[Users] Claws Mail 3.9.3 unleashed!!

[Michael Gmelin]

On Wed, 11 Jun 2014 17:26:40 +0200
Michael Gmelin <freebsd at grem.de> wrote:

> 
> 
> On Wed, 11 Jun 2014 17:19:09 +0200
> Colin Leroy <colin at colino.net> wrote:
> 
> > On Wed, 11 Jun 2014 16:47:21 +0200, Michael Gmelin <freebsd at grem.de>
> > wrote:
> > 
> > > > New in this release:
> > > > ~~~~~~~~~~~~~~~~~~~~
> > > > 
> > > > * The TAB address completion in the Compose window now matches
> > > > any part of the address and not just the beginning.  
> > > 
> > > I just updated to 3.10 (long overdue) and got this new feature, is
> > > there any chance to disable it by configuration?
> > 
> > address_search_wildcard=0 is what you want, a hidden preference that
> > is available starting from 3.10.1.
> > 
> 
> Cool, thank you.
> 

What I would like to see is matching any, but ordering by best match.
So if I type "colin" it would first show "colin at colino.net" (best
address match), then some other Colin (best name match) and then
"NameContainingColin" as the least best match.

Would this make sense to you?

Unrelated question:
I switched to 3.10.1 now and realized that cursor keys will jump
between fields in the compose window (in addition to the classic TAB /
Shift TAB method) - especially leaving the body to get to subject. Since
this is not really symmetric (moving up all the way to "Header" doesn't
change anything, but moving down will get stuck in the "From:" select
box and also change the selected account in there) I would like to
disable this feature - also because I have the habit of holding down
the cursor up key to get to the first line and I end up outside of
the body field now. Seems like unexpected behavior for a normal GTK
application to me.

Seems like this has been introduced unconditionally in 
 
http://git.claws-mail.org/?p=claws.git;a=blobdiff;f=src/compose.c;h=7a1f177319cdde802bf09b3450fb655646492588;hp=cad91a1ae8280416a09606100e7e46a585976c8f;hb=a5a06781957dde46bf24e75d0c2e9919a980ea54;hpb=b976c1dc267711cee1bfea10ab241b38b2c99a21

Would it be possible to add an option to disable this feature? If it's
too much hassle to implement, would you accept a patch?

-- 
Michael Gmelin