Re: Refactor handling of database attributes between pg_dump and pg_dumpall

On Fri, Oct 6, 2017 at 12:29 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:

> On Wed, Oct 4, 2017 at 3:40 AM, Haribabu Kommi <kommi(dot)haribabu(at)gmail(dot)com>
> wrote:
> > There are some differences in handling database objects
> > between pg_dump and pg_dumpall, To retain both pg_dump
> > and pg_dumpall behavior even after refactoring, this option
> > is added. Currently this option is used mainly for the three
> > purposes.
> >
> > 1. Don't print unnecessary CREATE DATABASE options like
> > ENCODING, LC_COLLATE and LC_CTYPE options if the
> > default encoding is same with the above values.
> >
> > The above behavior is as per the pg_dumpall, but it can be
> > changed to print irrespective of the default encoding.
> >
> > 2. Do not dump postgres and template0 databases.
> >
> > 3. Set default_transaction_read_only = off.
>
> To me it seems that a refactoring which requires pg_dump to behave
> differently in small ways like this based on whether it is being
> called by pg_dumpall or not is probably not a good refactoring. And I
> don't see why the proposal from Tom that started this thread would
> require such a thing to be true.
>

Before refactoring, pg_dumpall doesn't print "create database" commands
for both tempalte1 and postgres database, but on the other hand pg_dump
dump the create database commands with --create option.

To keep the behavior of all the database attributes in the dump of both
pg_dump and pg_dumpall, the code is unified and moved into pg_dump.
But to retain the pg_dumpall behavior of not dumping the "create database"
commands, a new option is added to pg_dump to skip dumping the
create database commands.

The new option name is now "--skip-create-default-db", this can be used
normal user also when try to dump the postgres database to not let create
the database commands in the dump.

From your list, I would say that (1) and (3) seem like behaviors that
> we either want or do not want. Whether pg_dump or pg_dumpall is
> involved seems irrelevant. (2) seems like it might need some special
> handling, but that could be handled in pg_dumpall by just not calling
> pg_dump at all for those database.

I didn't any better way other than creating a new option to not let the
default db create database commands to dump, so I renamed the
older option to better one and change the behavior to use by the
normal users also.

Updated patch attached.

Regards,
Hari Babu
Fujitsu Australia