+1. We should seperate the implementations of Saga and TCC, when starting the alpha server, users should be able to configure the modules as they want.
Willem Jiang Twitter: willemjiang Weibo: 姜宁willem On Sat, Mar 20, 2021 at 1:29 PM Lei Zhang <zhang...@apache.org> wrote: > > Hi, Park Team > > Currently, Park's module division is unclear, difficult for contributors to > participate. > > I suggest dividing the modules again and use SPI to provide these > components. > > The operations can be quite involved, requiring many steps to complete the > reconstruction. I think we can start with the division of modules for > example: > > * state machine mode does not require DB, > * state machine mode persistence layer should be easily replaced > * non-state machine mode Saga and TCC Should be split into two modules, etc. > > This is just my initial thoughts, any suggestions? > > alpha-benchmark > alpha-server-ui > alpha-server-core > alpha-server-saga-db > alpha-server-saga-fsm > alpha-server-tcc-db > alpha-server-tcc-fsm (maybe) > alpha-server-bootstrap (I expect that only this module depends on the > framework infrastructure in the future) > alpha-server-discovery-plugin-consul > alpha-server-discovery-plugin-eureka > alpha-server-discovery-plugin-nacos > alpha-server-discovery-plugin-zookeeper > > Best regards, > Lei Zhang