Skip to main content.
home | support | download

Back to List Archive

RE: Fuzzy Indexing with Double Metaphone

From: Bill Moseley <moseley(at)not-real.hank.org>
Date: Wed Sep 18 2002 - 17:36:56 GMT
At 10:12 AM 09/18/02 -0700, Jody Cleveland wrote:
>There are a few things I don't like though. Like, I installed an rpm the
>other day (Comanche), but I have no clue where the hell it went.

I run SuSE on a few machines and I have never figured out RPM.  I run
Debian on my laptop and a desktop and use apt-get and love it.  I still
install from source for the things that I care about -- like Perl,
mod_perl, and Apache.

>> Do you need Apache 2?
>
>I got Apache 2 ok, but I wanted to use an updated config gui thing for it.
>The one with redhat is for 1.3. Is the httpd.conf complicated at all? Do you
>know of a good gui for it?

Argh -- Personal opinion, but I really do not like a GUI front-end for
something as simple as the Apache config file.  It's a layer of abstraction
that only seems to complicate things.  I build Apache statically (no DSO
support) and my config files can be about 20 lines long.  My advice is to
not start with the default Apache config file and instead start from
something very small.

For example (not tested) but this is something small.

AccessConfig   /dev/null
ResourceConfig /dev/null

ServerRoot /home/moseley/apache
ErrorLog logs/error_log
TransferLog logs/access_log

User nobody
Group users

Listen 12345

# -- Start out very restrictive - applies to virtual hosts -----
<Directory />
    AllowOverride none
    order deny,allow
    Deny from all
</Directory>

# This is for performance
Options FollowSymLinks

# Block files example
<Files ~ "^\.ht">
  deny from all
</Files>

<VirtualHost *:12345>

    ErrorLog logs/error_log.12345
    TransferLog logs/access_log.12345

    AddType application/x-httpd-cgi .cgi
    AddHandler cgi-script .cgi

    DocumentRoot /home/moseley/apache/12345

    <Directory /home/moseley/apache/12345>
       Allow from all

       Options +ExecCGI
    </Directory>

</VirtualHost>

Now, add in mod_perl to apache, which is not hard, and you will be very
pleased with the speed of your searches.

It's basically something like:

<perl>
    use lib '/home/moseley/swish-e/example/modules';
    use lib '/home/moseley/swish-e/example';
    require "swish.cgi"
</perl>

Then in your virtual host:

    <location /search>
        SetHandler perl-script
        allow from all
        PerlHandler SwishSearch
    </location>

And to make it even faster you can use the SWIHSE.pm perl module -- then
all the perl code and swish-e becomes part of the Apache server.

I'll write all that up some day.


-- 
Bill Moseley
mailto:moseley@hank.org
Received on Wed Sep 18 17:40:30 2002