Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 19 Mar 2012 06:17:50 +0400
From: Solar Designer <solar@...nwall.com>
To: owl-dev@...ts.openwall.com
Subject: Re: blists status?

Hi Lauri,

Sorry for the delayed response.

On Sun, Feb 05, 2012 at 05:47:49PM +0200, Lauri Kokkonen wrote:
> Okay. I would like to be able to search messages by author, by subject
> and by message body.
> 
> This is my initial idea for building a search index from scratch: For
> each word store pointers to all messages that contain that word and
> encode these pointers as ranges of messages in time or in thread. So
> store the first message and the number of following messages that also
> contain that word.
> 
> For an example, searches like the following would be quite
> straightforward:
>   word1 AND word2 AND NOT word3
> 
> For searches like "word1 word2" (paragraph) we would need to do:
>   word1 AND word2
> and then fetch the message bodies to check if the words are next to
> each other.
> 
> But this would certainly need a lot more time, consideration, testing,
> statistics and so on to become something useful.

Right.  So we'd need to decide on using a homebrew approach like this
vs. just going with Xapian or the like.

> So, I could start by doing something small first, like writing comments
> before few functions in mailbox.c, then proceed to write a flat threaded
> view and maybe remove duplicate code in html.c if needed.

Yes, please feel free to contribute patches for that.

Thanks,

Alexander

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.