Re: ALTER command reworks

Kohei KaiGai <kaigai@kaigai.gr.jp> writes:
> The attached patch is a revised version.

I think you fixed all the problems I could see with your patch. It
applies cleanly (with some offsets), compiles cleanly (I have a custom
Makefile with CUSTOM_COPT=-Werror) and passes regression tests.

I'll go mark it as ready for commiter. Thanks!

> It follows the manner in ExecAlterObjectSchemaStmt(); that tries to check
> name duplication for object classes that support catcache with name-key.
> Elsewhere, it calls individual routines for each object class to solve object
> name and check namespace conflicts.
> For example, RenameFunction() is still called from ExecRenameStmt(),
> however, it looks up the given function name and checks namespace
> conflict only, then it just invokes AlterObjectRename_internal() in spite
> of system catalog updates by itself.

I think that's much better this way.

> It allows to use this consolidated routine by object classes with special
> rule for namespace conflicts, such as collation.
> In addition, this design allowed to eliminate get_object_type() to pull
> OBJECT_* enum from class_id/object_id pair.

Thanks for having done this refactoring.

Regards,
-- 
Dimitri Fontaine
http://2ndQuadrant.fr     PostgreSQL : Expertise, Formation et Support