
Stray on AnyPS5
Boots Reaches its first-boot brightness screen on Linux with a stack of unmerged patches; nothing past that screen has been tried.
- Nothing
- Boots
- Menus
- In game
- Playable
Linux: reaches the first-boot brightness screen (rendered correctly, 57-58 fps) only on a local stack of about a dozen PRs, many of which conflict; plain main stops 9 seconds after launch with no window.
Running it and it stopped on an error? Check the error against every report and fix for this game.
Screenshots
No screenshot of this game running has been posted in the project's threads yet. Testers who reach a new screen usually post one in the discussion linked on the right.
Where it stands
Stray, an Unreal Engine 4 game, gets as far as the first-boot brightness calibration screen on Linux. That is the screen that says to adjust the value until the image on the left disappears. It draws correctly and holds 57 to 58 fps, with two 10-minute runs and no crash. The tester sent no input, so nothing past it is tested.
This needs about twelve pull requests applied to main in a certain order, and several of them conflict, so the tester resolved the conflicts by hand. There is no public branch for it. Without them, plain main stops 9 seconds in, before any window, on a shader call that cannot be resolved ahead of time. A fix that defers those shaders has since been merged, as have two others (one for unsupported shader types and one for a packed pixel word). Others are still open.
It also takes a long time to start. Preparing the shaders takes most of the first 3.5 minutes before the window appears. On Windows a second tester got a window and the Stray FPS counter showing 0.00, with 402 shaders never ready, so nothing was drawn.
A change merged on 2026-10-09 broke two of the shaders that had been fixed earlier, which is tracked as a regression.
| Windows | Window opens and the FPS counter reads 0.00; 402 shaders fail to prepare, so nothing is drawn. |
|---|---|
| Linux | Reaches the first-boot brightness screen at 57-58 fps with a stack of unmerged PRs; plain main stops at 9 seconds with no window. |
| macOS | No test report yet |
What is stopping it
- Shaders that jump to a computed target (s_swappc_b64) cannot be translated. 371 of 402 failing shaders on Windows are of this kind. Merged work only defers them. Thread
- A packed pixel value that a shader reuses is rejected as an unsupported live use. 26 shaders fail on this; a recent merge brought back the failure for two shaders that were fixed. Thread
- Since #2083 two pixel shaders that compiled before fail again with the same packed pixel error. Thread
- Image heap limits, depth/stencil storage and layered draws each have open PRs that conflict with each other. Thread
- Nothing past the brightness screen has been tested yet because the tester sent no input. Thread
How it got here
- Reported: the game aborts at start-up on 13 shaders of an unknown type. A fix that skips preparing them at registration was merged on 2026-10-10. Source
- Reported: a draw with no pixel program aborts before the first frame. Fixed by preparing the null program at draw time. Source
- Reported: a shader that reads the render target index fails to compile. A fix was merged on 2026-10-10. Source
- Linux write-up: the game boots to the first-boot brightness calibration screen at 57-58 fps on a stack of PRs. Source
- A second packed-pixel shader failure (26 shaders) is reported. Source
- Windows report: window opens, FPS counter reads 0.00, 402 shaders never prepare. Source
- A regression from #2083 breaks two Stray pixel shaders again. Source
Tested hardware
| PC | Result | |
|---|---|---|
| Nobara Linux 44 (kernel 7.2), Ryzen 5 7600X, RTX 5070 Ti (NVIDIA 595.91.07), 32 GB | Brightness calibration screen at 57-58 fps for two 10-minute runs, with a PR stack; plain main stops at 9 s. | Report |
| Windows, MinGW-w64 GCC 16.2.0, RTX 4060 Laptop GPU | Window opens, FPS counter at 0.00, 5.8 GiB used, nothing drawn. | Report |