Re: Using MessageComposer to create invitation emails

Daniel Vrátil <[email protected]> Wed, 19 Apr 2023 20:49:56 +0200
Newsgroups	gmane.comp.kde.devel.pim
Message-ID	<3223312.aeNJFYEL58@brokkr>
On Wednesday, 19 April 2023 19:21:42 CEST Carl Schwan wrote:
> On Wednesday, 19 April 2023 17:42:33 CEST Daniel Vrátil wrote:
> > Hi all,
> 
> Hi,
> 
> > I'm looking at  bug #297956 [0] - invitation emails from KOrganizer do not
> > honor signing/encryption preferences.
> > 
> > Currently iTip messages are built in MailClient in akonadi-calendar repo,
> > which is a fairly simplistic and straightforward piece of code.
> > 
> > After a brief look at how signing/encryption is implemented in KMail it's
> > obvious that creating signed/encrypted messages requires non-trivial
> > amount
> > of logic and code.
> > 
> > Luckily most of this code and logic lives in  MessageComposer library
> > inside messagelib. It makes a lot of sense to me to reuse this code
> > inside akonadi-calendar to generate signed/encrypted iTip messages with
> > behavior and UI consistent with KMail. However it requires making
> > akonadi-calendar to depend on messagelib, which has non-trivial
> > dependency chain.
> > 
> > What do you folks think about this? Do you think it's OK to introduce this
> > dependency? Or would a simplified sign/encrypt implementation directly in
> > akonadi-calendar suffice?
> 
> We tried in the last PIM sprint to actually reduce the number of
> dependencies in Kalendar which resulted in splitting some libs in a
> core/widget parts and moving some stuff around.
> 
> But Kalendar already depends on mailcommon which depends on messagelib so it
> won't make it worse.

Good :) I'll proceed, then.

> 
> Regarding the encryption handling in messagecomposer I wish this could would
> not depends on QWidgets as it would simplify a few other things in Kalendar
> as well as make it easier to add to akonadi-calendar. I tried to do that in
> the past but converting all the blocking calls from KMessageBox to be more
> async is a bit complicated.

Yeah, the code seems quite spaghetti-like to me :) Major refactoring of the 
composer is out of scope of what I'm working on right now, but I can look at 
that independently later on.

/Dan


> Cheers,
> Carl
> 
> > Cheers,
> > Dan
> > 
> > [0] https://bugs.kde.org/show_bug.cgi?id=297956
signature.asc (application/pgp-signature, 833 B)