Add CMake support

Fixes #333

CHANGELOG.md

@@ -2,6 +2,11 @@
 
 ### Parsing
 
+Added support for CMake.
+
+Improved comment detection using tree-sitter syntax highlighting
+queries.
+
 Fixed an issue with language detection when file names were very
 short.
 

build.rs

@@ -88,6 +88,11 @@ fn main() {
             src_dir: "vendor/tree-sitter-clojure-src",
             extra_files: vec![],
         },
+        TreeSitterParser {
+            name: "tree-sitter-cmake",
+            src_dir: "vendor/tree-sitter-cmake-src",
+            extra_files: vec!["scanner.cc"],
+        },
         TreeSitterParser {
             name: "tree-sitter-commonlisp",
             src_dir: "vendor/tree-sitter-commonlisp-src",

manual/src/languages_supported.md

@@ -9,6 +9,7 @@ Difftastic supports the following programming languages.
 | C++             | [tree-sitter/tree-sitter-cpp](https://github.com/tree-sitter/tree-sitter-cpp)                                                                                  |
 | C#              | [tree-sitter/tree-sitter-c-sharp](https://github.com/tree-sitter/tree-sitter-c-sharp)                                                                          |
 | Clojure         | [sogaiu/tree-sitter-clojure](https://github.com/sogaiu/tree-sitter-clojure) ([branched](https://github.com/sogaiu/tree-sitter-clojure/tree/issue-21))          |
+| CMake           | [uyha/tree-sitter-cmake](https://github.com/uyha/tree-sitter-cmake)                                                                                            |
 | Common Lisp     | [theHamsta/tree-sitter-commonlisp](https://github.com/theHamsta/tree-sitter-commonlisp)                                                                        |
 | Dart            | [UserNobody14/tree-sitter-dart](https://github.com/UserNobody14/tree-sitter-dart)                                                                              |
 | Elixir          | [elixir-lang/tree-sitter-elixir](https://github.com/elixir-lang/tree-sitter-elixir)                                                                            |

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 | Hack
+        Bash | C | CMake | 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,

src/parse/guess_language.rs

@@ -21,6 +21,7 @@ pub enum Language {
     Bash,
     C,
     Clojure,
+    CMake,
     CommonLisp,
     CPlusPlus,
     CSharp,
@@ -196,6 +197,7 @@ fn from_name(path: &Path) -> Option<Language> {
             | "PKGBUILD" | "bash_aliases" | "bash_logout" | "bash_profile" | "bashrc" | "cshrc"
             | "gradlew" | "kshrc" | "login" | "man" | "profile" | "zlogin" | "zlogout"
             | "zprofile" | "zshenv" | "zshrc" => Some(Bash),
+            "CMakeLists.txt" => Some(CMake),
             ".emacs" | "_emacs" | "Cask" => Some(EmacsLisp),
             ".arcconfig" | ".auto-changelog" | ".c8rc" | ".htmlhintrc" | ".imgbotconfig"
             | ".nycrc" | ".tern-config" | ".tern-project" | ".watchmanconfig" | "Pipfile.lock"
@@ -222,6 +224,7 @@ pub fn from_extension(extension: &OsStr) -> Option<Language> {
             Some(Clojure)
         }
         "lisp" | "lsp" | "asd" => Some(CommonLisp),
+        "cmake" | "cmake.in" => Some(CMake),
         "cs" => Some(CSharp),
         "css" => Some(Css),
         "dart" => Some(Dart),

src/parse/tree_sitter_parser.rs

@@ -47,6 +47,7 @@ extern "C" {
     fn tree_sitter_c() -> ts::Language;
     fn tree_sitter_c_sharp() -> ts::Language;
     fn tree_sitter_clojure() -> ts::Language;
+    fn tree_sitter_cmake() -> ts::Language;
     fn tree_sitter_cpp() -> ts::Language;
     fn tree_sitter_commonlisp() -> ts::Language;
     fn tree_sitter_css() -> ts::Language;
@@ -164,6 +165,20 @@ pub fn from_language(language: guess::Language) -> TreeSitterConfig {
                 .unwrap(),
             }
         }
+        CMake => {
+            let language = unsafe { tree_sitter_cmake() };
+            TreeSitterConfig {
+                name: "CMake",
+                language,
+                atom_nodes: vec!["argument"].into_iter().collect(),
+                delimiter_tokens: vec![("(", ")")].into_iter().collect(),
+                highlight_query: ts::Query::new(
+                    language,
+                    include_str!("../../vendor/highlights/cmake.scm"),
+                )
+                .unwrap(),
+            }
+        }
         CommonLisp => {
             let language = unsafe { tree_sitter_commonlisp() };
             TreeSitterConfig {

vendor/highlights/cmake.scm

@@ -0,0 +1,128 @@
+;; Based on the nvim-treesitter highlighting, which is under the Apache license.
+;; https://github.com/nvim-treesitter/nvim-treesitter/blob/d76b0de6536c2461f97cfeca0550f8cb89793935/queries/cmake/highlights.scm
+
+[
+ (quoted_argument)
+ (bracket_argument)
+] @string
+
+(variable_ref) @none
+(variable) @variable
+
+[
+ (bracket_comment)
+ (line_comment)
+] @comment
+
+(normal_command (identifier) @function)
+
+["ENV" "CACHE"] @symbol
+["$" "{" "}" "<" ">"] @punctuation.special
+["(" ")"] @punctuation.bracket
+
+[
+ (function)
+ (endfunction)
+ (macro)
+ (endmacro)
+] @keyword.function
+
+[
+ (if)
+ (elseif)
+ (else)
+ (endif)
+] @conditional
+
+[
+ (foreach)
+ (endforeach)
+ (while)
+ (endwhile)
+] @repeat
+
+(function_command
+  (function)
+  . (argument) @function
+  (argument)* @parameter
+)
+
+(macro_command
+  (macro)
+  . (argument) @function.macro
+  (argument)* @parameter
+)
+
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   . (argument) @variable
+;;   (#match? @function.builtin "\\c^(set)$")
+;; )
+
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   (#match? @function.builtin "\\c^(set)$")
+;;   (
+;;     (argument) @constant
+;;     (#any-of? @constant "PARENT_SCOPE")
+;;   ) .
+;; )
+
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   (#match? @function.builtin "\\c^(set)$")
+;;   . (argument)
+;;   (
+;;     (argument) @_cache @constant
+;;     .
+;;     (argument) @_type @constant
+;;     (#any-of? @_cache "CACHE")
+;;     (#any-of? @_type "BOOL" "FILEPATH" "PATH" "STRING" "INTERNAL")
+;;   )
+;; )
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   (#match? @function.builtin "\\c^(set)$")
+;;   . (argument)
+;;   (argument) @_cache
+;;   (#any-of? @_cache "CACHE")
+;;   (
+;;     (argument) @_force @constant
+;;     (#any-of? @_force "FORCE")
+;;   ) .
+;; )
+
+;; ((argument) @boolean
+;;   (#match? @boolean "\\c^(1|on|yes|true|y|0|off|no|false|n|ignore|notfound|.*-notfound)$")
+;; )
+
+(if_command
+  (if)
+  (argument) @keyword.operator
+  (#any-of? @keyword.operator "NOT" "AND" "OR"
+                              "COMMAND" "POLICY" "TARGET" "TEST" "DEFINED" "IN_LIST"
+                              "EXISTS" "IS_NEWER_THAN" "IS_DIRECTORY" "IS_SYMLINK" "IS_ABSOLUTE"
+                              "MATCHES"
+                              "LESS" "GREATER" "EQUAL" "LESS_EQUAL" "GREATER_EQUAL"
+                              "STRLESS" "STRGREATER" "STREQUAL" "STRLESS_EQUAL" "STRGREATER_EQUAL"
+                              "VERSION_LESS" "VERSION_GREATER" "VERSION_EQUAL" "VERSION_LESS_EQUAL" "VERSION_GREATER_EQUAL"
+  )
+)
+
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   . (argument)
+;;   (argument) @constant
+;;   (#any-of? @constant "ALL" "COMMAND" "DEPENDS" "BYPRODUCTS" "WORKING_DIRECTORY" "COMMENT"
+;;                    "JOB_POOL" "VERBATIM" "USES_TERMINAL" "COMMAND_EXPAND_LISTS" "SOURCES")
+;;   (#match? @function.builtin "\\c^(add_custom_target)$")
+;; )
+
+;; (normal_command
+;;   (identifier) @function.builtin
+;;   (argument) @constant
+;;   (#any-of? @constant "OUTPUT" "COMMAND" "MAIN_DEPENDENCY" "DEPENDS" "BYPRODUCTS" "IMPLICIT_DEPENDS" "WORKING_DIRECTORY"
+;;                       "COMMENT" "DEPFILE" "JOB_POOL" "VERBATIM" "APPEND" "USES_TERMINAL" "COMMAND_EXPAND_LISTS")
+;;   (#match? @function.builtin "\\c^(add_custom_command)$")
+;; )
+

vendor/tree-sitter-cmake-src

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