fix(xtask): restore stripped sourcesContent when building the wasm module

crates/xtask/src/build_wasm.rs

@@ -3,7 +3,7 @@ use std::{
     ffi::{OsStr, OsString},
     fmt::Write,
     fs,
-    path::PathBuf,
+    path::{Path, PathBuf},
     process::Command,
     time::Duration,
 };
@@ -16,7 +16,9 @@ use notify::{
 };
 use notify_debouncer_full::new_debouncer;
 
-use crate::{bail_on_err, watch_wasm, BuildWasm, EMSCRIPTEN_TAG};
+use crate::{
+    bail_on_err, embed_sources::embed_sources_in_map, watch_wasm, BuildWasm, EMSCRIPTEN_TAG,
+};
 
 #[derive(PartialEq, Eq)]
 enum EmccSource {
@@ -48,10 +50,14 @@ const EXPORTED_RUNTIME_METHODS: [&str; 19] = [
 ];
 
 pub fn run_wasm(args: &BuildWasm) -> Result<()> {
-    let mut emscripten_flags = vec!["-O3", "--minify", "0"];
+    let mut emscripten_flags = if args.debug {
+        vec!["-O0", "--minify", "0"]
+    } else {
+        vec!["-O3", "--minify", "0"]
+    };
 
     if args.debug {
-        emscripten_flags.extend(["-s", "ASSERTIONS=1", "-s", "SAFE_HEAP=1", "-O0", "-g"]);
+        emscripten_flags.extend(["-s", "ASSERTIONS=1", "-s", "SAFE_HEAP=1", "-g"]);
     }
 
     if args.verbose {
@@ -287,6 +293,17 @@ fn build_wasm(cmd: &mut Command, edit_tsd: bool) -> Result<()> {
         fs::write(file, content)?;
     }
 
+    // Post-process the source map to embed source content for optimized builds
+    let map_path = Path::new("lib")
+        .join("binding_web")
+        .join("lib")
+        .join("web-tree-sitter.wasm.map");
+    if map_path.exists() {
+        if let Err(e) = embed_sources_in_map(&map_path) {
+            eprintln!("Warning: Failed to embed sources in source map: {e}");
+        }
+    }
+
     Ok(())
 }
 

crates/xtask/src/embed_sources.rs

@@ -0,0 +1,61 @@
+use anyhow::Result;
+use std::fs;
+use std::path::Path;
+
+/// Restores sourcesContent if it was stripped by Binaryen.
+///
+/// This is a workaround for Binaryen where `wasm-opt -O2` and higher
+/// optimization levels strip the `sourcesContent` field from source maps,
+/// even when the source map was generated with `--sources` flag.
+///
+/// This is fixed upstream in Binaryen as of Apr 9, 2025, but there hasn't been a release with the fix yet.
+/// See: <https://github.com/WebAssembly/binaryen/issues/6805>
+///
+/// This reads the original source files and embeds them in the
+/// source map's `sourcesContent` field, making debugging possible even
+/// with optimized builds.
+///
+/// TODO: Once Binaryen releases a version with the fix, and emscripten updates to that
+/// version, and we update our emscripten version, this function can be removed.
+pub fn embed_sources_in_map(map_path: &Path) -> Result<()> {
+    let map_content = fs::read_to_string(map_path)?;
+    let mut map: serde_json::Value = serde_json::from_str(&map_content)?;
+
+    if let Some(sources_content) = map.get("sourcesContent") {
+        if let Some(arr) = sources_content.as_array() {
+            if !arr.is_empty() && arr.iter().any(|v| !v.is_null()) {
+                return Ok(());
+            }
+        }
+    }
+
+    let sources = map["sources"]
+        .as_array()
+        .ok_or_else(|| anyhow::anyhow!("No sources array in source map"))?;
+
+    let map_dir = map_path.parent().unwrap_or(Path::new("."));
+    let mut sources_content = Vec::new();
+
+    for source in sources {
+        let source_path = source.as_str().unwrap_or("");
+        let full_path = map_dir.join(source_path);
+
+        let content = if full_path.exists() {
+            match fs::read_to_string(&full_path) {
+                Ok(content) => serde_json::Value::String(content),
+                Err(_) => serde_json::Value::Null,
+            }
+        } else {
+            serde_json::Value::Null
+        };
+
+        sources_content.push(content);
+    }
+
+    map["sourcesContent"] = serde_json::Value::Array(sources_content);
+
+    let output = serde_json::to_string(&map)?;
+    fs::write(map_path, output)?;
+
+    Ok(())
+}

crates/xtask/src/main.rs

@@ -3,6 +3,7 @@ mod build_wasm;
 mod bump;
 mod check_wasm_exports;
 mod clippy;
+mod embed_sources;
 mod fetch;
 mod generate;
 mod test;