Re: new version of allout.el - patch and ChangeLog

[Ken Manheimer]

On 9/30/05, Ken Manheimer <address@hidden> wrote:
> On 9/30/05, Sascha Wilde <address@hidden> wrote:
> > On Fri, Sep 30, 2005 at 01:24:14PM -0400, Ken Manheimer wrote:
> > [...]
> > >           topic encryption functionality depends on both mailcrypt and
> > >           crypt++ packages.  (non-encryption allout functionality is not
> > >           disrupted if either mailcrypt or crypt++ are absent.)
> >
> > why do you relay on these packages, which aren't part of gnu emacs,
> > when we have pgg?
>
> good question - i wasn't aware of pgg.  can you tell me more about it?
>  i'd be very happy to find a single package that cleanly provides the
> encryption interface that allout needs!

andreas pointed me at pgg's bundling in gnus, for which gentoo has an
ebuild.  i emerged it and got access to gpp and the info file.  it
looks nice - clean - but appears to have the same drawback, for my
purposes, as mailcrypt.  it will decrypt messages encoded with a
symmetric key, but provides no api for encrypting with a symmetric
key.  that lack is a show-stopper for my purposes.

that mailcrypt and pgg would concentrate on public-key mode makes a
lot of sense, considering the ultimate purpose is encryption and/or
signing of messages to be exchanged between remote parties - email,
news, software packaging, etc.  my primary concern for encryption in
allout is with conveniently securing notes for oneself.  the lattitude
of symmetric keys offers much better convenience in that context.

i actually inquired on the mailcrypt developers list about the lack of
provision for symmetric-key encryption, and the general concensus was
that it just was an oversight, not a deliberate avoidance.  i presume
(as i suggest above) because it's not useful for the message-exchange
realm.

it may well be that pgg is preferable, somehow, to mailcrypt - they
both offer similar functionality, including key caching, as well as
the lack of an interface for encrypting with a symmetric key.  but the
changeover won't gain me anything i can yet see - i'll still need to
also use crypt++ for the symmetric key encoding.  and the cost in
effort of switching from working code would not be trivial, so
until/unless i see a compelling gain, or someone else does the
integration (and convinces me that pgg is more widespread than
mailcrypt), i'm think i'm fine with mailcrypt.

ken
address@hidden