code does not insert to the editor always

[distributed-clouds]

Hi everyone, thanks for this great plugin
I have the following config for accept the diff like this

-- Accept the diff
accept_diff = {
  normal = "<C-y>",
  insert = "<C-y>",
},

but it is sometimes working and sometimes not, for example when I touch a file lets call pod.yaml and ask to generate a manifest it creates the manifest in the buffer but when I press the nothing is inserted to my nvim editor

I would appreciate it if you can help me thanks 😊

[deathbeam]

Does gj work at least? Or that one doesnt work either?

[distributed-clouds]

wow Tomas thanks interesting now when I press gj then press now the it inserted to the IDE, so but I don't know if I have to do it everytime first gj? or it show work just with

[distributed-clouds]

something funny it creates an open buffer (empty file) with the same name as the Copilot Chat suggest. In this case
file:nginx-pod.yaml
and after inserting I need to save it

but I want the generated code to be inserted in the same pod.yaml file I have already created!

[deathbeam]

Yea C-y currently has safeguard that requires you to press gj first in case it needs to create new buffer first (or switch to different buffer in general). I am considering removing this because safeguarding that probably isnt that useful.

and for the nginx-pod.yaml and pod.yaml, it looks like you created pod.yaml first, but did not told the chat about pod.yaml so it tries to create nginx-pod.yaml instead. You need to either select something in pod.yaml so it gets sent like selection or use #buffer context at least so the chat always knows what file you are working on (setting default context to buffer is pretty useful for this)

But i think the behaviour you are experiencing atm is expected at least, i was worried that it wasnt matching filenames properly but your filenames are different so makes sense (it tries to match the markdown header against files)

[distributed-clouds]

perfect thanks a lot.
Can you please help me with this

use #buffer context at least so the chat always knows what file you are working on (setting default context to buffer is pretty useful for this)

as Im pretty beginner in nvim. thanks again

[deathbeam]

perfect thanks a lot. Can you please help me with this

use #buffer context at least so the chat always knows what file you are working on (setting default context to buffer is pretty useful for this)

as Im pretty beginner in nvim. thanks again

When configuring the plugin through lazy or just calling .setup, pass context = 'buffer' to the config:

https://github.com/CopilotC-Nvim/CopilotChat.nvim?tab=readme-ov-file#default-configuration

I think using lazy.nvim it would look like this:

return {
  {
    "CopilotC-Nvim/CopilotChat.nvim",
    dependencies = {
      { "github/copilot.vim" }, -- or zbirenbaum/copilot.lua
      { "nvim-lua/plenary.nvim", branch = "master" }, -- for curl, log and async functions
    },
    build = "make tiktoken", -- Only on MacOS or Linux
    opts = {
      context = 'buffer'
    },
  },
}

alternatively you can use sticky so you can visually see the context in buffer and remove it when you want, so like this:

return {
  {
    "CopilotC-Nvim/CopilotChat.nvim",
    dependencies = {
      { "github/copilot.vim" }, -- or zbirenbaum/copilot.lua
      { "nvim-lua/plenary.nvim", branch = "master" }, -- for curl, log and async functions
    },
    build = "make tiktoken", -- Only on MacOS or Linux
    opts = {
      sticky = '#buffer'
    },
  },
}

I personally have #buffers in my sticky as most of the time i just want to send all buffers i have open as context and because of the context filtering, caching and stuff it isnt that big overhead really

[deathbeam]

by the way gj is no longer required, c-y will now also move the buffer/create it when needed

[distributed-clouds]

you mean the plugin has been updated? or because of the sticky = '#buffer' one I don't need to press gj

[deathbeam]

plugin has been updated and you still want to keep the sticky