Added to TODO:

>       o Have ALTER TABLE rename SERIAL sequences

Seems we at least need this.  Doesn't dependency tracking make this
easy to do now?

---------------------------------------------------------------------------

Tom Lane wrote:
> "Christopher Kings-Lynne" <chriskl(at)familyhealth(dot)com(dot)au> writes:
> >> I think we should stick with the existing naming convention.
>                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> > Non-colliding?
> 
> No; see above.
> 
> > Otherwise, it'd be ludicrous to fail a table rename because
> > a sequence with the new name already exists...
> 
> Why?  We already rename the table's rowtype, ergo you can fail a table
> rename because there is a conflicting datatype name.  I don't see
> anything much wrong with failing a table or column rename because there
> is a conflicting sequence name.  The whole point here is to have a
> non-surprising mapping between the names of serial columns and the names
> of their associated sequences.
> 
> 			regards, tom lane
> 

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman(at)candle(dot)pha(dot)pa(dot)us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

In response to

• Re: [PATCHES] Non-colliding auto generated names at 2003-03-06 02:08:59 from Tom Lane

pgsql-hackers by date

Next:	From: Bruce Momjian	Date: 2003-03-06 02:43:28
	Subject: Purpose of core group
Previous:	From: Christopher Kings-Lynne	Date: 2003-03-06 02:21:36
	Subject: Re: [GENERAL] problems with dropped columns

pgsql-patches by date

Next:	From: greg	Date: 2003-03-06 15:31:41
	Subject: Re: XML ouput for psql
Previous:	From: Christopher Kings-Lynne	Date: 2003-03-06 02:27:03
	Subject: Re: new format for describe (repost #2)