Return an error from build-wasm if scanner uses unavailable functions

Cargo.lock

@@ -306,7 +306,7 @@ dependencies = [
  "itertools",
  "log",
  "smallvec",
- "wasmparser",
+ "wasmparser 0.118.1",
  "wasmtime-types",
 ]
 
@@ -1274,6 +1274,7 @@ dependencies = [
  "tree-sitter-tests-proc-macro",
  "unindent",
  "walkdir",
+ "wasmparser 0.121.0",
  "webbrowser",
  "which",
 ]
@@ -1487,6 +1488,17 @@ dependencies = [
  "semver",
 ]
 
+[[package]]
+name = "wasmparser"
+version = "0.121.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "953cf6a7606ab31382cb1caa5ae403e77ba70c7f8e12eeda167e7040d42bfda8"
+dependencies = [
+ "bitflags 2.4.1",
+ "indexmap",
+ "semver",
+]
+
 [[package]]
 name = "wasmtime"
 version = "16.0.0"
@@ -1506,7 +1518,7 @@ dependencies = [
  "serde_derive",
  "serde_json",
  "target-lexicon",
- "wasmparser",
+ "wasmparser 0.118.1",
  "wasmtime-cranelift",
  "wasmtime-environ",
  "wasmtime-jit",
@@ -1562,7 +1574,7 @@ dependencies = [
  "object",
  "target-lexicon",
  "thiserror",
- "wasmparser",
+ "wasmparser 0.118.1",
  "wasmtime-cranelift-shared",
  "wasmtime-environ",
  "wasmtime-versioned-export-macros",
@@ -1598,7 +1610,7 @@ dependencies = [
  "serde_derive",
  "target-lexicon",
  "thiserror",
- "wasmparser",
+ "wasmparser 0.118.1",
  "wasmtime-types",
 ]
 
@@ -1668,7 +1680,7 @@ dependencies = [
  "serde",
  "serde_derive",
  "thiserror",
- "wasmparser",
+ "wasmparser 0.118.1",
 ]
 
 [[package]]

cli/Cargo.toml

@@ -47,6 +47,7 @@ serde_derive = "1.0"
 smallbitvec = "2.5.1"
 tiny_http = "0.12.0"
 walkdir = "2.3.3"
+wasmparser = "0.121"
 webbrowser = "0.8.10"
 which = "4.4.0"
 

cli/src/tests/wasm_language_test.rs

@@ -9,6 +9,25 @@ lazy_static! {
     static ref ENGINE: Engine = Engine::default();
 }
 
+#[test]
+fn test_wasm_stdlib_symbols() {
+    let symbols = tree_sitter::wasm_stdlib_symbols().collect::<Vec<_>>();
+    assert_eq!(
+        symbols,
+        {
+            let mut symbols = symbols.clone();
+            symbols.sort();
+            symbols
+        },
+        "symbols aren't sorted"
+    );
+
+    assert!(symbols.contains(&"malloc"));
+    assert!(symbols.contains(&"free"));
+    assert!(symbols.contains(&"memset"));
+    assert!(symbols.contains(&"memcpy"));
+}
+
 #[test]
 fn test_load_wasm_language() {
     allocations::record(|| {

cli/src/wasm.rs

@@ -1,7 +1,9 @@
 use super::generate::parse_grammar::GrammarJSON;
-use anyhow::{Context, Result};
+use anyhow::{anyhow, Context, Result};
 use std::{fs, path::Path};
+use tree_sitter::wasm_stdlib_symbols;
 use tree_sitter_loader::Loader;
+use wasmparser::Parser;
 
 pub fn load_language_wasm_file(language_dir: &Path) -> Result<(String, Vec<u8>)> {
     let grammar_name = get_grammar_name(&language_dir)
@@ -46,5 +48,46 @@ pub fn compile_language_to_wasm(
         &output_filename,
         force_docker,
     )?;
+
+    // Exit with an error if the external scanner uses symbols from the
+    // C or C++ standard libraries that aren't available to wasm parsers.
+    let stdlib_symbols: Vec<_> = wasm_stdlib_symbols().collect();
+    let builtin_symbols = [
+        "__indirect_function_table",
+        "__memory_base",
+        "__table_base",
+        "memory",
+    ];
+
+    let mut missing_symbols = Vec::new();
+    let wasm_bytes = fs::read(&output_filename)?;
+    let parser = Parser::new(0);
+    for payload in parser.parse_all(&wasm_bytes) {
+        if let wasmparser::Payload::ImportSection(imports) = payload? {
+            for import in imports {
+                let import = import?.name;
+                if !builtin_symbols.contains(&import) && !stdlib_symbols.contains(&import) {
+                    missing_symbols.push(import);
+                }
+            }
+        }
+    }
+
+    if !missing_symbols.is_empty() {
+        Err(anyhow!(
+            concat!(
+                "This external scanner uses a symbol that isn't available to wasm parsers.\n",
+                "\n",
+                "Missing symbols:\n",
+                "    {}\n",
+                "\n",
+                "Available symbols:\n",
+                "    {}",
+            ),
+            missing_symbols.join("\n    "),
+            stdlib_symbols.join("\n    ")
+        ))?;
+    }
+
     Ok(())
 }

lib/binding_rust/build.rs

@@ -2,6 +2,8 @@ use std::path::{Path, PathBuf};
 use std::{env, fs};
 
 fn main() {
+    let out_dir = PathBuf::from(env::var("OUT_DIR").unwrap());
+
     println!("cargo:rerun-if-env-changed=TREE_SITTER_STATIC_ANALYSIS");
     if env::var("TREE_SITTER_STATIC_ANALYSIS").is_ok() {
         if let (Some(clang_path), Some(scan_build_path)) = (which("clang"), which("scan-build")) {
@@ -18,7 +20,13 @@ fn main() {
     }
 
     #[cfg(feature = "bindgen")]
-    generate_bindings();
+    generate_bindings(&out_dir);
+
+    fs::copy(
+        "src/wasm/stdlib-symbols.txt",
+        out_dir.join("stdlib-symbols.txt"),
+    )
+    .unwrap();
 
     let mut config = cc::Build::new();
 
@@ -47,7 +55,7 @@ fn main() {
 }
 
 #[cfg(feature = "bindgen")]
-fn generate_bindings() {
+fn generate_bindings(out_dir: &PathBuf) {
     const HEADER_PATH: &str = "include/tree_sitter/api.h";
 
     println!("cargo:rerun-if-changed={}", HEADER_PATH);
@@ -77,9 +85,7 @@ fn generate_bindings() {
         .generate()
         .expect("Failed to generate bindings");
 
-    let out_dir = PathBuf::from(env::var("OUT_DIR").unwrap());
     let bindings_rs = out_dir.join("bindings.rs");
-
     bindings.write_to_file(&bindings_rs).expect(&*format!(
         "Failed to write bindings into path: {bindings_rs:?}"
     ));

lib/binding_rust/lib.rs

@@ -2706,6 +2706,15 @@ impl fmt::Display for QueryError {
     }
 }
 
+pub fn wasm_stdlib_symbols() -> impl Iterator<Item = &'static str> {
+    const WASM_STDLIB_SYMBOLS: &'static str =
+        include_str!(concat!(env!("OUT_DIR"), "/stdlib-symbols.txt"));
+
+    WASM_STDLIB_SYMBOLS
+        .lines()
+        .map(|s| s.trim_matches(|c| c == '"' || c == ','))
+}
+
 extern "C" {
     fn free(ptr: *mut c_void);
 }

lib/binding_rust/wasm_language.rs

@@ -8,10 +8,10 @@ use std::{
 };
 pub use wasmtime;
 
-#[cfg(feature = "wasm")]
-pub fn test() {
-    // Force Cargo to include wasmtime-c-api as a dependency of this crate,
-    // even though it is only used by the C code.
+// Force Cargo to include wasmtime-c-api as a dependency of this crate,
+// even though it is only used by the C code.
+#[allow(unused)]
+fn _use_wasmtime() {
     wasmtime_c_api::wasm_engine_new();
 }