Re: pg_upgrade: allow multiple -o/-O options

Hello,

RFE: Consider that you want to run pg_upgrade via some script with some
default '-o' option. But then you also want to give the script's user a
chance to specify the old-server's options according user's needs.
Then something like the following is not possible:

$ cat script
...
pg_upgrade ... -o 'sth' $PG_UPGRADE_OPT ...
...

I know that this problem is still script-able, but the fix should be
innocent and it would simplify things. Thanks for considering,

Pavel

On Tue, Mar 4, 2014 at 04:52:56PM +0100, Pavel Raiskup wrote:
> Hello,
>
> RFE: Consider that you want to run pg_upgrade via some script with some
> default '-o' option. But then you also want to give the script's user a
> chance to specify the old-server's options according user's needs.
> Then something like the following is not possible:
>
> $ cat script
> ...
> pg_upgrade ... -o 'sth' $PG_UPGRADE_OPT ...
> ...
>
> I know that this problem is still script-able, but the fix should be
> innocent and it would simplify things. Thanks for considering,

Attached is a patch that makes multiple -o options append their
arguments for pg_upgrade and pg_ctl, and documents this and the append
behavior of postmaster/postgres. This covers all the -o behaviors.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ Everyone has their own god. +

On Thursday 21 of August 2014 18:26:37 Bruce Momjian wrote:
> On Tue, Mar 4, 2014 at 04:52:56PM +0100, Pavel Raiskup wrote:
> > RFE: Consider that you want to run pg_upgrade via some script with some
> > default '-o' option. But then you also want to give the script's user a
> > chance to specify the old-server's options according user's needs.
> > Then something like the following is not possible:
> >
> > $ cat script
> > ...
> > pg_upgrade ... -o 'sth' $PG_UPGRADE_OPT ...
> > ...
> >
> > I know that this problem is still script-able, but the fix should be
> > innocent and it would simplify things. Thanks for considering,
>
> Attached is a patch that makes multiple -o options append their
> arguments for pg_upgrade and pg_ctl, and documents this and the append
> behavior of postmaster/postgres. This covers all the -o behaviors.

Thanks! Seems to be OK to me, one nit - why you did not go the
append_optiton way (there could be probably better name like arg_cat)?
Because this is just about few lines, it is probably OK from PostgreSQL
policy POV, so "review? ~> review+", thanks again!

Pavel

On Fri, Aug 22, 2014 at 10:52:12AM +0200, Pavel Raiskup wrote:
> On Thursday 21 of August 2014 18:26:37 Bruce Momjian wrote:
> > On Tue, Mar 4, 2014 at 04:52:56PM +0100, Pavel Raiskup wrote:
> > > RFE: Consider that you want to run pg_upgrade via some script with some
> > > default '-o' option. But then you also want to give the script's user a
> > > chance to specify the old-server's options according user's needs.
> > > Then something like the following is not possible:
> > >
> > > $ cat script
> > > ...
> > > pg_upgrade ... -o 'sth' $PG_UPGRADE_OPT ...
> > > ...
> > >
> > > I know that this problem is still script-able, but the fix should be
> > > innocent and it would simplify things. Thanks for considering,
> >
> > Attached is a patch that makes multiple -o options append their
> > arguments for pg_upgrade and pg_ctl, and documents this and the append
> > behavior of postmaster/postgres. This covers all the -o behaviors.
>
> Thanks! Seems to be OK to me, one nit - why you did not go the
> append_optiton way (there could be probably better name like arg_cat)?
> Because this is just about few lines, it is probably OK from PostgreSQL
> policy POV, so "review? ~> review+", thanks again!

Well, I found append_optiton() to be an extra function that wasn't
necessary --- the psprintf() use was short enough not to need a separate
function.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ Everyone has their own god. +

On Fri, Aug 22, 2014 at 10:02:11AM -0400, Bruce Momjian wrote:
> On Fri, Aug 22, 2014 at 10:52:12AM +0200, Pavel Raiskup wrote:
> > On Thursday 21 of August 2014 18:26:37 Bruce Momjian wrote:
> > > On Tue, Mar 4, 2014 at 04:52:56PM +0100, Pavel Raiskup wrote:
> > > > RFE: Consider that you want to run pg_upgrade via some script with some
> > > > default '-o' option. But then you also want to give the script's user a
> > > > chance to specify the old-server's options according user's needs.
> > > > Then something like the following is not possible:
> > > >
> > > > $ cat script
> > > > ...
> > > > pg_upgrade ... -o 'sth' $PG_UPGRADE_OPT ...
> > > > ...
> > > >
> > > > I know that this problem is still script-able, but the fix should be
> > > > innocent and it would simplify things. Thanks for considering,
> > >
> > > Attached is a patch that makes multiple -o options append their
> > > arguments for pg_upgrade and pg_ctl, and documents this and the append
> > > behavior of postmaster/postgres. This covers all the -o behaviors.
> >
> > Thanks! Seems to be OK to me, one nit - why you did not go the
> > append_optiton way (there could be probably better name like arg_cat)?
> > Because this is just about few lines, it is probably OK from PostgreSQL
> > policy POV, so "review? ~> review+", thanks again!
>
> Well, I found append_optiton() to be an extra function that wasn't
> necessary --- the psprintf() use was short enough not to need a separate
> function.

Patch applied; this will appear in PG 9.5.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ Everyone has their own god. +