" />
    
            <sequential>
                <echo>regi checkout trackedPackages --force</echo>
                <echo>workspacedir=@{workspacedir}</echo>
    
                <!-- Checkout all packages that have been tracked. -->
                <regi workspacedir="@{workspacedir}" command="checkout trackedPackages --force" />
            </sequential>
        </macrodef>
    
        <macrodef name="regi_delete_packages">
            <attribute name="workspacedir" />
            <attribute name="packages" />
    
            <sequential>
                <echo>regi delete packages @{packages}</echo>
                <echo>workspacedir=@{workspacedir}</echo>
    
                <!-- Delete the desired package -->
                <regi workspacedir="@{workspacedir}" command="delete packages @{packages}" />
            </sequential>
        </macrodef>
    
    </project>
  • 相关阅读:
    metasploitable实践
    免杀技术
    SQL注入
    ARP欺骗
    使用Nmap攻击靶机和使用Wireshark进行嗅探、分析
    Nmap
    搭建网络攻防环境
    20139216网络攻防技术第七次作业
    Android代码实现求和运算
    命令行输入参数,并实现排序
  • 原文地址:https://www.cnblogs.com/wanghonggang/p/largescale_hana.html
Copyright © 2020-2023  润新知