Re: Hydrant(mailutils's MUA) ideas

[Jeff Bailey]

On Tue, Oct 16, 2001 at 02:24:59PM -0400, Alain Magloire wrote:

> > Ah, such is life.  Perhaps it will change to include C++ bindings. =)
> 
> 8-)  Good try.

> > > I got more ideas on Hydrant (can we change name again? 8)
> > 
> > You don't like the name? =)
> > 
> > > Language C++
> > 
> > That was the plan from the beginning.
> 
> It was for me, but there was talk about scheme, python etc ..

I think that most of the layout and stuff could be exported to scheme
trivially and without loss.  The underlying infrastructure I want to
be C++.

> > > Can I have write permission to the repository since it Hydrant will
> > > not be included in mailutils.
> > 
> > Sure, I'll sort that out tonight.  I've been getting more annoyed with
> > mutt recently.
> 
> Thanks.

Fini.

> > > Things are a little sketchy and I'm not in to writin docs 8(
> > 
> > Just create a notes file and jot stuff down.  There's no point in
> > writing the docs for this until something's started to be written.
> 
> OK.

Just create NOTES in doc/

> > How is Jakob?  I wonder if he's in Afghanistan.

> That would be to early, he's still doing basic training.  Jakob has
> always strike as a smart sensitive person that does not fit the
> profile of the typical GI 8-)

Oh good.  I hope he's well wherever they've put him.

> > The extensibiilty is very important to me.  GNU for a long time
> > couldn't do anything with encryption because of US export regulations.
> > I suspect they will get stronger again, and I want anyone to be able
> > to just load a simple mod to get it.  Preferrably without recompiling.

> hmm ... that it is something I did not approach at all in mailutils
> encryption, PGP, signatures etc ... I suspect it will possible to
> added cleanly whithout shaking the foundations too much.

I think in many ways mailutils shouldn't have those.  signatures and
other hooks are simply part of the message body normally, and PGP just
needs to act on individual MIME objects which are already exported to
the app.

> The real(visible) change is in memory management, where it will
> be required that you xx_destroy() any object you xx_create ()
> or xxx_get_yy ().

Ah, even more reason to have c++ bindings.  =)

-- 
Abandon the search for truth; settle for a good fantasy.