tree-sitter-batch 0.8.0 → 0.11.0

package/README.md

@@ -6,11 +6,15 @@ Parses `.bat` and `.cmd` files into a concrete syntax tree for syntax highlighti
 
 ## Features
 
-- **Control flow** — `IF`/`ELSE` (EXIST, DEFINED, ERRORLEVEL, comparison with NOT), `FOR` (/D /R /L /F), `GOTO`, `CALL`
-- **Variables** — `SET` (plain, `/A` arithmetic, `/P` prompt), `%VAR%`, `!VAR!`, `%%i`, `%~dp0`, `%VAR:old=new%`
-- **Operators** — pipes `|`, redirects `>` `>>` `2>` `2>&1`, conditional `&&` `||`
-- **Structure** — labels `:name`, comments `REM` `::`, parenthesized blocks, `@ECHO OFF`
+- **Control flow** — `IF`/`ELSE` (EXIST, DEFINED, ERRORLEVEL, comparison with NOT), `FOR` (/D /R /L /F), `GOTO`, `CALL`, `EXIT /B`
+- **Variables** — `SET` (plain, `/A` arithmetic, `/P` prompt, display-only without `=`), `%VAR%`, `!VAR!`, `%%i`, `%~dp0`, `%VAR:old=new%`, subscripted delayed expansion `!ARR[%%i]!`, for-variable modifiers `%%~dpnxf` `%%~zS` `%%~$PATH:F`, escaped forms `%%VAR%%` `%%%%i`
+- **Echo** — free-form text with literal `(` `)` `!` `%`, inline strings, and variable references
+- **Operators** &mdash; pipes `|`, redirects `>` `>>` `<` `2>` `2>&1` (fds 0-9, including variable handles `>&%FD%` / `<&%FD%`), conditional `&&` `||`, separator `&`
+- **Line continuation** &mdash; trailing caret `^` joins the current line with the next (e.g. `"%JAVACMD%" ^` followed by indented arguments)
+- **Commands** &mdash; bare names, variable references as command, and quoted paths (`"C:\path\app.exe" args`, `call "%SCRIPT%" %*`)
+- **Structure** &mdash; labels `:name`, comments `REM` `::`, parenthesized blocks (including `@(...)`), `@ECHO OFF`, macro invocations, DosTips idioms `(call,)` / `(call;)` / `(call)` for ERRORLEVEL manipulation
 - **Scope** &mdash; `SETLOCAL`/`ENDLOCAL` with `ENABLEDELAYEDEXPANSION`
+- **Polyglot headers** &mdash; tolerates batch/PowerShell `<# ... #>` header lines and batch/VBScript lines marked with a trailing `'VBS` so SysToolsLib-style dual-language scripts parse cleanly
 - **Case-insensitive** &mdash; all keywords match regardless of casing
 
 ## Example
@@ -51,20 +55,20 @@ Parsed tree:
   (variable_assignment
     (set_keyword) (variable_name) (assignment_value))
   (variable_assignment
-    (set_keyword) (set_option) (variable_name) (assignment_value))
+    (set_keyword) (arithmetic_assignment (set_option) (arithmetic_expression)))
   (if_stmt
     (string)
     (parenthesized
       (cmd (command_name) (argument_list (argument_value)))))
   (for_stmt
     (for_variable)
-    (for_set)
+    (for_set (for_set_literal))
     (parenthesized
       (cmd (command_name) (argument_list (string) (string)))))
   (if_stmt
     (variable_reference)
     (comparison_op)
-    (integer)
+    (argument_value)
     (parenthesized
       (cmd (command_name) (argument_list (argument_value) (argument_value))))
     (else_clause

package/grammar.js

@@ -1,9 +1,33 @@
-const ci = (word) => new RegExp(word.split('').map((c) => /[a-zA-Z]/.test(c) ? `[${c.toLowerCase()}${c.toUpperCase()}]` : c).join(''));
+/// <reference types="tree-sitter-cli/dsl"/>
+// @ts-check
+
+/** @param {string} word */
+const ci = (word) => new RegExp(word.split('').map((/** @type {string} */ c) => /[a-zA-Z]/.test(c) ? `[${c.toLowerCase()}${c.toUpperCase()}]` : c).join(''));
+/** @param {string} word */
 const kw = (word) => token(prec(10, ci(word)));
+const varRefChoice = () => choice(
+  seq('%%', /[$@a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/, '%%'),
+  seq('%', /[$@+\-a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/, '%'),
+  seq('%~', /[a-zA-Z]*/, /[0-9]/),
+  seq('%', /[0-9]/),
+  '%*',
+  seq('%%%%%%', optional(seq('~', /[a-zA-Z$:_]*/)), /[a-zA-Z0-9.]/),
+  seq('%%%%', optional(seq('~', /[a-zA-Z$:_]*/)), /[a-zA-Z0-9.]/),
+  seq('%%', optional(seq('~', /[a-zA-Z$:_]*/)), /[a-zA-Z0-9.]/),
+  seq('%%', /[0-9]/),
+  seq('!', /[$%a-zA-Z_][$%a-zA-Z0-9_.#]*(?:\[[^!\r\n\]]*\])?/, '!'),
+  seq('%', /[$@a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/, ':', /[^%\r\n]+/, '%'),
+  seq('%', /[^%=\s\r\n]/, '%'),
+  seq('!', /[%$a-zA-Z_][%$a-zA-Z0-9_.#()\[\]]*/, /:[^!\r\n]+/, '!'),
+  seq('%', /[<>\/]+[@a-zA-Z_0-9.]*/, '%'),
+  seq('%', /\\[@a-zA-Z_0-9.]+/, '%'),
+  seq('%', /"[^"%\r\n]+"/, '%'),
+);
+/** @param {GrammarSymbols<string>} $ */
 const operand = ($) => [
   $.cond_exec, $.pipe_stmt, $.redirect_stmt, $.call_stmt, $.cmd, $.parenthesized,
   $.variable_assignment, $.goto_stmt, $.exit_stmt, $.setlocal_stmt, $.endlocal_stmt,
-  $.if_stmt, $.for_stmt, $.macro_invocation,
+  $.if_stmt, $.for_stmt, $.macro_invocation, $.echo_off,
 ];
 
 export default grammar({
@@ -27,6 +51,14 @@ export default grammar({
       seq(optional('@'), /[rR][eE][mM]/, optional(seq(/[ \t]/, /[^\r\n]*/))),
       seq(':', /[^$a-zA-Z_\r\n]/, /[^\r\n]*/),
       seq('%#', /[^\r\n]*/, '#%'),
+      // Polyglot batch/PowerShell header line: `<# ...` is a PS block-comment open
+      // that batch interprets as a (failing) redirect; we treat it as a comment.
+      seq('<#', /[^\r\n]*/),
+      // Batch+VBScript polyglot: lines containing a whitespace-prefixed `'VBS`
+      // marker are extracted by the script itself as VBScript; the batch
+      // interpreter never runs them. Consume the rest of the line after the
+      // marker so any trailing VBScript tokens don't leak into the batch parse.
+      seq(/[^'\r\n][^\r\n]*[ \t]'VBS/, /[^\r\n]*/),
     ))),
     label: () => token(seq(':', /[$a-zA-Z_][$a-zA-Z0-9_.#-]*/, optional(seq(/[ \t]/, /[^\r\n]*/)))),
     variable_assignment: ($) => prec(8, seq(
@@ -37,9 +69,9 @@ export default grammar({
         seq(
           /[ \t]+/,
           choice(
-            seq('"', repeat1(choice($.variable_reference, alias($._quoted_var_name_pattern, $.variable_name))), '=', optional($.quoted_assignment_value), '"', optional($.argument_list)),
-            seq('^"', repeat1(choice($.variable_reference, alias($._quoted_var_name_pattern, $.variable_name))), '=', optional($.caret_quoted_assignment_value), optional('^"')),
-            seq(choice($.variable_reference, alias($._var_name_pattern, $.variable_name)), '=', optional($.assignment_value)),
+            seq('"', repeat1(choice($.variable_reference, alias($._quoted_var_name_pattern, $.variable_name))), optional(seq('=', optional($.quoted_assignment_value))), '"', optional($.argument_list)),
+            seq(token(prec(2, '^"')), repeat1(choice($.variable_reference, alias($._quoted_var_name_pattern, $.variable_name))), optional(seq('=', optional($.caret_quoted_assignment_value))), optional(token(prec(2, '^"')))),
+            seq(repeat1(choice($.variable_reference, alias($._var_name_pattern, $.variable_name))), optional(seq('=', optional($.assignment_value)))),
           ),
         ),
       ),
@@ -53,7 +85,7 @@ export default grammar({
       alias('^', $.assignment_literal),
     ))),
     _var_name_pattern: () => token(choice(
-      /[$@a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/,
+      /[$@a-zA-Z_.][$@a-zA-Z0-9_.#()\[\]]*/,
       /\/[@a-zA-Z_][@a-zA-Z0-9_.#()\[\]]+/,
     )),
     _quoted_var_name_pattern: () => token(prec(1, choice(/[^\s="%][^="%]*/, /%%[a-zA-Z]?/, /"[^="\r\n]+"/))),
@@ -64,7 +96,10 @@ export default grammar({
     prompt_assignment: ($) => seq(
       optional(/[ \t]+/), alias(ci('/p'), $.set_option),
       optional(/[ \t]+/),
-      alias(/[@a-zA-Z_][@a-zA-Z0-9_()\[\]]*/, $.variable_name), '=', optional($.assignment_value),
+      choice(
+        seq(optional(alias($._var_name_pattern, $.variable_name)), '=', optional($.assignment_value)),
+        seq('"', optional(alias($._var_name_pattern, $.variable_name)), '=', optional($.quoted_assignment_value), '"', optional($.argument_list)),
+      ),
     ),
     arithmetic_expression: () => token(choice(
       seq('"', /[^"\r\n]*/, '"'),
@@ -76,6 +111,7 @@ export default grammar({
       alias(/[^%!()\r\n]+/, $.assignment_literal),
       alias('%', $.assignment_literal),
       alias('!', $.assignment_literal),
+      $._line_continuation,
     ))),
     assignment_paren_group: ($) => seq('(', repeat(choice(
       $.variable_reference,
@@ -93,11 +129,11 @@ export default grammar({
     ))),
     if_stmt: ($) => prec.right(8, seq(
       optional('@'), kw('if'),
-      optional(alias(ci('/i'), $.if_option)),
+      optional(alias(token(prec(2, ci('/i'))), $.if_option)),
       optional(kw('not')),
       choice(
-        seq(kw('exist'), choice($.string, $.variable_reference, $.argument_value)),
-        seq(kw('defined'), choice(/[$a-zA-Z_][$a-zA-Z0-9_.]*/, $.string, $.variable_reference)),
+        seq(kw('exist'), $._if_operand),
+        seq(kw('defined'), choice(/[$a-zA-Z_][$a-zA-Z0-9_.]*/, $._if_operand)),
         seq(kw('errorlevel'), $.integer),
         seq(
           $._if_operand,
@@ -122,10 +158,18 @@ export default grammar({
       $.redirect_stmt, $.pipe_stmt, $.comment,
     ),
     _if_operand: ($) => choice(
-      $.string, $.variable_reference, $.integer, $.bracketed_value,
-      alias($._if_word, $.argument_value),
+      prec.right(repeat1($.string)),
+      $.bracketed_value, $.paren_expression,
+      prec.right(seq(
+        choice($.variable_reference, alias($._if_word, $.argument_value), $.integer),
+        repeat(choice(
+          alias($._variable_reference_immediate, $.variable_reference),
+          alias($._if_word_rest, $.argument_value),
+        )),
+      )),
     ),
-    _if_word: () => token(prec(-1, /[^=<>\s\[\]"|&()][^=<>\s"|&()]*/)),
+    _if_word: () => token(prec(1, /[^=<>\s\[\]"|&()%!][^=<>\s"|&()%!]*/)),
+    _if_word_rest: () => token.immediate(/[^=<>\s\[\]"|&()%!][^=<>\s"|&()%!]*/),
     comparison_op: () => token(prec(10, choice('==', ci('equ'), ci('neq'), ci('lss'), ci('leq'), ci('gtr'), ci('geq')))),
     goto_stmt: ($) => prec(8, seq(
       optional('@'), kw('goto'),
@@ -137,8 +181,14 @@ export default grammar({
     )),
     call_stmt: ($) => prec(8, seq(
       optional('@'), kw('call'),
-      choice(token(seq(':', /[$a-zA-Z_][$a-zA-Z0-9_.#-]*/)), $.command_name, $.variable_reference),
-      optional($.argument_list),
+      optional(choice(
+        token.immediate(','),
+        token.immediate(';'),
+        seq(
+          choice(token(seq(':', /[$a-zA-Z_][$a-zA-Z0-9_.#-]*/)), $.command_name, $.variable_reference, $.string),
+          optional($.argument_list),
+        ),
+      )),
     )),
     exit_stmt: ($) => prec(8, seq(
       optional('@'), kw('exit'),
@@ -163,7 +213,7 @@ export default grammar({
       choice($.parenthesized, $._body_stmt),
     )),
     for_options: () => token(prec(10, choice(ci('/d'), seq(ci('/r'), optional(seq(/[ \t]+/, /(%[^\s%]|[^\s%])+%?/))), ci('/l'), seq(ci('/f'), optional(seq(/[ \t]+/, '"', /[^"]*/, '"')))))),
-    for_variable: () => token(seq('%%', optional('~'), /[a-zA-Z]/)),
+    for_variable: () => token(seq('%%', /[a-zA-Z0-9.]/)),
     for_set: ($) => prec.right(repeat1(choice(
       $.variable_reference,
       $.for_set_group,
@@ -179,18 +229,21 @@ export default grammar({
       alias('%', $.for_set_literal),
       alias('!', $.for_set_literal),
     )), ')'),
-    parenthesized: ($) => seq('(', repeat(choice(seq($._stmt, /\r?\n/), /\r?\n/)), optional($._stmt), ')'),
+    parenthesized: ($) => seq(optional('@'), '(', repeat(choice(seq($._stmt, /\r?\n/), /\r?\n/)), optional($._stmt), ')'),
     redirect_stmt: ($) => prec.right(4, choice(
-      seq(choice($.call_stmt, $.cmd, $.parenthesized), $.redirection),
-      seq($.redirection, choice($.call_stmt, $.cmd, $.parenthesized)),
+      seq(choice($.call_stmt, $.cmd, $.parenthesized, $.variable_assignment), $.redirection),
+      seq($.redirection, choice($.call_stmt, $.cmd, $.parenthesized, $.variable_assignment)),
     )),
     redirection: ($) => {
-      const file_redir = seq(optional(/[0-2]/), $.redirect_op, $.redirect_target);
+      const file_redir = seq($.redirect_op, $.redirect_target);
       const one_redir = choice(file_redir, $.fd_redirect);
       return prec.right(repeat1(one_redir));
     },
-    fd_redirect: () => token(seq(optional(/[0-2]/), '>&', /[0-9]/)),
-    redirect_op: () => token(choice('2>>', '2>', '>>', '>', '<')),
+    fd_redirect: ($) => choice(
+      token(seq(optional(/[0-9]/), />&|<&/, /[0-9]/)),
+      seq(alias(token(seq(optional(/[0-9]/), />&|<&/)), $.fd_redirect_op), $.variable_reference),
+    ),
+    redirect_op: () => token(choice(/[0-9]?>>/, /[0-9]?>/, /[0-9]?</)),
     redirect_target: () => token(choice(ci('nul'), ci('con'), /[^\s|&><\r\n]+/)),
     pipe_stmt: ($) => prec.left(3, seq(choice($.pipe_stmt, $.redirect_stmt, $.call_stmt, $.cmd, $.parenthesized), '|', choice($.redirect_stmt, $.call_stmt, $.cmd, $.parenthesized))),
     cond_exec: ($) => choice(
@@ -202,30 +255,31 @@ export default grammar({
       '&',
       choice(...operand($), $.comment),
     )),
-    variable_reference: () => token(choice(
-      seq('%', /[$@a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/, '%'),
-      seq('%~', /[a-zA-Z]*/, /[0-9]/),
-      seq('%', /[0-9]/),
-      '%*',
-      seq('%%', optional('~'), /[a-zA-Z]/),
-      seq('!', /[$%a-zA-Z_][$a-zA-Z0-9_.#]*/, '!'),
-      seq('%', /[$@a-zA-Z_][$@a-zA-Z0-9_.#()\[\]]*/, ':', /[^%\r\n]+/, '%'),
-      seq('%', /[^%=\s\r\n]/, '%'),
-      seq('!', /[%$a-zA-Z_][%$a-zA-Z0-9_.#()\[\]]*/, /:[^!\r\n]+/, '!'),
-      seq('%', /[<>\/]+[@a-zA-Z_0-9.]*/, '%'),
-      seq('%', /\\[@a-zA-Z_0-9.]+/, '%'),
-      seq('%', /"[^"%\r\n]+"/, '%'),
-    )),
+    variable_reference: () => token(varRefChoice()),
+    _variable_reference_immediate: () => token.immediate(varRefChoice()),
     string: () => token(seq('"', /[^"\r\n]*/, '"')),
     bracketed_value: ($) => seq('[', repeat(choice($.variable_reference, alias(token(/[^%!\[\]\r\n]+/), $.bracketed_literal))), ']'),
-    cmd: ($) => prec.right(5, seq(optional('@'), choice($.command_name, $.variable_reference), optional($.argument_list))),
+    cmd: ($) => prec.right(5, choice(
+      seq(optional('@'), alias(kw('echo'), $.command_name), optional(alias($._echo_text, $.argument_list))),
+      seq(optional('@'), choice($.command_name, $.variable_reference, $.string), optional($.argument_list)),
+    )),
+    _echo_text: ($) => prec.right(repeat1(choice(
+      $.variable_reference,
+      $.string,
+      alias($._echo_literal, $.argument_value),
+      $._line_continuation,
+    ))),
+    // Trailing `^` escapes the following newline, joining the next physical
+    // line onto the current logical one. Treated as whitespace between tokens.
+    _line_continuation: () => token(/\^\r?\n[ \t]*/),
+    _echo_literal: () => token(/(?:\^[&|<>^()]|[^\s|&><"\r\n%!()])+|[()!%]/),
     macro_invocation: ($) => prec.right(6, seq($.variable_reference, $.parenthesized, optional($.else_clause))),
-    command_name: () => /[$a-zA-Z_][$a-zA-Z0-9_.#-]*/,
+    command_name: () => /[$a-zA-Z_0-9][$a-zA-Z0-9_.#-]*/,
     argument_list: ($) => prec.right(repeat1($._arg)),
-    _arg: ($) => choice($.string, $.variable_reference, $.command_option, $.paren_expression, $.argument_value),
+    _arg: ($) => choice($.string, $.variable_reference, $.command_option, $.paren_expression, $.argument_value, $._line_continuation),
     command_option: () => token(seq('/', /[a-zA-Z_?][a-zA-Z0-9_:]*/)),
     paren_expression: ($) => seq('(', repeat($._arg), ')'),
-    argument_value: () => /(?:\^[&|<>^()]|[^\s|&><"\r\n%!()])(?:\^[&|<>^()]|[^\s|&><"\r\n()])*/,
+    argument_value: () => /(?:\^[&|<>^()]|[^\s|&><"\r\n%!()])(?:\^[&|<>^()]|[^\s|&><"\r\n()])*|[!%]/,
     integer: () => /[0-9]+/,
   },
 });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "tree-sitter-batch",
-  "version": "0.8.0",
+  "version": "0.11.0",
   "description": "A Windows Batch/CMD grammar for tree-sitter",
   "type": "module",
   "repository": {
@@ -34,11 +34,16 @@
     "node-gyp-build": "^4.8.4"
   },
   "devDependencies": {
-    "eslint": "^9.15.0",
+    "eslint": "^10.2.0",
     "eslint-config-treesitter": "^1.0.2",
+    "globals": "^17.5.0",
     "prebuildify": "^6.0.1",
-    "tree-sitter-cli": "0.24.7",
-    "tree-sitter-go-types": "^0.1.0"
+    "tree-sitter-cli": "0.26.8",
+    "tree-sitter-go-types": "^0.1.0",
+    "typescript": "^6.0.3"
+  },
+  "overrides": {
+    "eslint-plugin-jsdoc": "^62.9.0"
   },
   "peerDependencies": {
     "tree-sitter": ">=0.25.0"
@@ -52,8 +57,9 @@
     "install": "node-gyp-build",
     "prestart": "tree-sitter build --wasm",
     "start": "tree-sitter playground",
-    "generate": "tree-sitter generate && tree-sitter-go-types",
+    "generate": "tree-sitter generate --abi 14 && tree-sitter-go-types",
     "lint": "eslint grammar.js",
+    "typecheck": "tsc --noEmit -p tsconfig.json",
     "test": "node --test bindings/node/*_test.js"
   }
 }