Re: [PATCH] Skip ALTER x SET SCHEMA if the schema didn't change

On Thu, Nov 5, 2015 at 10:20 PM, Haribabu Kommi
<kommi(dot)haribabu(at)gmail(dot)com> wrote:
> On Tue, Sep 29, 2015 at 12:17 AM, Marti Raudsepp <marti(at)juffo(dot)org> wrote:
>> Hi list
>>
>> The attached patch changes the behavior of multiple ALTER x SET SCHEMA
>> commands, to skip, rather than fail, when the old and new schema is
>> the same.
>>
>> The advantage is that it's now easier to write DDL scripts that are indempotent.
>>
>> This already matches the behavior of ALTER EXTENSION SET SCHEMA in
>> earlier versions, as well as many other SET-ish commands, e.g. ALTER
>> TABLE SET TABLESPACE, OWNER TO, CLUSTER ON, SET (storage_parameter...)
>> etc. I don't see why SET SCHEMA should be treated any differently.
>>
>> The code is written such that object_access_hook is still called for
>> each object.
>>
>> Regression tests included. I couldn't find any documentation that
>> needs changing.
>
> I went through the patch, following are my observations,
>
> Patch applied with hunks and compiled with out warnings.
> Basic tests are passed.
>
> In AlterTableNamespaceInternal function, if a table or matview called
> for set schema,
> If the object contains any constraints, the constraint gets updated
> with new schema.
>
> In AlterTypeNamespaceInternal function, the InvokeObjectPostAlterHook function
> doesn't get called if the type is of composite type, domain and array
> types as because
> it just returns from top of the function.

Most of the community members didn't find any problem in changing the
behavior, so here I attached updated patch with the above two corrections.

Regards,
Hari Babu
Fujitsu Australia