Detect filetype from shebang line

helix-core/src/indent.rs

@@ -450,6 +450,7 @@ where
             language: vec![LanguageConfiguration {
                 scope: "source.rust".to_string(),
                 file_types: vec!["rs".to_string()],
+                shebangs: vec![],
                 language_id: "Rust".to_string(),
                 highlight_config: OnceCell::new(),
                 config: None,

helix-core/src/syntax.rs

@@ -14,6 +14,7 @@ use std::{
     cell::RefCell,
     collections::{HashMap, HashSet},
     fmt,
+    fs::File,
     path::Path,
     sync::Arc,
 };
@@ -52,6 +53,7 @@ pub struct LanguageConfiguration {
     pub language_id: String,
     pub scope: String,           // source.rust
     pub file_types: Vec<String>, // filename ends_with? <Gemfile, rb, etc>
+    pub shebangs: Vec<String>,   // interpreter(s) associated with language
     pub roots: Vec<String>,      // these indicate project roots <.git, Cargo.toml>
     pub comment_token: Option<String>,
 
@@ -254,13 +256,15 @@ pub struct Loader {
     // highlight_names ?
     language_configs: Vec<Arc<LanguageConfiguration>>,
     language_config_ids_by_file_type: HashMap<String, usize>, // Vec<usize>
+    language_config_ids_by_shebang: HashMap<String, usize>,
 }
 
 impl Loader {
     pub fn new(config: Configuration) -> Self {
         let mut loader = Self {
             language_configs: Vec::new(),
             language_config_ids_by_file_type: HashMap::new(),
+            language_config_ids_by_shebang: HashMap::new(),
         };
 
         for config in config.language {
@@ -273,6 +277,11 @@ impl Loader {
                     .language_config_ids_by_file_type
                     .insert(file_type.clone(), language_id);
             }
+            for shebang in &config.shebangs {
+                loader
+                    .language_config_ids_by_shebang
+                    .insert(shebang.clone(), language_id);
+            }
 
             loader.language_configs.push(Arc::new(config));
         }
@@ -283,7 +292,7 @@ impl Loader {
     pub fn language_config_for_file_name(&self, path: &Path) -> Option<Arc<LanguageConfiguration>> {
         // Find all the language configurations that match this file name
         // or a suffix of the file name.
-        let configuration_id = path
+        let mut configuration_id = path
             .file_name()
             .and_then(|n| n.to_str())
             .and_then(|file_name| self.language_config_ids_by_file_type.get(file_name))
@@ -293,9 +302,25 @@ impl Loader {
                     .and_then(|extension| self.language_config_ids_by_file_type.get(extension))
             });
 
-        configuration_id.and_then(|&id| self.language_configs.get(id).cloned())
+        // If we have not found the configuration_id, see if we can get it from a shebang line
+        if configuration_id.is_none() {
+            if let Ok(mut file) = File::open(path) {
+                let mut buf = [0; 100];
+                if std::io::Read::read(&mut file, &mut buf[..]).is_ok() {
+                    if let Ok(str) = str::from_utf8(&buf) {
+                        static SHEBANG_REGEX: Lazy<Regex> = Lazy::new(|| {
+                            Regex::new(r"#!/[^\s]*/(env\s)*([_a-zA-Z0-9-]+)").unwrap()
+                        });
+                        configuration_id = SHEBANG_REGEX
+                            .captures(str)
+                            .and_then(|cap| cap.get(2))
+                            .and_then(|cap| self.language_config_ids_by_shebang.get(cap.as_str()))
+                    }
+                }
+            }
+        };
 
-        // TODO: content_regex handling conflict resolution
+        configuration_id.and_then(|&id| self.language_configs.get(id).cloned())
     }
 
     pub fn language_config_for_scope(&self, scope: &str) -> Option<Arc<LanguageConfiguration>> {

languages.toml

@@ -3,6 +3,7 @@ name = "rust"
 scope = "source.rust"
 injection-regex = "rust"
 file-types = ["rs"]
+shebangs = []
 roots = []
 auto-format = true
 comment-token = "//"
@@ -17,6 +18,7 @@ name = "toml"
 scope = "source.toml"
 injection-regex = "toml"
 file-types = ["toml"]
+shebangs = []
 roots = []
 comment-token = "#"
 
@@ -27,6 +29,7 @@ name = "protobuf"
 scope = "source.proto"
 injection-regex = "protobuf"
 file-types = ["proto"]
+shebangs = []
 roots = []
 comment-token = "//"
 
@@ -37,6 +40,7 @@ name = "elixir"
 scope = "source.elixir"
 injection-regex = "elixir"
 file-types = ["ex", "exs"]
+shebangs = []
 roots = []
 comment-token = "#"
 
@@ -48,6 +52,7 @@ name = "json"
 scope = "source.json"
 injection-regex = "json"
 file-types = ["json"]
+shebangs = []
 roots = []
 
 indent = { tab-width = 2, unit = "  " }
@@ -57,6 +62,7 @@ name = "c"
 scope = "source.c"
 injection-regex = "c"
 file-types = ["c"] # TODO: ["h"]
+shebangs = []
 roots = []
 comment-token = "//"
 
@@ -68,6 +74,7 @@ name = "cpp"
 scope = "source.cpp"
 injection-regex = "cpp"
 file-types = ["cc", "hh", "cpp", "hpp", "h", "ipp", "tpp", "cxx", "hxx", "ixx", "txx", "ino"]
+shebangs = []
 roots = []
 comment-token = "//"
 
@@ -79,6 +86,7 @@ name = "c-sharp"
 scope = "source.csharp"
 injection-regex = "c-?sharp"
 file-types = ["cs"]
+shebangs = []
 roots = []
 comment-token = "//"
 
@@ -89,6 +97,7 @@ name = "go"
 scope = "source.go"
 injection-regex = "go"
 file-types = ["go"]
+shebangs = []
 roots = ["Gopkg.toml", "go.mod"]
 auto-format = true
 comment-token = "//"
@@ -102,6 +111,7 @@ name = "javascript"
 scope = "source.js"
 injection-regex = "^(js|javascript)$"
 file-types = ["js", "mjs"]
+shebangs = []
 roots = []
 comment-token = "//"
 # TODO: highlights-jsx, highlights-params
@@ -113,6 +123,7 @@ name = "typescript"
 scope = "source.ts"
 injection-regex = "^(ts|typescript)$"
 file-types = ["ts"]
+shebangs = []
 roots = []
 # TODO: highlights-jsx, highlights-params
 
@@ -124,6 +135,7 @@ name = "tsx"
 scope = "source.tsx"
 injection-regex = "^(tsx)$" # |typescript
 file-types = ["tsx"]
+shebangs = []
 roots = []
 # TODO: highlights-jsx, highlights-params
 
@@ -135,6 +147,7 @@ name = "css"
 scope = "source.css"
 injection-regex = "css"
 file-types = ["css"]
+shebangs = []
 roots = []
 
 indent = { tab-width = 2, unit = "  " }
@@ -144,6 +157,7 @@ name = "html"
 scope = "text.html.basic"
 injection-regex = "html"
 file-types = ["html"]
+shebangs = []
 roots = []
 
 indent = { tab-width = 2, unit = "  " }
@@ -153,6 +167,7 @@ name = "python"
 scope = "source.python"
 injection-regex = "python"
 file-types = ["py"]
+shebangs = ["python", "python2", "python3"]
 roots = []
 comment-token = "#"
 
@@ -165,6 +180,7 @@ name = "nix"
 scope = "source.nix"
 injection-regex = "nix"
 file-types = ["nix"]
+shebangs = []
 roots = []
 comment-token = "#"
 
@@ -176,6 +192,7 @@ name = "ruby"
 scope = "source.ruby"
 injection-regex = "ruby"
 file-types = ["rb"]
+shebangs = ["ruby"]
 roots = []
 comment-token = "#"
 
@@ -187,6 +204,7 @@ name = "bash"
 scope = "source.bash"
 injection-regex = "bash"
 file-types = ["sh", "bash"]
+shebangs = ["sh", "bash", "dash"]
 roots = []
 comment-token = "#"
 
@@ -198,6 +216,7 @@ name = "php"
 scope = "source.php"
 injection-regex = "php"
 file-types = ["php"]
+shebangs = ["php"]
 roots = []
 
 indent = { tab-width = 4, unit = "    " }
@@ -207,6 +226,7 @@ name = "latex"
 scope = "source.tex"
 injection-regex = "tex"
 file-types = ["tex"]
+shebangs = []
 roots = []
 comment-token = "%"
 
@@ -217,6 +237,7 @@ name = "julia"
 scope = "source.julia"
 injection-regex = "julia"
 file-types = ["jl"]
+shebangs = []
 roots = []
 comment-token = "#"
 language-server = { command = "julia", args = [
@@ -242,6 +263,7 @@ name = "java"
 scope = "source.java"
 injection-regex = "java"
 file-types = ["java"]
+shebangs = []
 roots = []
 indent = { tab-width = 4, unit = "    " }
 
@@ -250,6 +272,7 @@ name = "ledger"
 scope = "source.ledger"
 injection-regex = "ledger"
 file-types = ["ldg", "ledger", "journal"]
+shebangs = []
 roots = []
 comment-token = ";"
 indent = { tab-width = 4, unit = "    " }
@@ -259,6 +282,7 @@ name = "ocaml"
 scope = "source.ocaml"
 injection-regex = "ocaml"
 file-types = ["ml"]
+shebangs = []
 roots = []
 comment-token = "(**)"
 indent = { tab-width = 2, unit = "  " }
@@ -267,6 +291,7 @@ indent = { tab-width = 2, unit = "  " }
 name = "ocaml-interface"
 scope = "source.ocaml.interface"
 file-types = ["mli"]
+shebangs = []
 roots = []
 comment-token = "(**)"
 indent = { tab-width = 2, unit = "  "}
@@ -275,6 +300,7 @@ indent = { tab-width = 2, unit = "  "}
 name = "lua"
 scope = "source.lua"
 file-types = ["lua"]
+shebangs = []
 roots = []
 comment-token = "--"
 indent = { tab-width = 2, unit = "  " }
@@ -284,6 +310,7 @@ name = "svelte"
 scope = "source.svelte"
 injection-regex = "svelte"
 file-types = ["svelte"]
+shebangs = []
 roots = []
 indent = { tab-width = 2, unit = "  " }
 language-server = { command = "svelteserver", args = ["--stdio"] }
@@ -294,13 +321,15 @@ name = "vue"
 scope = "source.vue"
 injection-regex = "vue"
 file-types = ["vue"]
+shebangs = []
 roots = []
 indent = { tab-width = 2, unit = "  " }
 
 [[language]]
 name = "yaml"
 scope = "source.yaml"
 file-types = ["yml", "yaml"]
+shebangs = []
 roots = []
 comment-token = "#"
 indent = { tab-width = 2, unit = "  " }
@@ -320,6 +349,7 @@ name = "zig"
 scope = "source.zig"
 injection-regex = "zig"
 file-types = ["zig"]
+shebangs = []
 roots = ["build.zig"]
 auto-format = true
 comment-token = "//"
@@ -332,6 +362,7 @@ name = "prolog"
 scope = "source.prolog"
 roots = []
 file-types = ["pl", "prolog"]
+shebangs = ["swipl"]
 comment-token = "%"
 
 language-server = { command = "swipl", args = [
@@ -343,6 +374,7 @@ language-server = { command = "swipl", args = [
 name = "tsq"
 scope = "source.tsq"
 file-types = ["scm"]
+shebangs = []
 roots = []
 comment-token = ";"
 indent = { tab-width = 2, unit = "  " }
@@ -351,6 +383,7 @@ indent = { tab-width = 2, unit = "  " }
 name = "cmake"
 scope = "source.cmake"
 file-types = ["cmake", "CMakeLists.txt"]
+shebangs = []
 roots = []
 comment-token = "#"
 indent = { tab-width = 2, unit = "  " }