Re: Mime part viewers, support for PGP, ICS files and others

[Philip Van Hoof <spam pvanhoof be>]

On Thu, 2006-10-12 at 17:05 +0000, Nate Nielsen wrote:

> > Another example will be a PGP mime-part. The action to handle (viewing
> > it) the PGP mime-part might simply be asking Seahorse whether or not the
> > PGP key is a valid one. And if not, Seahorse would probably return
> > something bad so that the view can show a red label telling the user
> > that no, the key is not okay. Or a green label telling the user yes, the
> > key is okay.
> 
> The concept is cool.
> 
> However, this specific example won't work as expected. There's no way to
> return information about a non-imported key. The GPG developers state
> (and they are technically correct) that importing a key into the user's
> public keyring is a non-destructive and non-security sensitive operation
> by nature. This is true when the user uses 'PGP owner trust' and the
> 'web of trust' properly.
> 
> I imagine that when this PGP mime-part is implemented, the seahorse DBUS
> interface (libcryptui actually) will probably need some upgrades. It's
> currently in its infancy. I'd be willing to do work with you on what's
> needed.

Yes. That would be great indeed.

I'm probably first going to design the interfaces in such a way that
this concept can be turned into something real and touchable.

Then I will implement a few implementations of the TnyMimePartView type,
for example one for displaying a normal attachment mime-part, one for
displaying the text body of an E-mail, one for displaying a text/html
body of an E-mail. Refactor the existing code to the concept, etc etc.

And then I will probably implement the specific TnyMimePartView
implementations. For example the one for viewing a PGP key, another for
displaying a ".ics" mime-part. For example so that I can have a button
that says: "Register this item with Dates" or whatever.

It would indeed be great to implement those with the aid of you guys
(Chris and Nate) and perhaps get an interface into your infrastructure
(like libecal or a D-BUS interface for adding items to Dates, and an
easy entry into Seahorse for verifying a PGP key).

My idea is really to push "implementation" to your infrastructure. I
don't want tinymail to force a developer to use a specific piece of
implementation to get things like, registering a meeting request with
the calendar software or verifying the PGP key, done.

I believe it doesn't have to be hard, and that we (this means, Chris,
you, me and other people who are working on this type of stuff) can
actually get it right and do it without having to spend huge amounts of
time on it.

We, indeed, all have huge todo lists ;-). I know. Let's not translate
that into "bad quality" APIs. I'd rather have a very good platform
tomorrow after tomorrow, than a piece of shit platform tomorrow.

Anyway, let's talk about it. Right? :)


-- 
Philip Van Hoof, software developer at x-tend 
home: me at pvanhoof dot be 
gnome: pvanhoof at gnome dot org 
work: vanhoof at x-tend dot be 
http://www.pvanhoof.be - http://www.x-tend.be