localize_mf2_treesitter

defmodule Mix.Tasks.LocalizeMf2Treesitter.Sync do
  @moduledoc """
  Sync the vendored grammar + queries from the published
  [`tree-sitter-mf2`](https://www.npmjs.com/package/tree-sitter-mf2)
  npm package.

  This package embeds the grammar source as a C build-time input:

    * `c_src/grammar/parser.c` + `c_src/grammar/tree_sitter/parser.h`
      — compiled alongside the tree-sitter runtime into the NIF `.so`.
    * `priv/queries/*.scm` — loaded at runtime by `Query.load/1`.

  The canonical source is the npm package, pinned to an exact
  version at the top of this module (`@tree_sitter_mf2_version`).
  Bump that string and re-run the task to move to a new grammar
  release. **Keep the pin in step with `mf2_wasm_editor`'s own sync
  task** — tree shape is the API boundary between server-side parse
  (this NIF) and browser-side parse (WASM editor); a version skew
  can produce different trees for the same input, breaking the
  canonicalisation round-trip.

  ## Usage

      # Fetch from npm at the pinned version and update local files.
      mix localize_mf2_treesitter.sync

      # CI check — exit non-zero if any vendored file has drifted
      # from the pinned version. Does not modify files.
      mix localize_mf2_treesitter.sync --check

  ## Offline / local-iteration override

  If you're iterating on the grammar locally and want this task to
  read from a sibling checkout rather than hit the network, set
  `MF2_TREESITTER_DIR`:

      MF2_TREESITTER_DIR=/path/to/mf2_treesitter mix localize_mf2_treesitter.sync

  The checkout's layout must match the npm package layout (it does,
  if you're pointing at a working tree of the
  [`mf2_treesitter`](https://github.com/elixir-localize/mf2_treesitter)
  repo).
  """
  use Mix.Task

  @shortdoc "Sync vendored tree-sitter-mf2 grammar + queries from npm."

  # Pinned tree-sitter-mf2 version. Bump and re-run the task to pull
  # a newer grammar. Keep in sync with `mf2_wasm_editor`'s pin.
  @tree_sitter_mf2_version "0.1.4"

  # unpkg.com proxies the npm registry and serves individual files
  # from a package tarball over HTTPS at stable URLs.
  @cdn_base "https://unpkg.com/tree-sitter-mf2@#{@tree_sitter_mf2_version}"

  # {source_rel, dest_rel}:
  #   source relative to the tree-sitter-mf2 package root
  #   dest relative to this package's root
  @files [
    {"src/parser.c", "c_src/grammar/parser.c"},
    {"src/tree_sitter/parser.h", "c_src/grammar/tree_sitter/parser.h"},
    {"queries/highlights.scm", "priv/queries/highlights.scm"}
  ]

  @switches [check: :boolean]

  @impl Mix.Task
  def run(argv) do
    {opts, _rest} = OptionParser.parse!(argv, switches: @switches)
    check? = opts[:check] == true

    source = resolve_source()
    package_root = File.cwd!()

    case source do
      {:cdn, base} ->
        Mix.shell().info("[sync] grammar source: #{base}")

      {:local, path} ->
        Mix.shell().info("[sync] grammar source: #{path} (local override)")
    end

    Mix.shell().info("[sync] package root:   #{package_root}")

    drift = sync_files(source, package_root, @files, check?)

    cond do
      check? and drift != [] ->
        Mix.shell().error("[sync] drift detected in #{length(drift)} file(s):")

        Enum.each(drift, fn path ->
          Mix.shell().error("  - #{path}")
        end)

        Mix.shell().error("Run `mix localize_mf2_treesitter.sync` to update.")
        exit({:shutdown, 1})

      check? ->
        Mix.shell().info(
          "[sync] check: all vendored files match tree-sitter-mf2@#{@tree_sitter_mf2_version}."
        )

      drift == [] ->
        Mix.shell().info("[sync] already in sync; no files changed.")

      true ->
        Mix.shell().info("[sync] updated #{length(drift)} file(s).")
        Mix.shell().info("[sync] run `mix compile --force` to rebuild the NIF.")
    end

    :ok
  end

  defp sync_files(source, dst_root, pairs, check?) do
    Enum.reduce(pairs, [], fn {src_rel, dst_rel}, acc ->
      bytes = fetch_bytes!(source, src_rel)
      dst_path = Path.join(dst_root, dst_rel)

      if changed?(bytes, dst_path) do
        unless check? do
          File.mkdir_p!(Path.dirname(dst_path))
          File.write!(dst_path, bytes)
          Mix.shell().info("[sync] wrote #{dst_rel}")
        end

        [dst_rel | acc]
      else
        acc
      end
    end)
  end

  defp changed?(new_bytes, dst) do
    cond do
      not File.exists?(dst) -> true
      File.read!(dst) != new_bytes -> true
      true -> false
    end
  end

  # Fetch raw bytes for one file under the grammar package root.
  # Dispatches on source kind — local checkout copies from disk,
  # CDN mode does a verified HTTPS GET.
  defp fetch_bytes!({:local, dir}, rel) do
    path = Path.join(dir, rel)

    unless File.exists?(path) do
      Mix.raise("missing source file at local override: #{path}")
    end

    File.read!(path)
  end

  defp fetch_bytes!({:cdn, base}, rel) do
    url = "#{base}/#{rel}"
    http_get!(url)
  end

  defp http_get!(url) do
    ensure_http_started()

    url_charlist = String.to_charlist(url)
    request = {url_charlist, [{~c"accept", ~c"*/*"}]}

    case :httpc.request(:get, request, [ssl: ssl_options()], body_format: :binary) do
      {:ok, {{_http, status, _reason}, _headers, body}} when status in 200..299 ->
        body

      {:ok, {{_http, status, reason}, _headers, _body}} ->
        Mix.raise("""
        HTTP #{status} #{reason} from #{url}

        Check that tree-sitter-mf2@#{@tree_sitter_mf2_version} exists on npm
        (https://www.npmjs.com/package/tree-sitter-mf2), or set
        MF2_TREESITTER_DIR to a local checkout to bypass the CDN.
        """)

      {:error, reason} ->
        Mix.raise("""
        HTTP request to #{url} failed: #{inspect(reason)}

        Check your network connection, or set MF2_TREESITTER_DIR to a
        local checkout to bypass the CDN.
        """)
    end
  end

  defp ensure_http_started do
    :ssl.start()
    :inets.start()
  end

  defp ssl_options do
    [
      verify: :verify_peer,
      cacerts: :public_key.cacerts_get(),
      customize_hostname_check: [
        match_fun: :public_key.pkix_verify_hostname_match_fun(:https)
      ]
    ]
  end

  # Returns the source to read files from:
  #   {:local, path} when MF2_TREESITTER_DIR is set (local iteration)
  #   {:cdn, base_url} otherwise (the default, reproducible path)
  defp resolve_source do
    case System.get_env("MF2_TREESITTER_DIR") do
      nil ->
        {:cdn, @cdn_base}

      "" ->
        {:cdn, @cdn_base}

      override ->
        path = Path.expand(override)

        unless File.dir?(path) do
          Mix.raise("MF2_TREESITTER_DIR set but not a directory: #{path}")
        end

        {:local, path}
    end
  end
end