Re: Extensions support for pg_dump, patch v27 - Mailing list pgsql-hackers

From Dimitri Fontaine
Subject Re: Extensions support for pg_dump, patch v27
Date
Msg-id m2fwsfqsxg.fsf@2ndQuadrant.fr
Whole thread Raw
In response to Re: Extensions support for pg_dump, patch v27  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane <tgl@sss.pgh.pa.us> writes:
> Oh: then you're doing it wrong.  If you want to remember that WITH
> SCHEMA was specified, you need to explicitly store that as another
> column in pg_extension.  You should not be depending on the dependency
> mechanism to remember that for you, any more than we'd use pg_depend to
> remember a table's relnamespace.  The dependency mechanism is there
> to figure out the consequences of a DROP command, it's not there to
> remember arbitrary facts.  (And yes, I know that we've cheated on that
> principle a few times before; but you can't do it here.)

The thinking is that we need to have the dependency registered too, so
that DROP SCHEMA will cascade to the extension.  So while at it, I also
used the dependency for tracking the schema.

Even if I get to use a column to track the schema, I will have to
maintain registering the dependency.  Should I do that?

Regards,
-- 
Dimitri Fontaine
http://2ndQuadrant.fr     PostgreSQL : Expertise, Formation et Support


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: ALTER TYPE 3: add facility to identify further no-work cases
Next
From: "Kevin Grittner"
Date:
Subject: .gitignore patch for coverage builds