kitty @ kitten mykitten123.py vs. kitten mykitten123.py?

[SpicyRicecaker]

Hi,

Thanks for making kitty, never ceases to amaze me how low its latency is compared to other terminals (while having a ton of features and scriptability!).

Today I was trying to get custom kittens working following the instructions in https://sw.kovidgoyal.net/kitty/kittens/custom/#using-kittens-to-script-kitty-without-any-terminal-ui.

I initially tested my scripts using kitten mykitten123.py, but this resulted in the handle_result never getting called at all. After 2 hours of digging through the source, reading kitty help commands, and some gpt, I finally decided to try kitty @ kitten mykitten123.py and the handle_result callback fired instantly.

I don't understand at all why kitty @ kitten works when kitten doesn't... my understanding is that kitty @ opens the kitty shell which has some special defined commands, including the kitten command for launching kittens. But as kitten is also defined in main kitty's scope, I'm not sure why kitty @ kitten works when kitten doesn't. Is there something in the docs that I'm missing?

Thanks again.

[kovidgoyal]

kitten is a standalone binary written in Go meant for running the builtin kittens which are *almost* all implemented in Go. If you want to run a custom kitten you run it as kitty +kitten mykitten.py This runs the main() from mykitten and *does not* run handle_result() If you want to use handle_result() then you need to run your kitten in either of two ways: 1) Via a mapping in kitty.conf map f1 kitten mykittenpy 2) Via remote control at a shell or outside kitty kitten @ kitten mykitten.py or (the same thing but with kitty exe) kitty @ kitten mykitten.py