Contrib Schemas

From: David Fetter <david(at)fetter(dot)org>
To: PG Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Contrib Schemas
Date: 2006-01-12 22:46:04
Message-ID: 20060112224603.GA21105@fetter.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Folks,

For the same reason they feel that *n*x installs shouldn't be messing
around in /usr/local, some people also feel that contrib modules
shouldn't install in the public schema by default. Before I go
a-patching, I'd like to see what people think about how to handle
this. A few of possibilities come to mind:

1. One schema per contrib module. In some sense, this is a clean way
to do things because the modules can't step on each other's
namespaces, and it makes it easy to remove a module cleanly.

A downside of this is the increased maintenance overhead of adding
those schemas to search_paths as appropriate, although this might be
worked around by having the install script generate some search_path
modifications that the person installing could choose to run.

2. One big contrib schema. This avoids the above search_path
problem, but leaves the namespace collision of putting things in
public.

3. Name a schema on installation. This could help people decide
exactly where they want things to go.

I'm picturing something like this:

make install # Option 0 Leave as-is
make install --auto_schema [--generate_path_mod] # Option 1
make install --schema=contrib # Option 2
make install --schema=foo # Option 3

What do you think?

Cheers,
D
--
David Fetter david(at)fetter(dot)org http://fetter.org/
phone: +1 415 235 3778

Remember to vote!

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Anjan Kumar. A. 2006-01-12 23:05:45 Re: Please Help: PostgreSQL Query Optimizer
Previous Message Qingqing Zhou 2006-01-12 22:00:49 Re: Checkpoint question