Re: pg_restore depending on user functions

Дмитрий Иванов <firstdismay@gmail.com>

From: Дмитрий Иванов <firstdismay@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-bugs@lists.postgresql.org
Date: 2021-11-17T01:33:40Z
Lists: pgsql-bugs, pgsql-general
Ok, I'll do it.
Am I correct in understanding that I need to dump --schema-only in SQL
format and then delete everything, leaving one chain for example? If so, it
will have to be postponed until the weekend.

ср, 17 нояб. 2021 г. в 01:04, Tom Lane <tgl@sss.pgh.pa.us>:

> =?UTF-8?B?0JTQvNC40YLRgNC40Lkg0JjQstCw0L3QvtCy?= <firstdismay@gmail.com>
> writes:
> > --Line 4048:
> > CREATE TABLE bpd.class (
>
> There are still a lot of problems in this example:
>
> * references to nonexistent columns val_text, val_bytea, val_json
>
> * int_class_ext refers to int_class_ready, int_class_path,
> bpd.object, which weren't supplied
>
> I figured maybe I didn't need int_class_ext, since it doesn't appear
> to be referenced elsewhere.  But with the objects I have, pg_dump
> doesn't do anything wrong; the output can be loaded just fine.
>
> Please, send a self-contained SQL script that you have actually
> tested to be loadable, and which produces a database that
> causes pg_dump to do the wrong thing.
>
>                         regards, tom lane
>

Commits

  1. Adjust pg_dump's priority ordering for casts.