On 09/11/2007 02:36 PM, Bill Traynor wrote:
> On 9/10/07, Jordan Hayes <firstname.lastname@example.org> wrote:
>>> Do I need multiple configs to accomplish this?
>> Yep :-)
>> Minimally you'll need a ReplaceRules difference in swish-e.conf. You'll
>> also need a different title in .swishcgi.conf ...
> Here's what I've done thus far:
> In each mailing list archive directory I've created a swish.conf with
> the following:
> ReplaceRules remove <path to the archive>
> ReplaceRules prepend <URL to archive files>
> IndexDir <path to this directory>
> IdexFile <index.archivename>
> Then, in /usr/lib/cgi-bin/.swishcgi.conf I've put:
> sish_index =>
> etc. etc.
> and then a select_indexes statement for all possible archives.
> Following this, I run:
> $swish-e -c <pathto archive1 swish.conf>
> I run that command for each individual archive.
> This works. However, is there a more efficient way to do this? I've
> got 65+ archives and I'd prefer to issue only one $swish-e command to
> index them all.
> Also, am I following swish-e best practices here? That is, should I
> be putting all of my directives in swish.conf? Or somewhere else?
I like having all my data in a single index, storing things like archive_name
as a metaname. That way I can search like:
swish-e -w foo archive_name=bar
swish-e -w foo -f path/to/archive_name.index
That said, I don't know that there's built-in support for that kind of thing in
swish.cgi. I don't use that script myself.
But by storing archive info as a metaname, there is only one .conf file at
indexing time, and one indexing run to make. The only time I don't use a single
index is if the sheer number of docs in the collection starts to make indexing
(and re-indexing) prohibitive, either because of memory or time constraints.
Peter Karman . peter(at)not-real.peknet.com . http://peknet.com/
Users mailing list
Received on Wed Sep 12 11:28:09 2007