tree-sitter-delphi13 1.2.1 → 1.2.2

package/README.md

@@ -2,15 +2,22 @@
 
 [![npm](https://img.shields.io/npm/v/tree-sitter-delphi13.svg)](https://www.npmjs.com/package/tree-sitter-delphi13)
 [![license](https://img.shields.io/npm/l/tree-sitter-delphi13.svg)](LICENSE)
+[![corpus](https://img.shields.io/badge/valid_Delphi_13-100.000%25-brightgreen.svg)](CORPUS-CEILING-REPORT.md)
+[![corpus raw](https://img.shields.io/badge/full_corpus_raw-99.82%25-green.svg)](CORPUS-CEILING-REPORT.md)
 
 Tree-sitter grammar pipeline focused on **Delphi 13 (RAD Studio 37.0 / Florence)** and the surrounding modern Delphi ecosystem.
 
-**v1.2.1: 100.000% pass rate on compilable Delphi 13** (zero ERROR nodes), measured
-on an 11,722-unique-file corpus spanning the RAD Studio 13 RTL/VCL (including
-`System.pas`), DevExpress, Spring4D, EurekaLog, Indy, FireDAC, AsyncPro/Orpheus,
-mORMot-era community code and a full production application. Every failure that
-remains is individually diagnosed, dcc32-cross-checked, and *supposed* to fail
-(see [CORPUS-CEILING-REPORT.md](CORPUS-CEILING-REPORT.md)).
+**v1.2.2: 100.000% pass rate on compilable Delphi 13** (zero ERROR nodes), measured
+on a corpus spanning the RAD Studio 13 RTL/VCL (including `System.pas`), DevExpress,
+Spring4D, EurekaLog, Indy, FireDAC, AsyncPro/Orpheus, **JCL/JVCL (JEDI)**, mORMot-era
+community code and a full production application — 23,385 deduplicated files. Every
+failure that remains is individually diagnosed, dcc32-cross-checked, and *supposed*
+to fail (see [CORPUS-CEILING-REPORT.md](CORPUS-CEILING-REPORT.md)).
+
+v1.2.2 added the JEDI (JCL + JVCL) libraries to the corpus and closed the three
+valid-Delphi-13 gaps they surfaced: a subrange bound that is a nested constant-expr
+call (`2..Succ(High(T))`), the `at` soft keyword as an `inherited` method name
+(`inherited At(...)`), and `Operator` used as a record/class field name.
 
 | measurement basis | pass rate |
 |---|---|
@@ -31,9 +38,9 @@ watchlist).
 |---|---|
 | **`tree-sitter-delphi13`** *(this directory)* | Master grammar — Delphi-13 syntax, all proven fixes from upstream merged + new extensions. Self-contained "THEN-wins" parser (98.6% raw on its own). |
 | **[`tree-sitter-delphi13-pure`](pure/)** | Simpler sub-grammar that drops `pp_*` tokens entirely. Expects preprocessor-resolved source. |
-| **[`delphi13-preprocessor`](preprocessor/)** | Standalone JavaScript library that resolves `{$IFDEF}` / `{$IF}` / `{$DEFINE}` / `{$I X.inc}` directives as a text transformation. |
+| **[`delphi13-preprocessor`](preprocessor/)** | Standalone JavaScript library that resolves `{$IFDEF}` / `{$IF}` / `{$DEFINE}` / `{$I X.inc}` directives as a text transformation. The **canonical** preprocessor is now a pure-Delphi port of this library (shipping inside drag-lint); this JS package remains the byte-for-byte oracle and the drop-in choice for JavaScript consumers. |
 
-The full pipeline `preprocessor → pure → AST` is what reaches **99.97%** on valid
+The full pipeline `preprocessor → pure → AST` is what reaches **100.000%** on valid
 Delphi 13. Use the master grammar alone (98.6% raw) if you don't need
 conditional-compilation awareness; use the pipeline if you do.
 
@@ -138,7 +145,9 @@ node-gyp rebuild
 
 - **Inline `asm` blocks** treated as opaque text by design. Not tree-sitter-asm.
 - **`{$IF declared(SymName)}`** always evaluates to `false` — no symbol table. ~10 corpus files use this. A lightweight symbol pre-scan would address most cases.
-- **Include search path** is currently same-directory only. Cross-directory `{$I ..\Common\X.inc}` requires a search-path option (see TODO.md).
+- **Include resolution** is nearest-first (same directory, its subdirectories, then
+  up to 3 parent levels with their subdirectories — v1.1.0 of the preprocessor).
+  Arbitrary project search paths still need the `includePaths` option.
 - **FreePascal-specific extensions** (operator overloading variants, `{$mode ObjFPC}` directives, macOS PasCocoa namespaces) are not covered. A separate `tree-sitter-delphi-plus-fp` repo is planned.
 
 ## Acknowledgements

package/grammar.js

@@ -756,7 +756,16 @@ module.exports = grammar({
 			field('body', choice(tr($, 'block'), tr($, 'asm'))),
 		),
 
-		inherited:       $ => prec.right(seq($.kInherited, optional($.identifier))),
+		inherited:       $ => prec.right(seq($.kInherited, optional(choice(
+			$.identifier,
+			// Soft keyword `at` (the raise-with-address `raise E at Addr`
+			// keyword) used as an INHERITED method name: `inherited At(Offset)`
+			// (JCL JclCLR.pas). `At` already works as a plain/dotted call and a
+			// field name; only the inherited slot lost it to the kAtWord token.
+			// Aliased to identifier so the call-args attach exactly as for
+			// `inherited Foo(x)`. Narrow — only the corpus-confirmed kAtWord.
+			alias($.kAtWord, $.identifier),
+		)))),
 
 		exprDot:         $ => choice(
 			op.infix(5, $._ref, $.kDot, $._ref),
@@ -1288,6 +1297,13 @@ module.exports = grammar({
 			// conflicts with typerefDot in this position.)
 			seq($.identifier, '.', $.identifier),
 			seq($.identifier, '(', $.identifier, ')'),
+			// Iter 2026-07-17: nested const-expr call bound —
+			// `2 .. Succ(High(TDigitValue))` (JCL JclSysUtils TNumericSystemBase).
+			// dcc accepts any constant ordinal expression as a bound; the grammar
+			// stays with the narrow corpus-driven enumeration (a wide `_expr`
+			// bound blows the parser tables, see above), adding just the one
+			// real shape: `Ident( Ident(Ident) )`, one level of call nesting.
+			seq($.identifier, '(', seq($.identifier, '(', $.identifier, ')'), ')'),
 			// `hid - 1`, `MaxN + 2` — narrow `identifier OP integer` form for
 			// real-corpus patterns like `TNmbrRange = 0 .. hid - 1;`.
 			// Kept narrow (no full _expr) to avoid the conflict cascade hit
@@ -1663,6 +1679,12 @@ module.exports = grammar({
 			field('name', delimited1(choice(
 				$.identifier,
 				alias($.kRegister, $.identifier),
+				// Iter 2026-07-17: `Operator` as a FIELD NAME
+				// (`Operator: TJvXmlSQLOperator;` — JVCL JvXmlDatabase). The
+				// `operator` token otherwise starts _declOperator; GLR splits
+				// on the follow token (`:` -> field, operator-name -> class
+				// operator). Same narrow one-keyword alias as kRegister above.
+				alias($.kOperator, $.identifier),
 			))),
 			':',
 			// Iter 2026-07-06: a record/class field type may be a subrange, same

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "tree-sitter-delphi13",
-  "version": "1.2.1",
+  "version": "1.2.2",
   "description": "Tree-sitter grammar focused on Delphi 13 (RAD Studio 37.0)",
   "main": "bindings/node",
   "types": "bindings/node",

package/src/grammar.json

@@ -3671,8 +3671,22 @@
             "type": "CHOICE",
             "members": [
               {
-                "type": "SYMBOL",
-                "name": "identifier"
+                "type": "CHOICE",
+                "members": [
+                  {
+                    "type": "SYMBOL",
+                    "name": "identifier"
+                  },
+                  {
+                    "type": "ALIAS",
+                    "content": {
+                      "type": "SYMBOL",
+                      "name": "kAtWord"
+                    },
+                    "named": true,
+                    "value": "identifier"
+                  }
+                ]
               },
               {
                 "type": "BLANK"
@@ -7195,6 +7209,44 @@
             }
           ]
         },
+        {
+          "type": "SEQ",
+          "members": [
+            {
+              "type": "SYMBOL",
+              "name": "identifier"
+            },
+            {
+              "type": "STRING",
+              "value": "("
+            },
+            {
+              "type": "SEQ",
+              "members": [
+                {
+                  "type": "SYMBOL",
+                  "name": "identifier"
+                },
+                {
+                  "type": "STRING",
+                  "value": "("
+                },
+                {
+                  "type": "SYMBOL",
+                  "name": "identifier"
+                },
+                {
+                  "type": "STRING",
+                  "value": ")"
+                }
+              ]
+            },
+            {
+              "type": "STRING",
+              "value": ")"
+            }
+          ]
+        },
         {
           "type": "SEQ",
           "members": [
@@ -9658,6 +9710,15 @@
                                 },
                                 "named": true,
                                 "value": "identifier"
+                              },
+                              {
+                                "type": "ALIAS",
+                                "content": {
+                                  "type": "SYMBOL",
+                                  "name": "kOperator"
+                                },
+                                "named": true,
+                                "value": "identifier"
                               }
                             ]
                           },
@@ -9689,6 +9750,15 @@
                     },
                     "named": true,
                     "value": "identifier"
+                  },
+                  {
+                    "type": "ALIAS",
+                    "content": {
+                      "type": "SYMBOL",
+                      "name": "kOperator"
+                    },
+                    "named": true,
+                    "value": "identifier"
                   }
                 ]
               }