yes same when launching direct from file explorer. yes same computer as DC. But I did recently change my GPU from amd RX 580 to AMD RX 6700 XT, I am not sure if that was before or after. Ok I just launched DC and yea it does the same thing.
idk about the integrated chip, like i believe it exists but there is no reason my computer would be using it. I am plugged directly into my GPU via displayport? no idea
Ah, a change in GPU would definitely explain why you didn't have any issue before (there was no change at all to the display code between Dear Clockmaker and Zendo).
Are your drivers up to date? Because that really looks like a driver bug...
I've tried a small change to the display code that you could download independently from the game (can't update during the voting period), to see if perchance it fixes the issue: https://drive.google.com/file/d/1JDK0C7XWdVYmRBNPWErXcoVtMRglyu17/view?usp=drive... (download that file and drop it in the same folder as Zendo.exe, overwriting the existing version). Though I'm not very confident it will fix anything, because the code looks correct and does work on every other GPU I've encountered so far :/
If you try it, please tell me if it changes anything. Thanks!
The T-Pose is consistent with positions being randomly set to zero, just like the environment. I notice the color overlay is also impacted, which makes sense. Though it doesn't seem entirely random, there's a pattern behind this madness... Those values come from somewhere, but I don't know where yet.
It's very difficult to troubleshoot on my side without being able to reproduce it on any of the GPUs I have access to. But if you have a few minutes to spare, we can play a game on ping-pong where I send you a new version of the drawing module and you test it for me. I've uploaded a new version at the same link: https://drive.google.com/file/d/1JDK0C7XWdVYmRBNPWErXcoVtMRglyu17/view?usp=drive...
I would be very interested to know if this version works (though the framerate may not be good). Thanks for your help! ;)
That's good news! At least you can play the game now, though I'm not ready to officially commit to this fix just yet.
So, thanks to your video I've noticed the positions of the meshes were not random at all. It felt like at least some parts of the environments were placed, scaled and rotated just like the character parts should! But on the other hand, the character parts were never scaled, despite the fact a lot of the environment is. So I assumed some meshes were drawn using the previous content of the buffer. I'm not sure why. From what I understand of it, it goes completely against the OpenGL specs, but that wouldn't be the first time a driver optimization goes wrong. My "fix" was to ask for a new buffer every single time, to make sure I start from something clean. That's completely overkill and why I was concerned with the framerate (and I do notice a general slow-down on my 11-years old GPU). There's probably a better workaround... I have a few leads if you're willing to test them for me?
And yes, the map is revealed from the start (your character already knows this place well!) and there's no marker for your current position. It was a (very) late addition to the game and I assumed the map was detailed enough that you could quickly locate yourself from your surroundings. The most important part about this map is that you can basically write on it ;)
