Re: pgpool versus sequences

On Thu, Jun 2, 2011 at 10:31 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> On Wed, Jun 1, 2011 at 7:47 PM, Alvaro Herrera
>> <alvherre(at)commandprompt(dot)com> wrote:
>>> Yeah -- why is LOCK SEQUENCE foo_seq not allowed?  Seems a simple thing
>>> to have.
>
>> It cause a grammar conflict.
>
> That's a lot of work for a purely cosmetic issue, though.  What would be
> trivial is to let this work:
>
> regression=# create sequence s1;
> CREATE SEQUENCE
> regression=# begin;
> BEGIN
> regression=# lock table s1;
> ERROR:  "s1" is not a table
>
> We should do that anyway, even if we put in the effort to support the
> other syntax.

Ugh. We are already stuck supporting all kinds of backward
compatibility cruft in tablecmds.c as a result of the fact that you
used to have to use ALTER TABLE to operate on views and sequences.
The whole thing is confusing and a mess. -1 from me on extending that
mess to more places.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company