Add Hacklang support

Wilfred · Jul 12, 2022 · 8547987 · 8547987
1 parent 470585e
commit 8547987
Show file tree

Hide file tree

Showing 6 changed files with 31 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,7 +2,7 @@
 
 ### Parsing
 
-Added SQL support.
+Added Hacklang and SQL support.
 
 Updated to the latest tree-sitter parsers for C#, Dart, Elm, Gleam,
 Haskell, HCL, Java, JSON, OCaml, PHP, Python, Ruby, Scala and

diff --git a/build.rs b/build.rs
@@ -133,6 +133,11 @@ fn main() {
             src_dir: "vendor/tree-sitter-go-src",
             extra_files: vec![],
         },
+        TreeSitterParser {
+            name: "tree-sitter-hack",
+            src_dir: "vendor/tree-sitter-hack-src",
+            extra_files: vec!["scanner.cc"],
+        },
         TreeSitterParser {
             name: "tree-sitter-haskell",
             src_dir: "vendor/tree-sitter-haskell-src",

diff --git a/manual/src/languages_supported.md b/manual/src/languages_supported.md
@@ -16,6 +16,7 @@ Difftastic supports the following programming languages.
 | Emacs Lisp      | [wilfred/tree-sitter-elisp](https://github.com/Wilfred/tree-sitter-elisp)                                                                                      |
 | Gleam           | [gleam-lang/tree-sitter-gleam](https://github.com/gleam-lang/tree-sitter-gleam)                                                                                |
 | Go              | [tree-sitter/tree-sitter-go](https://github.com/tree-sitter/tree-sitter-go)                                                                                    |
+| Hack            | [slackhq/tree-sitter-hack](https://github.com/slackhq/tree-sitter-hack)                                                                                        |
 | Haskell         | [tree-sitter/tree-sitter-haskell](https://github.com/tree-sitter/tree-sitter-haskell)                                                                          |
 | Janet           | [sogaiu/tree-sitter-janet-simple](https://github.com/sogaiu/tree-sitter-janet-simple)                                                                          |
 | Java            | [tree-sitter/tree-sitter-java](https://github.com/tree-sitter/tree-sitter-java)                                                                                |

diff --git a/src/diff/sliders.rs b/src/diff/sliders.rs
@@ -67,7 +67,7 @@ fn prefer_outer_delimiter(language: guess_language::Language) -> bool {
         // For everything else, prefer the inner delimiter. These
         // languages have syntax like `foo(bar)` or `foo[bar]` where
         // the inner delimiter is more relevant.
-        Bash | C | CPlusPlus | CSharp | Css | Dart | Elixir | Elm | Elvish | Gleam | Go
+        Bash | C | CPlusPlus | CSharp | Css | Dart | Elixir | Elm | Elvish | Gleam | Go | Hack
         | Haskell | Html | Java | JavaScript | Jsx | Julia | Kotlin | Lua | Nix | OCaml
         | OCamlInterface | Perl | Php | Python | Ruby | Rust | Scala | Swift | Tsx | TypeScript
         | Yaml | Zig => false,

diff --git a/src/parse/guess_language.rs b/src/parse/guess_language.rs
@@ -32,6 +32,7 @@ pub enum Language {
     EmacsLisp,
     Gleam,
     Go,
+    Hack,
     Haskell,
     Hcl,
     Html,
@@ -162,6 +163,7 @@ fn from_shebang(src: &str) -> Option<Language> {
                     "lisp" | "sbc" | "ccl" | "clisp" | "ecl" => return Some(CommonLisp),
                     "elixir" => return Some(Elixir),
                     "elvish" => return Some(Elvish),
+                    "hhvm" => return Some(Hack),
                     "runghc" | "runhaskell" | "runhugs" => return Some(Haskell),
                     "chakra" | "d8" | "gjs" | "js" | "node" | "nodejs" | "qjs" | "rhino" | "v8"
                     | "v8-shell" => return Some(JavaScript),
@@ -175,6 +177,11 @@ fn from_shebang(src: &str) -> Option<Language> {
                 }
             }
         }
+
+        // Hack can use <?hh in files with a .php extension.
+        if first_line.starts_with("<?hh") {
+            return Some(Hack);
+        }
     }
 
     None
@@ -224,6 +231,7 @@ pub fn from_extension(extension: &OsStr) -> Option<Language> {
         "elv" => Some(Elvish),
         "gleam" => Some(Gleam),
         "go" => Some(Go),
+        "hack" | "hck" | "hhi" => Some(Hack),
         "hs" => Some(Haskell),
         "hcl" | "nomad" | "tf" | "tfvars" | "worfklow" => Some(Hcl),
         "html" | "htm" | "xhtml" => Some(Html),

diff --git a/src/parse/tree_sitter_parser.rs b/src/parse/tree_sitter_parser.rs
@@ -57,6 +57,7 @@ extern "C" {
     fn tree_sitter_elvish() -> ts::Language;
     fn tree_sitter_gleam() -> ts::Language;
     fn tree_sitter_go() -> ts::Language;
+    fn tree_sitter_hack() -> ts::Language;
     fn tree_sitter_haskell() -> ts::Language;
     fn tree_sitter_hcl() -> ts::Language;
     fn tree_sitter_html() -> ts::Language;
@@ -313,6 +314,20 @@ pub fn from_language(language: guess::Language) -> TreeSitterConfig {
                 .unwrap(),
             }
         }
+        Hack => {
+            let language = unsafe { tree_sitter_hack() };
+            TreeSitterConfig {
+                name: "Hack",
+                language,
+                atom_nodes: vec!["prefixed_string", "heredoc"].into_iter().collect(),
+                delimiter_tokens: vec![("[", "]"), ("(", ")"), ("<", ">"), ("{", "}")],
+                highlight_query: ts::Query::new(
+                    language,
+                    include_str!("../../vendor/highlights/hack.scm"),
+                )
+                .unwrap(),
+            }
+        }
         Haskell => {
             let language = unsafe { tree_sitter_haskell() };
             TreeSitterConfig {