Add the machinery to load syntax config from TOML.

It's embedded into the binary at build time for now, but it's progress.

12 files changed, 86 insertions, 55 deletions

diff --git a/Cargo.lock b/Cargo.lock
index 4d4d4554..40c2aa13 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -489,6 +489,7 @@ dependencies = [
  "once_cell",
  "regex",
  "ropey",
+ "serde",
  "smallvec",
  "tendril",
  "tree-sitter",
@@ -523,6 +524,7 @@ name = "helix-syntax"
 version = "0.1.0"
 dependencies = [
  "cc",
+ "serde",
  "threadpool",
  "tree-sitter",
 ]
@@ -549,6 +551,7 @@ dependencies = [
  "pulldown-cmark",
  "smol",
  "smol-timeout",
+ "toml",
  "tui",
 ]
 
@@ -1208,6 +1211,15 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "cda74da7e1a664f795bb1f8a87ec406fb89a02522cf6e50620d016add6dbbf5c"
 
 [[package]]
+name = "toml"
+version = "0.5.8"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a31142970826733df8241ef35dc040ef98c679ab14d7c3e54d827099b3acecaa"
+dependencies = [
+ "serde",
+]
+
+[[package]]
 name = "tree-sitter"
 version = "0.17.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
diff --git a/helix-core/Cargo.toml b/helix-core/Cargo.toml
index 49641ad6..d95c0f47 100644
--- a/helix-core/Cargo.toml
+++ b/helix-core/Cargo.toml
@@ -19,3 +19,5 @@ unicode-width = "0.1"
 tree-sitter = "0.17"
 once_cell = "1.4"
 regex = "1"
+
+serde = { version = "1.0", features = ["derive"] }
diff --git a/helix-core/src/syntax.rs b/helix-core/src/syntax.rs
index 3640e0cf..8a8d649a 100644
--- a/helix-core/src/syntax.rs
+++ b/helix-core/src/syntax.rs
@@ -10,9 +10,19 @@ use std::{
 };
 
 use once_cell::sync::{Lazy, OnceCell};
+use serde::{Deserialize, Serialize};
+
+#[derive(Serialize, Deserialize)]
+pub struct Configuration {
+    language: Vec<LanguageConfiguration>,
+}
 
 // largely based on tree-sitter/cli/src/loader.rs
+#[derive(Serialize, Deserialize)]
+#[serde(rename_all = "kebab-case")]
 pub struct LanguageConfiguration {
+    #[serde(rename = "name")]
+    pub(crate) language_id: Lang,
     pub scope: String,           // source.rust
     pub file_types: Vec<String>, // filename ends_with? <Gemfile, rb, etc>
     pub roots: Vec<String>,      // these indicate project roots <.git, Cargo.toml>
@@ -24,22 +34,29 @@ pub struct LanguageConfiguration {
     // injection_regex
     // first_line_regex
     //
-    //
-    pub(crate) language_id: Lang,
+    #[serde(skip)]
     pub(crate) highlight_config: OnceCell<Option<Arc<HighlightConfiguration>>>,
     // tags_config OnceCell<> https://github.com/tree-sitter/tree-sitter/pull/583
-    pub language_server_config: Option<LanguageServerConfiguration>,
-    pub indent_config: Option<IndentationConfiguration>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub language_server: Option<LanguageServerConfiguration>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub indent: Option<IndentationConfiguration>,
 }
 
+#[derive(Serialize, Deserialize)]
+#[serde(rename_all = "kebab-case")]
 pub struct LanguageServerConfiguration {
     pub command: String,
+    #[serde(default)]
+    #[serde(skip_serializing_if = "Vec::is_empty")]
     pub args: Vec<String>,
 }
 
+#[derive(Serialize, Deserialize)]
+#[serde(rename_all = "kebab-case")]
 pub struct IndentationConfiguration {
     pub tab_width: usize,
-    pub indent_unit: String,
+    pub unit: String,
 }
 
 impl LanguageConfiguration {
@@ -81,7 +98,7 @@ impl LanguageConfiguration {
     }
 }
 
-pub static LOADER: Lazy<Loader> = Lazy::new(Loader::init);
+pub static LOADER: OnceCell<Loader> = OnceCell::new();
 
 pub struct Loader {
     // highlight_names ?
@@ -90,48 +107,13 @@ pub struct Loader {
 }
 
 impl Loader {
-    fn init() -> Self {
+    pub fn new(config: Configuration) -> Self {
         let mut loader = Self {
             language_configs: Vec::new(),
             language_config_ids_by_file_type: HashMap::new(),
         };
 
-        // hardcoded from now, might load from toml
-        let configs = vec![
-            LanguageConfiguration {
-                scope: "source.rust".to_string(),
-                file_types: vec!["rs".to_string()],
-                language_id: Lang::Rust,
-                highlight_config: OnceCell::new(),
-                //
-                path: "../helix-syntax/languages/tree-sitter-rust".into(),
-                roots: vec![],
-                language_server_config: Some(LanguageServerConfiguration {
-                    command: "rust-analyzer".to_string(),
-                    args: vec![],
-                }),
-                indent_config: Some(IndentationConfiguration {
-                    tab_width: 4,
-                    indent_unit: String::from("    "),
-                }),
-            },
-            LanguageConfiguration {
-                scope: "source.toml".to_string(),
-                file_types: vec!["toml".to_string()],
-                language_id: Lang::Toml,
-                highlight_config: OnceCell::new(),
-                //
-                path: "../helix-syntax/languages/tree-sitter-toml".into(),
-                roots: vec![],
-                language_server_config: None,
-                indent_config: Some(IndentationConfiguration {
-                    tab_width: 2,
-                    indent_unit: String::from("  "),
-                }),
-            },
-        ];
-
-        for config in configs {
+        for config in config.language {
             // get the next id
             let language_id = loader.language_configs.len();
 
diff --git a/helix-lsp/src/lib.rs b/helix-lsp/src/lib.rs
index 64aa38f1..e7fe816a 100644
--- a/helix-lsp/src/lib.rs
+++ b/helix-lsp/src/lib.rs
@@ -127,7 +127,7 @@ impl Registry {
         ex: &smol::Executor,
     ) -> Option<Arc<Client>> {
         // TODO: propagate the error
-        if let Some(config) = &language_config.language_server_config {
+        if let Some(config) = &language_config.language_server {
             // avoid borrow issues
             let inner = &mut self.inner;
             let s_incoming = &self.incoming;
diff --git a/helix-syntax/Cargo.toml b/helix-syntax/Cargo.toml
index c50e3091..f145b6c0 100644
--- a/helix-syntax/Cargo.toml
+++ b/helix-syntax/Cargo.toml
@@ -8,6 +8,7 @@ edition = "2018"
 
 [dependencies]
 tree-sitter = "0.17"
+serde = { version = "1.0", features = ["derive"] }
 
 [build-dependencies]
 cc = { version = "1",  features = ["parallel"] }
diff --git a/helix-syntax/languages.toml b/helix-syntax/languages.toml
deleted file mode 100644
index dc4fcf6f..00000000
--- a/helix-syntax/languages.toml
+++ /dev/null
@@ -1,5 +0,0 @@
-[[language]]
-name = "rust"
-scope = "source.rust"
-injection-regex = "rust"
-file-types = ["rs"]
diff --git a/helix-syntax/src/lib.rs b/helix-syntax/src/lib.rs
index 64fe1023..5d3e0d99 100644
--- a/helix-syntax/src/lib.rs
+++ b/helix-syntax/src/lib.rs
@@ -1,3 +1,4 @@
+use serde::{Deserialize, Serialize};
 use tree_sitter::Language;
 
 #[macro_export]
@@ -12,7 +13,8 @@ macro_rules! mk_extern {
 #[macro_export]
 macro_rules! mk_enum {
     ( $( $camel:ident ),* ) => {
-        #[derive(Clone, Copy, Debug, PartialEq)]
+        #[derive(Clone, Copy, Debug, PartialEq, Serialize, Deserialize)]
+        #[serde(rename_all = "lowercase")]
         pub enum Lang {
             $(
                 $camel,
diff --git a/helix-term/Cargo.toml b/helix-term/Cargo.toml
index eaf8ebd7..2f3aa384 100644
--- a/helix-term/Cargo.toml
+++ b/helix-term/Cargo.toml
@@ -41,3 +41,6 @@ ignore = "0.4"
 dirs-next = "2.0"
 # markdown doc rendering
 pulldown-cmark = { version = "0.8", default-features = false }
+
+# config
+toml = "0.5"
diff --git a/helix-term/src/main.rs b/helix-term/src/main.rs
index 84b52ea2..dc6f31f4 100644
--- a/helix-term/src/main.rs
+++ b/helix-term/src/main.rs
@@ -73,6 +73,14 @@ fn main() {
 
     setup_logging(verbosity).expect("failed to initialize logging.");
 
+    // initialize language registry
+    use helix_core::syntax::{Loader, LOADER};
+    let toml = include_str!("../../languages.toml");
+    LOADER.get_or_init(|| {
+        let config = toml::from_str(&toml).expect("Could not parse languages.toml");
+        Loader::new(config)
+    });
+
     for _ in 0..num_cpus::get() {
         std::thread::spawn(move || smol::block_on(EX.run(smol::future::pending::<()>())));
     }
diff --git a/helix-term/src/ui/markdown.rs b/helix-term/src/ui/markdown.rs
index c4f6039b..55b90ad4 100644
--- a/helix-term/src/ui/markdown.rs
+++ b/helix-term/src/ui/markdown.rs
@@ -80,6 +80,8 @@ fn parse<'a>(contents: &'a str, theme: Option<&Theme>) -> tui::text::Text<'a> {
 
                         let rope = Rope::from(text.as_ref());
                         let syntax = syntax::LOADER
+                            .get()
+                            .unwrap()
                             .language_config_for_scope(&format!("source.{}", language))
                             .and_then(|config| config.highlight_config(theme.scopes()))
                             .map(|config| Syntax::new(&rope, config));
diff --git a/helix-view/src/document.rs b/helix-view/src/document.rs
index 0286b2b9..0ab8abbb 100644
--- a/helix-view/src/document.rs
+++ b/helix-view/src/document.rs
@@ -97,7 +97,10 @@ impl Document {
 
         let mut doc = Self::new(doc);
 
-        let language_config = LOADER.language_config_for_file_name(path.as_path());
+        let language_config = LOADER
+            .get()
+            .unwrap()
+            .language_config_for_file_name(path.as_path());
         doc.set_language(language_config, scopes);
 
         // canonicalize path to absolute value
@@ -161,7 +164,8 @@ impl Document {
     }
 
     pub fn set_language2(&mut self, scope: &str, scopes: &[String]) {
-        let language_config = LOADER.language_config_for_scope(scope);
+        let language_config = LOADER.get().unwrap().language_config_for_scope(scope);
+
         self.set_language(language_config, scopes);
     }
 
@@ -304,7 +308,7 @@ impl Document {
     pub fn tab_width(&self) -> usize {
         self.language
             .as_ref()
-            .and_then(|config| config.indent_config.as_ref())
+            .and_then(|config| config.indent.as_ref())
             .map(|config| config.tab_width)
             .unwrap_or(4) // fallback to 4 columns
     }
@@ -313,8 +317,8 @@ impl Document {
     pub fn indent_unit(&self) -> &str {
         self.language
             .as_ref()
-            .and_then(|config| config.indent_config.as_ref())
-            .map(|config| config.indent_unit.as_str())
+            .and_then(|config| config.indent.as_ref())
+            .map(|config| config.unit.as_str())
             .unwrap_or("  ") // fallback to 2 spaces
 
         // " ".repeat(TAB_WIDTH)
diff --git a/languages.toml b/languages.toml
new file mode 100644
index 00000000..eca55a13
--- /dev/null
+++ b/languages.toml
@@ -0,0 +1,20 @@
+[[language]]
+name = "rust"
+scope = "source.rust"
+injection-regex = "rust"
+file-types = ["rs"]
+roots = []
+path = "../helix-syntax/languages/tree-sitter-rust"
+
+language-server = { command = "rust-analyzer" }
+indent = { tab-width = 4, unit = "    " }
+
+[[language]]
+name = "toml"
+scope = "source.toml"
+injection-regex = "toml"
+file-types = ["toml"]
+roots = []
+path = "../helix-syntax/languages/tree-sitter-toml"
+
+indent = { tab-width = 2, unit = "  " }