The issue for the light-codegen for master branch (JDK11)

If user use the supportDB config value in the config.json

"supportDb": true,

The light-codegen will generate datasource config in the service.yml:

'''

• javax.sql.DataSource:
  • com.zaxxer.hikari.HikariDataSource:
    DriverClassName:
    jdbcUrl:
    username:
    password:
    maximumPoolSize: 10
    useServerPrepStmts: true
    cachePrepStmts: true
    cacheCallableStmts: true
    prepStmtCacheSize: 10
    prepStmtCacheSqlLimit: 2048
    connectionTimeout: 2000

'''

And in the values.yml, light-codegen output:

service.singletons:

I think user may chose different ways for datasource, if user use light--4j data-source module，then it will use datasource.yml file and user can flexibly choose the way for database process.

So I would like suggest to remove the datasource in the setting in the service yam and change it to:

singletons: ${service.singletons:}

And user can only deal with on config file: values.yml

Trace how the supportDb value from config.json produces the datasource section in service.yml and service.singletons in values.yml. Compare that output with the datasource.yml approach described in the issue, then verify that generated configurations use the requested singletons setting without the datasource block.