[htdig] RE: [htdig3-dev] Perl interface


Tillman, James (JamesTillman@fdle.state.fl.us)
Mon, 20 Sep 1999 07:04:57 -0400


>-----Original Message-----
>From: Geoff Hutchison [mailto:ghutchis@wso.williams.edu]
>Sent: Friday, September 17, 1999 6:56 PM
>To: Tillman, James
>Cc: htdig3-dev@htdig.org
>Subject: Re: [htdig3-dev] Perl interface
>
>
>At 2:04 PM -0400 9/17/99, Tillman, James wrote:
>>Is someone already working on a perl interface to htsearch? I and a
>>friend of mine are interested in doing the work, but don't want to
>>duplicate anyone else's effort.
>>What we really want is an XS module.
>
>There are some Perl wrappers in the contrib/ directory, notably the
>ewswrap script. An XS interface would certainly be a logical step up,
>especially if the XS code hooked into the htcommon/ code directly.
>
>-Geoff

We had checked out the perl-wrappers, and had the same thoughts. Looking
over the source of htsearch.c, we're thinking it would be easy to work up a
C++ class version of it that we could use from our XS code. Then all we'd
have to do when the main htdig project brings out a new release is merge the
new htsearch.c changes into our htsearch class. We want to avoid
duplicating the logic in htsearch.c in our own code (esp. since it's likely
to change as time goes by). Comments?

We're really pleased to see an open-source content indexing system under
development. We've even got an open-source project in mind that we want to
use ht://Dig in sometime soon. What we'd really like to do is a perl DBI
driver for htdig, but even just a regular module allowing queries and
results would be a boon to perl programmers.

One other concept we had in mind was a database indexing system. Since
ht://Dig requires a web server's http 1.0 responses to index a site (this
_is_ still true, right?), we thought the ideal solution would be to set up a
database-reader daemon that spoke http 1.0 and returned the database results
as HTML. I've seen that Microsoft's Index Server does something like this,
only much less efficiently. Does anyone see immediate problems with this?
My colleague and I think doing a daemon like this Perl using DBI, CGI, and
the LWP library would be a snap. The urls would look something like this, I
think:

http://hostname:8081/dbserver/db/dbtable/

That would return a list of links back to the same url, but with a record
indicator appended, like this:

http://hostname:8081/dbserver/db/dbtable/person_nbr=134&last_update=12/31/19
98

Let me know what you think of that. If there's interest we can work on
that, as well. Unless there's dissent, we'll get cracking on the XS module.

Jamie

------------------------------------
To unsubscribe from the htdig mailing list, send a message to
htdig@htdig.org containing the single word unsubscribe in
the SUBJECT of the message.



This archive was generated by hypermail 2.0b3 on Mon Sep 20 1999 - 04:09:16 PDT