topic Re: splitting persistence into modules in Alfresco Archive

splitting persistence into modules

tombaeyens — Mon, 07 Feb 2011 07:26:34 GMT

as a side track of http://jira.codehaus.org/browse/ACT-540 (Distill db name from the datasource), I split up the persistence modules. create, drop and update of tables is done per component. The 4 components are * engine* history* identity* cycleIn the ProcessEngineConfigurationImpl, those 4 compo

Re: splitting persistence into modules

meyerd — Mon, 07 Feb 2011 10:35:00 GMT

Hi tom,

Thx for the pointers. I will get back to you about the conceptional stuff later.
ATM I am still struggling with the build.

Currently looking at distro/src/setup/build.xml Line 209

should it be changed to

      <javaresource name="org/activiti/db/create/activiti.${db}.create.engine.sql" 
                    classpathref="classpath.libs" />

?

(same thing in line 222)

Regards,
daniel

Re: splitting persistence into modules

meyerd — Mon, 07 Feb 2011 13:13:45 GMT

Hi Tom,

could you please have a look at http://fisheye.codehaus.org/changelog/activiti/?cs=2035 ?

Is this how you intended it?

I also noticed that you took out the isDbEngineUsed field in your last commit. Is it coming back ?

Regards,
daniel

Re: splitting persistence into modules

meyerd — Mon, 07 Feb 2011 13:32:47 GMT

Maybe "org.activiti.engine.impl.db.DbManagementSession" needs to be revised now, specifically the field


protected static String[] tableNames = new String[]{
    "ACT_GE_PROPERTY",
    "ACT_GE_BYTEARRAY",
    "ACT_RE_DEPLOYMENT",
    "ACT_RU_EXECUTION",
    "ACT_ID_GROUP",
    "ACT_ID_MEMBERSHIP",
    "ACT_ID_USER",
    "ACT_RU_JOB",
    "ACT_RE_PROCDEF",
    "ACT_RU_TASK",
    "ACT_RU_IDENTITYLINK",
    "ACT_RU_VARIABLE",
    "ACT_HI_PROCINST",
    "ACT_HI_ACTINST",
    "ACT_HI_TASKINST",
    "ACT_HI_DETAIL"
  };

Daniel

Re: splitting persistence into modules

tombaeyens — Wed, 09 Feb 2011 07:40:59 GMT

sorry for the late response. only now trying to catch up with forums.

i reviewed the cycle code. you're now building a process engine and then taking the pieces that you need. that is exactly how it should be. great work!

i also refactored the scripts you mentioned and the tableNames now also are obtained through metadata.