No description
Previously, if `expected_node` couldn't import a buffer, it would print an error, then try the next node. There shouldn't really be a reason to attempt import if the format/modifier isn't in `dmabuf_texture_formats`. It seems the issue I've been seeing with cosmic-workspaces crashing the Nvidia driver is fixed by removing this, and disabling `dma_shadow_copy` (which was producing the same error). Importing Nvidia buffers on the AMD GPU seems to be causing issues. Not sure how `dma_shadow_copy` should be fixed, but a test for supported formats is easy to add here anyway. |
||
|---|---|---|
| .github/workflows | ||
| cosmic-comp-config | ||
| data | ||
| debian | ||
| resources | ||
| src | ||
| .gitignore | ||
| build.rs | ||
| Cargo.lock | ||
| Cargo.toml | ||
| flake.lock | ||
| flake.nix | ||
| i18n.toml | ||
| LICENSE | ||
| Makefile | ||
| rust-toolchain.toml | ||
| rustfmt.toml | ||