Re: Testing the async-commit patch - Mailing list pgsql-hackers

From Greg Smith
Subject Re: Testing the async-commit patch
Date
Msg-id Pine.GSO.4.64.0708141417370.11206@westnet.com
Whole thread Raw
In response to Re: Testing the async-commit patch  ("Simon Riggs" <simon@2ndquadrant.com>)
List pgsql-hackers
On Mon, 2007-08-13 at 16:27 -0400, Andrew Dunstan wrote:
> [asynch_commit]
> synchronous_commit = off
> [no_fsync]
>  fsync = off

This is the Windows INI file format.  As such, it's easy to find code 
samples in almost any language that parse this format for you.  For 
example, Python has a core library called ConfigParser that will read 
these in, and somewhere around here I even have some UNIX shell code that 
parses it.  There's already a PostgreSQL-related project using this 
format--even on UNIX systems the odbc.ini config files look like this.

I already have a program that generates multiple postgresql.conf files 
using this format around here for exactly this sort of test (compare 
benchmark results with two different configurations), just never had a 
reason to package it up for anybody else to use.  It's trivial code if you 
use the Python parser.

On Tue, 14 Aug 2007, Simon Riggs wrote:

> Sounds fine, though I'd prefer this in XML to allow further
> extensibility in the future.

Putting this in XML significantly raises the bar for how complicated tools 
that work on these files must be, with the implicit dependencies that go 
with that.  And as Andrew already pointed out, there is very little 
tree-structure to this data that justifies the extra complexity.

--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD


pgsql-hackers by date:

Previous
From: Heikki Linnakangas
Date:
Subject: Re: default_text_search_config and expression indexes
Next
From: Heikki Linnakangas
Date:
Subject: Re: change name of redirect_stderr?