But it makes no difference. I think that your explanation actually makes kinda sense. But I still fail to see, where EXACTLY is the problem here, actually.
EDIT: If I understand you correctly, you think actually the opposite is true: That the flag is set, while it shoudn't be. I don't think this is the case. From QDBusMessage documentation:
By default this flag is false and the other end is expected to make any authorization decisions non-interactively and promptly.
This flag sets on the message itself, so it has to be in the default state, as the message is created just one line above.
From the error description you provided, I understood it the other way around:
** Access to the requested operation is not permitted. However, it might be available after interactive authentication. **
Anyway, I've tried to set it both ways, no difference there :(.
Actually, while reading the bottom documentation snippet again, I feel like that flag only applies to use cases, where the user is already in a "rooted session" (I don't know if something like this even exists, but if you for example run pacman and install a package, it prompts you for a password, but it doesn't if you run it again shortly after).
Well I guess the flag is set to true, so it actually wants an interactive authorisation.
Also: In which file and on which line is the function that you edited? I would like to look around and maybe try some compilation attempts myself
EDIT: Okay, so let's maybe try something mildly stupid: Could you try commenting the return xyz line and replace it with return true;?
EDIT: Well, this is interesting. I actually see a mouse cursor for a moment, but then it fails anyway. Interestingly, when I try to redirect the logging output to a file, I get something completely different than what I see on the screen, there, in the end. Let me investigate it a bit more....
Anyway, I wasn't able to extract the logs, yet. As I was trying that, I've run kwin directly without the 'dbus-run-session' part.
The result was different ... well. This time it didn't terminated. So I was staring on black screen with a mouse cursor, and I wasn't able to exit or kill it in any way. Not even switch into a different tty. I had to hard-restart my computer XD. So, beware, if you're going to try this
Exactly. It is a progress. I will run it again with the dbus session, and will study the logs. Meybe there's some clue.
But step by step, we are getting there
EDIT:
Ok, a bit rudimentary, but desperate times ask for desperate actions
The "logind: openRestricted" and "so far so good" messages are mine, so don't be confused if you don't see them
Anyway, the messages I've added show, that kwin goes through DrmBackend::addGpu method multiple times, like it retries it over and over....
BTW, I've exceeded a maximum number of replies (as a new user}. So I probably have to stick to extending this post for the time being
FINAL EDIT FOR TODAY:
Nice progress!
I have to let it be, for today, unfortunately (family duties ;)), but I will continue in the following days for sure, as this is getting really interesting. I'm neither on Telegram or Matrix so far, but I might try to create an account there, and we will defeat it, finally!
Thanks for a pleasant cooperation, have a nice evening, and hopefully see you again soon
1: Well the filesize is 0, but it's a kind of harware "link" I guess. -1 Probably means the file does not exist
2: That prints a message, which doesn't happen here
3: The file is not ... loaded? No clue
4: The file ... loading fails?
Could you try to insert a lot of debug prints into the openRestricted function? It seems to compile for you faster than for me. I will try the same in the meantime
EDIT: I tried to run wayland, and wouldn't you know, I got the same black screen as you, but I can move the mouse and even open programs! So close!
BTW: Are you in the Garuda Telegram or Matrix channel? As you can't write stuff here anymore, maybe we can talk over at that end
I find it rather confusing, that dbus-run-session ./kwin_wayland works, but dbus-run-session kwin_wayland doesn't .... Is the kwin-git package working and the stable package not? I can't really replace kde with the git stuff, because something interferes, but this is weird.
might try pacman -Rdd kwin and pacmasn -Sdd kwin-git
I had the same issue. I found that disabling ananicity cpp in the garuda assistant fixed it. I have no clue why but it did. Is there someone somewhere we could report this issue to ?
Wooot @aviallon is also in the forum, dunno if he checks it though. Ananicy-Cpp just adjust nice levels, does this also occur when doing this manually?
let's go for it 
