@felixtensor/tree-sitter-mlir 0.1.0 → 0.1.2

package/README.md

@@ -3,6 +3,7 @@
 [![CI](https://github.com/felixtensor/tree-sitter-mlir/actions/workflows/ci.yml/badge.svg)](https://github.com/felixtensor/tree-sitter-mlir/actions/workflows/ci.yml)
 [![crates.io](https://img.shields.io/crates/v/tree-sitter-mlir.svg?logo=rust)](https://crates.io/crates/tree-sitter-mlir)
 [![PyPI](https://img.shields.io/pypi/v/tree-sitter-mlir.svg?logo=pypi&logoColor=white)](https://pypi.org/project/tree-sitter-mlir/)
+[![npm](https://img.shields.io/npm/v/@felixtensor/tree-sitter-mlir.svg?logo=npm)](https://www.npmjs.com/package/@felixtensor/tree-sitter-mlir)
 
 [MLIR](https://mlir.llvm.org) grammar for [tree-sitter](https://github.com/tree-sitter/tree-sitter).
 

package/grammar.js

@@ -50,6 +50,8 @@ export default grammar({
     _digit: $ => /[0-9]/,
     integer_literal: $ => choice($._decimal_literal, $._hexadecimal_literal),
     _decimal_literal: $ => token(seq(optional(/[-+]/), repeat1(/[0-9]/))),
+    _unsigned_decimal_literal: $ => token(repeat1(/[0-9]/)),
+    _unsigned_integer_literal: $ => choice($._unsigned_decimal_literal, $._hexadecimal_literal),
     _hexadecimal_literal: $ => token(seq('0x', repeat1(/[0-9a-fA-F]/))),
     float_literal: $ => token(seq(
       optional(/[-+]/), repeat1(/[0-9]/), '.', repeat(/[0-9]/),
@@ -182,6 +184,15 @@ export default grammar({
     func_operation: $ => prec.right(seq(
       field('name', choice(token(prec(20, 'func.func')), token(prec(20, 'llvm.func')))),
       field('visibility', optional(choice('private', 'public'))),
+      // Optional leading specifier keywords between the function name and its
+      // symbol. Covers MLIR symbol visibility (`nested`) and the LLVM dialect's
+      // llvm.func linkage / calling-convention / unnamed_addr / visibility
+      // keywords (internal, external, linkonce, fastcc, amdgpu_kernelcc,
+      // local_unnamed_addr, hidden, ...). MLIR's CConv enum alone has ~50
+      // keywords, so rather than enumerate a brittle list we accept bare_id-
+      // shaped specifiers here; the trailing `@symbol` (symbol_ref_id) is an
+      // unambiguous terminator.
+      repeat(field('specifier', alias($.bare_id, $.function_specifier))),
       field('sym_name', $.symbol_ref_id),
       field('arguments', $.func_arg_list),
       field('return', optional($.func_return)),
@@ -250,18 +261,22 @@ export default grammar({
       prec(2, $.type),              // !type, i32, memref<...>, etc.
       $.attribute,                  // #attr, {dict}, affine_map<...>
       $.region,                     // { ... } (regions with operations)
+      $._custom_body_value_group,   // {%v : type, ...}
       $._custom_body_paren,         // ( ... )
       $._custom_body_bracket,       // [ ... ]
       $._custom_body_angle_group,   // < ... >
       $._literal,                   // 42, 3.14, "string", true, dense<...>
       'array',                      // property names may collide with array<...>
+      'vector',                     // OpenACC keyword may collide with vector<...>
       'ceildiv', 'floordiv', 'mod', // inline affine keywords
       $.bare_id,                    // keywords: to, from, step, ins, outs, etc.
-      ',', '=', ':', '->', '*', '+', '-', '/', '&', '|', '~',
+      ',', '=', ':', '->', '*', '?', $.dimension_separator, '+', '-', '/', '&', '|', '~',
     ),
 
     _custom_body_paren: $ => seq('(', repeat($._nested_custom_body_element), ')'),
     _custom_body_bracket: $ => seq('[', repeat($._nested_custom_body_element), ']'),
+    _custom_body_value_group: $ => seq('{', $.value_use, ':', $.type,
+      repeat(seq(',', $.value_use, ':', $.type)), '}'),
     _custom_body_angle_group: $ => seq('<', repeat($._nested_custom_body_element), '>'),
     // Only nested groups accept `trailing_location` as a body element.
     // At the top level it is omitted on purpose so the operation rule
@@ -271,11 +286,11 @@ export default grammar({
 
     // =========================================================================
     // Blocks
-    //   block       ::= block-label operation+
+    //   block       ::= block-label operation*
     //   block-label ::= block-id block-arg-list? `:`
     //   caret-id    ::= `^` suffix-id
     // =========================================================================
-    block: $ => seq($.block_label, repeat1($.operation)),
+    block: $ => seq($.block_label, repeat($.operation)),
     block_label: $ => seq($._block_id, optional($.block_arg_list), ':'),
     _block_id: $ => $.caret_id,
     caret_id: $ => seq('^', $._suffix_id),
@@ -336,9 +351,9 @@ export default grammar({
       $.type,
       prec(2, $.attribute),
       $._literal,
-      'dense', 'sparse', 'array',
+      'dense', 'sparse', 'array', 'vector', 'tensor', 'opaque',
       $.bare_id,
-      ',', ':', '=', '->', '(', ')', '[', ']', '{', '}', '*',
+      ',', ':', '=', '->', '(', ')', '[', ']', '{', '}', '*', '?',
       '@', '#',
       token(prec(-1, /[^<>]/))
     )),
@@ -357,9 +372,13 @@ export default grammar({
       repeat1(seq($.dimension_separator, $._dialect_dim_primitive)))),
     _dialect_dim_primitive: $ => choice(
       prec(1, $.type),
+      prec(1, $._pretty_dialect_body_type),
       alias($.integer_literal, $.dimension_size),
       '?',
       '*'),
+    _pretty_dialect_body_type: $ => seq(
+      choice($.bare_id, 'array'),
+      $.pretty_dialect_item_body),
 
     // Builtin types
     builtin_type: $ => choice(
@@ -400,7 +419,7 @@ export default grammar({
       optional(seq(',', $.tensor_encoding)), '>'),
     tensor_encoding: $ => $.attribute_value,
 
-    vector_type: $ => seq(token('vector'), '<', repeat($.vector_dim_list), $._prim_type, '>'),
+    vector_type: $ => prec(1, seq(token('vector'), '<', repeat($.vector_dim_list), $._prim_type, '>')),
     vector_dim_list: $ => prec.left(choice(seq($._static_dim_list, 'x',
       optional(seq('[', $._static_dim_list, ']', 'x'))), seq('[', $._static_dim_list, ']', 'x'))),
     _static_dim_list: $ => seq($.dimension_size, repeat(seq('x', $.dimension_size))),
@@ -420,11 +439,12 @@ export default grammar({
     attribute_entry: $ => choice(
       seq(choice($.bare_id, $.string_literal), optional(seq('=', $.attribute_value))),
       // Array-valued entry (e.g. {["op.name"]} in transform dialect)
-      seq('[', optional($._attribute_value_nobracket),
-        repeat(seq(',', $._attribute_value_nobracket)), ']'),
+      $._attribute_array,
     ),
-    attribute_value: $ => choice(seq('[', optional($._attribute_value_nobracket),
-      repeat(seq(',', $._attribute_value_nobracket)), ']'), $._attribute_value_nobracket),
+    attribute_value: $ => choice($._attribute_array, $._attribute_value_nobracket),
+    _attribute_array: $ => seq('[', optional($._attribute_array_element),
+      repeat(seq(',', $._attribute_array_element)), ']'),
+    _attribute_array_element: $ => choice($._attribute_array, $._attribute_value_nobracket),
     _attribute_value_nobracket: $ => choice($.attribute_alias, $.dialect_attribute,
       $.builtin_attribute, $.dictionary_attribute, $._literal_and_type, $.type,
       $.function_type, $._affine_map_like, $.symbol_ref_id, $.trailing_location,
@@ -445,9 +465,13 @@ export default grammar({
       prec(1, $.affine_map),
       prec(1, $.affine_set),
       $.dense_resource_literal,
+      $.distinct_attribute,
     ),
     dense_resource_literal: $ => seq(token('dense_resource'), '<',
       choice($.bare_id, $.string_literal), '>'),
+    distinct_attribute: $ => seq(token('distinct'), '[',
+      alias($._unsigned_integer_literal, $.integer_literal), ']',
+      '<', optional($.attribute_value), '>'),
     strided_layout: $ => seq(token('strided'), '<', '[', optional($._stride_list_comma), ']',
       optional(seq(',', token('offset'), ':', $._stride_primitive)), '>'),
     _stride_list_comma: $ => seq($._stride_primitive, repeat(seq(',', $._stride_primitive))),
@@ -485,7 +509,7 @@ export default grammar({
     func_arg_list: $ => seq('(', optional(choice($.variadic,
       $._value_id_and_type_attr_list)), ')'),
     _value_id_and_type_attr_list: $ => seq($._value_id_and_type_attr,
-      repeat(seq(',', $._value_id_and_type_attr)), optional(seq(',', $.variadic))),
+      repeat(seq(',', choice($._value_id_and_type_attr, $.variadic)))),
     _value_id_and_type_attr: $ => seq($._function_arg, optional($.attribute),
       optional($.trailing_location)),
     _function_arg: $ => choice(seq($.value_use, ':', choice($.type, $.function_type)), $.value_use, $.type, $.function_type),

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@felixtensor/tree-sitter-mlir",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "MLIR grammar for tree-sitter",
   "type": "module",
   "repository": {