GitEdvard/trigger-commands.nvim

Show python program output within neovim. Send stacktrace to quickfix

trigger-commands

Trigger any command within vim, so that stdout and stderr is shown in a scratch window. Stderr is also directed to the quickfix window.

Motivation

1. Show errors in quickfix window
2. Reduce swapping between vim and terminal when doing a code-run-validate loop. Specificly when developing a rest service, which has this extra start and stop a service before trigger a rest call.

Requirements

https://github.com/GitEdvard/read-settings.nvim

You have to setup the errorformat variable for your programming language.

Setup

Create a .runsettings.json file at the root of your repository. A run setting must have the fields "type" and "command". The field "args" is optional.

Example single command setup:

{
  "type": "python",
  "program": "${workspaceFolder}/hello.py",
  "args": {"world!"}
}

Corresponding init.vim

local run_settings = require'read-settings'.read_json('.runsettings.json')
local trigger_command = function()
    local cmd = require'trigger_commands'.generate_command(run_settings)
    require'trigger-commands'.run_single(cmd)
end
local opts = { noremap = true, silent = true }
vim.keymap.set('n', '<leader>ur', trigger_command, opts)

Note the difference from rest call commands, regarding run_settings instead of cmd.

Example rest service command setup:

[
  {
    "type": "python",
    "program": "${workspaceFolder}/myserver.py",
    "args": { "--port", "2000" }
  },
  {
    "type": "curl",
    "program": "localhost:2000/api/hello"
  }
]

Corresponding init.lua setup (note the "run_rest_call" command):

local run_settings = require'read-settings'.read_json('.runsettings.json')
local trigger_command = function()
    require'trigger-commands'.run_rest_call(run_settings)
end
local opts = { noremap = true, silent = true }
vim.keymap.set('n', '<leader>ur', trigger_command, opts)

Run setup keywords

• Some variables are supported:
• ${file}: Active filename
• ${fileBasename}: The current file's basename
• ${fileBasenameNoExtension}: The current file's basename without extension
• ${fileDirname}: The current file's dirname
• ${fileExtname}: The current file's extension
• ${relativeFile}: The current file relative to |getcwd()|
• ${relativeFileDirname}: The current file's dirname relative to |getcwd()|
• ${workspaceFolder}: The current working directory of Neovim
• ${workspaceFolderBasename}: The name of the folder opened in Neovim
• ${env:Name}: Environment variable named Name, for example: ${env:HOME}.

These variable names are taken from nvim-dap: https://github.com/mfussenegger/nvim-dap/blob/master/doc/dap.txt

For a single command setting, one can use the same setting as nvim-dap. However, trigger-commands will only look for the arguments "type", "program" and "args".

Example using an enviroment variable in run settings:

{
  "type": "python",
  "program": "${workspaceFolder}/hello.py",
  "args": {"${env:HELLO}"}
}

in init.lua:

local run_settings = require'read-settings'.read_json('.runsettings.json')
local trigger_command = function()
    vim.cmd("let $HELLO = 'world!'")
    require'trigger-commands'.run_single(run_settings)
end
local opts = { noremap = true, silent = true }
vim.keymap.set('n', '<leader>ur', trigger_command, opts)