Message-ID: <3A6D1FAB.26528738@alumni.caltech.edu>
Date: Tue, 23 Jan 2001 06:07:39 +0000
From: Thomas Lockhart <lockhart@alumni.caltech.edu>
Reply-To: thomas@pgsql.com
Organization: PostgreSQL
MIME-Version: 1.0
To: otto.hirr@olabinc.com
Cc: pgsql-hackers@postgresql.org, pgsql-interfaces@postgresql.org
Subject: Re: Newbe hacker, interface question on includes...
References: <000001c084ff$8fc5b680$7bf7663f@xcdt.olabinc.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

"Otto A. Hirr, Jr." wrote:
> 
> I have begun perusing the source, eventually hoping to
> contribute to the development effort.

Welcome!

>  I have noticed that
> long ago the include files for a "component" were in the
> component directory, eg src/backend/catalog/Attic/catalog.h,v
> (here I'm assuming that Attic means a storage place for
> items that use to be there but are no longer there...)

Actually, you are looking at CVS directory structures, not the
checked-out tree. CVS uses the attic to retain the history of deleted
files and of files which are on branches. Check out the cvs docs for
more info.

> and now this file resides in src/include/catalog/catalog.h,v.  A quick
> scan seems that this was a common transformation.
> This seems to be a mighty step backwards, especially if
> one considers the recommendations on large systems development
> where you want to minimize dependancies...

We support well over 20 platforms now, and the platform dependencies are
pretty well isolated, though not perfectly.

> Could someone point out the history of code relocation and rationale.

We have a *long* history. But since you have access to a copy of the CVS
tree, you can check the comments and status on every file and every
change made since circa 1996.

                          - Thomas