@willbooster/tree-sitter-c-sharp 2.0.1 → 3.0.0

package/README.md

@@ -28,16 +28,12 @@ The grammar is based upon the Roslyn grammar with changes in order to:
 Known gaps:
 
 - `var` and `await` cannot be used as identifiers everywhere they are valid
-- A pointer dereference can be assigned to only when its operand is a variable, an address, a parenthesized or
-  postfix expression, or a cast of one of these or of an invocation (`*p = 1`, `*(p + 1) = 1`, `*(int*)p = 1`,
-  `*dst++ = 1`); `*++p = 1` and `*f() = 1` are not recognized
 
 ## Usage
 
 The npm package ships `tree-sitter-c_sharp.wasm` for
 [@willbooster/web-tree-sitter](https://www.npmjs.com/package/@willbooster/web-tree-sitter), which runs in Node.js, Bun,
-browsers, and Cloudflare Workers. Use runtime version 1.1.2 or later: browser loading requires its asynchronous
-instantiation support for this grammar. Install the runtime alongside the grammar when using the Wasm parser; it
+browsers, and Cloudflare Workers. Use runtime version 1.3.0 or later for the compact ABI 16 parser. Install the runtime alongside the grammar when using the Wasm parser; it
 remains an optional peer for consumers that only use the grammar source or queries.
 
 In Node.js and Bun, load it from the package:
@@ -83,13 +79,12 @@ The package also ships the queries in `queries/` and the node types in `src/node
 
 In Rust, depend on the [crate](https://crates.io/crates/willbooster-tree-sitter-c-sharp) and on
 [willbooster-tree-sitter](https://crates.io/crates/willbooster-tree-sitter), the runtime this package is tested and
-fuzzed with, whose fixes keep incremental reparses consistent with fresh parses (the grammar also loads in the upstream
-`tree-sitter` crate 0.27, whose error recovery never ends on some malformed input):
+fuzzed with. The compact ABI 16 parser requires runtime 1.3.0 or later:
 
 ```toml
 [dependencies]
-tree-sitter = { package = "willbooster-tree-sitter", version = "1.1.2" }
-tree-sitter-c-sharp = { package = "willbooster-tree-sitter-c-sharp", version = "2" }
+tree-sitter = { package = "willbooster-tree-sitter", version = "1.3.0" }
+tree-sitter-c-sharp = { package = "willbooster-tree-sitter-c-sharp", version = "3" }
 ```
 
 ```rust
@@ -114,16 +109,28 @@ runtime this package ships. The first run downloads that CLI from its GitHub Rel
 `cargo` (with the `cmake` that `mise.toml` pins) when the download fails or the release has no binary that runs here. The `tree-sitter-cli` package provides
 only the types of the grammar DSL that `grammar.js` checks against; its `tree-sitter` binary is upstream's.
 
+`bun run generate` records a fresh ABI 16 generation profile from the applicable `test/corpus` cases and Git-tracked
+files in `examples/`, then generates compact parser tables. The parser also embeds metadata from `tree-sitter.json`. After changing a grammar,
+`tree-sitter.json`, a corpus case, or a tracked example,
+regenerate and commit `src/`. Stage added or removed examples with `git add -A examples` before generation so the profile uses the intended file list.
+Profiles in `.tmp/generation-profiles/` are temporary and must not be committed. `bun run build-wasm`, `bun run build/ci`,
+and the release build regenerate the parsers before compiling them.
+
+The tracked `examples/JsonWriter.cs` and `examples/JsonSerializerInternalWriter.cs` are verbatim MIT-licensed
+snapshots from Newtonsoft.Json commit `52fa3aef1f2cadcd3a3f874251eddc98d3efbbaa`, under `Src/Newtonsoft.Json/`
+and its `Serialization/` directory respectively. Refresh these training inputs independently of the clone pins in
+`script/parse-examples` so changing a test checkout does not change profile training.
+
 `bun run test` runs:
 
 - the corpus in `test/corpus`, with the native build and with the Wasm build (the first run downloads the WASI SDK);
 - an incremental-parsing check (`test/unit/incremental.test.ts`): `script/fuzz-corpus` runs `tree-sitter fuzz`, which
   edits each corpus case at random, reparses it, undoes the edits, and reparses again. `TREE_SITTER_SEED`,
   `TREE_SITTER_ITERATIONS`, and `TREE_SITTER_EDITS` run other or more edits;
-- a check that the real-world C# files cloned into `examples/` fail to parse exactly as listed in
-  `script/known-failures.txt`. The first run clones them. The example repositories are pinned to commits in
-  `script/parse-examples`. After a grammar change or a moved pin alters that list, `script/parse-examples` rewrites
-  it; review its diff before committing;
+- a check that the C# files under `examples/` that fail to parse are exactly those listed in
+  `script/known-failures.txt`. Inputs include Git-tracked generation examples and pinned clones. The first run fetches the pinned repositories listed in
+  `script/parse-examples`. After adding, removing, or changing an example, changing the grammar, or moving a pin,
+  run `script/parse-examples` and review the failure-list diff before committing;
 - a performance check (`test/unit/performance.test.ts`) that recovering from an error on each line takes linear time
   (ten times the lines take about ten times the CPU time, under a ceiling), since consumers parse files while they are
   being edited. It loads the Wasm build through @willbooster/web-tree-sitter, which `bun run build/ci` rebuilds after

package/grammar.js

@@ -49,6 +49,8 @@ module.exports = grammar({
   name: 'c_sharp',
 
   conflicts: ($) => [
+    [$._assignment_pointer_cast, $._lvalue_cast_expression],
+    [$._assignment_pointer_expression, $._pointer_indirection_expression],
     [$.modifier, $._constructor_declaration_initializer, $._reserved_identifier],
     [$._simple_name, $.generic_name],
     [$._simple_name, $.type_parameter],
@@ -1209,11 +1211,67 @@ module.exports = grammar({
       seq(
         // An invocation can return a reference (`Unsafe.Add(ref x, i) = v`). It ends with `)`, so `=` after it adds few
         // parse states, unlike making any expression assignable.
-        field('left', choice($.lvalue_expression, $.invocation_expression)),
+        field(
+          'left',
+          choice(
+            $.lvalue_expression,
+            $.invocation_expression,
+            alias($._assignment_pointer_expression, $.prefix_unary_expression)
+          )
+        ),
         field('operator', choice('=', '+=', '-=', '*=', '/=', '%=', '&=', '^=', '|=', '<<=', '>>=', '>>>=', '??=')),
         field('right', $.expression)
       ),
 
+    // Keep this alternative confined to assignment targets: sharing it with ordinary dereferences changes hidden
+    // expression-supertype captures in nested reads. -3 prefers the existing lvalue path (-2) when both fit.
+    _assignment_pointer_expression: ($) =>
+      prec.dynamic(
+        -3,
+        prec.right(
+          PREC.UNARY,
+          seq(
+            '*',
+            choice(
+              $.lvalue_expression,
+              $.parenthesized_expression,
+              $.invocation_expression,
+              $.postfix_unary_expression,
+              alias($._assignment_pointer_expression, $.prefix_unary_expression),
+              alias($._pointer_prefix_update, $.prefix_unary_expression),
+              alias($._assignment_pointer_cast, $.cast_expression)
+            )
+          )
+        )
+      ),
+
+    _pointer_prefix_update: ($) =>
+      prec(
+        PREC.UNARY,
+        seq(choice('++', '--'), choice($.lvalue_expression, $.invocation_expression, $.parenthesized_expression))
+      ),
+
+    _assignment_pointer_cast: ($) =>
+      prec(
+        PREC.CAST,
+        seq(
+          '(',
+          field('type', $.type),
+          ')',
+          field(
+            'value',
+            choice(
+              $.lvalue_expression,
+              $.parenthesized_expression,
+              $.invocation_expression,
+              $.postfix_unary_expression,
+              alias($._assignment_pointer_expression, $.prefix_unary_expression),
+              alias($._pointer_prefix_update, $.prefix_unary_expression)
+            )
+          )
+        )
+      ),
+
     binary_expression: ($) =>
       choice(
         ...[
@@ -1267,14 +1325,9 @@ module.exports = grammar({
         )
       ),
 
-    // Pointer indirection is split by its operand. The assignable form, surfaced in `lvalue_expression`, takes an
-    // lvalue (`*p`, `*(p)`, `**pp`) or an operand that ends in a closing token or an lvalue: a parenthesized
-    // expression (`*(p + 1)`), a postfix expression (`*p++`), an address (`*&v`), or a cast of an lvalue, a
-    // parenthesized, postfix, or invocation expression, or an address (`*(int*)p`, `*(long*)(p + 4)`, `*(T*)f()`,
-    // `*(long*)&v`). Dereferencing any other expression (`*++p`) is only read: letting any expression be the assignable
-    // form's operand lets `=` follow every expression that can be the operand, which grows the parser by about 4,800
-    // states and nearly doubles the Wasm build, while these operands add about 550.
-    //
+    // Keep the lvalue operand set restricted: a general expression also permits assignment after every operand,
+    // substantially increasing parser states. Additional call and prefix-update operands use the assignment-only
+    // `_assignment_pointer_expression` path to preserve the shared read derivations and their query captures.
     // Dynamic precedences: an assignable dereference -2, a read-only one -4, a read-only cast +1, an assignable cast -1,
     // and a unary `+`, `-`, `^`, or `&` -2. C# reads a parenthesized name as a cast only when the token after `)` is
     // `~`, `!`, `(`, an identifier, a literal, or a keyword (C# spec §12.9.8 Cast expressions), and these values keep

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@willbooster/tree-sitter-c-sharp",
-  "version": "2.0.1",
+  "version": "3.0.0",
   "description": "C# grammar for tree-sitter",
   "keywords": [
     "incremental",
@@ -22,10 +22,11 @@
     "*.wasm"
   ],
   "scripts": {
-    "build-wasm": "script/tree-sitter build --wasm",
-    "build/ci": "script/tree-sitter generate && script/tree-sitter build --wasm",
+    "build-wasm": "script/generate --build-wasm",
+    "build/ci": "bun run build-wasm",
     "cleanup": "bun wb lint --fix --format",
     "format": "bun wb lint --format",
+    "generate": "script/generate",
     "lint": "bun wb lint",
     "lint-fix": "bun wb lint --fix",
     "prepare": "(bun lefthook install || true) && (test ! -f .lefthook/normalize-bun-lockfile.sh || bash .lefthook/normalize-bun-lockfile.sh)",
@@ -47,7 +48,7 @@
     "@willbooster/oxfmt-config": "1.2.2",
     "@willbooster/oxlint-config": "1.4.8",
     "@willbooster/wb": "22.13.1",
-    "@willbooster/web-tree-sitter": "1.1.2",
+    "@willbooster/web-tree-sitter": "1.4.0",
     "conventional-changelog-conventionalcommits": "10.4.0",
     "lefthook": "2.1.14",
     "oxfmt": "0.70.0",
@@ -63,7 +64,7 @@
     "wrangler": "4.135.0"
   },
   "peerDependencies": {
-    "@willbooster/web-tree-sitter": "^1.1.2"
+    "@willbooster/web-tree-sitter": "^1.3.0"
   },
   "peerDependenciesMeta": {
     "@willbooster/web-tree-sitter": {

package/tree-sitter.json

@@ -19,7 +19,7 @@
     }
   ],
   "metadata": {
-    "version": "2.0.1",
+    "version": "3.0.0",
     "license": "MIT",
     "description": "C# grammar for tree-sitter",
     "authors": [