Re: wanderlust is sooo slooowww...

[Ken Gunderson <kgunders@teamcool.net>]

At Mon, 27 Oct 2014 10:06:44 -0700,
Erik Hetzner wrote:
> 
> At Mon, 27 Oct 2014 16:58:08 +0100,
> David Maus wrote:
> > 
> > I think it does not have anything to do with server load but with the
> > server software's search implementation and the search queries sent by
> > Wanderlust. For example some older versions of Courier IMAP didn't
> > perform searches for arbitrary header fields very well (IIRC).
> > 
> > @Erik: Do you see a chance to ressurrect the old client-side search
> > logic and provide it on an optional basis? I.e.: How hard would it be?
> 
> Hi David,
> 
> I believe the only thing that really moved server-side was the OR
> logic (or was it AND?). Iʼd be curious to see what is so slow on Kenʼs
> IMAP server.
> 
> I donʼt really understand Archlinux packaging, but it seems the change
> Ken is referring to is:
> 
>   https://projects.archlinux.org/svntogit/community.git/commit/trunk?h=packages/wanderlust&id=6516664e9a8423f2749316cab302b257fa1cefba
> 
> and the packaging switched from a rather old version of WL (found
> here: https://github.com/p5n/wanderlust) to one built from the github
> wanderlust/wanderlust master branch. This previous version does seem
> to not include my search changes.

btw-  Wanderlust from Erik's e6h repo exhibits same behavior.

IMAP server log returns:

"Fatal error: word too long"

Emacs messages buffer returns:

Searching...
elmo-imap4-process-bye: IMAP4 session was terminated: "Fatal error: word too long", "Try Again"
Saving...done
IMAP4 session was terminated: "Fatal error: word too long", "Try Again"
Saving...done

So at least that's consistent, lol....

I'll send elmo debug in a bit.  Got to take care of something first.

Best-- Ken