[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: Kmail slow
From:       Don Sanders <don () sanders ! org>
Date:       2000-06-10 14:07:18
[Download RAW message or body]

On Sat, 10 Jun 2000, Michael Matz wrote:
> Hi,
>
> On Fri, 9 Jun 2000, Rik Hemsley wrote:
> > > Thats easy too :) In each message store a pointer (or msg-ID) to the
> > > sibling message. So if you have a parent P, you get the first child C
> > > by P->child, and all other children by following C->sibling. (like for
> > > (c=p->child; c; c=c->sibling) {do_something_with(c);} )
> >
> > Are you volunteering to write the algorithm to set those sibling
> > references up ? :)
>
> Yes. After KDE 2.0, like you ;) (At least if I remember by then, or
> somebody reminds me)
>
> > Actually, it's not really a good plan, as messages don't have siblings,
> > they only have parents (specified in 'References' headers), so it would
> > be a weird model to use.
>
> You can never have enough data about a graph and relationship between
> nodes ;) With this one could implement a fairly fast threading system.
> Note that I never used empath or kmail, so I don't know anything about the
> implementation as it is now, may be its efficient enough.

Both could be faster, threading especially.

The difficult problem is working out how to set up (and incrementally update) 
the database so that you can resolve certain queries quickly.

Eg. Return messages 60,000 to 60,100 in the correct order for a threaded view 
of messages sorted by subject.

This is not an easy problem to solve efficiently.

BFN,
Don.
 
>> Visit http://master.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

[prev in list] [next in list] [prev in thread] [next in thread]