Dynamically load grammar libraries at runtime

.gitignore

@@ -3,3 +3,4 @@ target
 helix-term/rustfmt.toml
 helix-syntax/languages/
 result
+runtime/grammars

helix-core/src/indent.rs

@@ -253,14 +253,14 @@ where
 
         let doc = Rope::from(doc);
         use crate::syntax::{
-            Configuration, IndentationConfiguration, Lang, LanguageConfiguration, Loader,
+            Configuration, IndentationConfiguration, LanguageConfiguration, Loader,
         };
         use once_cell::sync::OnceCell;
         let loader = Loader::new(Configuration {
             language: vec![LanguageConfiguration {
                 scope: "source.rust".to_string(),
                 file_types: vec!["rs".to_string()],
-                language_id: Lang::Rust,
+                language_id: "Rust".to_string(),
                 highlight_config: OnceCell::new(),
                 //
                 roots: vec![],

helix-core/src/syntax.rs

@@ -5,7 +5,7 @@ use crate::{
     Rope, RopeSlice, Tendril,
 };
 
-pub use helix_syntax::{get_language, get_language_name, Lang};
+pub use helix_syntax::get_language;
 
 use arc_swap::ArcSwap;
 
@@ -31,7 +31,7 @@ pub struct Configuration {
 #[serde(rename_all = "kebab-case")]
 pub struct LanguageConfiguration {
     #[serde(rename = "name")]
-    pub(crate) language_id: Lang,
+    pub(crate) language_id: String,
     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>
@@ -153,7 +153,7 @@ fn read_query(language: &str, filename: &str) -> String {
 
 impl LanguageConfiguration {
     fn initialize_highlight(&self, scopes: &[String]) -> Option<Arc<HighlightConfiguration>> {
-        let language = get_language_name(self.language_id).to_ascii_lowercase();
+        let language = self.language_id.to_ascii_lowercase();
 
         let highlights_query = read_query(&language, "highlights.scm");
         // always highlight syntax errors
@@ -166,7 +166,7 @@ impl LanguageConfiguration {
         if highlights_query.is_empty() {
             None
         } else {
-            let language = get_language(self.language_id);
+            let language = get_language(&crate::RUNTIME_DIR, &self.language_id).ok()?;
             let config = HighlightConfiguration::new(
                 language,
                 &highlights_query,
@@ -198,7 +198,7 @@ impl LanguageConfiguration {
     pub fn indent_query(&self) -> Option<&IndentQuery> {
         self.indent_query
             .get_or_init(|| {
-                let language = get_language_name(self.language_id).to_ascii_lowercase();
+                let language = self.language_id.to_ascii_lowercase();
 
                 let toml = load_runtime_file(&language, "indents.toml").ok()?;
                 toml::from_slice(toml.as_bytes()).ok()
@@ -1802,7 +1802,7 @@ mod test {
         .map(String::from)
         .collect();
 
-        let language = get_language(Lang::Rust);
+        let language = get_language(&crate::RUNTIME_DIR, "Rust").unwrap();
         let config = HighlightConfiguration::new(
             language,
             &std::fs::read_to_string(

helix-syntax/Cargo.toml

@@ -12,8 +12,10 @@ include = ["src/**/*", "languages/**/*", "build.rs", "!**/docs/**/*", "!**/test/
 
 [dependencies]
 tree-sitter = "0.19"
-serde = { version = "1.0", features = ["derive"] }
+libloading = "0.7"
+anyhow = "1"
 
 [build-dependencies]
-cc = { version = "1",  features = ["parallel"] }
+cc = { version = "1" }
 threadpool = { version = "1.0" }
+anyhow = "1"

helix-syntax/build.rs

@@ -1,79 +1,147 @@
+use anyhow::{anyhow, Context, Result};
 use std::fs;
-use std::path::PathBuf;
+use std::time::SystemTime;
+use std::{
+    path::{Path, PathBuf},
+    process::Command,
+};
 
 use std::sync::mpsc::channel;
 
-fn collect_tree_sitter_dirs(ignore: &[String]) -> Vec<String> {
+fn collect_tree_sitter_dirs(ignore: &[String]) -> Result<Vec<String>> {
     let mut dirs = Vec::new();
-    for entry in fs::read_dir("languages").unwrap().flatten() {
+    let path = PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("languages");
+
+    for entry in fs::read_dir(path)? {
+        let entry = entry?;
         let path = entry.path();
-        let dir = path.file_name().unwrap().to_str().unwrap().to_string();
-        if !ignore.contains(&dir) {
-            dirs.push(dir);
+
+        if !entry.file_type()?.is_dir() {
+            continue;
         }
-    }
-    dirs
-}
 
-fn collect_src_files(dir: &str) -> (Vec<String>, Vec<String>) {
-    eprintln!("Collect files for {}", dir);
+        let dir = path.file_name().unwrap().to_str().unwrap().to_string();
 
-    let mut c_files = Vec::new();
-    let mut cpp_files = Vec::new();
-    let path = PathBuf::from("languages").join(&dir).join("src");
-    for entry in fs::read_dir(path).unwrap().flatten() {
-        let path = entry.path();
-        if path
-            .file_stem()
-            .unwrap()
-            .to_str()
-            .unwrap()
-            .starts_with("binding")
-        {
+        // filter ignores
+        if ignore.contains(&dir) {
             continue;
         }
-        if let Some(ext) = path.extension() {
-            if ext == "c" {
-                c_files.push(path.to_str().unwrap().to_string());
-            } else if ext == "cc" || ext == "cpp" || ext == "cxx" {
-                cpp_files.push(path.to_str().unwrap().to_string());
-            }
-        }
+        dirs.push(dir)
     }
-    (c_files, cpp_files)
-}
 
-fn build_c(files: Vec<String>, language: &str) {
-    let mut build = cc::Build::new();
-    for file in files {
-        build
-            .file(&file)
-            .include(PathBuf::from(file).parent().unwrap())
-            .pic(true)
-            .warnings(false);
-    }
-    build.compile(&format!("tree-sitter-{}-c", language));
+    Ok(dirs)
 }
 
-fn build_cpp(files: Vec<String>, language: &str) {
-    let mut build = cc::Build::new();
+#[cfg(unix)]
+const DYLIB_EXTENSION: &str = "so";
+
+#[cfg(windows)]
+const DYLIB_EXTENSION: &str = "dll";
 
-    let flag = if build.get_compiler().is_like_msvc() {
-        "/std:c++17"
+fn build_library(src_path: &Path, language: &str) -> Result<()> {
+    let header_path = src_path;
+    // let grammar_path = src_path.join("grammar.json");
+    let parser_path = src_path.join("parser.c");
+    let mut scanner_path = src_path.join("scanner.c");
+
+    let scanner_path = if scanner_path.exists() {
+        Some(scanner_path)
     } else {
-        "-std=c++14"
+        scanner_path.set_extension("cc");
+        if scanner_path.exists() {
+            Some(scanner_path)
+        } else {
+            None
+        }
     };
+    let parser_lib_path = PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("../runtime/grammars");
+    let mut library_path = parser_lib_path.join(language);
+    library_path.set_extension(DYLIB_EXTENSION);
 
-    for file in files {
-        build
-            .file(&file)
-            .include(PathBuf::from(file).parent().unwrap())
-            .pic(true)
-            .warnings(false)
-            .cpp(true)
-            .flag_if_supported(flag);
+    let recompile = needs_recompile(&library_path, &parser_path, &scanner_path)
+        .with_context(|| "Failed to compare source and binary timestamps")?;
+
+    if !recompile {
+        return Ok(());
+    }
+    let mut config = cc::Build::new();
+    config.cpp(true).opt_level(2).cargo_metadata(false);
+    let compiler = config.get_compiler();
+    let mut command = Command::new(compiler.path());
+    command.current_dir(src_path);
+    for (key, value) in compiler.env() {
+        command.env(key, value);
     }
-    build.compile(&format!("tree-sitter-{}-cpp", language));
+
+    if cfg!(windows) {
+        command
+            .args(&["/nologo", "/LD", "/I"])
+            .arg(header_path)
+            .arg("/Od");
+        if let Some(scanner_path) = scanner_path.as_ref() {
+            command.arg(scanner_path);
+        }
+
+        command
+            .arg(parser_path)
+            .arg("/link")
+            .arg(format!("/out:{}", library_path.to_str().unwrap()));
+    } else {
+        command
+            .arg("-shared")
+            .arg("-fPIC")
+            .arg("-fno-exceptions")
+            .arg("-g")
+            .arg("-I")
+            .arg(header_path)
+            .arg("-o")
+            .arg(&library_path)
+            .arg("-O2");
+        if let Some(scanner_path) = scanner_path.as_ref() {
+            if scanner_path.extension() == Some("c".as_ref()) {
+                command.arg("-xc").arg("-std=c99").arg(scanner_path);
+            } else {
+                command.arg(scanner_path);
+            }
+        }
+        command.arg("-xc").arg(parser_path);
+    }
+
+    let output = command
+        .output()
+        .with_context(|| "Failed to execute C compiler")?;
+    if !output.status.success() {
+        return Err(anyhow!(
+            "Parser compilation failed.\nStdout: {}\nStderr: {}",
+            String::from_utf8_lossy(&output.stdout),
+            String::from_utf8_lossy(&output.stderr)
+        ));
+    }
+
+    Ok(())
+}
+fn needs_recompile(
+    lib_path: &Path,
+    parser_c_path: &Path,
+    scanner_path: &Option<PathBuf>,
+) -> Result<bool> {
+    if !lib_path.exists() {
+        return Ok(true);
+    }
+    let lib_mtime = mtime(lib_path)?;
+    if mtime(parser_c_path)? > lib_mtime {
+        return Ok(true);
+    }
+    if let Some(scanner_path) = scanner_path {
+        if mtime(scanner_path)? > lib_mtime {
+            return Ok(true);
+        }
+    }
+    Ok(false)
+}
+
+fn mtime(path: &Path) -> Result<SystemTime> {
+    Ok(fs::metadata(path)?.modified()?)
 }
 
 fn build_dir(dir: &str, language: &str) {
@@ -92,22 +160,21 @@ fn build_dir(dir: &str, language: &str) {
         eprintln!("You can fix in using 'git submodule init && git submodule update --recursive'.");
         std::process::exit(1);
     }
-    let (c, cpp) = collect_src_files(dir);
-    if !c.is_empty() {
-        build_c(c, language);
-    }
-    if !cpp.is_empty() {
-        build_cpp(cpp, language);
-    }
+
+    let path = PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("languages")
+        .join(dir)
+        .join("src");
+
+    build_library(&path, language).unwrap();
 }
 
 fn main() {
     let ignore = vec![
         "tree-sitter-typescript".to_string(),
         "tree-sitter-haskell".to_string(), // aarch64 failures: https://github.com/tree-sitter/tree-sitter-haskell/issues/34
-        ".DS_Store".to_string(),
     ];
-    let dirs = collect_tree_sitter_dirs(&ignore);
+    let dirs = collect_tree_sitter_dirs(&ignore).unwrap();
 
     let mut n_jobs = 0;
     let pool = threadpool::Builder::new().build(); // by going through the builder, it'll use num_cpus
@@ -118,7 +185,7 @@ fn main() {
         n_jobs += 1;
 
         pool.execute(move || {
-            let language = &dir[12..]; // skip tree-sitter- prefix
+            let language = &dir.strip_prefix("tree-sitter-").unwrap();
             build_dir(&dir, language);
 
             // report progress