Re: proposal: possibility to read dumped table's name from file

On Thu, 19 Nov 2020 at 19:57, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> wrote:
>
> minor update - fixed handling of processing names with double quotes inside
>

I see this is marked RFC, but reading the thread it doesn't feel like
we have reached consensus on the design for this feature.

I agree that being able to configure pg_dump via a config file would
be very useful, but the syntax proposed here feels much more like a
hacked-up syntax designed to meet this one use case, rather than a
good general-purpose design that can be easily extended.

IMO, a pg_dump config file should be able to specify all options
currently supported through the command line, and vice versa (subject
to command line length limits), with a single common code path for
handling options. That way, any new options we add will work on the
command line and in config files. Likewise, the user should only need
to learn one set of options, and have the choice of specifying them on
the command line or in a config file (or a mix of both).

I can imagine eventually supporting multiple different file formats,
each just being a different representation of the same data, so
perhaps this could work with 2 new options:

--option-file-format=plain|yaml|json|...
--option-file=filename

with "plain" being the default initial implementation, which might be
something like our current postgresql.conf file format.

Also, I think we should allow multiple "--option-file" arguments
(e.g., to list different object types in different files), and for a
config file to contain its own "--option-file" arguments, to allow
config files to include other config files.

The current design feels far too limited to me, and requires new code
and new syntax to be added each time we extend it, so I'm -1 on this
patch as it stands.

Regards,
Dean