Demi Marie Obenour

> > @1cho1ce Better would be something like > > ```shell > > nft 'add chain ip qubes whonix-input > > flush chain ip qubes whonix-input > > add rule...

@walid-git I think so, yes. I’m no expert on authoring VMODs but I can at least make sure it compiles and passes some basic tests.

Generally I reommend using a GPU or one of the few NPUs that are supported by Mesa. Those are much more likely to be supported long into the future because...

> Some wild guesses based on the stacktrace - Wayfire keeps a reference of the surface's texture/buffer: > > https://github.com/WayfireWM/wayfire/blob/6796b08545b40c74f08242868b6287a6ede41550/src/view/wlr-surface-node.cpp#L61 > > Depending on how wlroots has implemented GPU reset...

@marmarta: I think this is the source of most “XYZ becomes temporarily unresponsive” issues.

`QubesAsync()` seems like a good option, unless there are ordering requirements that make a callback-based solution more attractive.

Instead of relying on block-layer compression, have you considered doing compression in user-space before writing the files out?

> You can already implement something almost as efficient with the existing primitives (only one additional subtraction): > > ```ocaml > (* A statically allocated string will always have the...

The bootstrap can be eliminated by replacing the `%` primitives with normal primitives that happen to be in the primitive table. The boostrap and post-bootstrap `ocamlc` will both compile them...

> The bootstrap can be eliminated by replacing the `%` primitives with normal primitives that happen to be in the primitive table. The boostrap and post-bootstrap `ocamlc` will both compile...