Does
Works with 25.1.0-rc1.
imply that this is a mesa regression? If so, why is this being reported here?
25.1.0-rc1 was the solution too #2445. This is #2450. I felt it was important to include proper history and it also serves to provide system and application details.
Sorry, my autism makes it difficult for ppl to understand me... Witch I find so confusing because gcc seems to understand exactly what I'm getting at.
Ohh, I get what you are saying now... Yes, it could be a bug in mesa. What I would do in cases where "I don't have any evidence either way and there for don't know." is file bugs in both projects. Someone from vkd3d and someone from mesa should work out what the best solutions should be. There are many times where a fix can be placed on either side of an interface and sometimes even if it violates the spec it simply makes more sense to have 30 lines on one side instead of 300 on the other.
But if it's already fixed in 25.1, then there isn't any further action to take on either side unless it's considered very important to make it work on 25.0, which I don't think this issue is.
That's what I thought doitsujin was saying, but as I said previously there are multiple issues here and switching to 25.1 fixes one... but CREATES the other. I get thinking because 25.1 is so new it's likely been changed to create this issue, that fails though because Oblivion Remastered is the newest thing here. and if that's that's we are going by to determine who is responsible for fixing the issue, that's insane because it might take 300 lines of code in Oblivion to do the same thing a one line change in vkd3d could accomplish, so going by "newest release must apply patch" is just a poor idea.
Ok, I see, so once you got past the hurdle of missing 64-bit atomics pre-25.1, then you start running into this issue? I.e. it was never actually fixed in 25.1, which is what was confusing about the report.
A mesa issue for the problem has been created https://gitlab.freedesktop.org/mesa/mesa/-/issues/13063. I assume it is the same problem at least
Edit: Don't know if it is related to this other issue i have reproduced https://gitlab.freedesktop.org/mesa/mesa/-/issues/13046
That one is also present in other UE5 games with anv
Sounds like it's just an ANV bug really. If it doesn't reproduce on other GPUs, not sure what vkd3d-proton can do unless it's debugged exactly what is wrong.
I think it's wrong to assume the same code in vkd3d is run across different devices, that assumes games render the same across different devices and I hope you already know that's not true.
I'm not in any position to select the correct project, like this for example: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13049#note_2886860
Did you want a issue for it?
That looks like a vkd3d-proton issue that can be filed separately.
Originally posted by @cheako in [#2445](https://github.com/HansKristian-Work/vkd3d-proton/issues/2445#issuecomment-2831398270)
Behind that boulder right where the Xhair is, there should be a building!!! IDK what the other junk is, but it's persistent throughout the game. Sometimes I get the feeling the things that are being misplaced can even be outside the view frustum.
This is near the sewer entrance(that black circle, that could be just for the screen shot as I don't remember that looking bad) in the market district of the capitol... looking at Rinder's Staffs.