Delivery-Date: Tue, 27 Jul 93 23:07:03 PDT
Date: Wed, 28 Jul 1993 16:05:40 +1000 (EST)
From: "David J. Hughes" <bambi@Bond.edu.au>
Subject: Re: Swelling backend. 
To: "Paul M. Aoki" <aoki@postgres.Berkeley.EDU>
In-Reply-To: <199307280553.WAA15532@faerie.CS.Berkeley.EDU>
Message-Id: <Pine.3.07.9307281628.E176-b100000@fiddich.its.Bond.edu.au>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Tue, 27 Jul 1993, Paul M. Aoki wrote:

> the one thing Purify tells me about is the fact that the yacc-generated
> parser doesn't free its memory if an elog(WARN,...) is generated from
> within the parser.
> 
> i wouldn't think this would happen in your environment, though.


Thanks for the quick response Paul.

Is there anything other than PQclear() that I need to call from time to
time to free any internal buffers?  This process is the event manager for
the Network Management System I'm writing so the one proc should run
24hours * 7days.  Waiting for a new backend after a reset just kills the
performance of my stuff.  I guess I could reset it every 100 or so events
(ie. after 500-800 queries) but that'd be a shocking hack.

On another topic, would you guys be interested in an SQL to Postquel
translator?  Several people want to run my code over Oracle and Progress
so I'm moving to SQL internally.  I wrote a grammer for a subset of SQL
that offered the facilities I need and coded a yacc-based parser for it. 
It has backend hooks for generating other QL's and I've written the code
for Postquel.

It's very limited (doesn't allow joins or nested selects or anything other
than stock simple queries) but it may be of use as a starting point.


   ___                                  David J. Hughes    bambi@Bond.edu.au
  /   \                /  /    /        
 /  __/ __   __   ____/  /    / __            Senior Network Programmer
/    \ /  \ /  \ /   /  /    / /  \  /     Information Technology Services
\____/ \__//   / \__/   \___/ /   / /   Qld. 4229  AUSTRALIA  (+61 75 951450)