Add lua support

CHANGELOG.md

@@ -15,6 +15,8 @@ syntactic changes".
 Fixed an issue in C and C++ where blank lines were highlighted after
 novel preprocessor lines.
 
+Added support for Lua.
+
 ## 0.24 (release 26th March 2022)
 
 ### Diffing

README.md

@@ -53,6 +53,7 @@ Difftastic supports the following languages:
 * Java
 * JavaScript (and JSX)
 * JSON
+* Lua
 * OCaml
 * PHP
 * Python

build.rs

@@ -129,6 +129,11 @@ fn main() {
             src_dir: "vendor/tree-sitter-json-src",
             extra_files: vec![],
         },
+        TreeSitterParser {
+            name: "tree-sitter-lua",
+            src_dir: "vendor/tree-sitter-lua-src",
+            extra_files: vec!["scanner.cc"],
+        },
         TreeSitterParser {
             name: "tree-sitter-ocaml",
             src_dir: "vendor/tree-sitter-ocaml-src/ocaml/src",

manual/src/upstream_parsers.md

@@ -19,6 +19,7 @@ Difftastic uses the following tree-sitter parsers:
 | Java            | [tree-sitter/tree-sitter-java](https://github.com/tree-sitter/tree-sitter-java)                                                         |
 | JavaScript, JSX | [tree-sitter/tree-sitter-javascript](https://github.com/tree-sitter/tree-sitter-javascript)                                             |
 | JSON            | [tree-sitter/tree-sitter-json](https://github.com/tree-sitter/tree-sitter-json)                                                         |
+| Lua             | [nvim-treesitter/tree-sitter-lua](https://github.com/nvim-treesitter/tree-sitter-lua)                                                   |
 | OCaml           | [tree-sitter/tree-sitter-ocaml](https://github.com/tree-sitter/tree-sitter-ocaml)                                                       |
 | PHP             | [tree-sitter/tree-sitter-php](https://github.com/tree-sitter/tree-sitter-php)                                                           |
 | Python          | [tree-sitter/tree-sitter-python](https://github.com/tree-sitter/tree-sitter-python)                                                     |

sample_files/compare.expected

@@ -64,6 +64,9 @@ f0befdc2c221d021a03acd6c0e662e37  -
 sample_files/load_before.js sample_files/load_after.js
 c5a352313dfb4ae1818190f144c4874b  -
 
+sample_files/lua_before.lua sample_files/lua_after.lua
+2899a7fcb6bd0edf247a2773a7443992  -
+
 sample_files/modules_before.ml sample_files/modules_after.ml
 4c0613c6038eaea73e10a91e7a86e916  -
 

sample_files/lua_after.lua

@@ -0,0 +1,16 @@
+function test(name)
+  print("testing"..name)
+  return 10
+end
+
+if test("myname") == 20 then
+  print("20")
+end
+
+local list = { "A", "B", "C", "D" }
+local table = { a = "A", b = "B", c = "C", d = "D" }
+
+for k, v in pairs(table) do
+  print("k: "..k.." v: "..v)
+end
+

sample_files/lua_before.lua

@@ -0,0 +1,16 @@
+function test()
+  print("testing")
+  return 10
+end
+
+if test() == 20 then
+  print("30")
+end
+
+local list = { "A", "B", "D" }
+local table = { a = "A", b = "B", d = "D" }
+
+for k, v in pairs(table) do
+  print("k: "..k.." v: "..v)
+end
+

src/guess_language.rs

@@ -34,6 +34,7 @@ pub enum Language {
     JavaScript,
     Json,
     Jsx,
+    Lua,
     OCaml,
     OCamlInterface,
     Php,
@@ -192,6 +193,7 @@ fn from_extension(extension: &OsStr, src: &str) -> Option<Language> {
         "cjs" | "js" | "mjs" => Some(JavaScript),
         "jsx" => Some(Jsx),
         "json" => Some(Json),
+        "lua" => Some(Lua),
         "ml" => Some(OCaml),
         "mli" => Some(OCamlInterface),
         "php" if !src.starts_with("<?hh") => Some(Php),

src/tree_sitter_parser.rs

@@ -58,6 +58,7 @@ extern "C" {
     fn tree_sitter_java() -> ts::Language;
     fn tree_sitter_javascript() -> ts::Language;
     fn tree_sitter_json() -> ts::Language;
+    fn tree_sitter_lua() -> ts::Language;
     fn tree_sitter_ocaml() -> ts::Language;
     fn tree_sitter_ocaml_interface() -> ts::Language;
     fn tree_sitter_php() -> ts::Language;
@@ -325,6 +326,22 @@ pub fn from_language(language: guess::Language) -> TreeSitterConfig {
                 .unwrap(),
             }
         }
+        Lua => {
+            let language = unsafe { tree_sitter_lua() };
+            TreeSitterConfig {
+                name: "Lua",
+                language,
+                atom_nodes: (vec!["string"]).into_iter().collect(),
+                delimiter_tokens: (vec![("(", ")"), ("{", "}"), ("[", "]")])
+                    .into_iter()
+                    .collect(),
+                highlight_query: ts::Query::new(
+                    language,
+                    include_str!("../vendor/highlights/lua.scm"),
+                )
+                .unwrap(),
+            }
+        }
         OCaml => {
             let language = unsafe { tree_sitter_ocaml() };
             TreeSitterConfig {

vendor/highlights/lua.scm

@@ -0,0 +1,76 @@
+;; Keywords
+
+[
+ "goto"
+ "in"
+ "local"
+ "return"
+  "and"
+  "not"
+  "or"
+  "function"
+  "if"
+  "elseif"
+  "else"
+  "then"
+  "for"
+  "do"
+  "end"
+  "repeat"
+  "until"
+  "while"
+] @keyword
+
+[
+  "+"
+  "-"
+  "*"
+  "/"
+  "%"
+  "^"
+  "#"
+  "=="
+  "~="
+  "<="
+  ">="
+  "<"
+  ">"
+  "="
+  "&"
+  "~"
+  "|"
+  "<<"
+  ">>"
+  "//"
+  ".."
+] @operator
+
+;; Variables
+
+(identifier) @variable
+
+((identifier) @variable.builtin
+ (#match? @variable.builtin "self"))
+
+;; Constants
+
+((identifier) @constant
+ (#match? @constant "^[A-Z][A-Z_0-9]*$"))
+
+(nil) @constant.builtin
+
+[
+  (false)
+  (true)
+] @boolean
+
+;; Others
+
+(comment) @comment
+
+(number) @number
+
+(string) @string
+
+;; Error
+(ERROR) @error

vendor/tree-sitter-lua-src

@@ -0,0 +1 @@
+tree-sitter-lua/src/