about server side insert

[Piao songmei]

about server side insert, at function of locator_attribute_info_force() there're following codes:

locator_attribute_info_force()
{
  ......
  switch (operation)
  {
    case LC_FLUSH_UPDATE:
      ...
      break;
    case LC_FLUSH_INSERT:
      ...
      if (operation == LC_FLUSH_INSERT)
      {
        ...
      }
      else
      {
        assert (operation == LC_FLUSH_UPDATE)
        ...
      }
  }
}

in case of LC_FLUSH_INSERT why judged again operation?

[Kieun Park]

You can see that there is no "break" statement in the end of the first "case LC_FLUSH_UPDATE" block.
See the comment there… /* Fall through */

Thanks.

[Kieun Park]

One more thing.
Actually, those code is not related with "server-side insert" feature. They are related with "client-side insert".
As the CuBRID's terminology, "server-side insert" means that INSERT_PROC XASL.
"Client-side insert" means that calling locator_force() to flush out the object to the server.

[Piao songmei]

oops, my mistake.
thank you for your explain.

and for client side insert / server side insert:
in case of server side insert, it's also called locator_insert_force().  the stack is as bellow:
          qexec_execute_insert() -> locator_attribute_info_force() -> locator_insert_force()
in case of client side insert, the stack is:
          xlocator_force() -> locator_insert_force()

in server side insert, it's also prepared copy area and flushed to server.

[Kieun Park]

Yes. You're right. "locator_attribute_info_force()" is called also during server-side insert. I'm confused with somewhat.
Thanks.

[Piao songmei]

thank you