feat: portable kitty.conf without hardcoded path about kitty-scrollback.nvim

[mikesmithgh]

Hey @tummetott, thanks! I agree the configuration for kitty-scrollback.nvim can be improved. I added the details of his ticket to #69 where I plan on redesigning kitty-scrollback.nvim's configuration.

Have you seen any issues with kitty-scrollback.nvim after this change to your config? If you have any other ideas on how the configuration can be improved, feel free to share 👍

[tummetott]

Have you seen any issues with kitty-scrollback.nvim after this change to your config?

Initially, I tried symlinking just the kitty_scrollback_nvim.py file, but I encountered Python errors due to this script loading other Python files in the same directory. As a result, I opted to symlink the entire plugin directory, which has been working seamlessly without any issues so far.

If you have any other ideas on how the configuration can be improved, feel free to share 👍

I have an additional suggestion to further simplify the configuration process. It would be highly beneficial if configuration entries could accept both functions and tables. For instance, consider the following:

require('kitty-scrollback').setup {
    global = function()
        return {
            paste_window = { ... },
            kitty_get_text = { ... },
        }
    end,
}

can also be configured as

require('kitty-scrollback').setup {
    global = {
        paste_window = { ... },
        kitty_get_text = { ... },
    },
}

or

winopts_overrides = function()
    return {
        border = 'rounded',
        row = ...
        col = ...
    }
end

could also be configured as:

winopts_overrides = {
    border = 'rounded',
    row = ...
    col = ...
}

There are scenarios where using a function is the most appropriate choice, especially when specific configuration values need to be computed dynamically at runtime. However, there are equally valid situations where the table returned by the function consists of static values that do not require any dynamic computation. In such cases, the function wrapper may appear to introduce unnecessary boilerplate code, potentially complicating the configuration unnecessarily. Thus, offering the option to directly provide a table can greatly simplify the configuration process for these static values, making it more intuitive and concise.