From mboxrd@z Thu Jan  1 00:00:00 1970
From: gsf@fruct.us (Gabriel Sean Farrell)
Date: Wed, 23 Jan 2008 14:25:21 -0500
Subject: [sup-talk] sup crash on tab completion
In-Reply-To: <1201072070-sup-2323@south>
References: <ad5e29c50801182121k7299b14anf352653d648d4342@mail.gmail.com>
	<1200987490-sup-7392@south> <1200994656-sup-4998@tomsk>
	<1201072070-sup-2323@south>
Message-ID: <20080123192521.GA30955@manheim.library.drexel.edu>

On Tue, Jan 22, 2008 at 11:09:10PM -0800, William Morgan wrote:
> Reformatted excerpts from Marcus Williams's message of 2008-01-22:
> > Note that you can use wildcards... "from:joe*" will find emails from
> > joey/joe.bloggs/joel etc.
> 
> Wow, I totally forgot about that.
> 
> > Although it must be said I was pondering yesterday whether it was
> > possible to get searches assume wildcards some how so when you
> > searched for from:joe you would actually be searching for from:*joe*
> > or maybe from:joe* automagically. I dug around in the ferret docs to
> > see if this was possible but only came up with the fuzzy operators and
> > you might as well use a wildcard instead.
> 
> I don't recall seeing anything like this in the Ferret docs, so it's
> probably not possible. What I'm thinking of doing, which will kind of
> approximate this, is adding the LHS of the email address and the
> sender's name to the from: field at index time, so that email from you
> would match e.g. from:marcus, from:williams, and from:marcus-sup. 

Using Ferret's LetterAnalyzer on the "from" and "to" fields should do
this for you (i.e., create the appropriate tokens).  I've stuck it in
the code, but I haven't been able to make it work.  Something to do with
the twice-over query parsing, I think.

gsf