franKenMail? :)

[René J.V. Bertin]

On Friday August 25 2017 12:19:44 Martin Steigerwald wrote:

>Im am pretty much sure that it won´t work without considerable development 
>effort. For the simple reason of KF5 versions of Akonadi + Akonadi agents 
>being Qt5, while KMail 4.14/4.13 being Qt4. Also Akonadi in KF5 has switched 

As I said, that's a moot issue as long as the whole system depends on inTER process communication and not on plugins.
As an example, KDE4's kwallet API works happily with KF5's kwalletd.

>to a faster binary protocol (of course, you could try the versions before 
>that… but then you are out of support either…)

A change in the communication protocol would be a blocking problem of course.

>resource performance and that was with lots of help… but I am pretty much sure 
>that Dan would advise not to try this.

Oh, of course (but he's not against experimenting, I think :))

>I didn´t follow the interface issues thread… but I really wonder what you want 
>to achieve by going that route. 

In short: improvements under the hood with no visible change to the user. The way progress ought to be ;)
Oh, and I do use KNode. A lot. It's really a pity it had to go, there are no modern alternatives for it...

>I´d rather recommend to help tackling whatever 
>interface issues you faced with current KMail.

Hah, current KMail for me is 4.13 on Linux and 4.14 on Mac. I have a few issues with them, but nothing compared to what I keep reading here. I thought things were heading the right direction, but now with the interface "modernisation" going on I'm again very much put off from investing all the required effort in adding PIM5 to my personal builds. Effort and risk: email is a crucial productivity component for me, so I cannot afford starting an upgrade that might bring the whole PIM card house down if you see what I mean.
I would have started getting involved long ago if PIM5 could be installed in parallel to a functional PIM4, like I did with KDevelop. There I could continue to be productive in KDevelop4 while I ported and contributed things I considered crucial to KDevelop5. With PIM5 the only way to achieve that is using a separate system. I do have a VM set up but it's just good enough to evaluate how KF5 behaves in a full-fludged Plasma5 install, not to do serious development.

Currently the only other way for me to get involved with the user interface would be the opposite to the idea I just raised here: use Kontact5, KMail5 etc. with the KDE5 Akonadi backend. I'd expect it's even less likely that would ever work.

R.