meh.

Because there's no way to avoid reading messages from `handle_info/2` and some receives are done in a certain order. Right now I'm rewriting it to use the new elixir-socket API...

As soon as it works, which could be hours or a few days.

I pushed the rewrite in its own branch. Apparently it's 2x slower than the original code which surprised me quite a bit. Less suprising is that using the new protocol...

Any update on this? It's very annoying right now to bubble up errors from argonautica with stuff like `anyhow` because of this.

I've been encountering the same pains that the async variant solves.

tesseract-ocr uses my own version of ffi-inliner, which hasn't been merged yet, and I don't even know if it will.

@crocket sixels are unrelated to 24-bit SGR, I think I talked to @saitoha about an extension for 24-bit sixels that my terminal emulator supports, but then I didn't care enough...

#39 is surely a missing headers/libraries issue (for whatever reason I never got notifications about it, hence no responses). This looks like a bug in FFI itself, if it were...

Yeah, it looks like they changed quite some stuff, especially regarding output. It will take some time, in the meantime you can use `downgrade` or `downgrader` from the AUR.

@acrogenesis it looks like they added a `TessRenderResult` class which is used in place of `STRING` for `ProcessPages`.