LightDM-KDE 0.1.0 Released

[Hirager]

LightDM is currently not cross-platform. Or have you seen a port to FreeBSD or any other non-Linux OS anywhere? LightDM is ? within the Linux ecosystem ? cross-desktop but that's different from cross-platform.
The blog post from Matthew Garrett is correct in that regard that it makes no sense to replace a well maintained, fully working display manager with something LightDM-based. LightDM is indeed relatively new code. Simply throwing everything out just for fun is crazy. Granted, it's ironic to read that from a GNOME person, because the GNOME project did such stupid things several times (PulseAudio, GStreamer, Empathy). ;-)

In case of KDE the premise is entirely different: KDM ? while working well ? is relatively unmaintained and full of legacy technologies. An entirely new front-end (or Greeter in LightDM terminology) using QML/Plasma would be needed to be written anyway and ? as Matthew Garrett also wrote ? LightDM makes it easy to start from scratch.
From KDE?s POV using LightDM makes a lot of sense: The only required workload is the new Greeter. The rest is managed by other people.
KDM can easily be kept until LightDM-KDE is ready (incl. BSD ports).

Yup, my bad choice of word.

[Awesomeness]

The difference is, LightDM is cross-platform. Personally, I think replacement of all other login managers by this LightDM is a great idea.

LightDM is currently not cross-platform. Or have you seen a port to FreeBSD or any other non-Linux OS anywhere? LightDM is – within the Linux ecosystem – cross-desktop but that's different from cross-platform.
The blog post from Matthew Garrett is correct in that regard that it makes no sense to replace a well maintained, fully working display manager with something LightDM-based. LightDM is indeed relatively new code. Simply throwing everything out just for fun is crazy. Granted, it's ironic to read that from a GNOME person, because the GNOME project did such stupid things several times (PulseAudio, GStreamer, Empathy). ;-)

In case of KDE the premise is entirely different: KDM – while working well – is relatively unmaintained and full of legacy technologies. An entirely new front-end (or Greeter in LightDM terminology) using QML/Plasma would be needed to be written anyway and – as Matthew Garrett also wrote – LightDM makes it easy to start from scratch.
From KDE’s POV using LightDM makes a lot of sense: The only required workload is the new Greeter. The rest is managed by other people.
KDM can easily be kept until LightDM-KDE is ready (incl. BSD ports).

[steveriley]

Interesting replies, thanks. I'll admit I was operating under the presumption that KDM is roughly equivalent to GDM, and thus thinking that Matthew's argument against replacing GDM with LightDM was applicable to KDM vs. LightDM-KDE. Seems to be not so.

[Hirager]

KDM is so damn unmaintained, I doubt that KDM even uses relatively recent Plasma Workspace features. So whether to write code to make KDM use those features or do the same for LightDM-KDE shouldn't matter much.

The difference is, LightDM is cross-platform. Personally, I think replacement of all other login managers by this LightDM is a great idea.

[Awesomeness]

KDM is so damn unmaintained, I doubt that KDM even uses relatively recent Plasma Workspace features. So whether to write code to make KDM use those features or do the same for LightDM-KDE shouldn't matter much.

[Teho]

But itsn't the issue more than just easier porting KDM to QML so that we can have pretty greeters?

Maybe if someone actually did the job. The tasks itself is much easier with LightDM though. Plasma based KDM was planned for years but never took off. LightDM-KDE has been under active developement for a long time now. KDM hasn't seen much significant development and the thing with open source developement is that if someone isn't going to find it interesting no one will do it.

Matthew Garrett argues against LightDM regardless of what your DE is.

I think it should be quite obvious that not everone agrees.

LightDM appears to require re-implementing much of what the DE already does, so ultimately there's duplicate code hanging around, and potential for policy mismatches.

KDM doesn't seem to have any power management functionality and the policies are user defined anway so I don't see how that is so big loss. I don't know anything about these kinds of things but wouldn't it be possible to implement dbus API for LightDM powermanagement and implement that to PowerDevil or something to keep them in sync? The greeter itself shares code with Plasma. I guess there are some dublication but so is there now with KDM.

[steveriley]

But itsn't the issue more than just easier porting KDM to QML so that we can have pretty greeters? Matthew Garrett argues against LightDM regardless of what your DE is. LightDM appears to require re-implementing much of what the DE already does, so ultimately there's duplicate code hanging around, and potential for policy mismatches.

[Teho]

Why is this thing necessary? What will LightDM-KDE do for me better that KDM?

It takes less time to port to Wayland and has superior theming capability thanks to QML. Porting KDM to QML would have been a lot more difficult if I'm not mistaken. Thanks to lightDM it requires less maintance from KDE developers that has been a problem for KDM for a long time. KDE developers no longer have to build all features from scratch because lightDM is shared project with Ubuntu/Unity and possibly more in the future.

[steveriley]

Why is this thing necessary? What will LightDM-KDE do for me better that KDM?

[pejakm]

Well, it's almost usable.