Re: Various little fixes + my plans for balsa this summer

[Pawel Salek <pawsa TheoChem kth se>]

On 2000-06-25 21:08 Ian Campbell wrote:
> I am still waiting for CVS write access, so here is a patch to fix a
couple of
> small things:
> [..]
> I have added emacs mode hints to the top of files I touched, to force
emacs to
> the 2 space indenting. I intend to do this to any files I happen to touch
as I
> work (although really I would prefer to change everything to a more
normal
> style ;-)

Commited, thanks. Concerning the indentation style, after listening to all
the discussions concerning the indentation depth etc, _I_ tend to use in new
functions bsd style indentation with 4 space tab width which hopefully is a
reasonable compromise between 2 and 8 :-), and wrap lines at 80 characters
so I can print the code nicely. But this is my choice, of course. What do
you think gouys about discussion/voting for a common style, so balsa code
becomes more uniform?

> For example I would very much like to see some fixing of the name space
used. I
> think all libbalsa functions/structs should be prefixed with libbalsa_
or
> LibBalsa and balsa ones with balsa_ or Balsa, for example there are two
> BalsaMessages at the moment, one in src and one in libbalsa, it's
basically
> confusing. 

I am not sure about it. It is fine only If the function is not related to a
structure/object. For me, function that begins with message_ has to do
something with a message object and I know it's defined in libbalsa. Is it not
clear enough?

 
> I would also like to remove any usage of libmutt from balsa and
encapsulate it
> in libbalsa, since libmutt causes so many problems I think it would be
best to
> get it all in one place, it would also set the stage for removing the use
of
> libmutt for some things prehaps (a long term goal....)

Oh yes, that a very good idea! Actually, I think this has been followed to
a large extent but I wouldn't be surprised if there were some exceptions
that may turn out difficult to get rid of. The problem is balsa uses mutt data
structures and it is then difficult to access them and not use native
libmutt routines. Anyway, it is very noble idea and I fully support it.

> It seems to me that some of the things in libbalsa/balsa have had some
attempt
> at being made into GtkObject's. For example the mailboxes in libbalsa
are
> GtkObjects - but still use lists of watchers - I think they should really
use
> signals. I'd like to take those things and finish turning them into
proper
> GtkObjects.

Sounds good to me.

(Personally, I am in a opposite situation: it seems I will have less time
for balsa during next two months. It's a pity).

/Pawel
-- 
Pawel Salek (pawsa@theochem.kth.se) http://www.theochem.kth.se/~pawsa/
Theoretical Chemistry Division, KTH voice: +46 8 790-8202