Re: Add visibility map information to pg_freespace.

Thank you for the worthwhile additions.

At Tue, 16 Jul 2013 16:04:43 +0900, Satoshi Nagayasu <snaga(at)uptime(dot)jp> wrote in <51E4F08B(dot)3030307(at)uptime(dot)jp>
> > | postgres=# select * from pg_freespace_with_vminfo('t'::regclass) limit
> > | 10;
..
> I think we can simply add is_all_viible column to the existing
> pg_freespace(), because adding column would not break
> backward-compatibility in general. Any other thoughts?

I agree to you. I cannot guess any 'ordinary' application which
uses this function, or someone's craft critically affected by
this change. This decision was merely a safe bet.

I'll remerge _with_vminfo function to pg_freespace() in the next
patch if no objection is raised.

> > pgstattuple_vm_v1.patch:
...
> It seems working fine.
>
> And I added a regression test for pg_freespacemap and additional
> test cases for pgstattuple. Please take a look.

Thank you. This seems fine. I felt a bit uneasy with the absense
of regtests in pg_freespacemap, but I took advantage of the
absense not to add new ones.

I have simply merged the two regtests separately into two
original patches. You will find the two attached files.

pg_freespace_vm_v3.patch :
new patch for pg_freespace with regtests and _with_vminfo

pgstattuple_vm_v2.patch :
new patch for gstattuple with regtests

regards,

--
Kyotaro Horiguchi
NTT Open Source Software Center