Note that local hostnames are also excluded, and IMO that's also not helpful. My weekend-regexp-foo isn't sufficient to fix this right now, perhaps someone wants to give this a try?

@sitter here's a review sending the reparseConfiguration signal from the Colors KCM.

Done (assuming editing the phab description is enough).

• Qt4 compat

• use QApplication::paletteChanged as Kai suggests

I agree. I initially preferred this approach since the other one involves installing a very global event filter. Since you proposed a cleaner solution there, let's go with that. I'll abandon this one.

Plasma/Wayland supports this, I've got a prototype plasmoid which allows to autorotate on X. Needs a bit more polishing and can be offered separately then.

In principle yes, but the applet should of course react suitably to these constraints. A config option for this seems to duplicate the concept of constraints (which are indeed determined by the containment.)

I second that the icon should be made red, this mitigates the visibility issue and makes it consistent with the disconnected ethernet icon. @andreask, can you make this change?

• clean up now-dead code

OsdManager::showActionSelector() will delete existing OSDs before showing a new one, so we won't create multiple OSDs. That's arguably not as nice as tracking if it's already shown, but I think I'd rather avoid the added complexity of adding a check if it's already there, if that's OK from your POV.

When landing, can you merge the whole branch into master, the history here is relevant and it will make continuing work on my kscreen kcm qmlification branch a lot easier. Thanks!

That's what the "old" bookmarks runner was doing, it loaded the bookmarks from the file the browsers write out.

One minor thing to fix, otherwise the idea + approach look fine to me. Thanks!

I like, getting rid of C++ code is always welcome.

Yeah, so could you abandon this phabricator review? It makes it harder to sift through reviews that actually are interesting...

In D9414#193341, @dvratil wrote:

Please disregard the noise in this patch, sebas merged master into the dvratil/osd branch so I can't generate a proper diff against the base revision without the noise from master.

Alright, please abandon this patch then, it's definitely wrong.

Dingdong?

Okay, patch as-is now definitely breaks unplugging outputs, it eats these events. I'll mark it as such, if you fix it in another way, either update this or create a new review.