0.25.10

Co-authored-by: Amaan Qureshi <git@amaanq.com>
(cherry picked from commit 92678f0fc5)

.editorconfig

@@ -10,6 +10,9 @@ insert_final_newline = true
 [*.rs]
 indent_size = 4
 
+[*.{zig,zon}]
+indent_size = 4
+
 [Makefile]
 indent_style = tab
 indent_size = 8

.github/workflows/build.yml

@@ -139,6 +139,8 @@ jobs:
           [[ -n $runner ]] && printf 'CROSS_RUNNER=%s\n' "$runner" >> $GITHUB_ENV
         fi
 
+    # TODO: Remove RUSTFLAGS="--cap-lints allow" once we use a wasmtime release that addresses
+    # the `mismatched-lifetime-syntaxes` lint
     - name: Build wasmtime library
       if: ${{ !matrix.use-cross && contains(matrix.features, 'wasm') }}
       run: |
@@ -156,6 +158,7 @@ jobs:
         printf 'CMAKE_PREFIX_PATH=%s\n' "$PWD/artifacts" >> $GITHUB_ENV
       env:
         WASMTIME_REPO: https://github.com/bytecodealliance/wasmtime
+        RUSTFLAGS: "--cap-lints allow"
 
     - name: Build C library (make)
       if: ${{ runner.os != 'Windows' }}
@@ -195,6 +198,13 @@ jobs:
         npm run build
         npm run build:debug
 
+    - name: Check no_std builds
+      if: ${{ !matrix.no-run && inputs.run-test }}
+      shell: bash
+      run: |
+        cd lib
+        $BUILD_CMD check --no-default-features
+
     - name: Build target
       run: $BUILD_CMD build --release --target=${{ matrix.target }} --features=${{ matrix.features }}
 

.github/workflows/ci.yml

@@ -32,11 +32,6 @@ jobs:
         uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
           toolchain: stable
-
-      - name: Set up nightly Rust toolchain
-        uses: actions-rust-lang/setup-rust-toolchain@v1
-        with:
-          toolchain: nightly
           components: clippy, rustfmt
 
       - name: Lint files

Cargo.lock

@@ -10,14 +10,14 @@ checksum = "512761e0bb2578dd7380c6baaa0f4ce03e84f95e960231d1dec8bf4d7d6e2627"
 
 [[package]]
 name = "ahash"
-version = "0.8.11"
+version = "0.8.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e89da841a80418a9b391ebaea17f5c112ffaaa96f621d2c285b5174da76b9011"
+checksum = "5a15f179cd60c4584b8a8c596927aadc462e27f2ca70c04e0071964a73ba7a75"
 dependencies = [
  "cfg-if",
  "once_cell",
  "version_check",
- "zerocopy",
+ "zerocopy 0.8.27",
 ]
 
 [[package]]
@@ -1118,9 +1118,9 @@ checksum = "884e2677b40cc8c339eaefcb701c32ef1fd2493d71118dc0ca4b6a736c93bd67"
 
 [[package]]
 name = "libc"
-version = "0.2.169"
+version = "0.2.175"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b5aba8db14291edd000dfcc4d620c7ebfb122c613afb886ca8803fa4e128a20a"
+checksum = "6a82ae493e598baaea5209805c49bbf2ea7de956d50d7da0da1164f9c6d28543"
 
 [[package]]
 name = "libgit2-sys"
@@ -1460,7 +1460,7 @@ version = "0.2.20"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "77957b295656769bb8ad2b6a6b09d897d94f05c41b069aede1fcdaa675eaea04"
 dependencies = [
- "zerocopy",
+ "zerocopy 0.7.35",
 ]
 
 [[package]]
@@ -1563,9 +1563,9 @@ dependencies = [
 
 [[package]]
 name = "regalloc2"
-version = "0.11.1"
+version = "0.11.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "145c1c267e14f20fb0f88aa76a1c5ffec42d592c1d28b3cd9148ae35916158d3"
+checksum = "dc06e6b318142614e4a48bc725abbf08ff166694835c43c9dae5a9009704639a"
 dependencies = [
  "allocator-api2",
  "bumpalo",
@@ -1615,15 +1615,14 @@ dependencies = [
 
 [[package]]
 name = "ring"
-version = "0.17.8"
+version = "0.17.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c17fa4cb658e3583423e915b9f3acc01cceaee1860e33d59ebae66adc3a2dc0d"
+checksum = "70ac5d832aa16abd7d1def883a8545280c20a60f523a370aa3a9617c2b8550ee"
 dependencies = [
  "cc",
  "cfg-if",
  "getrandom",
  "libc",
- "spin",
  "untrusted",
  "windows-sys 0.52.0",
 ]
@@ -1787,12 +1786,6 @@ dependencies = [
  "serde",
 ]
 
-[[package]]
-name = "spin"
-version = "0.9.8"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6980e8d7511241f8acf4aebddbb1ff938df5eebe98691418c4468d0b72a96a67"
-
 [[package]]
 name = "sptr"
 version = "0.3.2"
@@ -1885,11 +1878,11 @@ dependencies = [
 
 [[package]]
 name = "thiserror"
-version = "2.0.11"
+version = "2.0.16"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d452f284b73e6d76dd36758a0c8684b1d5be31f92b89d07fd5822175732206fc"
+checksum = "3467d614147380f2e4e374161426ff399c91084acd2363eaf549172b3d5e60c0"
 dependencies = [
- "thiserror-impl 2.0.11",
+ "thiserror-impl 2.0.16",
 ]
 
 [[package]]
@@ -1905,9 +1898,9 @@ dependencies = [
 
 [[package]]
 name = "thiserror-impl"
-version = "2.0.11"
+version = "2.0.16"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "26afc1baea8a989337eeb52b6e72a039780ce45c3edfcc9c5b9d112feeb173c2"
+checksum = "6c5e1be1c48b9172ee610da68fd9cd2770e7a4056cb3fc98710ee6906f0c7960"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -2011,6 +2004,12 @@ dependencies = [
  "winnow",
 ]
 
+[[package]]
+name = "topological-sort"
+version = "0.2.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ea68304e134ecd095ac6c3574494fc62b909f416c4fca77e440530221e549d3d"
+
 [[package]]
 name = "tracing"
 version = "0.1.41"
@@ -2044,7 +2043,7 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "bindgen",
  "cc",
@@ -2058,7 +2057,7 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-cli"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "ansi_colours",
  "anstyle",
@@ -2093,6 +2092,7 @@ dependencies = [
  "streaming-iterator",
  "tempfile",
  "tiny_http",
+ "topological-sort",
  "tree-sitter",
  "tree-sitter-config",
  "tree-sitter-generate",
@@ -2110,7 +2110,7 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-config"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "anyhow",
  "etcetera",
@@ -2120,7 +2120,7 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-generate"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "anyhow",
  "heck",
@@ -2134,28 +2134,29 @@ dependencies = [
  "serde",
  "serde_json",
  "smallbitvec",
- "thiserror 2.0.11",
+ "thiserror 2.0.16",
+ "topological-sort",
  "tree-sitter",
  "url",
 ]
 
 [[package]]
 name = "tree-sitter-highlight"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "regex",
  "streaming-iterator",
- "thiserror 2.0.11",
+ "thiserror 2.0.16",
  "tree-sitter",
 ]
 
 [[package]]
 name = "tree-sitter-language"
-version = "0.1.4"
+version = "0.1.5"
 
 [[package]]
 name = "tree-sitter-loader"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "anyhow",
  "cc",
@@ -2178,12 +2179,12 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-tags"
-version = "0.25.1"
+version = "0.25.10"
 dependencies = [
  "memchr",
  "regex",
  "streaming-iterator",
- "thiserror 2.0.11",
+ "thiserror 2.0.16",
  "tree-sitter",
 ]
 
@@ -2391,19 +2392,19 @@ dependencies = [
 
 [[package]]
 name = "wasm-encoder"
-version = "0.221.2"
+version = "0.221.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c17a3bd88f2155da63a1f2fcb8a56377a24f0b6dfed12733bb5f544e86f690c5"
+checksum = "dc8444fe4920de80a4fe5ab564fff2ae58b6b73166b89751f8c6c93509da32e5"
 dependencies = [
  "leb128",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
 ]
 
 [[package]]
 name = "wasmparser"
-version = "0.221.2"
+version = "0.221.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9845c470a2e10b61dd42c385839cdd6496363ed63b5c9e420b5488b77bd22083"
+checksum = "d06bfa36ab3ac2be0dee563380147a5b81ba10dd8885d7fbbc9eb574be67d185"
 dependencies = [
  "bitflags 2.8.0",
  "hashbrown 0.15.2",
@@ -2427,13 +2428,13 @@ dependencies = [
 
 [[package]]
 name = "wasmprinter"
-version = "0.221.2"
+version = "0.221.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a80742ff1b9e6d8c231ac7c7247782c6fc5bce503af760bca071811e5fc9ee56"
+checksum = "7343c42a97f2926c7819ff81b64012092ae954c5d83ddd30c9fcdefd97d0b283"
 dependencies = [
  "anyhow",
  "termcolor",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
 ]
 
 [[package]]
@@ -2465,7 +2466,7 @@ dependencies = [
  "smallvec",
  "sptr",
  "target-lexicon",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
  "wasmtime-asm-macros",
  "wasmtime-component-macro",
  "wasmtime-cranelift",
@@ -2552,7 +2553,7 @@ dependencies = [
  "smallvec",
  "target-lexicon",
  "thiserror 1.0.69",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
  "wasmtime-environ",
  "wasmtime-versioned-export-macros",
 ]
@@ -2576,7 +2577,7 @@ dependencies = [
  "smallvec",
  "target-lexicon",
  "wasm-encoder",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
  "wasmprinter",
 ]
 
@@ -2644,7 +2645,7 @@ dependencies = [
  "gimli",
  "object",
  "target-lexicon",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
  "wasmtime-cranelift",
  "wasmtime-environ",
  "winch-codegen",
@@ -2727,7 +2728,7 @@ dependencies = [
  "smallvec",
  "target-lexicon",
  "thiserror 1.0.69",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
  "wasmtime-cranelift",
  "wasmtime-environ",
 ]
@@ -2957,9 +2958,9 @@ dependencies = [
 
 [[package]]
 name = "wit-parser"
-version = "0.221.2"
+version = "0.221.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fbe1538eea6ea5ddbe5defd0dc82539ad7ba751e1631e9185d24a931f0a5adc8"
+checksum = "896112579ed56b4a538b07a3d16e562d101ff6265c46b515ce0c701eef16b2ac"
 dependencies = [
  "anyhow",
  "id-arena",
@@ -2970,7 +2971,7 @@ dependencies = [
  "serde_derive",
  "serde_json",
  "unicode-xid",
- "wasmparser 0.221.2",
+ "wasmparser 0.221.3",
 ]
 
 [[package]]
@@ -3043,7 +3044,16 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1b9b4fd18abc82b8136838da5d50bae7bdea537c574d8dc1a34ed098d6c166f0"
 dependencies = [
  "byteorder",
- "zerocopy-derive",
+ "zerocopy-derive 0.7.35",
+]
+
+[[package]]
+name = "zerocopy"
+version = "0.8.27"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0894878a5fa3edfd6da3f88c4805f4c8558e2b996227a3d864f47fe11e38282c"
+dependencies = [
+ "zerocopy-derive 0.8.27",
 ]
 
 [[package]]
@@ -3057,6 +3067,17 @@ dependencies = [
  "syn",
 ]
 
+[[package]]
+name = "zerocopy-derive"
+version = "0.8.27"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "88d2b8d9c68ad2b9e4340d7832716a4d21a22a1154777ad56ea55c51a9cf3831"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
 [[package]]
 name = "zerofrom"
 version = "0.1.5"

Cargo.toml

@@ -3,6 +3,7 @@ default-members = ["cli"]
 members = [
   "cli",
   "cli/config",
+  "cli/generate",
   "cli/loader",
   "lib",
   "lib/language",
@@ -13,7 +14,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.25.1"
+version = "0.25.10"
 authors = [
   "Max Brunsfeld <maxbrunsfeld@gmail.com>",
   "Amaan Qureshi <amaanq12@gmail.com>",
@@ -59,6 +60,8 @@ missing_errors_doc = "allow"
 missing_panics_doc = "allow"
 module_name_repetitions = "allow"
 multiple_crate_versions = "allow"
+needless_for_each = "allow"
+obfuscated_if_else = "allow"
 option_if_let_else = "allow"
 or_fun_call = "allow"
 range_plus_one = "allow"
@@ -75,6 +78,9 @@ unnecessary_wraps = "allow"
 unused_self = "allow"
 used_underscore_items = "allow"
 
+[workspace.lints.rust]
+mismatched_lifetime_syntaxes = "allow"
+
 [profile.optimize]
 inherits = "release"
 strip = true         # Automatically strip symbols from the binary.
@@ -143,15 +149,16 @@ tempfile = "3.15.0"
 thiserror = "2.0.11"
 tiny_http = "0.12.0"
 toml = "0.8.19"
+topological-sort = "0.2.2"
 unindent = "0.2.3"
 url = { version = "2.5.4", features = ["serde"] }
 walkdir = "2.5.0"
 wasmparser = "0.224.0"
 webbrowser = "1.0.3"
 
-tree-sitter = { version = "0.25.1", path = "./lib" }
-tree-sitter-generate = { version = "0.25.1", path = "./cli/generate" }
-tree-sitter-loader = { version = "0.25.1", path = "./cli/loader" }
-tree-sitter-config = { version = "0.25.1", path = "./cli/config" }
-tree-sitter-highlight = { version = "0.25.1", path = "./highlight" }
-tree-sitter-tags = { version = "0.25.1", path = "./tags" }
+tree-sitter = { version = "0.25.10", path = "./lib" }
+tree-sitter-generate = { version = "0.25.10", path = "./cli/generate" }
+tree-sitter-loader = { version = "0.25.10", path = "./cli/loader" }
+tree-sitter-config = { version = "0.25.10", path = "./cli/config" }
+tree-sitter-highlight = { version = "0.25.10", path = "./highlight" }
+tree-sitter-tags = { version = "0.25.10", path = "./tags" }

Makefile

@@ -2,7 +2,7 @@ ifeq ($(OS),Windows_NT)
 $(error Windows is not supported)
 endif
 
-VERSION := 0.25.1
+VERSION := 0.25.10
 DESCRIPTION := An incremental parsing system for programming tools
 HOMEPAGE_URL := https://tree-sitter.github.io/tree-sitter/
 
@@ -27,7 +27,7 @@ OBJ := $(SRC:.c=.o)
 ARFLAGS := rcs
 CFLAGS ?= -O3 -Wall -Wextra -Wshadow -Wpedantic -Werror=incompatible-pointer-types
 override CFLAGS += -std=c11 -fPIC -fvisibility=hidden
-override CFLAGS += -D_POSIX_C_SOURCE=200112L -D_DEFAULT_SOURCE
+override CFLAGS += -D_POSIX_C_SOURCE=200112L -D_DEFAULT_SOURCE -D_DARWIN_C_SOURCE
 override CFLAGS += -Ilib/src -Ilib/src/wasm -Ilib/include
 
 # ABI versioning
@@ -106,15 +106,15 @@ test-wasm:
 lint:
 	cargo update --workspace --locked --quiet
 	cargo check --workspace --all-targets
-	cargo +nightly fmt --all --check
-	cargo +nightly clippy --workspace --all-targets -- -D warnings
+	cargo fmt --all --check
+	cargo clippy --workspace --all-targets -- -D warnings
 
 lint-web:
 	npm --prefix lib/binding_web ci
 	npm --prefix lib/binding_web run lint
 
 format:
-	cargo +nightly fmt --all
+	cargo fmt --all
 
 changelog:
 	@git-cliff --config .github/cliff.toml --prepend CHANGELOG.md --latest --github-token $(shell gh auth token)

Package.swift

@@ -27,6 +27,7 @@ let package = Package(
                         .headerSearchPath("src"),
                         .define("_POSIX_C_SOURCE", to: "200112L"),
                         .define("_DEFAULT_SOURCE"),
+                        .define("_DARWIN_C_SOURCE"),
                 ]),
     ],
     cLanguageStandard: .c11

build.zig

@@ -1,116 +1,122 @@
 const std = @import("std");
 
 pub fn build(b: *std.Build) !void {
-  const target = b.standardTargetOptions(.{});
-  const optimize = b.standardOptimizeOption(.{});
+    const target = b.standardTargetOptions(.{});
+    const optimize = b.standardOptimizeOption(.{});
 
-  const wasm = b.option(bool, "enable-wasm", "Enable Wasm support") orelse false;
-  const shared = b.option(bool, "build-shared", "Build a shared library") orelse false;
-  const amalgamated = b.option(bool, "amalgamated", "Build using an amalgamated source") orelse false;
+    const wasm = b.option(bool, "enable-wasm", "Enable Wasm support") orelse false;
+    const shared = b.option(bool, "build-shared", "Build a shared library") orelse false;
+    const amalgamated = b.option(bool, "amalgamated", "Build using an amalgamated source") orelse false;
 
-  const lib: *std.Build.Step.Compile = if (!shared) b.addStaticLibrary(.{
-    .name = "tree-sitter",
-    .target = target,
-    .optimize = optimize,
-    .link_libc = true,
-  }) else b.addSharedLibrary(.{
-    .name = "tree-sitter",
-    .pic = true,
-    .target = target,
-    .optimize = optimize,
-    .link_libc = true,
-  });
-
-  if (amalgamated) {
-    lib.addCSourceFile(.{
-      .file = b.path("lib/src/lib.c"),
-      .flags = &.{"-std=c11"},
+    const lib: *std.Build.Step.Compile = b.addLibrary(.{
+        .name = "tree-sitter",
+        .linkage = if (shared) .dynamic else .static,
+        .root_module = b.createModule(.{
+            .target = target,
+            .optimize = optimize,
+            .link_libc = true,
+            .pic = if (shared) true else null,
+        }),
     });
-  } else {
-    lib.addCSourceFiles(.{
-      .root = b.path("lib/src"),
-      .files = try findSourceFiles(b),
-      .flags = &.{"-std=c11"},
-    });
-  }
 
-  lib.addIncludePath(b.path("lib/include"));
-  lib.addIncludePath(b.path("lib/src"));
-  lib.addIncludePath(b.path("lib/src/wasm"));
-
-  lib.root_module.addCMacro("_POSIX_C_SOURCE", "200112L");
-  lib.root_module.addCMacro("_DEFAULT_SOURCE", "");
-
-  if (wasm) {
-    if (b.lazyDependency(wasmtimeDep(target.result), .{})) |wasmtime| {
-      lib.root_module.addCMacro("TREE_SITTER_FEATURE_WASM", "");
-      lib.addSystemIncludePath(wasmtime.path("include"));
-      lib.addLibraryPath(wasmtime.path("lib"));
-      lib.linkSystemLibrary("wasmtime");
+    if (amalgamated) {
+        lib.addCSourceFile(.{
+            .file = b.path("lib/src/lib.c"),
+            .flags = &.{"-std=c11"},
+        });
+    } else {
+        const files = try findSourceFiles(b);
+        defer b.allocator.free(files);
+        lib.addCSourceFiles(.{
+            .root = b.path("lib/src"),
+            .files = files,
+            .flags = &.{"-std=c11"},
+        });
     }
-  }
 
-  lib.installHeadersDirectory(b.path("lib/include"), ".", .{});
+    lib.addIncludePath(b.path("lib/include"));
+    lib.addIncludePath(b.path("lib/src"));
+    lib.addIncludePath(b.path("lib/src/wasm"));
 
-  b.installArtifact(lib);
+    lib.root_module.addCMacro("_POSIX_C_SOURCE", "200112L");
+    lib.root_module.addCMacro("_DEFAULT_SOURCE", "");
+    lib.root_module.addCMacro("_DARWIN_C_SOURCE", "");
+
+    if (wasm) {
+        if (b.lazyDependency(wasmtimeDep(target.result), .{})) |wasmtime| {
+            lib.root_module.addCMacro("TREE_SITTER_FEATURE_WASM", "");
+            lib.addSystemIncludePath(wasmtime.path("include"));
+            lib.addLibraryPath(wasmtime.path("lib"));
+            if (shared) lib.linkSystemLibrary("wasmtime");
+        }
+    }
+
+    lib.installHeadersDirectory(b.path("lib/include"), ".", .{});
+
+    b.installArtifact(lib);
 }
 
-fn wasmtimeDep(target: std.Target) []const u8 {
-  const arch = target.cpu.arch;
-  const os = target.os.tag;
-  const abi = target.abi;
-  return switch (os) {
-    .linux => switch (arch) {
-      .x86_64 => switch (abi) {
-        .gnu => "wasmtime_c_api_x86_64_linux",
-        .musl => "wasmtime_c_api_x86_64_musl",
-        .android => "wasmtime_c_api_x86_64_android",
-        else => null
-      },
-      .aarch64 => switch (abi) {
-        .gnu => "wasmtime_c_api_aarch64_linux",
-        .android => "wasmtime_c_api_aarch64_android",
-        else => null
-      },
-      .s390x => "wasmtime_c_api_s390x_linux",
-      .riscv64 => "wasmtime_c_api_riscv64gc_linux",
-      else => null
-    },
-    .windows => switch (arch) {
-      .x86_64 => switch (abi) {
-        .gnu => "wasmtime_c_api_x86_64_mingw",
-        .msvc => "wasmtime_c_api_x86_64_windows",
-        else => null
-      },
-      else => null
-    },
-    .macos => switch (arch) {
-      .x86_64 => "wasmtime_c_api_x86_64_macos",
-      .aarch64 => "wasmtime_c_api_aarch64_macos",
-      else => null
-    },
-    else => null
-  } orelse std.debug.panic(
-    "Unsupported target for wasmtime: {s}-{s}-{s}",
-    .{ @tagName(arch), @tagName(os), @tagName(abi) }
-  );
+/// Get the name of the wasmtime dependency for this target.
+pub fn wasmtimeDep(target: std.Target) []const u8 {
+    const arch = target.cpu.arch;
+    const os = target.os.tag;
+    const abi = target.abi;
+    return switch (os) {
+        .linux => switch (arch) {
+            .x86_64 => switch (abi) {
+                .gnu => "wasmtime_c_api_x86_64_linux",
+                .musl => "wasmtime_c_api_x86_64_musl",
+                .android => "wasmtime_c_api_x86_64_android",
+                else => null,
+            },
+            .aarch64 => switch (abi) {
+                .gnu => "wasmtime_c_api_aarch64_linux",
+                .android => "wasmtime_c_api_aarch64_android",
+                else => null,
+            },
+            .s390x => "wasmtime_c_api_s390x_linux",
+            .riscv64 => "wasmtime_c_api_riscv64gc_linux",
+            else => null,
+        },
+        .windows => switch (arch) {
+            .x86_64 => switch (abi) {
+                .gnu => "wasmtime_c_api_x86_64_mingw",
+                .msvc => "wasmtime_c_api_x86_64_windows",
+                else => null,
+            },
+            .aarch64 => switch (abi) {
+                .msvc => "wasmtime_c_api_aarch64_windows",
+                else => null,
+            },
+            else => null,
+        },
+        .macos => switch (arch) {
+            .x86_64 => "wasmtime_c_api_x86_64_macos",
+            .aarch64 => "wasmtime_c_api_aarch64_macos",
+            else => null,
+        },
+        else => null,
+    } orelse std.debug.panic(
+        "Unsupported target for wasmtime: {s}-{s}-{s}",
+        .{ @tagName(arch), @tagName(os), @tagName(abi) },
+    );
 }
 
 fn findSourceFiles(b: *std.Build) ![]const []const u8 {
-  var sources = std.ArrayList([]const u8).init(b.allocator);
+    var sources: std.ArrayListUnmanaged([]const u8) = .empty;
 
-  var dir = try b.build_root.handle.openDir("lib/src", .{ .iterate = true });
-  var iter = dir.iterate();
-  defer dir.close();
+    var dir = try b.build_root.handle.openDir("lib/src", .{ .iterate = true });
+    var iter = dir.iterate();
+    defer dir.close();
 
-  while (try iter.next()) |entry| {
-    if (entry.kind != .file) continue;
-    const file = entry.name;
-    const ext = std.fs.path.extension(file);
-    if (std.mem.eql(u8, ext, ".c") and !std.mem.eql(u8, file, "lib.c")) {
-      try sources.append(b.dupe(file));
+    while (try iter.next()) |entry| {
+        if (entry.kind != .file) continue;
+        const file = entry.name;
+        const ext = std.fs.path.extension(file);
+        if (std.mem.eql(u8, ext, ".c") and !std.mem.eql(u8, file, "lib.c")) {
+            try sources.append(b.allocator, b.dupe(file));
+        }
     }
-  }
 
-  return sources.items;
+    return sources.toOwnedSlice(b.allocator);
 }

build.zig.zon

@@ -1,69 +1,76 @@
 .{
-  .name = "tree-sitter",
-  .version = "0.25.1",
-  .paths = .{
-    "build.zig",
-    "build.zig.zon",
-    "lib/src",
-    "lib/include",
-    "README.md",
-    "LICENSE",
-  },
-  .dependencies = .{
-    .wasmtime_c_api_aarch64_android = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-android-c-api.tar.xz",
-      .hash = "12204c77979ad8291c6e395d695a824fb053ffdfeb2cc21de95fffb09f77d77188d1",
-      .lazy = true,
+    .name = .tree_sitter,
+    .fingerprint = 0x841224b447ac0d4f,
+    .version = "0.25.9",
+    .minimum_zig_version = "0.14.1",
+    .paths = .{
+        "build.zig",
+        "build.zig.zon",
+        "lib/src",
+        "lib/include",
+        "README.md",
+        "LICENSE",
     },
-    .wasmtime_c_api_aarch64_linux = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-linux-c-api.tar.xz",
-      .hash = "12203a8e3d823490186fb1e230d54f575148713088e914926305ee5678790b731bba",
-      .lazy = true,
+    .dependencies = .{
+        .wasmtime_c_api_aarch64_android = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-android-c-api.tar.xz",
+            .hash = "N-V-__8AAC3KCQZMd5ea2CkcbjldaVqCT7BT_9_rLMId6V__",
+            .lazy = true,
+        },
+        .wasmtime_c_api_aarch64_linux = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-linux-c-api.tar.xz",
+            .hash = "N-V-__8AAGUY3gU6jj2CNJAYb7HiMNVPV1FIcTCI6RSSYwXu",
+            .lazy = true,
+        },
+        .wasmtime_c_api_aarch64_macos = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-macos-c-api.tar.xz",
+            .hash = "N-V-__8AAM1GMARD6LGQebhVsSZ0uePUoo3Fw5nEO2L764vf",
+            .lazy = true,
+        },
+        .wasmtime_c_api_aarch64_windows = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-windows-c-api.zip",
+            .hash = "N-V-__8AAH8a_wQ7oAeVVsaJcoOZhKTMkHIBc_XjDyLlHp2x",
+            .lazy = true,
+        },
+        .wasmtime_c_api_riscv64gc_linux = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-riscv64gc-linux-c-api.tar.xz",
+            .hash = "N-V-__8AAN2cuQadBwMc8zJxv0sMY99Ae1Nc1dZcZAK9b4DZ",
+            .lazy = true,
+        },
+        .wasmtime_c_api_s390x_linux = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-s390x-linux-c-api.tar.xz",
+            .hash = "N-V-__8AAPevngYz99mwT0KQY9my2ax1p6APzgLEJeV4II9U",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_android = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-android-c-api.tar.xz",
+            .hash = "N-V-__8AABHIEgaTyzPfjgnnCy0dwJiXoDiJFblCkYOJsQvy",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_linux = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-linux-c-api.tar.xz",
+            .hash = "N-V-__8AALUN5AWSEDRulL9u-OJJ-l0_GoT5UFDtGWZayEIq",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_macos = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-macos-c-api.tar.xz",
+            .hash = "N-V-__8AANUeXwSPh13TqJCSSFdi87GEcHs8zK6FqE4v_TjB",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_mingw = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-mingw-c-api.zip",
+            .hash = "N-V-__8AALundgW-p1ffOnd7bsYyL8SY5OziDUZu7cXio2EL",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_musl = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-musl-c-api.tar.xz",
+            .hash = "N-V-__8AALMZ5wXJWW5qY-3MMjTAYR0MusckvzCsmg-69ALH",
+            .lazy = true,
+        },
+        .wasmtime_c_api_x86_64_windows = .{
+            .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-windows-c-api.zip",
+            .hash = "N-V-__8AAG-uVQVEDMsB1ymJzxpHcoiXo1_I3TFnPM5Zjy1i",
+            .lazy = true,
+        },
     },
-    .wasmtime_c_api_aarch64_macos = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-aarch64-macos-c-api.tar.xz",
-      .hash = "122043e8b19079b855b12674b9e3d4a28dc5c399c43b62fbeb8bdf0fdb4ef2d1d38c",
-      .lazy = true,
-    },
-    .wasmtime_c_api_riscv64gc_linux = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-riscv64gc-linux-c-api.tar.xz",
-      .hash = "12209d07031cf33271bf4b0c63df407b535cd5d65c6402bd6f80d99de439d6feb89b",
-      .lazy = true,
-    },
-    .wasmtime_c_api_s390x_linux = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-s390x-linux-c-api.tar.xz",
-      .hash = "122033f7d9b04f429063d9b2d9ac75a7a00fce02c425e578208f54ddc40edaa1e355",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_android = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-android-c-api.tar.xz",
-      .hash = "122093cb33df8e09e70b2d1dc09897a0388915b942918389b10bf23f9684bdb6f047",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_linux = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-linux-c-api.tar.xz",
-      .hash = "12209210346e94bf6ef8e249fa5d3f1a84f95050ed19665ac8422a15b5f2246d83af",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_macos = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-macos-c-api.tar.xz",
-      .hash = "12208f875dd3a89092485762f3b184707b3cccae85a84e2ffd38c138cc3a3fd90447",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_mingw = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-mingw-c-api.zip",
-      .hash = "1220bea757df3a777b6ec6322fc498e4ece20d466eedc5e2a3610b338849553cd94d",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_musl = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-musl-c-api.tar.xz",
-      .hash = "1220c9596e6a63edcc3234c0611d0cbac724bf30ac9a0fbaf402c7da649b278b1322",
-      .lazy = true,
-    },
-    .wasmtime_c_api_x86_64_windows = .{
-      .url = "https://github.com/bytecodealliance/wasmtime/releases/download/v29.0.1/wasmtime-v29.0.1-x86_64-windows-c-api.zip",
-      .hash = "1220440ccb01d72989cf1a47728897a35fc8dd31673cce598f2d62c58e2c3228b0ed",
-      .lazy = true,
-    },
-  }
 }

cli/Cargo.toml

@@ -59,6 +59,7 @@ similar.workspace = true
 smallbitvec.workspace = true
 streaming-iterator.workspace = true
 tiny_http.workspace = true
+topological-sort.workspace = true
 url.workspace = true
 walkdir.workspace = true
 wasmparser.workspace = true

cli/benches/benchmark.rs

@@ -112,7 +112,7 @@ fn main() {
 
             parse(path, max_path_length, |source| {
                 Query::new(&language, str::from_utf8(source).unwrap())
-                    .with_context(|| format!("Query file path: {path:?}"))
+                    .with_context(|| format!("Query file path: {}", path.display()))
                     .expect("Failed to parse query");
             });
         }
@@ -201,7 +201,7 @@ fn parse(path: &Path, max_path_length: usize, mut action: impl FnMut(&[u8])) ->
     );
 
     let source_code = fs::read(path)
-        .with_context(|| format!("Failed to read {path:?}"))
+        .with_context(|| format!("Failed to read {}", path.display()))
         .unwrap();
     let time = Instant::now();
     for _ in 0..*REPETITION_COUNT {
@@ -221,6 +221,6 @@ fn get_language(path: &Path) -> Language {
     let src_path = GRAMMARS_DIR.join(path).join("src");
     TEST_LOADER
         .load_language_at_path(CompileConfig::new(&src_path, None, None))
-        .with_context(|| format!("Failed to load language at path {src_path:?}"))
+        .with_context(|| format!("Failed to load language at path {}", src_path.display()))
         .unwrap()
 }

cli/build.rs

@@ -60,8 +60,6 @@ fn web_playground_files_present() -> bool {
     paths.iter().all(|p| Path::new(p).exists())
 }
 
-// When updating this function, don't forget to also update generate/build.rs which has a
-// near-identical function.
 fn read_git_sha() -> Option<String> {
     let crate_path = PathBuf::from(env::var("CARGO_MANIFEST_DIR").unwrap());
 

cli/eslint/package.json

@@ -4,7 +4,8 @@
   "description": "Eslint configuration for Tree-sitter grammar files",
   "repository": {
     "type": "git",
-    "url": "git+https://github.com/tree-sitter/tree-sitter.git"
+    "url": "git+https://github.com/tree-sitter/tree-sitter.git",
+    "directory": "crates/cli/eslint"
   },
   "license": "MIT",
   "author": "Amaan Qureshi <amaanq12@gmail.com>",

cli/generate/Cargo.toml

@@ -29,6 +29,9 @@ serde.workspace = true
 serde_json.workspace = true
 smallbitvec.workspace = true
 thiserror.workspace = true
-url.workspace = true
+topological-sort.workspace = true
 
 tree-sitter.workspace = true
+
+[target.'cfg(windows)'.dependencies]
+url.workspace = true

cli/generate/build.rs

@@ -1,32 +0,0 @@
-use std::{env, path::PathBuf, process::Command};
-
-fn main() {
-    if let Some(git_sha) = read_git_sha() {
-        println!("cargo:rustc-env=BUILD_SHA={git_sha}");
-    }
-}
-
-// This is copied from the build.rs in parent directory. This should be updated if the
-// parent build.rs gets fixes.
-fn read_git_sha() -> Option<String> {
-    let crate_path = PathBuf::from(env::var("CARGO_MANIFEST_DIR").unwrap());
-
-    if !crate_path
-        .parent()?
-        .parent()
-        .is_some_and(|p| p.join(".git").exists())
-    {
-        return None;
-    }
-
-    Command::new("git")
-        .args(["rev-parse", "HEAD"])
-        .current_dir(crate_path)
-        .output()
-        .map_or(None, |output| {
-            if !output.status.success() {
-                return None;
-            }
-            Some(String::from_utf8_lossy(&output.stdout).to_string())
-        })
-}

cli/generate/src/build_tables/build_parse_table.rs

@@ -312,6 +312,12 @@ impl<'a> ParseTableBuilder<'a> {
             }
         }
 
+        let non_terminal_sets_len = non_terminal_extra_item_sets_by_first_terminal.len();
+        self.non_terminal_extra_states
+            .reserve(non_terminal_sets_len);
+        self.parse_state_info_by_id.reserve(non_terminal_sets_len);
+        self.parse_table.states.reserve(non_terminal_sets_len);
+        self.parse_state_queue.reserve(non_terminal_sets_len);
         // Add a state for each starting terminal of a non-terminal extra rule.
         for (terminal, item_set) in non_terminal_extra_item_sets_by_first_terminal {
             if terminal.is_non_terminal() {
@@ -320,9 +326,10 @@ impl<'a> ParseTableBuilder<'a> {
                 ))?;
             }
 
-            self.non_terminal_extra_states
-                .push((terminal, self.parse_table.states.len()));
-            self.add_parse_state(&Vec::new(), &Vec::new(), item_set);
+            // Add the parse state, and *then* push the terminal and the state id into the
+            // list of nonterminal extra states
+            let state_id = self.add_parse_state(&Vec::new(), &Vec::new(), item_set);
+            self.non_terminal_extra_states.push((terminal, state_id));
         }
 
         while let Some(entry) = self.parse_state_queue.pop_front() {
@@ -908,7 +915,7 @@ impl<'a> ParseTableBuilder<'a> {
 
         let get_rule_names = |items: &[&ParseItem]| -> Vec<String> {
             let mut last_rule_id = None;
-            let mut result = Vec::new();
+            let mut result = Vec::with_capacity(items.len());
             for item in items {
                 if last_rule_id == Some(item.variable_index) {
                     continue;

cli/generate/src/dsl.js

@@ -529,7 +529,7 @@ globalThis.optional = optional;
 globalThis.prec = prec;
 globalThis.repeat = repeat;
 globalThis.repeat1 = repeat1;
-global.reserved = reserved;
+globalThis.reserved = reserved;
 globalThis.seq = seq;
 globalThis.sym = sym;
 globalThis.token = token;

cli/generate/src/lib.rs

@@ -27,7 +27,7 @@ mod tables;
 use build_tables::build_tables;
 pub use build_tables::ParseTableBuilderError;
 use grammars::InputGrammar;
-pub use node_types::VariableInfoError;
+pub use node_types::{SuperTypeCycleError, VariableInfoError};
 use parse_grammar::parse_grammar;
 pub use parse_grammar::ParseGrammarError;
 use prepare_grammar::prepare_grammar;
@@ -70,6 +70,8 @@ pub enum GenerateError {
     BuildTables(#[from] ParseTableBuilderError),
     #[error(transparent)]
     ParseVersion(#[from] ParseVersionError),
+    #[error(transparent)]
+    SuperTypeCycle(#[from] SuperTypeCycleError),
 }
 
 impl From<std::io::Error> for GenerateError {
@@ -183,7 +185,8 @@ pub fn generate_parser_in_directory(
     if grammar_path.file_name().unwrap() != "grammar.json" {
         fs::write(src_path.join("grammar.json"), &grammar_json).map_err(|e| {
             GenerateError::IO(format!(
-                "Failed to write grammar.json to {src_path:?} -- {e}"
+                "Failed to write grammar.json to {} -- {e}",
+                src_path.display()
             ))
         })?;
     }
@@ -249,7 +252,7 @@ fn generate_parser_for_grammar_with_opts(
         &lexical_grammar,
         &simple_aliases,
         &variable_info,
-    );
+    )?;
     let supertype_symbol_map =
         node_types::get_supertype_symbol_map(&syntax_grammar, &simple_aliases, &variable_info);
     let tables = build_tables(

cli/generate/src/node_types.rs

@@ -1,7 +1,4 @@
-use std::{
-    cmp::Ordering,
-    collections::{BTreeMap, HashMap, HashSet},
-};
+use std::collections::{BTreeMap, HashMap, HashSet};
 
 use anyhow::Result;
 use serde::Serialize;
@@ -444,12 +441,33 @@ pub fn get_supertype_symbol_map(
     supertype_symbol_map
 }
 
+pub type SuperTypeCycleResult<T> = Result<T, SuperTypeCycleError>;
+
+#[derive(Debug, Error, Serialize)]
+pub struct SuperTypeCycleError {
+    items: Vec<String>,
+}
+
+impl std::fmt::Display for SuperTypeCycleError {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Dependency cycle detected in node types:")?;
+        for (i, item) in self.items.iter().enumerate() {
+            write!(f, " {item}")?;
+            if i < self.items.len() - 1 {
+                write!(f, ",")?;
+            }
+        }
+
+        Ok(())
+    }
+}
+
 pub fn generate_node_types_json(
     syntax_grammar: &SyntaxGrammar,
     lexical_grammar: &LexicalGrammar,
     default_aliases: &AliasMap,
     variable_info: &[VariableInfo],
-) -> Vec<NodeInfoJSON> {
+) -> SuperTypeCycleResult<Vec<NodeInfoJSON>> {
     let mut node_types_json = BTreeMap::new();
 
     let child_type_to_node_type = |child_type: &ChildType| match child_type {
@@ -507,6 +525,31 @@ pub fn generate_node_types_json(
 
     let aliases_by_symbol = get_aliases_by_symbol(syntax_grammar, default_aliases);
 
+    let empty = HashSet::new();
+    let extra_names = syntax_grammar
+        .extra_symbols
+        .iter()
+        .flat_map(|symbol| {
+            aliases_by_symbol
+                .get(symbol)
+                .unwrap_or(&empty)
+                .iter()
+                .map(|alias| {
+                    alias.as_ref().map_or(
+                        match symbol.kind {
+                            SymbolType::NonTerminal => &syntax_grammar.variables[symbol.index].name,
+                            SymbolType::Terminal => &lexical_grammar.variables[symbol.index].name,
+                            SymbolType::External => {
+                                &syntax_grammar.external_tokens[symbol.index].name
+                            }
+                            _ => unreachable!(),
+                        },
+                        |alias| &alias.value,
+                    )
+                })
+        })
+        .collect::<HashSet<_>>();
+
     let mut subtype_map = Vec::new();
     for (i, info) in variable_info.iter().enumerate() {
         let symbol = Symbol::non_terminal(i);
@@ -519,7 +562,7 @@ pub fn generate_node_types_json(
                         kind: variable.name.clone(),
                         named: true,
                         root: false,
-                        extra: false,
+                        extra: extra_names.contains(&variable.name),
                         fields: None,
                         children: None,
                         subtypes: None,
@@ -563,7 +606,7 @@ pub fn generate_node_types_json(
                         kind: kind.clone(),
                         named: is_named,
                         root: i == 0,
-                        extra: false,
+                        extra: extra_names.contains(&kind),
                         fields: Some(BTreeMap::new()),
                         children: None,
                         subtypes: None,
@@ -602,15 +645,33 @@ pub fn generate_node_types_json(
         }
     }
 
-    // Sort the subtype map so that subtypes are listed before their supertypes.
-    subtype_map.sort_by(|a, b| {
-        if b.1.contains(&a.0) {
-            Ordering::Less
-        } else if a.1.contains(&b.0) {
-            Ordering::Greater
-        } else {
-            Ordering::Equal
+    // Sort the subtype map topologically so that subtypes are listed before their supertypes.
+    let mut sorted_kinds = Vec::with_capacity(subtype_map.len());
+    let mut top_sort = topological_sort::TopologicalSort::<String>::new();
+    for (supertype, subtypes) in &subtype_map {
+        for subtype in subtypes {
+            top_sort.add_dependency(subtype.kind.clone(), supertype.kind.clone());
         }
+    }
+    loop {
+        let mut next_kinds = top_sort.pop_all();
+        match (next_kinds.is_empty(), top_sort.is_empty()) {
+            (true, true) => break,
+            (true, false) => {
+                let mut items = top_sort.collect::<Vec<String>>();
+                items.sort();
+                return Err(SuperTypeCycleError { items });
+            }
+            (false, _) => {
+                next_kinds.sort();
+                sorted_kinds.extend(next_kinds);
+            }
+        }
+    }
+    subtype_map.sort_by(|a, b| {
+        let a_idx = sorted_kinds.iter().position(|n| n.eq(&a.0.kind)).unwrap();
+        let b_idx = sorted_kinds.iter().position(|n| n.eq(&b.0.kind)).unwrap();
+        a_idx.cmp(&b_idx)
     });
 
     for node_type_json in node_types_json.values_mut() {
@@ -634,7 +695,6 @@ pub fn generate_node_types_json(
 
     let mut anonymous_node_types = Vec::new();
 
-    let empty = HashSet::new();
     let regular_tokens = lexical_grammar
         .variables
         .iter()
@@ -668,29 +728,6 @@ pub fn generate_node_types_json(
                         })
                     })
             });
-    let extra_names = syntax_grammar
-        .extra_symbols
-        .iter()
-        .flat_map(|symbol| {
-            aliases_by_symbol
-                .get(symbol)
-                .unwrap_or(&empty)
-                .iter()
-                .map(|alias| {
-                    alias.as_ref().map_or(
-                        match symbol.kind {
-                            SymbolType::NonTerminal => &syntax_grammar.variables[symbol.index].name,
-                            SymbolType::Terminal => &lexical_grammar.variables[symbol.index].name,
-                            SymbolType::External => {
-                                &syntax_grammar.external_tokens[symbol.index].name
-                            }
-                            _ => unreachable!(),
-                        },
-                        |alias| &alias.value,
-                    )
-                })
-        })
-        .collect::<HashSet<_>>();
 
     for (name, kind) in regular_tokens.chain(external_tokens) {
         match kind {
@@ -743,7 +780,7 @@ pub fn generate_node_types_json(
             .then_with(|| a.kind.cmp(&b.kind))
     });
     result.dedup();
-    result
+    Ok(result)
 }
 
 fn process_supertypes(info: &mut FieldInfoJSON, subtype_map: &[(NodeTypeJSON, Vec<NodeTypeJSON>)]) {
@@ -829,7 +866,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(node_types.len(), 3);
 
@@ -918,7 +956,9 @@ mod tests {
                 },
                 // This rule is not reachable from the start symbol, but
                 // it is reachable from the 'extra_symbols' so it
-                // should be present in the node_types
+                // should be present in the node_types.
+                // But because it's only a literal, it will get replaced by
+                // a lexical variable.
                 Variable {
                     name: "v3".to_string(),
                     kind: VariableType::Named,
@@ -926,7 +966,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(node_types.len(), 4);
 
@@ -1007,6 +1048,118 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_node_types_deeper_extras() {
+        let node_types = get_node_types(&InputGrammar {
+            extra_symbols: vec![Rule::named("v3")],
+            variables: vec![
+                Variable {
+                    name: "v1".to_string(),
+                    kind: VariableType::Named,
+                    rule: Rule::seq(vec![
+                        Rule::field("f1".to_string(), Rule::named("v2")),
+                        Rule::field("f2".to_string(), Rule::string(";")),
+                    ]),
+                },
+                Variable {
+                    name: "v2".to_string(),
+                    kind: VariableType::Named,
+                    rule: Rule::string("x"),
+                },
+                // This rule is not reachable from the start symbol, but
+                // it is reachable from the 'extra_symbols' so it
+                // should be present in the node_types.
+                // Because it is not just a literal, it won't get replaced
+                // by a lexical variable.
+                Variable {
+                    name: "v3".to_string(),
+                    kind: VariableType::Named,
+                    rule: Rule::seq(vec![Rule::string("y"), Rule::repeat(Rule::string("z"))]),
+                },
+            ],
+            ..Default::default()
+        })
+        .unwrap();
+
+        assert_eq!(node_types.len(), 6);
+
+        assert_eq!(
+            node_types[0],
+            NodeInfoJSON {
+                kind: "v1".to_string(),
+                named: true,
+                root: true,
+                extra: false,
+                subtypes: None,
+                children: None,
+                fields: Some(
+                    vec![
+                        (
+                            "f1".to_string(),
+                            FieldInfoJSON {
+                                multiple: false,
+                                required: true,
+                                types: vec![NodeTypeJSON {
+                                    kind: "v2".to_string(),
+                                    named: true,
+                                }]
+                            }
+                        ),
+                        (
+                            "f2".to_string(),
+                            FieldInfoJSON {
+                                multiple: false,
+                                required: true,
+                                types: vec![NodeTypeJSON {
+                                    kind: ";".to_string(),
+                                    named: false,
+                                }]
+                            }
+                        ),
+                    ]
+                    .into_iter()
+                    .collect()
+                )
+            }
+        );
+        assert_eq!(
+            node_types[1],
+            NodeInfoJSON {
+                kind: "v3".to_string(),
+                named: true,
+                root: false,
+                extra: true,
+                subtypes: None,
+                children: None,
+                fields: Some(BTreeMap::default())
+            }
+        );
+        assert_eq!(
+            node_types[2],
+            NodeInfoJSON {
+                kind: ";".to_string(),
+                named: false,
+                root: false,
+                extra: false,
+                subtypes: None,
+                children: None,
+                fields: None
+            }
+        );
+        assert_eq!(
+            node_types[3],
+            NodeInfoJSON {
+                kind: "v2".to_string(),
+                named: true,
+                root: false,
+                extra: false,
+                subtypes: None,
+                children: None,
+                fields: None
+            }
+        );
+    }
+
     #[test]
     fn test_node_types_with_supertypes() {
         let node_types = get_node_types(&InputGrammar {
@@ -1038,7 +1191,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types[0],
@@ -1127,7 +1281,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types[0],
@@ -1212,7 +1367,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types[0],
@@ -1286,7 +1442,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(node_types.iter().find(|t| t.kind == "foo_identifier"), None);
         assert_eq!(
@@ -1342,7 +1499,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types[0],
@@ -1391,7 +1549,8 @@ mod tests {
                 ]),
             }],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types,
@@ -1439,7 +1598,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             &node_types
@@ -1558,7 +1718,8 @@ mod tests {
                 },
             ],
             ..Default::default()
-        });
+        })
+        .unwrap();
 
         assert_eq!(
             node_types.iter().map(|n| &n.kind).collect::<Vec<_>>(),
@@ -1885,7 +2046,7 @@ mod tests {
         );
     }
 
-    fn get_node_types(grammar: &InputGrammar) -> Vec<NodeInfoJSON> {
+    fn get_node_types(grammar: &InputGrammar) -> SuperTypeCycleResult<Vec<NodeInfoJSON>> {
         let (syntax_grammar, lexical_grammar, _, default_aliases) =
             prepare_grammar(grammar).unwrap();
         let variable_info =

cli/generate/src/parse_grammar.rs

@@ -1,6 +1,7 @@
 use std::collections::HashSet;
 
 use anyhow::Result;
+use regex::Regex;
 use serde::{Deserialize, Serialize};
 use serde_json::{Map, Value};
 use thiserror::Error;
@@ -238,13 +239,14 @@ pub(crate) fn parse_grammar(input: &str) -> ParseGrammarResult<InputGrammar> {
     let mut in_progress = HashSet::new();
 
     for (name, rule) in &rules {
-        if !variable_is_used(
-            &rules,
-            &extra_symbols,
-            &external_tokens,
-            name,
-            &mut in_progress,
-        ) && grammar_json.word.as_ref().is_none_or(|w| w != name)
+        if grammar_json.word.as_ref().is_none_or(|w| w != name)
+            && !variable_is_used(
+                &rules,
+                &extra_symbols,
+                &external_tokens,
+                name,
+                &mut in_progress,
+            )
         {
             grammar_json.conflicts.retain(|r| !r.contains(name));
             grammar_json.supertypes.retain(|r| r != name);
@@ -261,6 +263,27 @@ pub(crate) fn parse_grammar(input: &str) -> ParseGrammarResult<InputGrammar> {
             });
             continue;
         }
+
+        if extra_symbols
+            .iter()
+            .any(|r| rule_is_referenced(r, name, false))
+        {
+            let inner_rule = if let Rule::Metadata { rule, .. } = rule {
+                rule
+            } else {
+                rule
+            };
+            let matches_empty = match inner_rule {
+                Rule::String(rule_str) => rule_str.is_empty(),
+                Rule::Pattern(ref value, _) => Regex::new(value)
+                    .map(|reg| reg.is_match(""))
+                    .unwrap_or(false),
+                _ => false,
+            };
+            if matches_empty {
+                eprintln!("Warning: Named extra rule `{name}` matches the empty string. Inline this to avoid infinite loops while parsing.");
+            }
+        }
         variables.push(Variable {
             name: name.clone(),
             kind: VariableType::Named,
@@ -272,12 +295,11 @@ pub(crate) fn parse_grammar(input: &str) -> ParseGrammarResult<InputGrammar> {
         .reserved
         .into_iter()
         .map(|(name, rule_values)| {
-            let mut reserved_words = Vec::new();
-
             let Value::Array(rule_values) = rule_values else {
                 Err(ParseGrammarError::InvalidReservedWordSet)?
             };
 
+            let mut reserved_words = Vec::with_capacity(rule_values.len());
             for value in rule_values {
                 reserved_words.push(parse_rule(serde_json::from_value(value)?, false)?);
             }

cli/generate/src/prepare_grammar/expand_tokens.rs

@@ -90,7 +90,7 @@ pub fn expand_tokens(mut grammar: ExtractedLexicalGrammar) -> ExpandTokensResult
         Rule::repeat(Rule::choice(grammar.separators))
     };
 
-    let mut variables = Vec::new();
+    let mut variables = Vec::with_capacity(grammar.variables.len());
     for (i, variable) in grammar.variables.into_iter().enumerate() {
         if variable.rule.is_empty() {
             Err(ExpandTokensError::EmptyString(variable.name.clone()))?;
@@ -195,7 +195,7 @@ impl NfaBuilder {
                 Ok(!s.is_empty())
             }
             Rule::Choice(elements) => {
-                let mut alternative_state_ids = Vec::new();
+                let mut alternative_state_ids = Vec::with_capacity(elements.len());
                 for element in elements {
                     if self.expand_rule(element, next_state_id)? {
                         alternative_state_ids.push(self.nfa.last_state_id());
@@ -338,7 +338,7 @@ impl NfaBuilder {
                 Ok(result)
             }
             HirKind::Alternation(alternations) => {
-                let mut alternative_state_ids = Vec::new();
+                let mut alternative_state_ids = Vec::with_capacity(alternations.len());
                 for hir in alternations {
                     if self.expand_regex(hir, next_state_id)? {
                         alternative_state_ids.push(self.nfa.last_state_id());

cli/generate/src/prepare_grammar/extract_tokens.rs

@@ -26,10 +26,34 @@ unless they are used only as the grammar's start rule.
     ExternalTokenNonTerminal(String),
     #[error("Non-symbol rules cannot be used as external tokens")]
     NonSymbolExternalToken,
-    #[error("Non-terminal symbol '{0}' cannot be used as the word token, because its rule is duplicated in '{1}'")]
-    NonTerminalWordToken(String, String),
-    #[error("Reserved words must be tokens")]
-    NonTokenReservedWord,
+    #[error(transparent)]
+    WordToken(NonTerminalWordTokenError),
+    #[error("Reserved word '{0}' must be a token")]
+    NonTokenReservedWord(String),
+}
+
+#[derive(Debug, Error, Serialize)]
+pub struct NonTerminalWordTokenError {
+    pub symbol_name: String,
+    pub conflicting_symbol_name: Option<String>,
+}
+
+impl std::fmt::Display for NonTerminalWordTokenError {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(
+            f,
+            "Non-terminal symbol '{}' cannot be used as the word token",
+            self.symbol_name
+        )?;
+        if let Some(conflicting_name) = &self.conflicting_symbol_name {
+            writeln!(
+                f,
+                ", because its rule is duplicated in '{conflicting_name}'",
+            )
+        } else {
+            writeln!(f)
+        }
+    }
 }
 
 pub(super) fn extract_tokens(
@@ -62,7 +86,7 @@ pub(super) fn extract_tokens(
     // that pointed to that variable will need to be updated to point to the
     // variable in the lexical grammar. Symbols that pointed to later variables
     // will need to have their indices decremented.
-    let mut variables = Vec::new();
+    let mut variables = Vec::with_capacity(grammar.variables.len());
     let mut symbol_replacer = SymbolReplacer {
         replacements: HashMap::new(),
     };
@@ -162,23 +186,23 @@ pub(super) fn extract_tokens(
         let token = symbol_replacer.replace_symbol(token);
         if token.is_non_terminal() {
             let word_token_variable = &variables[token.index];
-            let conflicting_variable = variables
+            let conflicting_symbol_name = variables
                 .iter()
                 .enumerate()
                 .find(|(i, v)| *i != token.index && v.rule == word_token_variable.rule)
-                .expect("Failed to find a variable with the same rule as the word token");
+                .map(|(_, v)| v.name.clone());
 
-            Err(ExtractTokensError::NonTerminalWordToken(
-                word_token_variable.name.clone(),
-                conflicting_variable.1.name.clone(),
-            ))?;
+            Err(ExtractTokensError::WordToken(NonTerminalWordTokenError {
+                symbol_name: word_token_variable.name.clone(),
+                conflicting_symbol_name,
+            }))?;
         }
         word_token = Some(token);
     }
 
-    let mut reserved_word_contexts = Vec::new();
+    let mut reserved_word_contexts = Vec::with_capacity(grammar.reserved_word_sets.len());
     for reserved_word_context in grammar.reserved_word_sets {
-        let mut reserved_words = Vec::new();
+        let mut reserved_words = Vec::with_capacity(reserved_word_contexts.len());
         for reserved_rule in reserved_word_context.reserved_words {
             if let Rule::Symbol(symbol) = reserved_rule {
                 reserved_words.push(symbol_replacer.replace_symbol(symbol));
@@ -188,7 +212,12 @@ pub(super) fn extract_tokens(
             {
                 reserved_words.push(Symbol::terminal(index));
             } else {
-                Err(ExtractTokensError::NonTokenReservedWord)?;
+                let token_name = match &reserved_rule {
+                    Rule::String(s) => s.clone(),
+                    Rule::Pattern(p, _) => p.clone(),
+                    _ => "unknown".to_string(),
+                };
+                Err(ExtractTokensError::NonTokenReservedWord(token_name))?;
             }
         }
         reserved_word_contexts.push(ReservedWordContext {

cli/generate/src/prepare_grammar/flatten_grammar.rs

@@ -57,8 +57,9 @@ impl RuleFlattener {
     }
 
     fn flatten_variable(&mut self, variable: Variable) -> FlattenGrammarResult<SyntaxVariable> {
-        let mut productions = Vec::new();
-        for rule in extract_choices(variable.rule) {
+        let choices = extract_choices(variable.rule);
+        let mut productions = Vec::with_capacity(choices.len());
+        for rule in choices {
             let production = self.flatten_rule(rule)?;
             if !productions.contains(&production) {
                 productions.push(production);
@@ -195,7 +196,7 @@ fn extract_choices(rule: Rule) -> Vec<Rule> {
             let mut result = vec![Rule::Blank];
             for element in elements {
                 let extraction = extract_choices(element);
-                let mut next_result = Vec::new();
+                let mut next_result = Vec::with_capacity(result.len());
                 for entry in result {
                     for extraction_entry in &extraction {
                         next_result.push(Rule::Seq(vec![entry.clone(), extraction_entry.clone()]));
@@ -206,7 +207,7 @@ fn extract_choices(rule: Rule) -> Vec<Rule> {
             result
         }
         Rule::Choice(elements) => {
-            let mut result = Vec::new();
+            let mut result = Vec::with_capacity(elements.len());
             for element in elements {
                 for rule in extract_choices(element) {
                     result.push(rule);
@@ -262,9 +263,10 @@ pub(super) fn flatten_grammar(
 
     for (i, variable) in variables.iter().enumerate() {
         let symbol = Symbol::non_terminal(i);
+        let used = symbol_is_used(&variables, symbol);
 
         for production in &variable.productions {
-            if production.steps.is_empty() && symbol_is_used(&variables, symbol) {
+            if used && production.steps.is_empty() {
                 Err(FlattenGrammarError::EmptyString(variable.name.clone()))?;
             }
 
@@ -533,7 +535,7 @@ mod tests {
 
         assert_eq!(
             result.unwrap_err().to_string(),
-            "Rule `test` cannot be inlined because it contains a reference to itself.",
+            "Rule `test` cannot be inlined because it contains a reference to itself",
         );
     }
 }

cli/generate/src/prepare_grammar/intern_symbols.rs

@@ -65,7 +65,7 @@ pub(super) fn intern_symbols(grammar: &InputGrammar) -> InternSymbolsResult<Inte
 
     let mut reserved_words = Vec::with_capacity(grammar.reserved_words.len());
     for reserved_word_set in &grammar.reserved_words {
-        let mut interned_set = Vec::new();
+        let mut interned_set = Vec::with_capacity(reserved_word_set.reserved_words.len());
         for rule in &reserved_word_set.reserved_words {
             interned_set.push(interner.intern_rule(rule, None)?);
         }
@@ -75,7 +75,7 @@ pub(super) fn intern_symbols(grammar: &InputGrammar) -> InternSymbolsResult<Inte
         });
     }
 
-    let mut expected_conflicts = Vec::new();
+    let mut expected_conflicts = Vec::with_capacity(grammar.expected_conflicts.len());
     for conflict in &grammar.expected_conflicts {
         let mut interned_conflict = Vec::with_capacity(conflict.len());
         for name in conflict {

cli/generate/src/prepare_grammar/mod.rs

@@ -8,7 +8,7 @@ mod process_inlines;
 
 use std::{
     cmp::Ordering,
-    collections::{hash_map, HashMap, HashSet},
+    collections::{hash_map, BTreeSet, HashMap, HashSet},
     mem,
 };
 
@@ -16,6 +16,7 @@ use anyhow::Result;
 pub use expand_tokens::ExpandTokensError;
 pub use extract_tokens::ExtractTokensError;
 pub use flatten_grammar::FlattenGrammarError;
+use indexmap::IndexMap;
 pub use intern_symbols::InternSymbolsError;
 pub use process_inlines::ProcessInlinesError;
 use serde::Serialize;
@@ -80,6 +81,7 @@ pub type PrepareGrammarResult<T> = Result<T, PrepareGrammarError>;
 #[error(transparent)]
 pub enum PrepareGrammarError {
     ValidatePrecedences(#[from] ValidatePrecedenceError),
+    ValidateIndirectRecursion(#[from] IndirectRecursionError),
     InternSymbols(#[from] InternSymbolsError),
     ExtractTokens(#[from] ExtractTokensError),
     FlattenGrammar(#[from] FlattenGrammarError),
@@ -96,6 +98,22 @@ pub enum ValidatePrecedenceError {
     Ordering(#[from] ConflictingPrecedenceOrderingError),
 }
 
+#[derive(Debug, Error, Serialize)]
+pub struct IndirectRecursionError(pub Vec<String>);
+
+impl std::fmt::Display for IndirectRecursionError {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Grammar contains an indirectly recursive rule: ")?;
+        for (i, symbol) in self.0.iter().enumerate() {
+            if i > 0 {
+                write!(f, " -> ")?;
+            }
+            write!(f, "{symbol}")?;
+        }
+        Ok(())
+    }
+}
+
 #[derive(Debug, Error, Serialize)]
 pub struct UndeclaredPrecedenceError {
     pub precedence: String,
@@ -141,6 +159,7 @@ pub fn prepare_grammar(
     AliasMap,
 )> {
     validate_precedences(input_grammar)?;
+    validate_indirect_recursion(input_grammar)?;
 
     let interned_grammar = intern_symbols(input_grammar)?;
     let (syntax_grammar, lexical_grammar) = extract_tokens(interned_grammar)?;
@@ -152,6 +171,83 @@ pub fn prepare_grammar(
     Ok((syntax_grammar, lexical_grammar, inlines, default_aliases))
 }
 
+/// Check for indirect recursion cycles in the grammar that can cause infinite loops while
+/// parsing. An indirect recursion cycle occurs when a non-terminal can derive itself through
+/// a chain of single-symbol productions (e.g., A -> B, B -> A).
+fn validate_indirect_recursion(grammar: &InputGrammar) -> Result<(), IndirectRecursionError> {
+    let mut epsilon_transitions: IndexMap<&str, BTreeSet<String>> = IndexMap::new();
+
+    for variable in &grammar.variables {
+        let productions = get_single_symbol_productions(&variable.rule);
+        // Filter out rules that *directly* reference themselves, as this doesn't
+        // cause a parsing loop.
+        let filtered: BTreeSet<String> = productions
+            .into_iter()
+            .filter(|s| s != &variable.name)
+            .collect();
+        epsilon_transitions.insert(variable.name.as_str(), filtered);
+    }
+
+    for start_symbol in epsilon_transitions.keys() {
+        let mut visited = BTreeSet::new();
+        let mut path = Vec::new();
+        if let Some((start_idx, end_idx)) =
+            get_cycle(start_symbol, &epsilon_transitions, &mut visited, &mut path)
+        {
+            let cycle_symbols = path[start_idx..=end_idx]
+                .iter()
+                .map(|s| (*s).to_string())
+                .collect();
+            return Err(IndirectRecursionError(cycle_symbols));
+        }
+    }
+
+    Ok(())
+}
+
+fn get_single_symbol_productions(rule: &Rule) -> BTreeSet<String> {
+    match rule {
+        Rule::NamedSymbol(name) => BTreeSet::from([name.clone()]),
+        Rule::Choice(choices) => choices
+            .iter()
+            .flat_map(get_single_symbol_productions)
+            .collect(),
+        Rule::Metadata { rule, .. } => get_single_symbol_productions(rule),
+        _ => BTreeSet::new(),
+    }
+}
+
+/// Perform a depth-first search to detect cycles in single state transitions.
+fn get_cycle<'a>(
+    current: &'a str,
+    transitions: &'a IndexMap<&'a str, BTreeSet<String>>,
+    visited: &mut BTreeSet<&'a str>,
+    path: &mut Vec<&'a str>,
+) -> Option<(usize, usize)> {
+    if let Some(first_idx) = path.iter().position(|s| *s == current) {
+        path.push(current);
+        return Some((first_idx, path.len() - 1));
+    }
+
+    if visited.contains(current) {
+        return None;
+    }
+
+    path.push(current);
+    visited.insert(current);
+
+    if let Some(next_symbols) = transitions.get(current) {
+        for next in next_symbols {
+            if let Some(cycle) = get_cycle(next, transitions, visited, path) {
+                return Some(cycle);
+            }
+        }
+    }
+
+    path.pop();
+    None
+}
+
 /// Check that all of the named precedences used in the grammar are declared
 /// within the `precedences` lists, and also that there are no conflicting
 /// precedence orderings declared in those lists.

cli/generate/src/render.rs

@@ -24,7 +24,6 @@ pub const ABI_VERSION_MIN: usize = 14;
 pub const ABI_VERSION_MAX: usize = tree_sitter::LANGUAGE_VERSION;
 const ABI_VERSION_WITH_RESERVED_WORDS: usize = 15;
 const BUILD_VERSION: &str = env!("CARGO_PKG_VERSION");
-const BUILD_SHA: Option<&'static str> = option_env!("BUILD_SHA");
 
 macro_rules! add {
     ($this: tt, $($arg: tt)*) => {{
@@ -322,13 +321,9 @@ impl Generator {
     }
 
     fn add_header(&mut self) {
-        let version = BUILD_SHA.map_or_else(
-            || BUILD_VERSION.to_string(),
-            |build_sha| format!("{BUILD_VERSION} ({build_sha})"),
-        );
         add_line!(
             self,
-            "/* Automatically generated by tree-sitter v{version} */",
+            "/* Automatically @generated by tree-sitter v{BUILD_VERSION} */",
         );
         add_line!(self, "");
     }
@@ -683,12 +678,12 @@ impl Generator {
             &mut next_flat_field_map_index,
         );
 
-        let mut field_map_ids = Vec::new();
+        let mut field_map_ids = Vec::with_capacity(self.parse_table.production_infos.len());
         for production_info in &self.parse_table.production_infos {
             if production_info.field_map.is_empty() {
                 field_map_ids.push((0, 0));
             } else {
-                let mut flat_field_map = Vec::new();
+                let mut flat_field_map = Vec::with_capacity(production_info.field_map.len());
                 for (field_name, locations) in &production_info.field_map {
                     for location in locations {
                         flat_field_map.push((field_name.clone(), *location));
@@ -1111,7 +1106,11 @@ impl Generator {
             return;
         }
 
-        add_line!(self, "const TSCharacterRange {}[] = {{", info.constant_name);
+        add_line!(
+            self,
+            "static const TSCharacterRange {}[] = {{",
+            info.constant_name
+        );
 
         indent!(self);
         for (ix, range) in characters.ranges().enumerate() {
@@ -1351,7 +1350,12 @@ impl Generator {
             indent!(self);
 
             let mut next_table_index = 0;
-            let mut small_state_indices = Vec::new();
+            let mut small_state_indices = Vec::with_capacity(
+                self.parse_table
+                    .states
+                    .len()
+                    .saturating_sub(self.large_state_count),
+            );
             let mut symbols_by_value = HashMap::<(usize, SymbolType), Vec<Symbol>>::new();
             for state in self.parse_table.states.iter().skip(self.large_state_count) {
                 small_state_indices.push(next_table_index);
@@ -1847,11 +1851,11 @@ impl Generator {
                         '\u{007F}' => "DEL",
                         '\u{FEFF}' => "BOM",
                         '\u{0080}'..='\u{FFFF}' => {
-                            result.push_str(&format!("u{:04x}", c as u32));
+                            write!(result, "u{:04x}", c as u32).unwrap();
                             break 'special_chars;
                         }
                         '\u{10000}'..='\u{10FFFF}' => {
-                            result.push_str(&format!("U{:08x}", c as u32));
+                            write!(result, "U{:08x}", c as u32).unwrap();
                             break 'special_chars;
                         }
                         '0'..='9' | 'a'..='z' | 'A'..='Z' | '_' => unreachable!(),
@@ -1882,11 +1886,9 @@ impl Generator {
                 '\r' => result += "\\r",
                 '\t' => result += "\\t",
                 '\0' => result += "\\0",
-                '\u{0001}'..='\u{001f}' => result += &format!("\\x{:02x}", c as u32),
-                '\u{007F}'..='\u{FFFF}' => result += &format!("\\u{:04x}", c as u32),
-                '\u{10000}'..='\u{10FFFF}' => {
-                    result.push_str(&format!("\\U{:08x}", c as u32));
-                }
+                '\u{0001}'..='\u{001f}' => write!(result, "\\x{:02x}", c as u32).unwrap(),
+                '\u{007F}'..='\u{FFFF}' => write!(result, "\\u{:04x}", c as u32).unwrap(),
+                '\u{10000}'..='\u{10FFFF}' => write!(result, "\\U{:08x}", c as u32).unwrap(),
                 _ => result.push(c),
             }
         }

cli/generate/src/rules.rs

@@ -306,7 +306,6 @@ impl Symbol {
 }
 
 impl From<Symbol> for Rule {
-    #[must_use]
     fn from(symbol: Symbol) -> Self {
         Self::Symbol(symbol)
     }

cli/loader/emscripten-version

@@ -1 +1 @@
-4.0.1
+4.0.4

cli/loader/src/lib.rs

@@ -11,6 +11,7 @@ use std::{
     ffi::{OsStr, OsString},
     fs,
     io::{BufRead, BufReader},
+    marker::PhantomData,
     mem,
     path::{Path, PathBuf},
     process::Command,
@@ -18,9 +19,7 @@ use std::{
     time::SystemTime,
 };
 
-#[cfg(any(feature = "tree-sitter-highlight", feature = "tree-sitter-tags"))]
-use anyhow::Error;
-use anyhow::{anyhow, Context, Result};
+use anyhow::{anyhow, Context, Error, Result};
 use etcetera::BaseStrategy as _;
 use fs4::fs_std::FileExt;
 use indoc::indoc;
@@ -327,6 +326,7 @@ pub struct LanguageConfiguration<'a> {
     highlight_names: &'a Mutex<Vec<String>>,
     #[cfg(feature = "tree-sitter-highlight")]
     use_all_highlight_names: bool,
+    _phantom: PhantomData<&'a ()>,
 }
 
 pub struct Loader {
@@ -561,8 +561,8 @@ impl Loader {
                 // If multiple language configurations match, then determine which
                 // one to use by applying the configurations' content regexes.
                 else {
-                    let file_contents =
-                        fs::read(path).with_context(|| format!("Failed to read path {path:?}"))?;
+                    let file_contents = fs::read(path)
+                        .with_context(|| format!("Failed to read path {}", path.display()))?;
                     let file_contents = String::from_utf8_lossy(&file_contents);
                     let mut best_score = -2isize;
                     let mut best_configuration_id = None;
@@ -780,8 +780,8 @@ impl Loader {
         if recompile {
             fs::create_dir_all(lock_path.parent().unwrap()).with_context(|| {
                 format!(
-                    "Failed to create directory {:?}",
-                    lock_path.parent().unwrap()
+                    "Failed to create directory {}",
+                    lock_path.parent().unwrap().display()
                 )
             })?;
             let lock_file = fs::OpenOptions::new()
@@ -799,7 +799,7 @@ impl Loader {
         }
 
         let library = unsafe { Library::new(&output_path) }
-            .with_context(|| format!("Error opening dynamic library {output_path:?}"))?;
+            .with_context(|| format!("Error opening dynamic library {}", output_path.display()))?;
         let language = unsafe {
             let language_fn = library
                 .get::<Symbol<unsafe extern "C" fn() -> Language>>(language_fn_name.as_bytes())
@@ -1214,6 +1214,7 @@ impl Loader {
                     highlight_names: &self.highlight_names,
                     #[cfg(feature = "tree-sitter-highlight")]
                     use_all_highlight_names: self.use_all_highlight_names,
+                    _phantom: PhantomData,
                 };
 
                 for file_type in &configuration.file_types {
@@ -1283,6 +1284,7 @@ impl Loader {
                 highlight_names: &self.highlight_names,
                 #[cfg(feature = "tree-sitter-highlight")]
                 use_all_highlight_names: self.use_all_highlight_names,
+                _phantom: PhantomData,
             };
             self.language_configurations.push(unsafe {
                 mem::transmute::<LanguageConfiguration<'_>, LanguageConfiguration<'static>>(
@@ -1564,7 +1566,7 @@ impl LanguageConfiguration<'_> {
         error.row = source[range.start..offset_within_section]
             .matches('\n')
             .count();
-        Error::from(error).context(format!("Error in query file {path:?}"))
+        Error::from(error).context(format!("Error in query file {}", path.display()))
     }
 
     #[allow(clippy::type_complexity)]
@@ -1581,7 +1583,7 @@ impl LanguageConfiguration<'_> {
                 let abs_path = self.root_path.join(path);
                 let prev_query_len = query.len();
                 query += &fs::read_to_string(&abs_path)
-                    .with_context(|| format!("Failed to read query file {path:?}"))?;
+                    .with_context(|| format!("Failed to read query file {}", path.display()))?;
                 path_ranges.push((path.clone(), prev_query_len..query.len()));
             }
         } else {
@@ -1599,7 +1601,7 @@ impl LanguageConfiguration<'_> {
             let path = queries_path.join(default_path);
             if path.exists() {
                 query = fs::read_to_string(&path)
-                    .with_context(|| format!("Failed to read query file {path:?}"))?;
+                    .with_context(|| format!("Failed to read query file {}", path.display()))?;
                 path_ranges.push((PathBuf::from(default_path), 0..query.len()));
             }
         }
@@ -1612,8 +1614,8 @@ fn needs_recompile(lib_path: &Path, paths_to_check: &[PathBuf]) -> Result<bool>
     if !lib_path.exists() {
         return Ok(true);
     }
-    let lib_mtime =
-        mtime(lib_path).with_context(|| format!("Failed to read mtime of {lib_path:?}"))?;
+    let lib_mtime = mtime(lib_path)
+        .with_context(|| format!("Failed to read mtime of {}", lib_path.display()))?;
     for path in paths_to_check {
         if mtime(path)? > lib_mtime {
             return Ok(true);

cli/npm/dsl.d.ts

@@ -10,6 +10,7 @@ type PrecRightRule = { type: 'PREC_RIGHT'; content: Rule; value: number };
 type PrecRule = { type: 'PREC'; content: Rule; value: number };
 type Repeat1Rule = { type: 'REPEAT1'; content: Rule };
 type RepeatRule = { type: 'REPEAT'; content: Rule };
+type ReservedRule = { type: 'RESERVED'; content: Rule; context_name: string };
 type SeqRule = { type: 'SEQ'; members: Rule[] };
 type StringRule = { type: 'STRING'; value: string };
 type SymbolRule<Name extends string> = { type: 'SYMBOL'; name: Name };
@@ -33,12 +34,10 @@ type Rule =
   | SymbolRule<string>
   | TokenRule;
 
-class RustRegex {
+declare class RustRegex {
   value: string;
 
-  constructor(pattern: string) {
-    this.value = pattern;
-  }
+  constructor(pattern: string);
 }
 
 type RuleOrLiteral = Rule | RegExp | RustRegex | string;
@@ -167,6 +166,17 @@ interface Grammar<
    * @see https://tree-sitter.github.io/tree-sitter/creating-parsers/3-writing-the-grammar#keyword-extraction
    */
   word?: ($: GrammarSymbols<RuleName | BaseGrammarRuleName>) => RuleOrLiteral;
+
+
+  /**
+   * Mapping of names to reserved word sets. The first reserved word set is the
+   * global word set, meaning it applies to every rule in every parse state.
+   * The other word sets can be used with the `reserved` function.
+   */
+  reserved?: Record<
+    string,
+    ($: GrammarSymbols<RuleName | BaseGrammarRuleName>) => RuleOrLiteral[]
+  >;
 }
 
 type GrammarSchema<RuleName extends string> = {
@@ -251,7 +261,7 @@ declare function optional(rule: RuleOrLiteral): ChoiceRule;
  * @see https://docs.oracle.com/cd/E19504-01/802-5880/6i9k05dh3/index.html
  */
 declare const prec: {
-  (value: String | number, rule: RuleOrLiteral): PrecRule;
+  (value: string | number, rule: RuleOrLiteral): PrecRule;
 
   /**
    * Marks the given rule as left-associative (and optionally applies a
@@ -267,7 +277,7 @@ declare const prec: {
    * @see https://docs.oracle.com/cd/E19504-01/802-5880/6i9k05dh3/index.html
    */
   left(rule: RuleOrLiteral): PrecLeftRule;
-  left(value: String | number, rule: RuleOrLiteral): PrecLeftRule;
+  left(value: string | number, rule: RuleOrLiteral): PrecLeftRule;
 
   /**
    * Marks the given rule as right-associative (and optionally applies a
@@ -283,7 +293,7 @@ declare const prec: {
    * @see https://docs.oracle.com/cd/E19504-01/802-5880/6i9k05dh3/index.html
    */
   right(rule: RuleOrLiteral): PrecRightRule;
-  right(value: String | number, rule: RuleOrLiteral): PrecRightRule;
+  right(value: string | number, rule: RuleOrLiteral): PrecRightRule;
 
   /**
    * Marks the given rule with a numerical precedence which will be used to
@@ -300,7 +310,7 @@ declare const prec: {
    *
    * @see https://www.gnu.org/software/bison/manual/html_node/Generalized-LR-Parsing.html
    */
-  dynamic(value: String | number, rule: RuleOrLiteral): PrecDynamicRule;
+  dynamic(value: string | number, rule: RuleOrLiteral): PrecDynamicRule;
 };
 
 /**
@@ -320,6 +330,15 @@ declare function repeat(rule: RuleOrLiteral): RepeatRule;
  */
 declare function repeat1(rule: RuleOrLiteral): Repeat1Rule;
 
+/**
+ * Overrides the global reserved word set for a given rule. The word set name
+ * should be defined in the `reserved` field in the grammar.
+ *
+ * @param wordset name of the reserved word set
+ * @param rule rule that will use the reserved word set
+ */
+declare function reserved(wordset: string, rule: RuleOrLiteral): ReservedRule;
+
 /**
  * Creates a rule that matches any number of other rules, one after another.
  * It is analogous to simply writing multiple symbols next to each other
@@ -338,7 +357,7 @@ declare function sym<Name extends string>(name: Name): SymbolRule<Name>;
 
 /**
  * Marks the given rule as producing only a single token. Tree-sitter's
- * default is to treat each String or RegExp literal in the grammar as a
+ * default is to treat each string or RegExp literal in the grammar as a
  * separate token. Each token is matched separately by the lexer and
  * returned as its own leaf node in the tree. The token function allows
  * you to express a complex rule using the DSL functions (rather

cli/npm/install.js

@@ -6,7 +6,8 @@ const http = require('http');
 const https = require('https');
 const packageJSON = require('./package.json');
 
-// Look to a results table in https://github.com/tree-sitter/tree-sitter/issues/2196
+https.globalAgent.keepAlive = false;
+
 const matrix = {
   platform: {
     'darwin': {

cli/npm/package.json

@@ -1,6 +1,6 @@
 {
   "name": "tree-sitter-cli",
-  "version": "0.25.1",
+  "version": "0.25.10",
   "author": {
     "name": "Max Brunsfeld",
     "email": "maxbrunsfeld@gmail.com"
@@ -14,14 +14,14 @@
   "license": "MIT",
   "repository": {
     "type": "git",
-    "url": "https://github.com/tree-sitter/tree-sitter.git"
+    "url": "git+https://github.com/tree-sitter/tree-sitter.git",
+    "directory": "crates/cli/npm"
   },
   "description": "CLI for generating fast incremental parsers",
   "keywords": [
     "parser",
     "lexer"
   ],
-  "main": "lib/api/index.js",
   "engines": {
     "node": ">=12.0.0"
   },

cli/src/fuzz/allocations.rs

@@ -109,7 +109,7 @@ unsafe extern "C" fn ts_record_realloc(ptr: *mut c_void, size: usize) -> *mut c_
     let result = realloc(ptr, size);
     if ptr.is_null() {
         record_alloc(result);
-    } else if ptr != result {
+    } else if !core::ptr::eq(ptr, result) {
         record_dealloc(ptr);
         record_alloc(result);
     }

cli/src/fuzz/mod.rs

@@ -56,7 +56,9 @@ fn regex_env_var(name: &'static str) -> Option<Regex> {
 pub fn new_seed() -> usize {
     int_env_var("TREE_SITTER_SEED").unwrap_or_else(|| {
         let mut rng = rand::thread_rng();
-        rng.gen::<usize>()
+        let seed = rng.gen::<usize>();
+        eprintln!("Seed: {seed}");
+        seed
     })
 }
 
@@ -213,8 +215,9 @@ pub fn fuzz_language_corpus(
                 }
 
                 // Perform a random series of edits and reparse.
-                let mut undo_stack = Vec::new();
-                for _ in 0..=rand.unsigned(*EDIT_COUNT) {
+                let edit_count = rand.unsigned(*EDIT_COUNT);
+                let mut undo_stack = Vec::with_capacity(edit_count);
+                for _ in 0..=edit_count {
                     let edit = get_random_edit(&mut rand, &input);
                     undo_stack.push(invert_edit(&input, &edit));
                     perform_edit(&mut tree, &mut input, &edit).unwrap();

cli/src/fuzz/random.rs

@@ -20,8 +20,8 @@ impl Rand {
     }
 
     pub fn words(&mut self, max_count: usize) -> Vec<u8> {
-        let mut result = Vec::new();
         let word_count = self.unsigned(max_count);
+        let mut result = Vec::with_capacity(2 * word_count);
         for i in 0..word_count {
             if i > 0 {
                 if self.unsigned(5) == 0 {

cli/src/highlight.rs

@@ -1,5 +1,5 @@
 use std::{
-    collections::{HashMap, HashSet},
+    collections::{BTreeMap, HashSet},
     fmt::Write,
     fs,
     io::{self, Write as _},
@@ -82,9 +82,9 @@ impl<'de> Deserialize<'de> for Theme {
     {
         let mut styles = Vec::new();
         let mut highlight_names = Vec::new();
-        if let Ok(colors) = HashMap::<String, Value>::deserialize(deserializer) {
-            highlight_names.reserve(colors.len());
+        if let Ok(colors) = BTreeMap::<String, Value>::deserialize(deserializer) {
             styles.reserve(colors.len());
+            highlight_names.reserve(colors.len());
             for (name, style_value) in colors {
                 let mut style = Style::default();
                 parse_style(&mut style, style_value);
@@ -127,7 +127,7 @@ impl Serialize for Theme {
                 || effects.contains(Effects::ITALIC)
                 || effects.contains(Effects::UNDERLINE)
             {
-                let mut style_json = HashMap::new();
+                let mut style_json = BTreeMap::new();
                 if let Some(color) = color {
                     style_json.insert("color", color);
                 }

cli/src/init.rs

@@ -98,6 +98,7 @@ const TESTS_SWIFT_TEMPLATE: &str = include_str!("./templates/tests.swift");
 const BUILD_ZIG_TEMPLATE: &str = include_str!("./templates/build.zig");
 const BUILD_ZIG_ZON_TEMPLATE: &str = include_str!("./templates/build.zig.zon");
 const ROOT_ZIG_TEMPLATE: &str = include_str!("./templates/root.zig");
+const TEST_ZIG_TEMPLATE: &str = include_str!("./templates/test.zig");
 
 const TREE_SITTER_JSON_SCHEMA: &str =
     "https://tree-sitter.github.io/tree-sitter/assets/schemas/config.schema.json";
@@ -301,14 +302,36 @@ pub fn generate_grammar_files(
     };
 
     // Create package.json
-    missing_path(repo_path.join("package.json"), |path| {
-        generate_file(
-            path,
-            PACKAGE_JSON_TEMPLATE,
-            dashed_language_name.as_str(),
-            &generate_opts,
-        )
-    })?;
+    missing_path_else(
+        repo_path.join("package.json"),
+        allow_update,
+        |path| {
+            generate_file(
+                path,
+                PACKAGE_JSON_TEMPLATE,
+                dashed_language_name.as_str(),
+                &generate_opts,
+            )
+        },
+        |path| {
+            let contents = fs::read_to_string(path)?
+                .replace(
+                    r#""node-addon-api": "^8.3.1"#,
+                    r#""node-addon-api": "^8.5.0""#,
+                )
+                .replace(
+                    indoc! {r#"
+                    "prebuildify": "^6.0.1",
+                    "tree-sitter-cli":"#},
+                    indoc! {r#"
+                    "prebuildify": "^6.0.1",
+                    "tree-sitter": "^0.22.4",
+                    "tree-sitter-cli":"#},
+                );
+            write_file(path, contents)?;
+            Ok(())
+        },
+    )?;
 
     // Do not create a grammar.js file in a repo with multiple language configs
     if !tree_sitter_config.has_multiple_language_configs() {
@@ -371,14 +394,25 @@ pub fn generate_grammar_files(
                 generate_file(path, BUILD_RS_TEMPLATE, language_name, &generate_opts)
             })?;
 
-            missing_path(repo_path.join("Cargo.toml"), |path| {
-                generate_file(
-                    path,
-                    CARGO_TOML_TEMPLATE,
-                    dashed_language_name.as_str(),
-                    &generate_opts,
-                )
-            })?;
+            missing_path_else(
+                repo_path.join("Cargo.toml"),
+                allow_update,
+                |path| {
+                    generate_file(
+                        path,
+                        CARGO_TOML_TEMPLATE,
+                        dashed_language_name.as_str(),
+                        &generate_opts,
+                    )
+                },
+                |path| {
+                    let contents = fs::read_to_string(path)?;
+                    if contents.contains("\"LICENSE\"") {
+                        write_file(path, contents.replace("\"LICENSE\"", "\"/LICENSE\""))?;
+                    }
+                    Ok(())
+                },
+            )?;
 
             Ok(())
         })?;
@@ -394,6 +428,7 @@ pub fn generate_grammar_files(
                 |path| {
                     let contents = fs::read_to_string(path)?;
                     if !contents.contains("bun") {
+                        eprintln!("Replacing index.js");
                         generate_file(path, INDEX_JS_TEMPLATE, language_name, &generate_opts)?;
                     }
                     Ok(())
@@ -597,14 +632,32 @@ pub fn generate_grammar_files(
             })?;
 
             missing_path(path.join("tests"), create_dir)?.apply(|path| {
-                missing_path(path.join("test_binding.py"), |path| {
-                    generate_file(
-                        path,
-                        TEST_BINDING_PY_TEMPLATE,
-                        language_name,
-                        &generate_opts,
-                    )
-                })?;
+                missing_path_else(
+                    path.join("test_binding.py"),
+                    allow_update,
+                    |path| {
+                        generate_file(
+                            path,
+                            TEST_BINDING_PY_TEMPLATE,
+                            language_name,
+                            &generate_opts,
+                        )
+                    },
+                    |path| {
+                        let mut contents = fs::read_to_string(path)?;
+                        if !contents.contains("Parser(Language(") {
+                            contents = contents
+                                .replace("tree_sitter.Language(", "Parser(Language(")
+                                .replace(".language())\n", ".language()))\n")
+                                .replace(
+                                    "import tree_sitter\n",
+                                    "from tree_sitter import Language, Parser\n",
+                                );
+                            write_file(path, contents)?;
+                        }
+                        Ok(())
+                    },
+                )?;
                 Ok(())
             })?;
 
@@ -614,7 +667,7 @@ pub fn generate_grammar_files(
                 |path| generate_file(path, SETUP_PY_TEMPLATE, language_name, &generate_opts),
                 |path| {
                     let contents = fs::read_to_string(path)?;
-                    if !contents.contains("egg_info") || !contents.contains("Py_GIL_DISABLED") {
+                    if !contents.contains("build_ext") {
                         eprintln!("Replacing setup.py");
                         generate_file(path, SETUP_PY_TEMPLATE, language_name, &generate_opts)?;
                     }
@@ -653,22 +706,17 @@ pub fn generate_grammar_files(
     // Generate Swift bindings
     if tree_sitter_config.bindings.swift {
         missing_path(bindings_dir.join("swift"), create_dir)?.apply(|path| {
-            let lang_path = path.join(format!("TreeSitter{camel_name}"));
+            let lang_path = path.join(&class_name);
             missing_path(&lang_path, create_dir)?;
 
             missing_path(lang_path.join(format!("{language_name}.h")), |path| {
                 generate_file(path, PARSER_NAME_H_TEMPLATE, language_name, &generate_opts)
             })?;
 
-            missing_path(
-                path.join(format!("TreeSitter{camel_name}Tests")),
-                create_dir,
-            )?
-            .apply(|path| {
-                missing_path(
-                    path.join(format!("TreeSitter{camel_name}Tests.swift")),
-                    |path| generate_file(path, TESTS_SWIFT_TEMPLATE, language_name, &generate_opts),
-                )?;
+            missing_path(path.join(format!("{class_name}Tests")), create_dir)?.apply(|path| {
+                missing_path(path.join(format!("{class_name}Tests.swift")), |path| {
+                    generate_file(path, TESTS_SWIFT_TEMPLATE, language_name, &generate_opts)
+                })?;
 
                 Ok(())
             })?;
@@ -679,10 +727,13 @@ pub fn generate_grammar_files(
                 |path| generate_file(path, PACKAGE_SWIFT_TEMPLATE, language_name, &generate_opts),
                 |path| {
                     let mut contents = fs::read_to_string(path)?;
-                    contents = contents.replace(
-                        "https://github.com/ChimeHQ/SwiftTreeSitter",
-                        "https://github.com/tree-sitter/swift-tree-sitter",
-                    );
+                    contents = contents
+                        .replace(
+                            "https://github.com/ChimeHQ/SwiftTreeSitter",
+                            "https://github.com/tree-sitter/swift-tree-sitter",
+                        )
+                        .replace("version: \"0.8.0\")", "version: \"0.9.0\")")
+                        .replace("(url:", "(name: \"SwiftTreeSitter\", url:");
                     write_file(path, contents)?;
                     Ok(())
                 },
@@ -694,17 +745,54 @@ pub fn generate_grammar_files(
 
     // Generate Zig bindings
     if tree_sitter_config.bindings.zig {
-        missing_path(repo_path.join("build.zig"), |path| {
-            generate_file(path, BUILD_ZIG_TEMPLATE, language_name, &generate_opts)
-        })?;
+        missing_path_else(
+            repo_path.join("build.zig"),
+            allow_update,
+            |path| generate_file(path, BUILD_ZIG_TEMPLATE, language_name, &generate_opts),
+            |path| {
+                let contents = fs::read_to_string(path)?;
+                if !contents.contains("b.pkg_hash.len") {
+                    eprintln!("Replacing build.zig");
+                    generate_file(path, BUILD_ZIG_TEMPLATE, language_name, &generate_opts)
+                } else {
+                    Ok(())
+                }
+            },
+        )?;
 
-        missing_path(repo_path.join("build.zig.zon"), |path| {
-            generate_file(path, BUILD_ZIG_ZON_TEMPLATE, language_name, &generate_opts)
-        })?;
+        missing_path_else(
+            repo_path.join("build.zig.zon"),
+            allow_update,
+            |path| generate_file(path, BUILD_ZIG_ZON_TEMPLATE, language_name, &generate_opts),
+            |path| {
+                let contents = fs::read_to_string(path)?;
+                if !contents.contains(".name = .tree_sitter_") {
+                    eprintln!("Replacing build.zig.zon");
+                    generate_file(path, BUILD_ZIG_ZON_TEMPLATE, language_name, &generate_opts)
+                } else {
+                    Ok(())
+                }
+            },
+        )?;
 
         missing_path(bindings_dir.join("zig"), create_dir)?.apply(|path| {
-            missing_path(path.join("root.zig"), |path| {
-                generate_file(path, ROOT_ZIG_TEMPLATE, language_name, &generate_opts)
+            missing_path_else(
+                path.join("root.zig"),
+                allow_update,
+                |path| generate_file(path, ROOT_ZIG_TEMPLATE, language_name, &generate_opts),
+                |path| {
+                    let contents = fs::read_to_string(path)?;
+                    if contents.contains("ts.Language") {
+                        eprintln!("Replacing root.zig");
+                        generate_file(path, ROOT_ZIG_TEMPLATE, language_name, &generate_opts)
+                    } else {
+                        Ok(())
+                    }
+                },
+            )?;
+
+            missing_path(path.join("test.zig"), |path| {
+                generate_file(path, TEST_ZIG_TEMPLATE, language_name, &generate_opts)
             })?;
 
             Ok(())

cli/src/input.rs

@@ -89,8 +89,8 @@ pub fn get_input(
                 let Some(path_str) = path.to_str() else {
                     bail!("Invalid path: {}", path.display());
                 };
-                let paths =
-                    glob(path_str).with_context(|| format!("Invalid glob pattern {path:?}"))?;
+                let paths = glob(path_str)
+                    .with_context(|| format!("Invalid glob pattern {}", path.display()))?;
                 for path in paths {
                     incorporate_path(path?, positive);
                 }

cli/src/main.rs

@@ -206,7 +206,8 @@ struct Parse {
     #[arg(long, short)]
     pub quiet: bool,
     #[allow(clippy::doc_markdown)]
-    /// Apply edits in the format: \"row, col delcount insert_text\"
+    /// Apply edits in the format: \"row,col|position delcount insert_text\", can be supplied
+    /// multiple times
     #[arg(
         long,
         num_args = 1..,
@@ -964,8 +965,11 @@ impl Parse {
 
                 for path in &paths {
                     let path = Path::new(&path);
-                    let language =
-                        loader.select_language(path, current_dir, self.scope.as_deref())?;
+                    let language = loader
+                        .select_language(path, current_dir, self.scope.as_deref())
+                        .with_context(|| {
+                            anyhow!("Failed to load langauge for path \"{}\"", path.display())
+                        })?;
 
                     parse::parse_file_at_path(
                         &mut parser,

cli/src/parse.rs

@@ -29,18 +29,28 @@ pub struct Stats {
 impl fmt::Display for Stats {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         let duration_us = self.total_duration.as_micros();
+        let success_rate = if self.total_parses > 0 {
+            format!(
+                "{:.2}%",
+                ((self.successful_parses as f64) / (self.total_parses as f64)) * 100.0,
+            )
+        } else {
+            "N/A".to_string()
+        };
+        let duration_str = match (self.total_parses, duration_us) {
+            (0, _) => "N/A".to_string(),
+            (_, 0) => "0 bytes/ms".to_string(),
+            (_, _) => format!(
+                "{} bytes/ms",
+                ((self.total_bytes as u128) * 1_000) / duration_us
+            ),
+        };
         writeln!(
             f,
-            "Total parses: {}; successful parses: {}; failed parses: {}; success percentage: {:.2}%; average speed: {} bytes/ms",
+            "Total parses: {}; successful parses: {}; failed parses: {}; success percentage: {success_rate}; average speed: {duration_str}",
             self.total_parses,
             self.successful_parses,
             self.total_parses - self.successful_parses,
-            ((self.successful_parses as f64) / (self.total_parses as f64)) * 100.0,
-            if duration_us != 0 {
-                ((self.total_bytes as u128) * 1_000) / duration_us
-            } else {
-                0
-            }
         )
     }
 }
@@ -225,7 +235,7 @@ pub struct ParseStats {
     pub cumulative_stats: Stats,
 }
 
-#[derive(Serialize, ValueEnum, Debug, Clone, Default, Eq, PartialEq)]
+#[derive(Serialize, ValueEnum, Debug, Copy, Clone, Default, Eq, PartialEq)]
 pub enum ParseDebugType {
     #[default]
     Quiet,
@@ -273,10 +283,11 @@ pub fn parse_file_at_path(
     }
     // Log to stderr if `--debug` was passed
     else if opts.debug != ParseDebugType::Quiet {
-        let mut curr_version: usize = 0usize;
+        let mut curr_version: usize = 0;
         let use_color = std::env::var("NO_COLOR").map_or(true, |v| v != "1");
-        parser.set_logger(Some(Box::new(|log_type, message| {
-            if opts.debug == ParseDebugType::Normal {
+        let debug = opts.debug;
+        parser.set_logger(Some(Box::new(move |log_type, message| {
+            if debug == ParseDebugType::Normal {
                 if log_type == LogType::Lex {
                     write!(&mut io::stderr(), "  ").unwrap();
                 }
@@ -686,19 +697,23 @@ pub fn parse_file_at_path(
             if let Some(node) = first_error {
                 let start = node.start_position();
                 let end = node.end_position();
+                let mut node_text = String::new();
+                for c in node.kind().chars() {
+                    if let Some(escaped) = escape_invisible(c) {
+                        node_text += escaped;
+                    } else {
+                        node_text.push(c);
+                    }
+                }
                 write!(&mut stdout, "\t(")?;
                 if node.is_missing() {
                     if node.is_named() {
-                        write!(&mut stdout, "MISSING {}", node.kind())?;
+                        write!(&mut stdout, "MISSING {node_text}")?;
                     } else {
-                        write!(
-                            &mut stdout,
-                            "MISSING \"{}\"",
-                            node.kind().replace('\n', "\\n")
-                        )?;
+                        write!(&mut stdout, "MISSING \"{node_text}\"")?;
                     }
                 } else {
-                    write!(&mut stdout, "{}", node.kind())?;
+                    write!(&mut stdout, "{node_text}")?;
                 }
                 write!(
                     &mut stdout,

cli/src/query.rs

@@ -34,7 +34,7 @@ pub fn query_file_at_path(
     let mut stdout = stdout.lock();
 
     let query_source = fs::read_to_string(query_path)
-        .with_context(|| format!("Error reading query file {query_path:?}"))?;
+        .with_context(|| format!("Error reading query file {}", query_path.display()))?;
     let query = Query::new(language, &query_source).with_context(|| "Query compilation failed")?;
 
     let mut query_cursor = QueryCursor::new();
@@ -55,7 +55,7 @@ pub fn query_file_at_path(
     }
 
     let source_code =
-        fs::read(path).with_context(|| format!("Error reading source file {path:?}"))?;
+        fs::read(path).with_context(|| format!("Error reading source file {}", path.display()))?;
     let tree = parser.parse(&source_code, None).unwrap();
 
     let start = Instant::now();

cli/src/templates/_cargo.toml

@@ -18,7 +18,7 @@ include = [
   "queries/*",
   "src/*",
   "tree-sitter.json",
-  "LICENSE",
+  "/LICENSE",
 ]
 
 [lib]

cli/src/templates/build.zig

@@ -7,24 +7,24 @@ pub fn build(b: *std.Build) !void {
     const shared = b.option(bool, "build-shared", "Build a shared library") orelse true;
     const reuse_alloc = b.option(bool, "reuse-allocator", "Reuse the library allocator") orelse false;
 
-    const lib: *std.Build.Step.Compile = if (shared) b.addSharedLibrary(.{
-        .name = "tree-sitter-PARSER_NAME",
-        .pic = true,
-        .target = target,
-        .optimize = optimize,
-        .link_libc = true,
-    }) else b.addStaticLibrary(.{
-        .name = "tree-sitter-PARSER_NAME",
-        .target = target,
-        .optimize = optimize,
-        .link_libc = true,
+    const library_name = "tree-sitter-PARSER_NAME";
+
+    const lib: *std.Build.Step.Compile = b.addLibrary(.{
+        .name = library_name,
+        .linkage = if (shared) .dynamic else .static,
+        .root_module = b.createModule(.{
+            .target = target,
+            .optimize = optimize,
+            .link_libc = true,
+            .pic = if (shared) true else null,
+        }),
     });
 
     lib.addCSourceFile(.{
         .file = b.path("src/parser.c"),
         .flags = &.{"-std=c11"},
     });
-    if (hasScanner(b.build_root.handle)) {
+    if (fileExists(b, "src/scanner.c")) {
         lib.addCSourceFile(.{
             .file = b.path("src/scanner.c"),
             .flags = &.{"-std=c11"},
@@ -42,38 +42,52 @@ pub fn build(b: *std.Build) !void {
 
     b.installArtifact(lib);
     b.installFile("src/node-types.json", "node-types.json");
-    b.installDirectory(.{ .source_dir = b.path("queries"), .install_dir = .prefix, .install_subdir = "queries", .include_extensions = &.{"scm"} });
 
-    const module = b.addModule("tree-sitter-PARSER_NAME", .{
+    if (fileExists(b, "queries")) {
+        b.installDirectory(.{
+            .source_dir = b.path("queries"),
+            .install_dir = .prefix,
+            .install_subdir = "queries",
+            .include_extensions = &.{"scm"},
+        });
+    }
+
+    const module = b.addModule(library_name, .{
         .root_source_file = b.path("bindings/zig/root.zig"),
         .target = target,
         .optimize = optimize,
     });
     module.linkLibrary(lib);
 
-    const ts_dep = b.dependency("tree-sitter", .{});
-    const ts_mod = ts_dep.module("tree-sitter");
-    module.addImport("tree-sitter", ts_mod);
-
-    // ╭─────────────────╮
-    // │      Tests      │
-    // ╰─────────────────╯
-
     const tests = b.addTest(.{
-        .root_source_file = b.path("bindings/zig/root.zig"),
-        .target = target,
-        .optimize = optimize,
+        .root_module = b.createModule(.{
+            .root_source_file = b.path("bindings/zig/test.zig"),
+            .target = target,
+            .optimize = optimize,
+        }),
     });
-    tests.linkLibrary(lib);
-    tests.root_module.addImport("tree-sitter", ts_mod);
+    tests.root_module.addImport(library_name, module);
+
+    // HACK: fetch tree-sitter dependency only when testing this module
+    if (b.pkg_hash.len == 0) {
+        var args = try std.process.argsWithAllocator(b.allocator);
+        defer args.deinit();
+        while (args.next()) |a| {
+            if (std.mem.eql(u8, a, "test")) {
+                const ts_dep = b.lazyDependency("tree_sitter", .{}) orelse continue;
+                tests.root_module.addImport("tree-sitter", ts_dep.module("tree-sitter"));
+                break;
+            }
+        }
+    }
 
     const run_tests = b.addRunArtifact(tests);
-
     const test_step = b.step("test", "Run unit tests");
     test_step.dependOn(&run_tests.step);
 }
 
-inline fn hasScanner(dir: std.fs.Dir) bool {
-    dir.access("src/scanner.c", .{}) catch return false;
+inline fn fileExists(b: *std.Build, filename: []const u8) bool {
+    const dir = b.build_root.handle;
+    dir.access(filename, .{}) catch return false;
     return true;
 }

cli/src/templates/build.zig.zon

@@ -1,10 +1,13 @@
 .{
-    .name = "tree-sitter-PARSER_NAME",
+    .name = .tree_sitter_PARSER_NAME,
     .version = "PARSER_VERSION",
-    .dependencies = .{ .@"tree-sitter" = .{
-        .url = "https://github.com/tree-sitter/zig-tree-sitter/archive/refs/tags/v0.25.0.tar.gz",
-        .hash = "12201a8d5e840678bbbf5128e605519c4024af422295d68e2ba2090e675328e5811d",
-    } },
+    .dependencies = .{
+        .tree_sitter = .{
+            .url = "git+https://github.com/tree-sitter/zig-tree-sitter#b4b72c903e69998fc88e27e154a5e3cc9166551b",
+            .hash = "tree_sitter-0.25.0-8heIf51vAQConvVIgvm-9mVIbqh7yabZYqPXfOpS3YoG",
+            .lazy = true,
+        },
+    },
     .paths = .{
         "build.zig",
         "build.zig.zon",

cli/src/templates/cmakelists.cmake

@@ -15,6 +15,8 @@ if(NOT ${TREE_SITTER_ABI_VERSION} MATCHES "^[0-9]+$")
     message(FATAL_ERROR "TREE_SITTER_ABI_VERSION must be an integer")
 endif()
 
+include(GNUInstallDirs)
+
 find_program(TREE_SITTER_CLI tree-sitter DOC "Tree-sitter CLI")
 
 add_custom_command(OUTPUT "${CMAKE_CURRENT_SOURCE_DIR}/src/parser.c"
@@ -47,13 +49,11 @@ set_target_properties(tree-sitter-KEBAB_PARSER_NAME
 configure_file(bindings/c/tree-sitter-KEBAB_PARSER_NAME.pc.in
                "${CMAKE_CURRENT_BINARY_DIR}/tree-sitter-KEBAB_PARSER_NAME.pc" @ONLY)
 
-include(GNUInstallDirs)
-
 install(DIRECTORY "${CMAKE_CURRENT_SOURCE_DIR}/bindings/c/tree_sitter"
         DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
         FILES_MATCHING PATTERN "*.h")
 install(FILES "${CMAKE_CURRENT_BINARY_DIR}/tree-sitter-KEBAB_PARSER_NAME.pc"
-        DESTINATION "${CMAKE_INSTALL_DATAROOTDIR}/pkgconfig")
+        DESTINATION "${CMAKE_INSTALL_LIBDIR}/pkgconfig")
 install(TARGETS tree-sitter-KEBAB_PARSER_NAME
         LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}")
 

cli/src/templates/gitattributes

@@ -37,5 +37,6 @@ Package.swift linguist-generated
 Package.resolved linguist-generated
 
 # Zig bindings
+bindings/zig/* linguist-generated
 build.zig linguist-generated
 build.zig.zon linguist-generated

cli/src/templates/gitignore

@@ -1,13 +1,16 @@
 # Rust artifacts
 target/
+Cargo.lock
 
 # Node artifacts
 build/
 prebuilds/
 node_modules/
+package-lock.json
 
 # Swift artifacts
 .build/
+Package.resolved
 
 # Go artifacts
 _obj/

cli/src/templates/lib.rs

@@ -1,7 +1,7 @@
-//! This crate provides CAMEL_PARSER_NAME language support for the [tree-sitter][] parsing library.
+//! This crate provides TITLE_PARSER_NAME language support for the [tree-sitter] parsing library.
 //!
-//! Typically, you will use the [LANGUAGE][] constant to add this language to a
-//! tree-sitter [Parser][], and then use the parser to parse some code:
+//! Typically, you will use the [`LANGUAGE`] constant to add this language to a
+//! tree-sitter [`Parser`], and then use the parser to parse some code:
 //!
 //! ```
 //! let code = r#"
@@ -15,7 +15,7 @@
 //! assert!(!tree.root_node().has_error());
 //! ```
 //!
-//! [Parser]: https://docs.rs/tree-sitter/*/tree_sitter/struct.Parser.html
+//! [`Parser`]: https://docs.rs/tree-sitter/RUST_BINDING_VERSION/tree_sitter/struct.Parser.html
 //! [tree-sitter]: https://tree-sitter.github.io/
 
 use tree_sitter_language::LanguageFn;
@@ -24,12 +24,10 @@ extern "C" {
     fn tree_sitter_PARSER_NAME() -> *const ();
 }
 
-/// The tree-sitter [`LanguageFn`][LanguageFn] for this grammar.
-///
-/// [LanguageFn]: https://docs.rs/tree-sitter-language/*/tree_sitter_language/struct.LanguageFn.html
+/// The tree-sitter [`LanguageFn`] for this grammar.
 pub const LANGUAGE: LanguageFn = unsafe { LanguageFn::from_raw(tree_sitter_PARSER_NAME) };
 
-/// The content of the [`node-types.json`][] file for this grammar.
+/// The content of the [`node-types.json`] file for this grammar.
 ///
 /// [`node-types.json`]: https://tree-sitter.github.io/tree-sitter/using-parsers/6-static-node-types
 pub const NODE_TYPES: &str = include_str!("../../src/node-types.json");

cli/src/templates/makefile

@@ -77,7 +77,9 @@ install: all
 	install -m755 lib$(LANGUAGE_NAME).$(SOEXT) '$(DESTDIR)$(LIBDIR)'/lib$(LANGUAGE_NAME).$(SOEXTVER)
 	ln -sf lib$(LANGUAGE_NAME).$(SOEXTVER) '$(DESTDIR)$(LIBDIR)'/lib$(LANGUAGE_NAME).$(SOEXTVER_MAJOR)
 	ln -sf lib$(LANGUAGE_NAME).$(SOEXTVER_MAJOR) '$(DESTDIR)$(LIBDIR)'/lib$(LANGUAGE_NAME).$(SOEXT)
+ifneq ($(wildcard queries/*.scm),)
 	install -m644 queries/*.scm '$(DESTDIR)$(DATADIR)'/tree-sitter/queries/KEBAB_PARSER_NAME
+endif
 
 uninstall:
 	$(RM) '$(DESTDIR)$(LIBDIR)'/lib$(LANGUAGE_NAME).a \

cli/src/templates/package.json

@@ -29,15 +29,16 @@
     "*.wasm"
   ],
   "dependencies": {
-    "node-addon-api": "^8.2.1",
-    "node-gyp-build": "^4.8.2"
+    "node-addon-api": "^8.5.0",
+    "node-gyp-build": "^4.8.4"
   },
   "devDependencies": {
     "prebuildify": "^6.0.1",
+    "tree-sitter": "^0.22.4",
     "tree-sitter-cli": "^CLI_VERSION"
   },
   "peerDependencies": {
-    "tree-sitter": "^0.21.1"
+    "tree-sitter": "^0.22.4"
   },
   "peerDependenciesMeta": {
     "tree-sitter": {

cli/src/templates/package.swift

@@ -14,7 +14,7 @@ let package = Package(
         .library(name: "PARSER_CLASS_NAME", targets: ["PARSER_CLASS_NAME"]),
     ],
     dependencies: [
-        .package(url: "https://github.com/tree-sitter/swift-tree-sitter", from: "0.8.0"),
+        .package(name: "SwiftTreeSitter", url: "https://github.com/tree-sitter/swift-tree-sitter", from: "0.9.0"),
     ],
     targets: [
         .target(

cli/src/templates/pyproject.toml

@@ -1,5 +1,5 @@
 [build-system]
-requires = ["setuptools>=42", "wheel"]
+requires = ["setuptools>=62.4.0", "wheel"]
 build-backend = "setuptools.build_meta"
 
 [project]

cli/src/templates/root.zig

@@ -1,19 +1,5 @@
-const testing = @import("std").testing;
+extern fn tree_sitter_PARSER_NAME() callconv(.c) *const anyopaque;
 
-const ts = @import("tree-sitter");
-const Language = ts.Language;
-const Parser = ts.Parser;
-
-pub extern fn tree_sitter_PARSER_NAME() callconv(.C) *const Language;
-
-pub export fn language() *const Language {
+pub fn language() *const anyopaque {
     return tree_sitter_PARSER_NAME();
 }
-
-test "can load grammar" {
-    const parser = Parser.create();
-    defer parser.destroy();
-    try testing.expectEqual(parser.setLanguage(language()), void{});
-    try testing.expectEqual(parser.getLanguage(), tree_sitter_PARSER_NAME());
-}
-

cli/src/templates/setup.py

@@ -1,31 +1,12 @@
 from os import path
-from platform import system
 from sysconfig import get_config_var
 
 from setuptools import Extension, find_packages, setup
 from setuptools.command.build import build
+from setuptools.command.build_ext import build_ext
 from setuptools.command.egg_info import egg_info
 from wheel.bdist_wheel import bdist_wheel
 
-sources = [
-    "bindings/python/tree_sitter_LOWER_PARSER_NAME/binding.c",
-    "src/parser.c",
-]
-if path.exists("src/scanner.c"):
-    sources.append("src/scanner.c")
-
-macros: list[tuple[str, str | None]] = [
-    ("PY_SSIZE_T_CLEAN", None),
-    ("TREE_SITTER_HIDE_SYMBOLS", None),
-]
-if limited_api := not get_config_var("Py_GIL_DISABLED"):
-    macros.append(("Py_LIMITED_API", "0x030A0000"))
-
-if system() != "Windows":
-    cflags = ["-std=c11", "-fvisibility=hidden"]
-else:
-    cflags = ["/std:c11", "/utf-8"]
-
 
 class Build(build):
     def run(self):
@@ -35,6 +16,19 @@ class Build(build):
         super().run()
 
 
+class BuildExt(build_ext):
+    def build_extension(self, ext: Extension):
+        if self.compiler.compiler_type != "msvc":
+            ext.extra_compile_args = ["-std=c11", "-fvisibility=hidden"]
+        else:
+            ext.extra_compile_args = ["/std:c11", "/utf-8"]
+        if path.exists("src/scanner.c"):
+            ext.sources.append("src/scanner.c")
+        if ext.py_limited_api:
+            ext.define_macros.append(("Py_LIMITED_API", "0x030A0000"))
+        super().build_extension(ext)
+
+
 class BdistWheel(bdist_wheel):
     def get_tag(self):
         python, abi, platform = super().get_tag()
@@ -61,15 +55,21 @@ setup(
     ext_modules=[
         Extension(
             name="_binding",
-            sources=sources,
-            extra_compile_args=cflags,
-            define_macros=macros,
+            sources=[
+                "bindings/python/tree_sitter_LOWER_PARSER_NAME/binding.c",
+                "src/parser.c",
+            ],
+            define_macros=[
+                ("PY_SSIZE_T_CLEAN", None),
+                ("TREE_SITTER_HIDE_SYMBOLS", None),
+            ],
             include_dirs=["src"],
-            py_limited_api=limited_api,
+            py_limited_api=not get_config_var("Py_GIL_DISABLED"),
         )
     ],
     cmdclass={
         "build": Build,
+        "build_ext": BuildExt,
         "bdist_wheel": BdistWheel,
         "egg_info": EggInfo,
     },

cli/src/templates/test.zig

@@ -0,0 +1,17 @@
+const testing = @import("std").testing;
+
+const ts = @import("tree-sitter");
+const root = @import("tree-sitter-PARSER_NAME");
+const Language = ts.Language;
+const Parser = ts.Parser;
+
+test "can load grammar" {
+    const parser = Parser.create();
+    defer parser.destroy();
+
+    const lang: *const ts.Language = @ptrCast(root.language());
+    defer lang.destroy();
+
+    try testing.expectEqual(void{}, parser.setLanguage(lang));
+    try testing.expectEqual(lang, parser.getLanguage());
+}

cli/src/templates/test_binding.py

@@ -1,12 +1,12 @@
 from unittest import TestCase
 
-import tree_sitter
+from tree_sitter import Language, Parser
 import tree_sitter_LOWER_PARSER_NAME
 
 
 class TestLanguage(TestCase):
     def test_can_load_grammar(self):
         try:
-            tree_sitter.Language(tree_sitter_LOWER_PARSER_NAME.language())
+            Parser(Language(tree_sitter_LOWER_PARSER_NAME.language()))
         except Exception:
             self.fail("Error loading TITLE_PARSER_NAME grammar")

cli/src/test_highlight.rs

@@ -172,7 +172,7 @@ pub fn iterate_assertions(
             let mut j = i;
             while let (false, Some(highlight)) = (passed, highlights.get(j)) {
                 end_column = position.column + length - 1;
-                if highlight.0.column > end_column {
+                if highlight.0.row >= position.row && highlight.0.column > end_column {
                     break 'highlight_loop;
                 }
 

cli/src/tests/async_context_test.rs

@@ -238,7 +238,7 @@ async fn yield_now() {
     SimpleYieldNow { yielded: false }.await;
 }
 
-pub fn noop_waker() -> Waker {
+pub const fn noop_waker() -> Waker {
     const VTABLE: RawWakerVTable = RawWakerVTable::new(
         // Cloning just returns a new no-op raw waker
         |_| RAW,

cli/src/tests/corpus_test.rs

@@ -23,7 +23,7 @@ use crate::{
 };
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_bash(seed: usize) {
+fn test_corpus_for_bash_language(seed: usize) {
     test_language_corpus(
         "bash",
         seed,
@@ -39,73 +39,77 @@ fn test_corpus_for_bash(seed: usize) {
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_c(seed: usize) {
+fn test_corpus_for_c_language(seed: usize) {
     test_language_corpus("c", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_cpp(seed: usize) {
+fn test_corpus_for_cpp_language(seed: usize) {
     test_language_corpus("cpp", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_embedded_template(seed: usize) {
+fn test_corpus_for_embedded_template_language(seed: usize) {
     test_language_corpus("embedded-template", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_go(seed: usize) {
+fn test_corpus_for_go_language(seed: usize) {
     test_language_corpus("go", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_html(seed: usize) {
+fn test_corpus_for_html_language(seed: usize) {
     test_language_corpus("html", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_java(seed: usize) {
-    test_language_corpus("java", seed, None, None);
+fn test_corpus_for_java_language(seed: usize) {
+    test_language_corpus(
+        "java",
+        seed,
+        Some(&["java - corpus - expressions - switch with unnamed pattern variable"]),
+        None,
+    );
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_javascript(seed: usize) {
+fn test_corpus_for_javascript_language(seed: usize) {
     test_language_corpus("javascript", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_json(seed: usize) {
+fn test_corpus_for_json_language(seed: usize) {
     test_language_corpus("json", seed, None, None);
 }
 
-#[ignore]
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_php(seed: usize) {
-    test_language_corpus("php", seed, None, None);
+fn test_corpus_for_php_language(seed: usize) {
+    test_language_corpus("php", seed, None, Some("php"));
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_python(seed: usize) {
+fn test_corpus_for_python_language(seed: usize) {
     test_language_corpus("python", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_ruby(seed: usize) {
+fn test_corpus_for_ruby_language(seed: usize) {
     test_language_corpus("ruby", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_rust(seed: usize) {
+fn test_corpus_for_rust_language(seed: usize) {
     test_language_corpus("rust", seed, None, None);
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_typescript(seed: usize) {
+fn test_corpus_for_typescript_language(seed: usize) {
     test_language_corpus("typescript", seed, None, Some("typescript"));
 }
 
 #[test_with_seed(retry=10, seed=*START_SEED, seed_fn=new_seed)]
-fn test_corpus_for_tsx(seed: usize) {
+fn test_corpus_for_tsx_language(seed: usize) {
     test_language_corpus("typescript", seed, None, Some("tsx"));
 }
 
@@ -239,8 +243,9 @@ pub fn test_language_corpus(
                 }
 
                 // Perform a random series of edits and reparse.
-                let mut undo_stack = Vec::new();
-                for _ in 0..=rand.unsigned(*EDIT_COUNT) {
+                let edit_count = rand.unsigned(*EDIT_COUNT);
+                let mut undo_stack = Vec::with_capacity(edit_count);
+                for _ in 0..=edit_count {
                     let edit = get_random_edit(&mut rand, &input);
                     undo_stack.push(invert_edit(&input, &edit));
                     perform_edit(&mut tree, &mut input, &edit).unwrap();
@@ -376,7 +381,7 @@ fn test_feature_corpus_files() {
                 let actual_message = e.to_string().replace("\r\n", "\n");
                 if expected_message != actual_message {
                     eprintln!(
-                        "Unexpected error message.\n\nExpected:\n\n{expected_message}\nActual:\n\n{actual_message}\n",
+                        "Unexpected error message.\n\nExpected:\n\n`{expected_message}`\nActual:\n\n`{actual_message}`\n",
                     );
                     failure_count += 1;
                 }

cli/src/tests/helpers/allocations.rs

@@ -108,7 +108,7 @@ unsafe extern "C" fn ts_record_realloc(ptr: *mut c_void, size: usize) -> *mut c_
     let result = realloc(ptr, size);
     if ptr.is_null() {
         record_alloc(result);
-    } else if ptr != result {
+    } else if !core::ptr::eq(ptr, result) {
         record_dealloc(ptr);
         record_alloc(result);
     }

cli/src/tests/helpers/fixtures.rs

@@ -6,11 +6,13 @@ use std::{
 
 use anyhow::Context;
 use tree_sitter::Language;
-use tree_sitter_generate::{ALLOC_HEADER, ARRAY_HEADER};
+use tree_sitter_generate::{load_grammar_file, ALLOC_HEADER, ARRAY_HEADER};
 use tree_sitter_highlight::HighlightConfiguration;
 use tree_sitter_loader::{CompileConfig, Loader};
 use tree_sitter_tags::TagsConfiguration;
 
+use crate::tests::generate_parser;
+
 include!("./dirs.rs");
 
 static TEST_LOADER: LazyLock<Loader> = LazyLock::new(|| {
@@ -21,6 +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
 }
@@ -40,6 +45,22 @@ pub fn get_language(name: &str) -> Language {
     TEST_LOADER.load_language_at_path(config).unwrap()
 }
 
+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_internal(&parser_name, &parser_code, Some(&grammar_dir_path), wasm)
+}
+
 pub fn get_language_queries_path(language_name: &str) -> PathBuf {
     GRAMMARS_DIR.join(language_name).join("queries")
 }
@@ -78,6 +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();
 
@@ -127,5 +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/highlight_test.rs

@@ -350,12 +350,11 @@ fn test_highlighting_empty_lines() {
 fn test_highlighting_carriage_returns() {
     let source = "a = \"a\rb\"\r\nb\r";
 
-    // FIXME(amaanq): figure why this changed w/ JS's grammar changes
     assert_eq!(
         &to_html(source, &JS_HIGHLIGHT).unwrap(),
         &[
-            "<span class=variable>a</span> <span class=operator>=</span> <span class=string>&quot;a<span class=variable>b</span>&quot;</span>\n",
-            "<span class=variable>b</span>\n",
+            "<span class=variable>a</span> <span class=operator>=</span> <span class=string>&quot;a<span class=carriage-return></span><span class=variable>b</span>&quot;</span>\n",
+            "<span class=variable>b</span><span class=carriage-return></span>\n",
         ],
     );
 }
@@ -598,7 +597,7 @@ fn test_highlighting_via_c_api() {
     let output_line_offsets =
         unsafe { slice::from_raw_parts(output_line_offsets, output_line_count as usize) };
 
-    let mut lines = Vec::new();
+    let mut lines = Vec::with_capacity(output_line_count as usize);
     for i in 0..(output_line_count as usize) {
         let line_start = output_line_offsets[i] as usize;
         let line_end = output_line_offsets

cli/src/tests/language_test.rs

@@ -152,6 +152,7 @@ fn test_supertypes() {
                         "_literal_pattern",
                         "captured_pattern",
                         "const_block",
+                        "generic_pattern",
                         "identifier",
                         "macro_invocation",
                         "mut_pattern",

cli/src/tests/node_test.rs

@@ -6,7 +6,10 @@ use super::{
     helpers::fixtures::{fixtures_dir, get_language, get_test_language},
     Rand,
 };
-use crate::{parse::perform_edit, tests::generate_parser};
+use crate::{
+    parse::perform_edit,
+    tests::{generate_parser, helpers::fixtures::get_test_fixture_language},
+};
 
 const JSON_EXAMPLE: &str = r#"
 
@@ -308,19 +311,8 @@ fn test_parent_of_zero_width_node() {
 
 #[test]
 fn test_next_sibling_of_zero_width_node() {
-    let grammar_json = load_grammar_file(
-        &fixtures_dir()
-            .join("test_grammars")
-            .join("next_sibling_from_zwt")
-            .join("grammar.js"),
-        None,
-    )
-    .unwrap();
-
-    let (parser_name, parser_code) = generate_parser(&grammar_json).unwrap();
-
     let mut parser = Parser::new();
-    let language = get_test_language(&parser_name, &parser_code, None);
+    let language = get_test_fixture_language("next_sibling_from_zwt");
     parser.set_language(&language).unwrap();
 
     let tree = parser.parse("abdef", None).unwrap();

cli/src/tests/parser_test.rs

@@ -6,7 +6,6 @@ use std::{
 use tree_sitter::{
     Decode, IncludedRangesError, InputEdit, LogType, ParseOptions, ParseState, Parser, Point, Range,
 };
-use tree_sitter_generate::load_grammar_file;
 use tree_sitter_proc_macro::retry;
 
 use super::helpers::{
@@ -17,7 +16,7 @@ use super::helpers::{
 use crate::{
     fuzz::edits::Edit,
     parse::perform_edit,
-    tests::{generate_parser, helpers::fixtures::fixtures_dir, invert_edit},
+    tests::{generate_parser, helpers::fixtures::get_test_fixture_language, invert_edit},
 };
 
 #[test]
@@ -88,7 +87,6 @@ fn test_parsing_with_logging() {
 }
 
 #[test]
-#[cfg(unix)]
 fn test_parsing_with_debug_graph_enabled() {
     use std::io::{BufRead, BufReader, Seek};
 
@@ -482,15 +480,9 @@ fn test_parsing_empty_file_with_reused_tree() {
 
 #[test]
 fn test_parsing_after_editing_tree_that_depends_on_column_values() {
-    let dir = fixtures_dir()
-        .join("test_grammars")
-        .join("uses_current_column");
-    let grammar_json = load_grammar_file(&dir.join("grammar.js"), None).unwrap();
-    let (grammar_name, parser_code) = generate_parser(&grammar_json).unwrap();
-
     let mut parser = Parser::new();
     parser
-        .set_language(&get_test_language(&grammar_name, &parser_code, Some(&dir)))
+        .set_language(&get_test_fixture_language("uses_current_column"))
         .unwrap();
 
     let mut code = b"
@@ -559,16 +551,9 @@ h + i
 
 #[test]
 fn test_parsing_after_editing_tree_that_depends_on_column_position() {
-    let dir = fixtures_dir()
-        .join("test_grammars")
-        .join("depends_on_column");
-
-    let grammar_json = load_grammar_file(&dir.join("grammar.js"), None).unwrap();
-    let (grammar_name, parser_code) = generate_parser(grammar_json.as_str()).unwrap();
-
     let mut parser = Parser::new();
     parser
-        .set_language(&get_test_language(&grammar_name, &parser_code, Some(&dir)))
+        .set_language(&get_test_fixture_language("depends_on_column"))
         .unwrap();
 
     let mut code = b"\n x".to_vec();
@@ -1702,13 +1687,9 @@ if foo && bar || baz {}
 
 #[test]
 fn test_parsing_with_scanner_logging() {
-    let dir = fixtures_dir().join("test_grammars").join("external_tokens");
-    let grammar_json = load_grammar_file(&dir.join("grammar.js"), None).unwrap();
-    let (grammar_name, parser_code) = generate_parser(&grammar_json).unwrap();
-
     let mut parser = Parser::new();
     parser
-        .set_language(&get_test_language(&grammar_name, &parser_code, Some(&dir)))
+        .set_language(&get_test_fixture_language("external_tokens"))
         .unwrap();
 
     let mut found = false;
@@ -1726,13 +1707,9 @@ fn test_parsing_with_scanner_logging() {
 
 #[test]
 fn test_parsing_get_column_at_eof() {
-    let dir = fixtures_dir().join("test_grammars").join("get_col_eof");
-    let grammar_json = load_grammar_file(&dir.join("grammar.js"), None).unwrap();
-    let (grammar_name, parser_code) = generate_parser(&grammar_json).unwrap();
-
     let mut parser = Parser::new();
     parser
-        .set_language(&get_test_language(&grammar_name, &parser_code, Some(&dir)))
+        .set_language(&get_test_fixture_language("get_col_eof"))
         .unwrap();
 
     parser.parse("a", None).unwrap();

cli/src/tests/query_test.rs

@@ -17,7 +17,10 @@ use super::helpers::{
 };
 use crate::tests::{
     generate_parser,
-    helpers::query_helpers::{collect_captures, collect_matches},
+    helpers::{
+        fixtures::get_test_fixture_language,
+        query_helpers::{collect_captures, collect_matches},
+    },
     ITERATION_COUNT,
 };
 
@@ -330,6 +333,16 @@ fn test_query_errors_on_invalid_symbols() {
                 message: "alternatives".to_string()
             }
         );
+        assert_eq!(
+            Query::new(&language, "fakefield: (identifier)").unwrap_err(),
+            QueryError {
+                row: 0,
+                offset: 0,
+                column: 0,
+                kind: QueryErrorKind::Field,
+                message: "fakefield".to_string()
+            }
+        );
     });
 }
 
@@ -2978,6 +2991,61 @@ fn test_query_matches_with_deeply_nested_patterns_with_fields() {
     });
 }
 
+#[test]
+fn test_query_matches_with_alternations_and_predicates() {
+    allocations::record(|| {
+        let language = get_language("java");
+        let query = Query::new(
+            &language,
+            "
+            (block
+                [
+                    (local_variable_declaration
+                        (variable_declarator
+                            (identifier) @def.a
+                            (string_literal) @lit.a
+                        )
+                    )
+                    (local_variable_declaration
+                        (variable_declarator
+                            (identifier) @def.b
+                            (null_literal) @lit.b
+                        )
+                    )
+                ]
+                (expression_statement
+                    (method_invocation [
+                        (argument_list
+                            (identifier) @ref.a
+                            (string_literal)
+                        )
+                        (argument_list
+                            (null_literal)
+                            (identifier) @ref.b
+                        )
+                    ])
+                )
+                (#eq? @def.a @ref.a )
+                (#eq? @def.b @ref.b )
+            )
+            ",
+        )
+        .unwrap();
+
+        assert_query_matches(
+            &language,
+            &query,
+            r#"
+            void test() {
+                int a = "foo";
+                f(null, b);
+            }
+            "#,
+            &[],
+        );
+    });
+}
+
 #[test]
 fn test_query_matches_with_indefinite_step_containing_no_captures() {
     allocations::record(|| {
@@ -5621,3 +5689,63 @@ const foo = [
     assert_eq!(matches.len(), 1);
     assert_eq!(matches[0].1, captures);
 }
+
+#[test]
+fn test_query_with_predicate_causing_oob_access() {
+    let language = get_language("rust");
+
+    let query = "(call_expression
+     function: (scoped_identifier
+       path: (scoped_identifier (identifier) @_regex (#any-of? @_regex \"Regex\" \"RegexBuilder\") .))
+     (#set! injection.language \"regex\"))";
+    Query::new(&language, query).unwrap();
+}
+
+#[test]
+fn test_query_with_anonymous_error_node() {
+    let language = get_test_fixture_language("anonymous_error");
+    let mut parser = Parser::new();
+    parser.set_language(&language).unwrap();
+
+    let source = "ERROR";
+
+    let tree = parser.parse(source, None).unwrap();
+    let query = Query::new(
+        &language,
+        r#"
+          "ERROR" @error
+          (document "ERROR" @error)
+        "#,
+    )
+    .unwrap();
+    let mut cursor = QueryCursor::new();
+    let matches = cursor.matches(&query, tree.root_node(), source.as_bytes());
+    let matches = collect_matches(matches, &query, source);
+
+    assert_eq!(
+        matches,
+        vec![(1, vec![("error", "ERROR")]), (0, vec![("error", "ERROR")])]
+    );
+}
+
+#[test]
+fn test_query_allows_error_nodes_with_children() {
+    allocations::record(|| {
+        let language = get_language("cpp");
+
+        let code = "SomeStruct foo{.bar{}};";
+
+        let mut parser = Parser::new();
+        parser.set_language(&language).unwrap();
+
+        let tree = parser.parse(code, None).unwrap();
+        let root = tree.root_node();
+
+        let query = Query::new(&language, "(initializer_list (ERROR) @error)").unwrap();
+        let mut cursor = QueryCursor::new();
+
+        let matches = cursor.matches(&query, root, code.as_bytes());
+        let matches = collect_matches(matches, &query, code);
+        assert_eq!(matches, &[(0, vec![("error", ".bar")])]);
+    });
+}

cli/src/tests/tags_test.rs

@@ -401,8 +401,11 @@ fn test_tags_via_c_api() {
 
         let syntax_types = unsafe {
             let mut len = 0;
-            let ptr =
-                c::ts_tagger_syntax_kinds_for_scope_name(tagger, c_scope_name.as_ptr(), &mut len);
+            let ptr = c::ts_tagger_syntax_kinds_for_scope_name(
+                tagger,
+                c_scope_name.as_ptr(),
+                &raw mut len,
+            );
             slice::from_raw_parts(ptr, len as usize)
                 .iter()
                 .map(|i| CStr::from_ptr(*i).to_str().unwrap())

cli/src/tests/text_provider_test.rs

@@ -107,6 +107,19 @@ fn test_text_provider_for_arc_of_bytes_slice() {
     check_parsing(text.clone(), text.as_ref());
 }
 
+#[test]
+fn test_text_provider_for_vec_utf16_text() {
+    let source_text = "你好".encode_utf16().collect::<Vec<_>>();
+
+    let language = get_language("c");
+    let mut parser = Parser::new();
+    parser.set_language(&language).unwrap();
+    let tree = parser.parse_utf16_le(&source_text, None).unwrap();
+
+    let tree_text = tree.root_node().utf16_text(&source_text);
+    assert_eq!(source_text, tree_text);
+}
+
 #[test]
 fn test_text_provider_callback_with_str_slice() {
     let text: &str = "// comment";

cli/src/tests/tree_test.rs

@@ -3,7 +3,11 @@ use std::str;
 use tree_sitter::{InputEdit, Parser, Point, Range, Tree};
 
 use super::helpers::fixtures::get_language;
-use crate::{fuzz::edits::Edit, parse::perform_edit, tests::invert_edit};
+use crate::{
+    fuzz::edits::Edit,
+    parse::perform_edit,
+    tests::{helpers::fixtures::get_test_fixture_language, invert_edit},
+};
 
 #[test]
 fn test_tree_edit() {
@@ -377,6 +381,40 @@ fn test_tree_cursor() {
     assert_eq!(copy.node().kind(), "struct_item");
 }
 
+#[test]
+fn test_tree_cursor_previous_sibling_with_aliases() {
+    let mut parser = Parser::new();
+    parser
+        .set_language(&get_test_fixture_language("aliases_in_root"))
+        .unwrap();
+
+    let text = "# comment\n# \nfoo foo";
+    let tree = parser.parse(text, None).unwrap();
+    let mut cursor = tree.walk();
+    assert_eq!(cursor.node().kind(), "document");
+
+    cursor.goto_first_child();
+    assert_eq!(cursor.node().kind(), "comment");
+
+    assert!(cursor.goto_next_sibling());
+    assert_eq!(cursor.node().kind(), "comment");
+
+    assert!(cursor.goto_next_sibling());
+    assert_eq!(cursor.node().kind(), "bar");
+
+    assert!(cursor.goto_previous_sibling());
+    assert_eq!(cursor.node().kind(), "comment");
+
+    assert!(cursor.goto_previous_sibling());
+    assert_eq!(cursor.node().kind(), "comment");
+
+    assert!(cursor.goto_next_sibling());
+    assert_eq!(cursor.node().kind(), "comment");
+
+    assert!(cursor.goto_next_sibling());
+    assert_eq!(cursor.node().kind(), "bar");
+}
+
 #[test]
 fn test_tree_cursor_previous_sibling() {
     let mut parser = Parser::new();

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;
 
@@ -23,10 +23,18 @@ pub fn load_language_wasm_file(language_dir: &Path) -> Result<(String, Vec<u8>)>
 pub fn get_grammar_name(language_dir: &Path) -> Result<String> {
     let src_dir = language_dir.join("src");
     let grammar_json_path = src_dir.join("grammar.json");
-    let grammar_json = fs::read_to_string(&grammar_json_path)
-        .with_context(|| format!("Failed to read grammar file {grammar_json_path:?}"))?;
-    let grammar: GrammarJSON = serde_json::from_str(&grammar_json)
-        .with_context(|| format!("Failed to parse grammar file {grammar_json_path:?}"))?;
+    let grammar_json = fs::read_to_string(&grammar_json_path).with_context(|| {
+        format!(
+            "Failed to read grammar file {}",
+            grammar_json_path.display()
+        )
+    })?;
+    let grammar: GrammarJSON = serde_json::from_str(&grammar_json).with_context(|| {
+        format!(
+            "Failed to parse grammar file {}",
+            grammar_json_path.display()
+        )
+    })?;
     Ok(grammar.name)
 }
 
@@ -38,7 +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");

docs/src/assets/schemas/config.schema.json

@@ -22,8 +22,15 @@
             "examples": [
               "Rust",
               "HTML"
-            ],
-            "$comment": "This is used in the description and the class names."
+            ]
+          },
+          "title": {
+            "type": "string",
+            "description": "The title of the language.",
+            "examples": [
+              "Rust",
+              "HTML"
+            ]
           },
           "scope": {
             "type": "string",
@@ -237,9 +244,7 @@
       "properties": {
         "c": {
           "type": "boolean",
-          "default": true,
-          "const": true,
-          "$comment": "Always generated"
+          "default": true
         },
         "go": {
           "type": "boolean",
@@ -255,9 +260,7 @@
         },
         "node": {
           "type": "boolean",
-          "default": true,
-          "const": true,
-          "$comment": "Always generated (for now)"
+          "default": true
         },
         "python": {
           "type": "boolean",
@@ -265,9 +268,7 @@
         },
         "rust": {
           "type": "boolean",
-          "default": true,
-          "const": true,
-          "$comment": "Always generated"
+          "default": true
         },
         "swift": {
           "type": "boolean",

docs/src/assets/schemas/grammar.schema.json

@@ -246,6 +246,21 @@
       "required": ["type", "content"]
     },
 
+    "reserved-rule": {
+      "type": "object",
+      "properties": {
+        "type": {
+          "type": "string",
+          "const": "RESERVED"
+        },
+        "context_name": { "type": "string" },
+        "content": {
+          "$ref": "#/definitions/rule"
+        }
+      },
+      "required": ["type", "context_name", "content"]
+    },
+
     "token-rule": {
       "type": "object",
       "properties": {
@@ -313,6 +328,7 @@
         { "$ref": "#/definitions/choice-rule" },
         { "$ref": "#/definitions/repeat1-rule" },
         { "$ref": "#/definitions/repeat-rule" },
+        { "$ref": "#/definitions/reserved-rule" },
         { "$ref": "#/definitions/token-rule" },
         { "$ref": "#/definitions/field-rule" },
         { "$ref": "#/definitions/prec-rule" }

docs/src/cli/parse.md

@@ -66,7 +66,7 @@ Suppress main output.
 
 ### `--edits <EDITS>...`
 
-Apply edits after parsing the file. Edits are in the form of `row, col delcount insert_text` where row and col are 0-indexed.
+Apply edits after parsing the file. Edits are in the form of `row,col|position delcount insert_text` where row and col, or position are 0-indexed.
 
 ### `--encoding <ENCODING>`
 

highlight/src/lib.rs

@@ -143,6 +143,8 @@ pub struct HtmlRenderer {
     pub html: Vec<u8>,
     pub line_offsets: Vec<u32>,
     carriage_return_highlight: Option<Highlight>,
+    // The offset in `self.html` of the last carriage return.
+    last_carriage_return: Option<usize>,
 }
 
 #[derive(Debug)]
@@ -1090,6 +1092,7 @@ impl HtmlRenderer {
             html: Vec::with_capacity(BUFFER_HTML_RESERVE_CAPACITY),
             line_offsets: Vec::with_capacity(BUFFER_LINES_RESERVE_CAPACITY),
             carriage_return_highlight: None,
+            last_carriage_return: None,
         };
         result.line_offsets.push(0);
         result
@@ -1131,6 +1134,9 @@ impl HtmlRenderer {
                 Err(a) => return Err(a),
             }
         }
+        if let Some(offset) = self.last_carriage_return.take() {
+            self.add_carriage_return(offset, attribute_callback);
+        }
         if self.html.last() != Some(&b'\n') {
             self.html.push(b'\n');
         }
@@ -1155,14 +1161,21 @@ impl HtmlRenderer {
             })
     }
 
-    fn add_carriage_return<F>(&mut self, attribute_callback: &F)
+    fn add_carriage_return<F>(&mut self, offset: usize, attribute_callback: &F)
     where
         F: Fn(Highlight, &mut Vec<u8>),
     {
         if let Some(highlight) = self.carriage_return_highlight {
+            // If a CR is the last character in a `HighlightEvent::Source`
+            // region, then we don't know until the next `Source` event or EOF
+            // whether it is part of CRLF or on its own. To avoid unbounded
+            // lookahead, save the offset of the CR and insert there now that we
+            // know.
+            let rest = self.html.split_off(offset);
             self.html.extend(b"<span ");
             (attribute_callback)(highlight, &mut self.html);
             self.html.extend(b"></span>");
+            self.html.extend(rest);
         }
     }
 
@@ -1194,19 +1207,17 @@ impl HtmlRenderer {
             }
         }
 
-        let mut last_char_was_cr = false;
         for c in LossyUtf8::new(src).flat_map(|p| p.bytes()) {
             // Don't render carriage return characters, but allow lone carriage returns (not
             // followed by line feeds) to be styled via the attribute callback.
             if c == b'\r' {
-                last_char_was_cr = true;
+                self.last_carriage_return = Some(self.html.len());
                 continue;
             }
-            if last_char_was_cr {
+            if let Some(offset) = self.last_carriage_return.take() {
                 if c != b'\n' {
-                    self.add_carriage_return(attribute_callback);
+                    self.add_carriage_return(offset, attribute_callback);
                 }
-                last_char_was_cr = false;
             }
 
             // At line boundaries, close and re-open all of the open tags.

lib/CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.13)
 
 project(tree-sitter
-        VERSION "0.25.1"
+        VERSION "0.25.10"
         DESCRIPTION "An incremental parsing system for programming tools"
         HOMEPAGE_URL "https://tree-sitter.github.io/tree-sitter/"
         LANGUAGES C)
@@ -81,15 +81,15 @@ set_target_properties(tree-sitter
                       SOVERSION "${PROJECT_VERSION_MAJOR}.${PROJECT_VERSION_MINOR}"
                       DEFINE_SYMBOL "")
 
-target_compile_definitions(tree-sitter PRIVATE _POSIX_C_SOURCE=200112L _DEFAULT_SOURCE)
-
-configure_file(tree-sitter.pc.in "${CMAKE_CURRENT_BINARY_DIR}/tree-sitter.pc" @ONLY)
+target_compile_definitions(tree-sitter PRIVATE _POSIX_C_SOURCE=200112L _DEFAULT_SOURCE _DARWIN_C_SOURCE)
 
 include(GNUInstallDirs)
 
+configure_file(tree-sitter.pc.in "${CMAKE_CURRENT_BINARY_DIR}/tree-sitter.pc" @ONLY)
+
 install(FILES include/tree_sitter/api.h
         DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}/tree_sitter")
 install(FILES "${CMAKE_CURRENT_BINARY_DIR}/tree-sitter.pc"
-        DESTINATION "${CMAKE_INSTALL_DATAROOTDIR}/pkgconfig")
+        DESTINATION "${CMAKE_INSTALL_LIBDIR}/pkgconfig")
 install(TARGETS tree-sitter
         LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}")

lib/binding_rust/build.rs

@@ -43,6 +43,7 @@ fn main() {
         .include(&include_path)
         .define("_POSIX_C_SOURCE", "200112L")
         .define("_DEFAULT_SOURCE", None)
+        .define("_DARWIN_C_SOURCE", None)
         .warnings(false)
         .file(src_path.join("lib.c"))
         .compile("tree-sitter");
@@ -112,7 +113,10 @@ fn generate_bindings(out_dir: &std::path::Path) {
         .expect("Failed to generate bindings");
 
     let bindings_rs = out_dir.join("bindings.rs");
-    bindings
-        .write_to_file(&bindings_rs)
-        .unwrap_or_else(|_| panic!("Failed to write bindings into path: {bindings_rs:?}"));
+    bindings.write_to_file(&bindings_rs).unwrap_or_else(|_| {
+        panic!(
+            "Failed to write bindings into path: {}",
+            bindings_rs.display()
+        )
+    });
 }

lib/binding_rust/ffi.rs

@@ -16,6 +16,7 @@ extern "C" {
 }
 
 #[cfg(windows)]
+#[cfg(feature = "std")]
 extern "C" {
     pub(crate) fn _ts_dup(handle: *mut std::os::raw::c_void) -> std::os::raw::c_int;
 }

lib/binding_rust/lib.rs

@@ -699,8 +699,7 @@ impl Parser {
             drop(unsafe { Box::from_raw(prev_logger.payload.cast::<Logger>()) });
         }
 
-        let c_logger;
-        if let Some(logger) = logger {
+        let c_logger = if let Some(logger) = logger {
             let container = Box::new(logger);
 
             unsafe extern "C" fn log(
@@ -721,16 +720,16 @@ impl Parser {
 
             let raw_container = Box::into_raw(container);
 
-            c_logger = ffi::TSLogger {
+            ffi::TSLogger {
                 payload: raw_container.cast::<c_void>(),
                 log: Some(log),
-            };
+            }
         } else {
-            c_logger = ffi::TSLogger {
+            ffi::TSLogger {
                 payload: ptr::null_mut(),
                 log: None,
-            };
-        }
+            }
+        };
 
         unsafe { ffi::ts_parser_set_logger(self.0.as_ptr(), c_logger) };
     }
@@ -1222,7 +1221,7 @@ impl Parser {
             len: u32,
             code_point: *mut i32,
         ) -> u32 {
-            let (c, len) = D::decode(std::slice::from_raw_parts(data, len as usize));
+            let (c, len) = D::decode(core::slice::from_raw_parts(data, len as usize));
             if let Some(code_point) = code_point.as_mut() {
                 *code_point = c;
             }
@@ -1422,7 +1421,7 @@ impl Parser {
         if let Some(flag) = flag {
             ffi::ts_parser_set_cancellation_flag(
                 self.0.as_ptr(),
-                std::ptr::from_ref::<AtomicUsize>(flag).cast::<usize>(),
+                core::ptr::from_ref::<AtomicUsize>(flag).cast::<usize>(),
             );
         } else {
             ffi::ts_parser_set_cancellation_flag(self.0.as_ptr(), ptr::null());
@@ -1432,12 +1431,21 @@ impl Parser {
 
 impl Drop for Parser {
     fn drop(&mut self) {
-        self.stop_printing_dot_graphs();
+        #[cfg(feature = "std")]
+        #[cfg(not(target_os = "wasi"))]
+        {
+            self.stop_printing_dot_graphs();
+        }
         self.set_logger(None);
         unsafe { ffi::ts_parser_delete(self.0.as_ptr()) }
     }
 }
 
+#[cfg(windows)]
+extern "C" {
+    fn _open_osfhandle(osfhandle: isize, flags: core::ffi::c_int) -> core::ffi::c_int;
+}
+
 impl Tree {
     /// Get the root node of the syntax tree.
     #[doc(alias = "ts_tree_root_node")]
@@ -1547,7 +1555,8 @@ impl Tree {
         #[cfg(windows)]
         {
             let handle = file.as_raw_handle();
-            unsafe { ffi::ts_tree_print_dot_graph(self.0.as_ptr(), handle as i32) }
+            let fd = unsafe { _open_osfhandle(handle as isize, 0) };
+            unsafe { ffi::ts_tree_print_dot_graph(self.0.as_ptr(), fd) }
         }
     }
 }
@@ -2058,7 +2067,7 @@ impl<'tree> Node<'tree> {
 
     #[must_use]
     pub fn utf16_text<'a>(&self, source: &'a [u16]) -> &'a [u16] {
-        &source[self.start_byte()..self.end_byte()]
+        &source[self.start_byte() / 2..self.end_byte() / 2]
     }
 
     /// Create a new [`TreeCursor`] starting from this node.
@@ -2087,7 +2096,7 @@ impl<'tree> Node<'tree> {
 
 impl PartialEq for Node<'_> {
     fn eq(&self, other: &Self) -> bool {
-        self.0.id == other.0.id
+        core::ptr::eq(self.0.id, other.0.id)
     }
 }
 
@@ -2440,7 +2449,7 @@ impl Query {
                 // Error types that report names
                 ffi::TSQueryErrorNodeType | ffi::TSQueryErrorField | ffi::TSQueryErrorCapture => {
                     let suffix = source.split_at(offset).1;
-                    let in_quotes = source.as_bytes()[offset - 1] == b'"';
+                    let in_quotes = offset > 0 && source.as_bytes()[offset - 1] == b'"';
                     let mut backslashes = 0;
                     let end_offset = suffix
                         .find(|c| {
@@ -3349,9 +3358,11 @@ impl<'tree> QueryMatch<'_, 'tree> {
             .iter()
             .all(|predicate| match predicate {
                 TextPredicateCapture::EqCapture(i, j, is_positive, match_all_nodes) => {
-                    let mut nodes_1 = self.nodes_for_capture_index(*i);
-                    let mut nodes_2 = self.nodes_for_capture_index(*j);
-                    while let (Some(node1), Some(node2)) = (nodes_1.next(), nodes_2.next()) {
+                    let mut nodes_1 = self.nodes_for_capture_index(*i).peekable();
+                    let mut nodes_2 = self.nodes_for_capture_index(*j).peekable();
+                    while nodes_1.peek().is_some() && nodes_2.peek().is_some() {
+                        let node1 = nodes_1.next().unwrap();
+                        let node2 = nodes_2.next().unwrap();
                         let mut text1 = text_provider.text(node1);
                         let mut text2 = text_provider.text(node2);
                         let text1 = node_text1.get_text(&mut text1);

lib/binding_web/lib/tree-sitter.c

@@ -44,17 +44,22 @@ static inline void marshal_node(const void **buffer, TSNode node) {
   buffer[4] = (const void *)node.context[3];
 }
 
-static inline TSNode unmarshal_node(const TSTree *tree) {
+static inline TSNode unmarshal_node_at(const TSTree *tree, uint32_t index) {
   TSNode node;
-  node.id = TRANSFER_BUFFER[0];
-  node.context[0] = code_unit_to_byte((uint32_t)TRANSFER_BUFFER[1]);
-  node.context[1] = (uint32_t)TRANSFER_BUFFER[2];
-  node.context[2] = code_unit_to_byte((uint32_t)TRANSFER_BUFFER[3]);
-  node.context[3] = (uint32_t)TRANSFER_BUFFER[4];
+  const void **buffer = TRANSFER_BUFFER + index * SIZE_OF_NODE;
+  node.id = buffer[0];
+  node.context[0] = code_unit_to_byte((uint32_t)buffer[1]);
+  node.context[1] = (uint32_t)buffer[2];
+  node.context[2] = code_unit_to_byte((uint32_t)buffer[3]);
+  node.context[3] = (uint32_t)buffer[4];
   node.tree = tree;
   return node;
 }
 
+static inline TSNode unmarshal_node(const TSTree *tree) {
+  return unmarshal_node_at(tree, 0);
+}
+
 static inline void marshal_cursor(const TSTreeCursor *cursor) {
   TRANSFER_BUFFER[0] = cursor->id;
   TRANSFER_BUFFER[1] = (const void *)cursor->context[0];
@@ -616,7 +621,7 @@ void ts_node_parent_wasm(const TSTree *tree) {
 
 void ts_node_child_with_descendant_wasm(const TSTree *tree) {
   TSNode node = unmarshal_node(tree);
-  TSNode descendant = unmarshal_node(tree);
+  TSNode descendant = unmarshal_node_at(tree, 1);
   marshal_node(TRANSFER_BUFFER, ts_node_child_with_descendant(node, descendant));
 }
 

lib/binding_web/lib/tree-sitter.d.ts

@@ -50,7 +50,7 @@ declare namespace RuntimeExports {
     function setValue(ptr: number, value: number, type?: string): void;
     let currentParseCallback: ((index: number, position: {row: number, column: number}) => string | undefined) | null;
     let currentLogCallback: ((message: string, isLex: boolean) => void) | null;
-    let currentProgressCallback: ((state: {currentOffset: number}) => void) | null;
+    let currentProgressCallback: ((state: {currentOffset: number, hasError: boolean}) => void) | null;
     let currentQueryProgressCallback: ((state: {currentOffset: number}) => void) | null;
     let HEAPF32: Float32Array;
     let HEAPF64: Float64Array;

lib/binding_web/package.json

@@ -1,9 +1,12 @@
 {
   "name": "web-tree-sitter",
-  "version": "0.25.1",
+  "version": "0.25.10",
   "description": "Tree-sitter bindings for the web",
-  "repository": "https://github.com/tree-sitter/tree-sitter",
-  "homepage": "https://github.com/tree-sitter/tree-sitter/tree/master/lib/binding_web",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/tree-sitter/tree-sitter.git",
+    "directory": "lib/binding_web"
+  },
   "license": "MIT",
   "author": {
     "name": "Max Brunsfeld",
@@ -19,12 +22,16 @@
   "exports": {
     ".": {
       "import": "./tree-sitter.js",
-      "require": "./tree-sitter.cjs"
+      "require": "./tree-sitter.cjs",
+      "types": "./web-tree-sitter.d.ts"
     },
+    "./tree-sitter.wasm": "./tree-sitter.wasm",
     "./debug": {
       "import": "./debug/tree-sitter.js",
-      "require": "./debug/tree-sitter.cjs"
-    }
+      "require": "./debug/tree-sitter.cjs",
+      "types": "./web-tree-sitter.d.ts"
+    },
+    "./debug/tree-sitter.wasm": "./debug/tree-sitter.wasm"
   },
   "types": "web-tree-sitter.d.ts",
   "keywords": [
@@ -54,10 +61,9 @@
     "lib/*.h"
   ],
   "devDependencies": {
-    "@eslint/js": "^9.19.0",
-    "@types/emscripten": "^1.40.0",
-    "@types/node": "^22.12.0",
-    "@vitest/coverage-v8": "^3.0.4",
+    "@eslint/js": "^9.20.0",
+    "@types/node": "^22.13.1",
+    "@vitest/coverage-v8": "^3.0.5",
     "dts-buddy": "^0.5.4",
     "esbuild": "^0.24.2",
     "eslint": "^9.19.0",
@@ -67,8 +73,16 @@
     "typescript-eslint": "^8.22.0",
     "vitest": "^3.0.4"
   },
+  "peerDependencies": {
+    "@types/emscripten": "^1.40.0"
+  },
+  "peerDependenciesMeta": {
+    "@types/emscripten": {
+      "optional": true
+    }
+  },
   "scripts": {
-    "build:ts": "node script/build.js",
+    "build:ts": "tsc -b . && node script/build.js",
     "build:wasm": "cd ../../ && cargo xtask build-wasm",
     "build:wasm:debug": "cd ../../ && cargo xtask build-wasm --debug",
     "build": "npm run build:wasm && npm run build:ts",

lib/binding_web/src/index.ts

@@ -1,4 +1,4 @@
-export {
+export type {
   Point,
   Range,
   Edit,
@@ -7,8 +7,8 @@ export {
   LogCallback,
 } from './constants';
 export {
-  ParseOptions,
-  ParseState,
+  type ParseOptions,
+  type ParseState,
   LANGUAGE_VERSION,
   MIN_COMPATIBLE_VERSION,
   Parser,
@@ -18,14 +18,14 @@ export { Tree } from './tree';
 export { Node } from './node';
 export { TreeCursor } from './tree_cursor';
 export {
-  QueryOptions,
-  QueryState,
-  QueryProperties,
-  QueryPredicate,
-  QueryCapture,
-  QueryMatch,
+  type QueryOptions,
+  type QueryState,
+  type QueryProperties,
+  type QueryPredicate,
+  type QueryCapture,
+  type QueryMatch,
   CaptureQuantifier,
-  PredicateStep,
+  type PredicateStep,
   Query,
-}  from './query';
+} from './query';
 export { LookaheadIterator } from './lookahead_iterator';

lib/binding_web/src/language.ts

@@ -6,7 +6,7 @@ import { Query } from './query';
 
 const LANGUAGE_FUNCTION_REGEX = /^tree_sitter_\w+$/;
 
-export class LanguageMetadata {
+export interface LanguageMetadata {
   readonly major_version: number;
   readonly minor_version: number;
   readonly patch_version: number;
@@ -261,8 +261,7 @@ export class Language {
     } else {
       // eslint-disable-next-line @typescript-eslint/no-unnecessary-condition
       if (globalThis.process?.versions.node) {
-        // eslint-disable-next-line @typescript-eslint/no-unsafe-assignment, @typescript-eslint/no-require-imports
-        const fs: typeof import('fs/promises') = require('fs/promises');
+        const fs: typeof import('fs/promises') = await import('fs/promises');
         bytes = fs.readFile(input);
       } else {
         bytes = fetch(input)

lib/binding_web/src/marshal.ts

@@ -34,8 +34,8 @@ export function unmarshalCaptures(
  *
  * Marshals a {@link Node} to the transfer buffer.
  */
-export function marshalNode(node: Node) {
-  let address = TRANSFER_BUFFER;
+export function marshalNode(node: Node, index = 0) {
+  let address = TRANSFER_BUFFER + index * SIZE_OF_NODE;
   C.setValue(address, node.id, 'i32');
   address += SIZE_OF_INT;
   C.setValue(address, node.startIndex, 'i32');
@@ -168,10 +168,9 @@ export function marshalEdit(edit: Edit, address = TRANSFER_BUFFER) {
  *
  * Unmarshals a {@link LanguageMetadata} from the transfer buffer.
  */
-export function unmarshalLanguageMetadata(address: number): LanguageMetadata {
-  const result = {} as LanguageMetadata;
-  result.major_version = C.getValue(address, 'i32'); address += SIZE_OF_INT;
-  result.minor_version = C.getValue(address, 'i32'); address += SIZE_OF_INT;
-  result.field_count = C.getValue(address, 'i32');
-  return result;
+export function unmarshalLanguageMetadata(address: number): LanguageMetadata {  
+  const major_version = C.getValue(address, 'i32');
+  const minor_version = C.getValue(address += SIZE_OF_INT, 'i32');
+  const patch_version = C.getValue(address += SIZE_OF_INT, 'i32');
+  return { major_version, minor_version, patch_version };
 }

lib/binding_web/src/node.ts

@@ -9,7 +9,8 @@ import { TRANSFER_BUFFER } from './parser';
 /** A single node within a syntax {@link Tree}. */
 export class Node {
   /** @internal */
-  private [0] = 0; // Internal handle for WASM
+  // @ts-expect-error: never read
+  private [0] = 0; // Internal handle for Wasm
 
   /** @internal */
   private _children?: (Node | null)[];
@@ -416,6 +417,11 @@ export class Node {
     // Convert the type strings to numeric type symbols
     const symbols: number[] = [];
     const typesBySymbol = this.tree.language.types;
+    for (const node_type of types) {
+      if (node_type == "ERROR") {
+        symbols.push(65535); // Internally, ts_builtin_sym_error is -1, which is UINT_16MAX
+      }
+    }
     for (let i = 0, n = typesBySymbol.length; i < n; i++) {
       if (types.includes(typesBySymbol[i])) {
         symbols.push(i);
@@ -517,7 +523,7 @@ export class Node {
    */
   childWithDescendant(descendant: Node): Node | null {
     marshalNode(this);
-    marshalNode(descendant);
+    marshalNode(descendant, 1);
     C._ts_node_child_with_descendant_wasm(this.tree[0]);
     return unmarshalNode(this.tree);
   }
@@ -630,7 +636,7 @@ export class Node {
   }
 
   /** Get the S-expression representation of this node. */
-  toString() {
+  toString(): string {
     marshalNode(this);
     const address = C._ts_node_to_string_wasm(this.tree[0]);
     const result = C.AsciiToString(address);

lib/binding_web/src/tree_cursor.ts

@@ -7,16 +7,20 @@ import { getText, Tree } from './tree';
 /** A stateful object for walking a syntax {@link Tree} efficiently. */
 export class TreeCursor {
   /** @internal */
-  private [0] = 0; // Internal handle for WASM
+  // @ts-expect-error: never read
+  private [0] = 0; // Internal handle for Wasm
 
   /** @internal */
-  private [1] = 0; // Internal handle for WASM
+  // @ts-expect-error: never read
+  private [1] = 0; // Internal handle for Wasm
 
   /** @internal */
-  private [2] = 0; // Internal handle for WASM
+  // @ts-expect-error: never read
+  private [2] = 0; // Internal handle for Wasm
 
   /** @internal */
-  private [3] = 0; // Internal handle for WASM
+  // @ts-expect-error: never read
+  private [3] = 0; // Internal handle for Wasm
 
   /** @internal */
   private tree: Tree;

lib/binding_web/test/language.test.ts

@@ -89,6 +89,7 @@ describe('Language', () => {
               '_literal_pattern',
               'captured_pattern',
               'const_block',
+              'generic_pattern',
               'identifier',
               'macro_invocation',
               'mut_pattern',

lib/binding_web/test/node.test.ts

@@ -63,7 +63,7 @@ describe('Node', () => {
       tree = parser.parse('x10 + 1000')!;
       expect(tree.rootNode.children).toHaveLength(1);
       const sumNode = tree.rootNode.firstChild!.firstChild!;
-      expect(sumNode.children.map(child => child!.type)).toEqual(['identifier', '+', 'number' ]);
+      expect(sumNode.children.map(child => child!.type)).toEqual(['identifier', '+', 'number']);
     });
   });
 
@@ -189,6 +189,21 @@ describe('Node', () => {
     });
   });
 
+  describe('.childWithDescendant()', () => {
+    it('correctly retrieves immediate children', () => {
+      const sourceCode = 'let x = 1; console.log(x);';
+      tree = parser.parse(sourceCode)!;
+      const root = tree.rootNode;
+      const child = root.children[0]!.children[0]!;
+      const a = root.childWithDescendant(child);
+      expect(a!.startIndex).toBe(0);
+      const b = a!.childWithDescendant(child);
+      expect(b).toEqual(child);
+      const c = b!.childWithDescendant(child);
+      expect(c).toBeNull();
+    });
+  });
+
   describe('.nextSibling and .previousSibling', () => {
     it('returns the node\'s next and previous sibling', () => {
       tree = parser.parse('x10 + 1000')!;
@@ -449,6 +464,24 @@ describe('Node', () => {
     });
   });
 
+  describe('.descendantsOfType("ERROR")', () => {
+    it('finds all of the descendants of an ERROR node', () => {
+      tree = parser.parse(
+        `if ({a: 'b'} {c: 'd'}) {
+          // ^ ERROR
+          x = function(a) { b; } function(c) { d; }
+        }`
+      )!;
+      const errorNode = tree.rootNode;
+      const descendants = errorNode.descendantsOfType('ERROR');
+      expect(
+        descendants.map((node) => node!.startIndex)
+      ).toEqual(
+        [4]
+      );
+    });
+  });
+
   describe('.descendantsOfType', () => {
     it('finds all descendants of a given type in the given range', () => {
       tree = parser.parse('a + 1 * b * 2 + c + 3')!;

lib/binding_web/test/parser.test.ts

@@ -256,7 +256,7 @@ describe('Parser', () => {
       expect(() => parser.parse({})).toThrow('Argument must be a string or a function');
     });
 
-    it('handles long input strings', { timeout: 5000 }, () => {
+    it('handles long input strings', { timeout: 10000 }, () => {
       const repeatCount = 10000;
       const inputString = `[${Array(repeatCount).fill('0').join(',')}]`;
 

lib/binding_web/test/query.test.ts

@@ -64,7 +64,7 @@ describe('Query', () => {
   });
 
   describe('.matches', () => {
-    it('returns all of the matches for the given query', () => {
+    it('returns all of the matches for the given query', { timeout: 10000 }, () => {
       tree = parser.parse('function one() { two(); function three() {} }')!;
       query = new Query(JavaScript, `
         (function_declaration name: (identifier) @fn-def)
@@ -462,7 +462,7 @@ describe('Query', () => {
   });
 
   describe('Set a timeout', () => {
-    it('returns less than the expected matches', () => {
+    it('returns less than the expected matches', { timeout: 10000 }, () => {
       tree = parser.parse('function foo() while (true) { } }\n'.repeat(1000))!;
       query = new Query(JavaScript, '(function_declaration name: (identifier) @function)');
       const matches = query.matches(tree.rootNode, { timeoutMicros: 1000 });
@@ -538,7 +538,7 @@ describe('Query', () => {
     });
   });
 
-  describe('Executes with a timeout', () => {
+  describe('Executes with a timeout', { timeout: 10000 }, () => {
     it('Returns less than the expected matches', () => {
       tree = parser.parse('function foo() while (true) { } }\n'.repeat(1000))!;
       query = new Query(JavaScript, '(function_declaration) @function');

lib/binding_web/tsconfig.json

@@ -25,11 +25,14 @@
     "esModuleInterop": true,
     "forceConsistentCasingInFileNames": true,
     "skipLibCheck": true,
+    "composite": true,
+    "isolatedModules": true,
   },
   "include": [
-    "src/**/*",
-    "script/**/*",
-    "test/**/*",
+    "src/*.ts",
+    "script/*",
+    "test/*",
+    "lib/*.ts"
   ],
   "exclude": [
     "node_modules",

lib/include/tree_sitter/api.h

@@ -42,7 +42,6 @@ typedef uint16_t TSStateId;
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
-typedef struct TSLanguageMetadata TSLanguageMetadata;
 typedef struct TSParser TSParser;
 typedef struct TSTree TSTree;
 typedef struct TSQuery TSQuery;

lib/language/Cargo.toml

@@ -1,10 +1,10 @@
 [package]
 name = "tree-sitter-language"
 description = "The tree-sitter Language type, used by the library and by language implementations"
-version = "0.1.4"
+version = "0.1.5"
 authors.workspace = true
 edition.workspace = true
-rust-version.workspace = true
+rust-version = "1.76"
 readme = "README.md"
 homepage.workspace = true
 repository.workspace = true

lib/src/get_changed_ranges.c

@@ -34,7 +34,7 @@ bool ts_range_array_intersects(
   uint32_t end_byte
 ) {
   for (unsigned i = start_index; i < self->size; i++) {
-    TSRange *range = &self->contents[i];
+    TSRange *range = array_get(self, i);
     if (range->end_byte > start_byte) {
       if (range->start_byte >= end_byte) break;
       return true;
@@ -108,6 +108,7 @@ typedef struct {
   const TSLanguage *language;
   unsigned visible_depth;
   bool in_padding;
+  Subtree prev_external_token;
 } Iterator;
 
 static Iterator iterator_new(
@@ -127,6 +128,7 @@ static Iterator iterator_new(
     .language = language,
     .visible_depth = 1,
     .in_padding = false,
+    .prev_external_token = NULL_SUBTREE,
   };
 }
 
@@ -157,7 +159,7 @@ static bool iterator_tree_is_visible(const Iterator *self) {
   TreeCursorEntry entry = *array_back(&self->cursor.stack);
   if (ts_subtree_visible(*entry.subtree)) return true;
   if (self->cursor.stack.size > 1) {
-    Subtree parent = *self->cursor.stack.contents[self->cursor.stack.size - 2].subtree;
+    Subtree parent = *array_get(&self->cursor.stack, self->cursor.stack.size - 2)->subtree;
     return ts_language_alias_at(
       self->language,
       parent.ptr->production_id,
@@ -181,10 +183,10 @@ static void iterator_get_visible_state(
   }
 
   for (; i + 1 > 0; i--) {
-    TreeCursorEntry entry = self->cursor.stack.contents[i];
+    TreeCursorEntry entry = *array_get(&self->cursor.stack, i);
 
     if (i > 0) {
-      const Subtree *parent = self->cursor.stack.contents[i - 1].subtree;
+      const Subtree *parent = array_get(&self->cursor.stack, i - 1)->subtree;
       *alias_symbol = ts_language_alias_at(
         self->language,
         parent->ptr->production_id,
@@ -244,6 +246,10 @@ static bool iterator_descend(Iterator *self, uint32_t goal_position) {
 
       position = child_right;
       if (!ts_subtree_extra(*child)) structural_child_index++;
+      Subtree last_external_token = ts_subtree_last_external_token(*child);
+      if (last_external_token.ptr) {
+        self->prev_external_token = last_external_token;
+      }
     }
   } while (did_descend);
 
@@ -268,6 +274,10 @@ static void iterator_advance(Iterator *self) {
 
     const Subtree *parent = array_back(&self->cursor.stack)->subtree;
     uint32_t child_index = entry.child_index + 1;
+    Subtree last_external_token = ts_subtree_last_external_token(*entry.subtree);
+    if (last_external_token.ptr) {
+      self->prev_external_token = last_external_token;
+    }
     if (ts_subtree_child_count(*parent) > child_index) {
       Length position = length_add(entry.position, ts_subtree_total_size(*entry.subtree));
       uint32_t structural_child_index = entry.structural_child_index;
@@ -313,29 +323,41 @@ static IteratorComparison iterator_compare(
   TSSymbol new_alias_symbol = 0;
   iterator_get_visible_state(old_iter, &old_tree, &old_alias_symbol, &old_start);
   iterator_get_visible_state(new_iter, &new_tree, &new_alias_symbol, &new_start);
+  TSSymbol old_symbol = ts_subtree_symbol(old_tree);
+  TSSymbol new_symbol = ts_subtree_symbol(new_tree);
 
   if (!old_tree.ptr && !new_tree.ptr) return IteratorMatches;
   if (!old_tree.ptr || !new_tree.ptr) return IteratorDiffers;
+  if (old_alias_symbol != new_alias_symbol || old_symbol != new_symbol) return IteratorDiffers;
+
+  uint32_t old_size = ts_subtree_size(old_tree).bytes;
+  uint32_t new_size = ts_subtree_size(new_tree).bytes;
+  TSStateId old_state = ts_subtree_parse_state(old_tree);
+  TSStateId new_state = ts_subtree_parse_state(new_tree);
+  bool old_has_external_tokens = ts_subtree_has_external_tokens(old_tree);
+  bool new_has_external_tokens = ts_subtree_has_external_tokens(new_tree);
+  uint32_t old_error_cost = ts_subtree_error_cost(old_tree);
+  uint32_t new_error_cost = ts_subtree_error_cost(new_tree);
 
   if (
-    old_alias_symbol == new_alias_symbol &&
-    ts_subtree_symbol(old_tree) == ts_subtree_symbol(new_tree)
+    old_start != new_start ||
+    old_symbol == ts_builtin_sym_error ||
+    old_size != new_size ||
+    old_state == TS_TREE_STATE_NONE ||
+    new_state == TS_TREE_STATE_NONE ||
+    ((old_state == ERROR_STATE) != (new_state == ERROR_STATE)) ||
+    old_error_cost != new_error_cost ||
+    old_has_external_tokens != new_has_external_tokens ||
+    ts_subtree_has_changes(old_tree) ||
+    (
+      old_has_external_tokens &&
+      !ts_subtree_external_scanner_state_eq(old_iter->prev_external_token, new_iter->prev_external_token)
+    )
   ) {
-    if (old_start == new_start &&
-        !ts_subtree_has_changes(old_tree) &&
-        ts_subtree_symbol(old_tree) != ts_builtin_sym_error &&
-        ts_subtree_size(old_tree).bytes == ts_subtree_size(new_tree).bytes &&
-        ts_subtree_parse_state(old_tree) != TS_TREE_STATE_NONE &&
-        ts_subtree_parse_state(new_tree) != TS_TREE_STATE_NONE &&
-        (ts_subtree_parse_state(old_tree) == ERROR_STATE) ==
-        (ts_subtree_parse_state(new_tree) == ERROR_STATE)) {
-      return IteratorMatches;
-    } else {
-      return IteratorMayDiffer;
-    }
+    return IteratorMayDiffer;
   }
 
-  return IteratorDiffers;
+  return IteratorMatches;
 }
 
 #ifdef DEBUG_GET_CHANGED_RANGES
@@ -348,8 +370,8 @@ static inline void iterator_print_state(Iterator *self) {
     "(%-25s %s\t depth:%u [%u, %u] - [%u, %u])",
     name, self->in_padding ? "(p)" : "   ",
     self->visible_depth,
-    start.row + 1, start.column,
-    end.row + 1, end.column
+    start.row, start.column,
+    end.row, end.column
   );
 }
 #endif
@@ -380,7 +402,7 @@ unsigned ts_subtree_get_changed_ranges(
 
   do {
     #ifdef DEBUG_GET_CHANGED_RANGES
-    printf("At [%-2u, %-2u] Compare ", position.extent.row + 1, position.extent.column);
+    printf("At [%-2u, %-2u] Compare ", position.extent.row, position.extent.column);
     iterator_print_state(&old_iter);
     printf("\tvs\t");
     iterator_print_state(&new_iter);
@@ -475,9 +497,9 @@ unsigned ts_subtree_get_changed_ranges(
     // Keep track of the current position in the included range differences
     // array in order to avoid scanning the entire array on each iteration.
     while (included_range_difference_index < included_range_differences->size) {
-      const TSRange *range = &included_range_differences->contents[
+      const TSRange *range = array_get(included_range_differences,
         included_range_difference_index
-      ];
+      );
       if (range->end_byte <= position.bytes) {
         included_range_difference_index++;
       } else {

lib/src/language.c

@@ -186,7 +186,7 @@ TSSymbol ts_language_symbol_for_name(
   uint32_t length,
   bool is_named
 ) {
-  if (!strncmp(string, "ERROR", length)) return ts_builtin_sym_error;
+  if (is_named && !strncmp(string, "ERROR", length)) return ts_builtin_sym_error;
   uint16_t count = (uint16_t)ts_language_symbol_count(self);
   for (TSSymbol i = 0; i < count; i++) {
     TSSymbolMetadata metadata = ts_language_symbol_metadata(self, i);

lib/src/parser.c

@@ -193,7 +193,7 @@ static bool ts_parser__breakdown_top_of_stack(
     did_break_down = true;
     pending = false;
     for (uint32_t i = 0; i < pop.size; i++) {
-      StackSlice slice = pop.contents[i];
+      StackSlice slice = *array_get(&pop, i);
       TSStateId state = ts_stack_state(self->stack, slice.version);
       Subtree parent = *array_front(&slice.subtrees);
 
@@ -212,7 +212,7 @@ static bool ts_parser__breakdown_top_of_stack(
       }
 
       for (uint32_t j = 1; j < slice.subtrees.size; j++) {
-        Subtree tree = slice.subtrees.contents[j];
+        Subtree tree = *array_get(&slice.subtrees, j);
         ts_stack_push(self->stack, slice.version, tree, false, state);
       }
 
@@ -396,20 +396,24 @@ static void ts_parser__external_scanner_destroy(
 static unsigned ts_parser__external_scanner_serialize(
   TSParser *self
 ) {
+  uint32_t length;
   if (ts_language_is_wasm(self->language)) {
-    return ts_wasm_store_call_scanner_serialize(
+    length = ts_wasm_store_call_scanner_serialize(
       self->wasm_store,
       (uintptr_t)self->external_scanner_payload,
       self->lexer.debug_buffer
     );
+    if (ts_wasm_store_has_error(self->wasm_store)) {
+      self->has_scanner_error = true;
+    }
   } else {
-    uint32_t length = self->language->external_scanner.serialize(
+    length = self->language->external_scanner.serialize(
       self->external_scanner_payload,
       self->lexer.debug_buffer
     );
-    ts_assert(length <= TREE_SITTER_SERIALIZATION_BUFFER_SIZE);
-    return length;
   }
+  ts_assert(length <= TREE_SITTER_SERIALIZATION_BUFFER_SIZE);
+  return length;
 }
 
 static void ts_parser__external_scanner_deserialize(
@@ -556,27 +560,29 @@ static Subtree ts_parser__lex(
           external_scanner_state_len
         );
 
-        // When recovering from an error, ignore any zero-length external tokens
-        // unless they have changed the external scanner's state. This helps to
-        // avoid infinite loops which could otherwise occur, because the lexer is
-        // looking for any possible token, instead of looking for the specific set of
-        // tokens that are valid in some parse state.
+        // Avoid infinite loops caused by the external scanner returning empty tokens.
+        // Empty tokens are needed in some circumstances, e.g. indent/dedent tokens
+        // in Python. Ignore the following classes of empty tokens:
         //
-        // Note that it's possible that the token end position may be *before* the
-        // original position of the lexer because of the way that tokens are positioned
-        // at included range boundaries: when a token is terminated at the start of
-        // an included range, it is marked as ending at the *end* of the preceding
-        // included range.
+        // * Tokens produced during error recovery. When recovering from an error,
+        //   all tokens are allowed, so it's easy to accidentally return unwanted
+        //   empty tokens.
+        // * Tokens that are marked as 'extra' in the grammar. These don't change
+        //   the parse state, so they would definitely cause an infinite loop.
         if (
           self->lexer.token_end_position.bytes <= current_position.bytes &&
-          (error_mode || !ts_stack_has_advanced_since_error(self->stack, version)) &&
           !external_scanner_state_changed
         ) {
-          LOG(
-            "ignore_empty_external_token symbol:%s",
-            SYM_NAME(self->language->external_scanner.symbol_map[self->lexer.data.result_symbol])
-          )
-          found_token = false;
+          TSSymbol symbol = self->language->external_scanner.symbol_map[self->lexer.data.result_symbol];
+          TSStateId next_parse_state = ts_language_next_state(self->language, parse_state, symbol);
+          bool token_is_extra = (next_parse_state == parse_state);
+          if (error_mode || !ts_stack_has_advanced_since_error(self->stack, version) || token_is_extra) {
+            LOG(
+              "ignore_empty_external_token symbol:%s",
+              SYM_NAME(self->language->external_scanner.symbol_map[self->lexer.data.result_symbol])
+            );
+            found_token = false;
+          }
         }
       }
 
@@ -947,20 +953,22 @@ static StackVersion ts_parser__reduce(
   // children.
   StackSliceArray pop = ts_stack_pop_count(self->stack, version, count);
   uint32_t removed_version_count = 0;
+  uint32_t halted_version_count = ts_stack_halted_version_count(self->stack);
   for (uint32_t i = 0; i < pop.size; i++) {
-    StackSlice slice = pop.contents[i];
+    StackSlice slice = *array_get(&pop, i);
     StackVersion slice_version = slice.version - removed_version_count;
 
     // This is where new versions are added to the parse stack. The versions
     // will all be sorted and truncated at the end of the outer parsing loop.
     // Allow the maximum version count to be temporarily exceeded, but only
     // by a limited threshold.
-    if (slice_version > MAX_VERSION_COUNT + MAX_VERSION_COUNT_OVERFLOW) {
+    if (slice_version > MAX_VERSION_COUNT + MAX_VERSION_COUNT_OVERFLOW + halted_version_count) {
       ts_stack_remove_version(self->stack, slice_version);
       ts_subtree_array_delete(&self->tree_pool, &slice.subtrees);
       removed_version_count++;
       while (i + 1 < pop.size) {
-        StackSlice next_slice = pop.contents[i + 1];
+        LOG("aborting reduce with too many versions")
+        StackSlice next_slice = *array_get(&pop, i + 1);
         if (next_slice.version != slice.version) break;
         ts_subtree_array_delete(&self->tree_pool, &next_slice.subtrees);
         i++;
@@ -983,7 +991,7 @@ static StackVersion ts_parser__reduce(
     // choose one of the arrays of trees to be the parent node's children, and
     // delete the rest of the tree arrays.
     while (i + 1 < pop.size) {
-      StackSlice next_slice = pop.contents[i + 1];
+      StackSlice next_slice = *array_get(&pop, i + 1);
       if (next_slice.version != slice.version) break;
       i++;
 
@@ -1025,7 +1033,7 @@ static StackVersion ts_parser__reduce(
     // were previously on top of the stack.
     ts_stack_push(self->stack, slice_version, ts_subtree_from_mut(parent), false, next_state);
     for (uint32_t j = 0; j < self->trailing_extras.size; j++) {
-      ts_stack_push(self->stack, slice_version, self->trailing_extras.contents[j], false, next_state);
+      ts_stack_push(self->stack, slice_version, *array_get(&self->trailing_extras, j), false, next_state);
     }
 
     for (StackVersion j = 0; j < slice_version; j++) {
@@ -1053,11 +1061,11 @@ static void ts_parser__accept(
 
   StackSliceArray pop = ts_stack_pop_all(self->stack, version);
   for (uint32_t i = 0; i < pop.size; i++) {
-    SubtreeArray trees = pop.contents[i].subtrees;
+    SubtreeArray trees = array_get(&pop, i)->subtrees;
 
     Subtree root = NULL_SUBTREE;
     for (uint32_t j = trees.size - 1; j + 1 > 0; j--) {
-      Subtree tree = trees.contents[j];
+      Subtree tree = *array_get(&trees, j);
       if (!ts_subtree_extra(tree)) {
         ts_assert(!tree.data.is_inline);
         uint32_t child_count = ts_subtree_child_count(tree);
@@ -1092,7 +1100,7 @@ static void ts_parser__accept(
     }
   }
 
-  ts_stack_remove_version(self->stack, pop.contents[0].version);
+  ts_stack_remove_version(self->stack, array_get(&pop, 0)->version);
   ts_stack_halt(self->stack, version);
 }
 
@@ -1158,7 +1166,7 @@ static bool ts_parser__do_all_potential_reductions(
 
     StackVersion reduction_version = STACK_VERSION_NONE;
     for (uint32_t j = 0; j < self->reduce_actions.size; j++) {
-      ReduceAction action = self->reduce_actions.contents[j];
+      ReduceAction action = *array_get(&self->reduce_actions, j);
 
       reduction_version = ts_parser__reduce(
         self, version, action.symbol, action.count,
@@ -1196,7 +1204,7 @@ static bool ts_parser__recover_to_state(
   StackVersion previous_version = STACK_VERSION_NONE;
 
   for (unsigned i = 0; i < pop.size; i++) {
-    StackSlice slice = pop.contents[i];
+    StackSlice slice = *array_get(&pop, i);
 
     if (slice.version == previous_version) {
       ts_subtree_array_delete(&self->tree_pool, &slice.subtrees);
@@ -1214,12 +1222,12 @@ static bool ts_parser__recover_to_state(
     SubtreeArray error_trees = ts_stack_pop_error(self->stack, slice.version);
     if (error_trees.size > 0) {
       ts_assert(error_trees.size == 1);
-      Subtree error_tree = error_trees.contents[0];
+      Subtree error_tree = *array_get(&error_trees, 0);
       uint32_t error_child_count = ts_subtree_child_count(error_tree);
       if (error_child_count > 0) {
         array_splice(&slice.subtrees, 0, 0, error_child_count, ts_subtree_children(error_tree));
         for (unsigned j = 0; j < error_child_count; j++) {
-          ts_subtree_retain(slice.subtrees.contents[j]);
+          ts_subtree_retain(*array_get(&slice.subtrees, j));
         }
       }
       ts_subtree_array_delete(&self->tree_pool, &error_trees);
@@ -1235,7 +1243,7 @@ static bool ts_parser__recover_to_state(
     }
 
     for (unsigned j = 0; j < self->trailing_extras.size; j++) {
-      Subtree tree = self->trailing_extras.contents[j];
+      Subtree tree = *array_get(&self->trailing_extras, j);
       ts_stack_push(self->stack, slice.version, tree, false, goal_state);
     }
 
@@ -1271,7 +1279,7 @@ static void ts_parser__recover(
   // if the current lookahead token would be valid in that state.
   if (summary && !ts_subtree_is_error(lookahead)) {
     for (unsigned i = 0; i < summary->size; i++) {
-      StackSummaryEntry entry = summary->contents[i];
+      StackSummaryEntry entry = *array_get(summary, i);
 
       if (entry.state == ERROR_STATE) continue;
       if (entry.position.bytes == position.bytes) continue;
@@ -1316,10 +1324,23 @@ static void ts_parser__recover(
   // and subsequently halted. Remove those versions.
   for (unsigned i = previous_version_count; i < ts_stack_version_count(self->stack); i++) {
     if (!ts_stack_is_active(self->stack, i)) {
+      LOG("removed paused version:%u", i);
       ts_stack_remove_version(self->stack, i--);
+      LOG_STACK();
     }
   }
 
+  // If the parser is still in the error state at the end of the file, just wrap everything
+  // in an ERROR node and terminate.
+  if (ts_subtree_is_eof(lookahead)) {
+    LOG("recover_eof");
+    SubtreeArray children = array_new();
+    Subtree parent = ts_subtree_new_error_node(&children, false, self->language);
+    ts_stack_push(self->stack, version, parent, false, 1);
+    ts_parser__accept(self, version, lookahead);
+    return;
+  }
+
   // If strategy 1 succeeded, a new stack version will have been created which is able to handle
   // the current lookahead token. Now, in addition, try strategy 2 described above: skip the
   // current lookahead token by wrapping it in an ERROR node.
@@ -1340,17 +1361,6 @@ static void ts_parser__recover(
     return;
   }
 
-  // If the parser is still in the error state at the end of the file, just wrap everything
-  // in an ERROR node and terminate.
-  if (ts_subtree_is_eof(lookahead)) {
-    LOG("recover_eof");
-    SubtreeArray children = array_new();
-    Subtree parent = ts_subtree_new_error_node(&children, false, self->language);
-    ts_stack_push(self->stack, version, parent, false, 1);
-    ts_parser__accept(self, version, lookahead);
-    return;
-  }
-
   // Do not recover if the result would clearly be worse than some existing stack version.
   unsigned new_cost =
     current_error_cost + ERROR_COST_PER_SKIPPED_TREE +
@@ -1396,18 +1406,18 @@ static void ts_parser__recover(
     // arbitrarily and discard the rest.
     if (pop.size > 1) {
       for (unsigned i = 1; i < pop.size; i++) {
-        ts_subtree_array_delete(&self->tree_pool, &pop.contents[i].subtrees);
+        ts_subtree_array_delete(&self->tree_pool, &array_get(&pop, i)->subtrees);
       }
-      while (ts_stack_version_count(self->stack) > pop.contents[0].version + 1) {
-        ts_stack_remove_version(self->stack, pop.contents[0].version + 1);
+      while (ts_stack_version_count(self->stack) > array_get(&pop, 0)->version + 1) {
+        ts_stack_remove_version(self->stack, array_get(&pop, 0)->version + 1);
       }
     }
 
-    ts_stack_renumber_version(self->stack, pop.contents[0].version, version);
-    array_push(&pop.contents[0].subtrees, ts_subtree_from_mut(error_repeat));
+    ts_stack_renumber_version(self->stack, array_get(&pop, 0)->version, version);
+    array_push(&array_get(&pop, 0)->subtrees, ts_subtree_from_mut(error_repeat));
     error_repeat = ts_subtree_new_node(
       ts_builtin_sym_error_repeat,
-      &pop.contents[0].subtrees,
+      &array_get(&pop, 0)->subtrees,
       0,
       self->language
     );
@@ -1534,7 +1544,7 @@ static bool ts_parser__check_progress(TSParser *self, Subtree *lookahead, const
   if (self->operation_count >= OP_COUNT_PER_PARSER_TIMEOUT_CHECK) {
     self->operation_count = 0;
   }
-  if (self->parse_options.progress_callback && position != NULL) {
+  if (position != NULL) {
     self->parse_state.current_byte_offset = *position;
     self->parse_state.has_error = self->has_error;
   }
@@ -1616,6 +1626,7 @@ static bool ts_parser__advance(
     // an ambiguous state. REDUCE actions always create a new stack
     // version, whereas SHIFT actions update the existing stack version
     // and terminate this loop.
+    bool did_reduce = false;
     StackVersion last_reduction_version = STACK_VERSION_NONE;
     for (uint32_t i = 0; i < table_entry.action_count; i++) {
       TSParseAction action = table_entry.actions[i];
@@ -1651,6 +1662,7 @@ static bool ts_parser__advance(
             action.reduce.dynamic_precedence, action.reduce.production_id,
             is_fragile, end_of_non_terminal_extra
           );
+          did_reduce = true;
           if (reduction_version != STACK_VERSION_NONE) {
             last_reduction_version = reduction_version;
           }
@@ -1702,9 +1714,12 @@ static bool ts_parser__advance(
       continue;
     }
 
-    // A non-terminal extra rule was reduced and merged into an existing
-    // stack version. This version can be discarded.
-    if (!lookahead.ptr) {
+    // A reduction was performed, but was merged into an existing stack version.
+    // This version can be discarded.
+    if (did_reduce) {
+      if (lookahead.ptr) {
+        ts_subtree_release(&self->tree_pool, lookahead);
+      }
       ts_stack_halt(self->stack, version);
       return true;
     }
@@ -1753,7 +1768,7 @@ static bool ts_parser__advance(
     // versions that exist. If some other version advances successfully, then
     // this version can simply be removed. But if all versions end up paused,
     // then error recovery is needed.
-    LOG("detect_error");
+    LOG("detect_error lookahead:%s", TREE_NAME(lookahead));
     ts_stack_pause(self->stack, version, lookahead);
     return true;
   }
@@ -1842,6 +1857,7 @@ static unsigned ts_parser__condense_stack(TSParser *self) {
           has_unpaused_version = true;
         } else {
           ts_stack_remove_version(self->stack, i);
+          made_changes = true;
           i--;
           n--;
         }
@@ -1877,9 +1893,9 @@ static bool ts_parser__balance_subtree(TSParser *self) {
       return false;
     }
 
-    MutableSubtree tree = self->tree_pool.tree_stack.contents[
+    MutableSubtree tree = *array_get(&self->tree_pool.tree_stack, 
       self->tree_pool.tree_stack.size - 1
-    ];
+    );
 
     if (tree.ptr->repeat_depth > 0) {
       Subtree child1 = ts_subtree_children(tree)[0];
@@ -2128,7 +2144,7 @@ TSTree *ts_parser_parse(
       LOG("parse_after_edit");
       LOG_TREE(self->old_tree);
       for (unsigned i = 0; i < self->included_range_differences.size; i++) {
-        TSRange *range = &self->included_range_differences.contents[i];
+        TSRange *range = array_get(&self->included_range_differences, i);
         LOG("different_included_range %u - %u", range->start_byte, range->end_byte);
       }
     } else {
@@ -2185,7 +2201,7 @@ TSTree *ts_parser_parse(
     }
 
     while (self->included_range_difference_index < self->included_range_differences.size) {
-      TSRange *range = &self->included_range_differences.contents[self->included_range_difference_index];
+      TSRange *range = array_get(&self->included_range_differences, self->included_range_difference_index);
       if (range->end_byte <= position) {
         self->included_range_difference_index++;
       } else {
@@ -2226,6 +2242,8 @@ TSTree *ts_parser_parse_with_options(
   self->parse_options = parse_options;
   self->parse_state.payload = parse_options.payload;
   TSTree *result = ts_parser_parse(self, old_tree, input);
+  // Reset parser options before further parse calls.
+  self->parse_options = (TSParseOptions) {0};
   return result;
 }
 

lib/src/parser.h

@@ -18,7 +18,6 @@ typedef uint16_t TSStateId;
 typedef uint16_t TSSymbol;
 typedef uint16_t TSFieldId;
 typedef struct TSLanguage TSLanguage;
-typedef struct TSLanguageMetadata TSLanguageMetadata;
 typedef struct TSLanguageMetadata {
   uint8_t major_version;
   uint8_t minor_version;

lib/src/portable/endian.h

@@ -18,16 +18,24 @@
 #if defined(HAVE_ENDIAN_H) || \
     defined(__linux__) || \
     defined(__GNU__) || \
+    defined(__HAIKU__) || \
+    defined(__illumos__) || \
+    defined(__NetBSD__) || \
     defined(__OpenBSD__) || \
     defined(__CYGWIN__) || \
     defined(__MSYS__) || \
-    defined(__EMSCRIPTEN__)
+    defined(__EMSCRIPTEN__) || \
+    defined(__wasi__) || \
+    defined(__wasm__)
+
+#if defined(__NetBSD__)
+#define _NETBSD_SOURCE 1
+#endif
 
 # include <endian.h>
 
 #elif defined(HAVE_SYS_ENDIAN_H) || \
     defined(__FreeBSD__) || \
-    defined(__NetBSD__) || \
     defined(__DragonFly__)
 
 # include <sys/endian.h>