fix(wasm): fix alias map size computation

This fixes a crash where parsing with certain languages can lead to a crash due to how the alias map was allocated and laid out in wasm memory

(cherry picked from commit f09dc3cf46)

cli/src/tests/helpers/fixtures.rs

@@ -23,6 +23,9 @@ static TEST_LOADER: LazyLock<Loader> = LazyLock::new(|| {
     loader
 });
 
+#[cfg(feature = "wasm")]
+pub static ENGINE: LazyLock<tree_sitter::wasmtime::Engine> = LazyLock::new(Default::default);
+
 pub fn test_loader() -> &'static Loader {
     &TEST_LOADER
 }
@@ -43,10 +46,19 @@ pub fn get_language(name: &str) -> Language {
 }
 
 pub fn get_test_fixture_language(name: &str) -> Language {
+    get_test_fixture_language_internal(name, false)
+}
+
+#[cfg(feature = "wasm")]
+pub fn get_test_fixture_language_wasm(name: &str) -> Language {
+    get_test_fixture_language_internal(name, true)
+}
+
+fn get_test_fixture_language_internal(name: &str, wasm: bool) -> Language {
     let grammar_dir_path = fixtures_dir().join("test_grammars").join(name);
     let grammar_json = load_grammar_file(&grammar_dir_path.join("grammar.js"), None).unwrap();
     let (parser_name, parser_code) = generate_parser(&grammar_json).unwrap();
-    get_test_language(&parser_name, &parser_code, Some(&grammar_dir_path))
+    get_test_language_internal(&parser_name, &parser_code, Some(&grammar_dir_path), wasm)
 }
 
 pub fn get_language_queries_path(language_name: &str) -> PathBuf {
@@ -87,6 +99,15 @@ pub fn get_tags_config(language_name: &str) -> TagsConfiguration {
 }
 
 pub fn get_test_language(name: &str, parser_code: &str, path: Option<&Path>) -> Language {
+    get_test_language_internal(name, parser_code, path, false)
+}
+
+fn get_test_language_internal(
+    name: &str,
+    parser_code: &str,
+    path: Option<&Path>,
+    wasm: bool,
+) -> Language {
     let src_dir = scratch_dir().join("src").join(name);
     fs::create_dir_all(&src_dir).unwrap();
 
@@ -136,5 +157,21 @@ pub fn get_test_language(name: &str, parser_code: &str, path: Option<&Path>) ->
     config.header_paths = vec![&HEADER_DIR];
     config.name = name.to_string();
 
-    TEST_LOADER.load_language_at_path_with_name(config).unwrap()
+    if wasm {
+        #[cfg(feature = "wasm")]
+        {
+            let mut loader = Loader::with_parser_lib_path(SCRATCH_DIR.clone());
+            loader.use_wasm(&ENGINE);
+            if env::var("TREE_SITTER_GRAMMAR_DEBUG").is_ok() {
+                loader.debug_build(true);
+            }
+            loader.load_language_at_path_with_name(config).unwrap()
+        }
+        #[cfg(not(feature = "wasm"))]
+        {
+            unimplemented!("Wasm feature is not enabled")
+        }
+    } else {
+        TEST_LOADER.load_language_at_path_with_name(config).unwrap()
+    }
 }

cli/src/tests/wasm_language_test.rs

@@ -1,14 +1,13 @@
-use std::{fs, sync::LazyLock};
+use std::fs;
 
 use streaming_iterator::StreamingIterator;
-use tree_sitter::{
-    wasmtime::Engine, Parser, Query, QueryCursor, WasmError, WasmErrorKind, WasmStore,
+use tree_sitter::{Parser, Query, QueryCursor, WasmError, WasmErrorKind, WasmStore};
+
+use crate::tests::helpers::{
+    allocations,
+    fixtures::{get_test_fixture_language_wasm, ENGINE, WASM_DIR},
 };
 
-use crate::tests::helpers::{allocations, fixtures::WASM_DIR};
-
-static ENGINE: LazyLock<Engine> = LazyLock::new(Engine::default);
-
 #[test]
 fn test_wasm_stdlib_symbols() {
     let symbols = tree_sitter::wasm_stdlib_symbols().collect::<Vec<_>>();
@@ -92,6 +91,33 @@ fn test_load_wasm_javascript_language() {
     });
 }
 
+#[test]
+fn test_load_wasm_python_language() {
+    allocations::record(|| {
+        let mut store = WasmStore::new(&ENGINE).unwrap();
+        let mut parser = Parser::new();
+        let wasm = fs::read(WASM_DIR.join("tree-sitter-python.wasm")).unwrap();
+        let language = store.load_language("python", &wasm).unwrap();
+        parser.set_wasm_store(store).unwrap();
+        parser.set_language(&language).unwrap();
+        let tree = parser.parse("a = b\nc = d", None).unwrap();
+        assert_eq!(tree.root_node().to_sexp(), "(module (expression_statement (assignment left: (identifier) right: (identifier))) (expression_statement (assignment left: (identifier) right: (identifier))))");
+    });
+}
+
+#[test]
+fn test_load_fixture_language_wasm() {
+    allocations::record(|| {
+        let store = WasmStore::new(&ENGINE).unwrap();
+        let mut parser = Parser::new();
+        let language = get_test_fixture_language_wasm("epsilon_external_tokens");
+        parser.set_wasm_store(store).unwrap();
+        parser.set_language(&language).unwrap();
+        let tree = parser.parse("hello", None).unwrap();
+        assert_eq!(tree.root_node().to_sexp(), "(document (zero_width))");
+    });
+}
+
 #[test]
 fn test_load_multiple_wasm_languages() {
     allocations::record(|| {

cli/src/wasm.rs

@@ -5,7 +5,7 @@ use std::{
 
 use anyhow::{anyhow, Context, Result};
 use tree_sitter::wasm_stdlib_symbols;
-use tree_sitter_generate::parse_grammar::GrammarJSON;
+use tree_sitter_generate::{load_grammar_file, parse_grammar::GrammarJSON};
 use tree_sitter_loader::Loader;
 use wasmparser::Parser;
 
@@ -46,7 +46,8 @@ pub fn compile_language_to_wasm(
     output_file: Option<PathBuf>,
     force_docker: bool,
 ) -> Result<()> {
-    let grammar_name = get_grammar_name(language_dir)?;
+    let grammar_name = get_grammar_name(language_dir)
+        .or_else(|_| load_grammar_file(&language_dir.join("grammar.js"), None))?;
     let output_filename =
         output_file.unwrap_or_else(|| output_dir.join(format!("tree-sitter-{grammar_name}.wasm")));
     let src_path = language_dir.join("src");

lib/src/wasm_store.c

@@ -1377,11 +1377,12 @@ const TSLanguage *ts_wasm_store_load_language(
       if (symbol == 0) break;
       uint16_t value_count;
       memcpy(&value_count, &memory[wasm_language.alias_map + alias_map_size], sizeof(value_count));
+      alias_map_size += sizeof(uint16_t);
       alias_map_size += value_count * sizeof(TSSymbol);
     }
     language->alias_map = copy(
       &memory[wasm_language.alias_map],
-      alias_map_size * sizeof(TSSymbol)
+      alias_map_size
     );
     language->alias_sequences = copy(
       &memory[wasm_language.alias_sequences],

test/fixtures/test_grammars/external_tokens/scanner.c

@@ -30,7 +30,7 @@ void tree_sitter_external_tokens_external_scanner_destroy(void *payload) {
 unsigned tree_sitter_external_tokens_external_scanner_serialize(
   void *payload,
   char *buffer
-) { return true; }
+) { return 0; }
 
 void tree_sitter_external_tokens_external_scanner_deserialize(
   void *payload,