Dynamically load grammar libraries at runtime

3 files changed, 180 insertions, 145 deletions

diff --git a/helix-syntax/Cargo.toml b/helix-syntax/Cargo.toml
index 140e3d24..7ad24488 100644
--- a/helix-syntax/Cargo.toml
+++ b/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"
diff --git a/helix-syntax/build.rs b/helix-syntax/build.rs
index 847f8a67..02a5bf49 100644
--- a/helix-syntax/build.rs
+++ b/helix-syntax/build.rs
@@ -1,5 +1,7 @@
+use anyhow::Result;
 use std::fs;
 use std::path::PathBuf;
+use std::time::SystemTime;
 
 use std::sync::mpsc::channel;
 
@@ -15,66 +17,156 @@ fn collect_tree_sitter_dirs(ignore: &[String]) -> Vec<String> {
     dirs
 }
 
-fn collect_src_files(dir: &str) -> (Vec<String>, Vec<String>) {
-    eprintln!("Collect files for {}", dir);
+#[cfg(unix)]
+const DYLIB_EXTENSION: &str = "so";
 
-    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")
-        {
-            continue;
+#[cfg(windows)]
+const DYLIB_EXTENSION: &str = "dll";
+
+// const BUILD_TARGET: &'static str = env!("BUILD_TARGET");
+
+use anyhow::{anyhow, Context};
+use std::{path::Path, process::Command};
+
+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 {
+        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);
+
+    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);
+    // .target(BUILD_TARGET)
+    // .host(BUILD_TARGET);
+    let compiler = config.get_compiler();
+    let mut command = Command::new(compiler.path());
+    for (key, value) in compiler.env() {
+        command.env(key, value);
+    }
+
+    if cfg!(windows) {
+        command
+            .args(&["/nologo", "/LD", "/I"])
+            .arg(header_path)
+            .arg("/Od")
+            .arg(parser_path);
+        if let Some(scanner_path) = scanner_path.as_ref() {
+            command.arg(scanner_path);
         }
-        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());
+        command
+            .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);
     }
-    (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);
+    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)
+        ));
     }
-    build.compile(&format!("tree-sitter-{}-c", language));
+
+    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 build_cpp(files: Vec<String>, language: &str) {
-    let mut build = cc::Build::new();
+fn mtime(path: &Path) -> Result<SystemTime> {
+    Ok(fs::metadata(path)?.modified()?)
+}
 
-    let flag = if build.get_compiler().is_like_msvc() {
-        "/std:c++17"
-    } else {
-        "-std=c++14"
-    };
+// 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));
+// }
 
-    for file in files {
-        build
-            .file(&file)
-            .include(PathBuf::from(file).parent().unwrap())
-            .pic(true)
-            .warnings(false)
-            .cpp(true)
-            .flag_if_supported(flag);
-    }
-    build.compile(&format!("tree-sitter-{}-cpp", language));
-}
+// fn build_cpp(files: Vec<String>, language: &str) {
+//     let mut build = cc::Build::new();
+
+//     let flag = if build.get_compiler().is_like_msvc() {
+//         "/std:c++17"
+//     } else {
+//         "-std=c++14"
+//     };
+
+//     for file in files {
+//         build
+//             .file(&file)
+//             .include(PathBuf::from(file).parent().unwrap())
+//             .pic(true)
+//             .warnings(false)
+//             .cpp(true)
+//             .flag_if_supported(flag);
+//     }
+//     build.compile(&format!("tree-sitter-{}-cpp", language));
+// }
 
 fn build_dir(dir: &str, language: &str) {
     println!("Build language {}", language);
@@ -92,13 +184,9 @@ 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 = Path::new("languages").join(dir).join("src");
+    build_library(&path, language).unwrap();
 }
 
 fn main() {
@@ -129,6 +217,6 @@ fn main() {
     // drop(tx);
     assert_eq!(rx.try_iter().sum::<usize>(), n_jobs);
 
-    build_dir("tree-sitter-typescript/tsx", "tsx");
-    build_dir("tree-sitter-typescript/typescript", "typescript");
+    // build_dir("tree-sitter-typescript/tsx", "tsx");
+    // build_dir("tree-sitter-typescript/typescript", "typescript");
 }
diff --git a/helix-syntax/src/lib.rs b/helix-syntax/src/lib.rs
index 5e3bb3ea..b6c0ecf3 100644
--- a/helix-syntax/src/lib.rs
+++ b/helix-syntax/src/lib.rs
@@ -1,94 +1,39 @@
-use serde::{Deserialize, Serialize};
+use anyhow::{Context, Result};
+use libloading::{Library, Symbol};
 use tree_sitter::Language;
 
-#[macro_export]
-macro_rules! mk_extern {
-    ( $( $name:ident ),* ) => {
-        $(
-            extern "C" { pub fn $name() -> Language; }
-        )*
-    };
-}
-
-#[macro_export]
-macro_rules! mk_enum {
-    ( $( $camel:ident ),* ) => {
-        #[derive(Clone, Copy, Debug, PartialEq, Serialize, Deserialize)]
-        #[serde(rename_all = "lowercase")]
-        pub enum Lang {
-            $(
-                $camel,
-            )*
+fn replace_dashes_with_underscores(name: &str) -> String {
+    let mut result = String::with_capacity(name.len());
+    for c in name.chars() {
+        if c == '-' {
+            result.push('_');
+        } else {
+            result.push(c);
         }
-    };
+    }
+    result
 }
+#[cfg(unix)]
+const DYLIB_EXTENSION: &str = "so";
 
-#[macro_export]
-macro_rules! mk_get_language {
-    ( $( ($camel:ident, $name:ident) ),* ) => {
-        #[must_use]
-        pub fn get_language(lang: Lang) -> Language {
-            unsafe {
-                match lang {
-                    $(
-                        Lang::$camel => $name(),
-                    )*
-                }
-            }
-        }
-    };
-}
+#[cfg(windows)]
+const DYLIB_EXTENSION: &str = "dll";
 
-#[macro_export]
-macro_rules! mk_get_language_name {
-    ( $( $camel:ident ),* ) => {
-        #[must_use]
-        pub const fn get_language_name(lang: Lang) -> &'static str {
-            match lang {
-                $(
-                    Lang::$camel => stringify!($camel),
-                )*
-            }
-        }
-    };
-}
+pub fn get_language(runtime_path: &std::path::Path, name: &str) -> Result<Language> {
+    let name = name.to_ascii_lowercase();
+    let mut library_path = runtime_path.join("grammars").join(&name);
+    // TODO: duplicated under build
+    library_path.set_extension(DYLIB_EXTENSION);
 
-#[macro_export]
-macro_rules! mk_langs {
-    ( $( ($camel:ident, $name:ident) ),* ) => {
-        mk_extern!($( $name ),*);
-        mk_enum!($( $camel ),*);
-        mk_get_language!($( ($camel, $name) ),*);
-        mk_get_language_name!($( $camel ),*);
+    let library = unsafe { Library::new(&library_path) }
+        .with_context(|| format!("Error opening dynamic library {:?}", &library_path))?;
+    let language_fn_name = format!("tree_sitter_{}", replace_dashes_with_underscores(&name));
+    let language = unsafe {
+        let language_fn: Symbol<unsafe extern "C" fn() -> Language> = library
+            .get(language_fn_name.as_bytes())
+            .with_context(|| format!("Failed to load symbol {}", language_fn_name))?;
+        language_fn()
     };
+    std::mem::forget(library);
+    Ok(language)
 }
-
-mk_langs!(
-    // 1) Name for enum
-    // 2) tree-sitter function to call to get a Language
-    (Agda, tree_sitter_agda),
-    (Bash, tree_sitter_bash),
-    (Cpp, tree_sitter_cpp),
-    (CSharp, tree_sitter_c_sharp),
-    (Css, tree_sitter_css),
-    (C, tree_sitter_c),
-    (Elixir, tree_sitter_elixir),
-    (Go, tree_sitter_go),
-    // (Haskell, tree_sitter_haskell),
-    (Html, tree_sitter_html),
-    (Javascript, tree_sitter_javascript),
-    (Java, tree_sitter_java),
-    (Json, tree_sitter_json),
-    (Julia, tree_sitter_julia),
-    (Latex, tree_sitter_latex),
-    (Nix, tree_sitter_nix),
-    (Php, tree_sitter_php),
-    (Python, tree_sitter_python),
-    (Ruby, tree_sitter_ruby),
-    (Rust, tree_sitter_rust),
-    (Scala, tree_sitter_scala),
-    (Swift, tree_sitter_swift),
-    (Toml, tree_sitter_toml),
-    (Tsx, tree_sitter_tsx),
-    (Typescript, tree_sitter_typescript)
-);