Skip to content

tjdevries/tree-sitter-lua

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

95 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

tree-sitter-(n)lua

TODO: Document how you can use make build_parser to automatically install this for lua.

Tree sitter grammar for Lua built to be used inside of Neovim.

Goal: Make a super great Lua grammar and tailor it to usage for Neovim.

Docgen

This grammar has a docgen module.

How to write emmy documentation

How to generate documentation

Write a lua script that looks like this:

local docgen = require('docgen')

local docs = {}

docs.test = function()
  -- Filepaths that should generate docs
  local input_files = {
    "./lua/telescope/init.lua",
    "./lua/telescope/builtin/init.lua",
    "./lua/telescope/pickers/layout_strategies.lua",
    "./lua/telescope/actions/init.lua",
    "./lua/telescope/previewers/init.lua",
  }

  -- Maybe sort them that depends what you want and need
  table.sort(input_files, function(a, b)
    return #a < #b
  end)

  -- Output file
  local output_file = "./doc/telescope.txt"
  local output_file_handle = io.open(output_file, "w")

  for _, input_file in ipairs(input_files) do
    docgen.write(input_file, output_file_handle)
  end

  output_file_handle:write(" vim:tw=78:ts=8:ft=help:norl:\n")
  output_file_handle:close()
  vim.cmd [[checktime]]
end

docs.test()

return docs

It is suggested to use a minimal_init.vim. Example:

set rtp+=.
set rtp+=../plenary.nvim/
set rtp+=../tree-sitter-lua/

runtime! plugin/plenary.vim

After that you can run docgen like this:

nvim --headless --noplugin -u scripts/minimal_init.vim -c "luafile ./scripts/gendocs.lua" -c 'qa'

Thoughts on LSP:

OK, so here's what I'm thinking:

  1. We have tree sitter for Lua (that we wrote)
  2. Can use tree sitter + queries to get information about one file
    • This is like, what a variable is in the file, where it's defined, references, etc.
  3. We can use "programming" to link up multiple files w/ our tree sitter info
    • Use package.path, package.searchers, etc. to find where requires lead to.
  4. Now we have "project" level knowledge.
  5. Can give smarts based on those things.

ALSO!

We can cheat :) :) :)

Let's say we run our LSP in another neovim instance (just nvim --headless -c 'some stuff') ... OK

this means we can ask vim about itself (and other globals, and your package.path, etc.)

Part 2 of cheating:

we can re-use vim.lsp in our implementation

Status

  • Grammar
    • [ ]

How did I get this to be da one for nvim-treesitter

  1. It helps if you made @vigoux the person he is today.
    • (Not a firm requirement, but it's very helpful)
  2. Make a grammar.js and then be able to generate a parser.so
  3. Write a bunch of tests for the language and provide somethjing the other parser doesnt.
    • Mine is mostly that it parses docs
    • It also parses more specifically many aspects of the language for smarter highlighting (hopefully)
    • I also like the names more
  4. To test with nvim-treesitter, you need to make a lua.so (or whatever your filetype is) available somewhere in rtp in a folder called parser.
    • Then you need to write the .scm files, like highlight and all that good stuff.
  5. ???
  6. Profit!