Improves Primitive Types (#508)

* adds bigint as built-in type, changes 'date' to 'Date', adds val converters

* tests added for bigint, Date, and boolean, noted distinct effect from mixed prop
type w/ bools

* omitting BIGINT & DATE from rule auto-converters, not needed at this time

packages/langium/src/grammar/generated/ast.ts

@@ -34,7 +34,7 @@ export function isCondition(item: unknown): item is Condition {
 
 export type FeatureName = string;
 
-export type PrimitiveType = 'boolean' | 'date' | 'number' | 'string';
+export type PrimitiveType = 'Date' | 'bigint' | 'boolean' | 'number' | 'string';
 
 export interface AbstractElement extends AstNode {
     readonly $container: Alternatives | Assignment | AtomType | CharacterRange | CrossReference | Group | NegatedToken | ParserRule | TerminalAlternatives | TerminalGroup | TerminalRule | UnorderedGroup | UntilToken;

packages/langium/src/grammar/generated/grammar.ts

@@ -576,7 +576,13 @@ export const LangiumGrammarGrammar = (): Grammar => loadedLangiumGrammarGrammar
           },
           {
             "$type": "Keyword",
-            "value": "date"
+            "value": "Date",
+            "elements": []
+          },
+          {
+            "$type": "Keyword",
+            "value": "bigint",
+            "elements": []
           }
         ]
       },

packages/langium/src/grammar/langium-grammar-validator.ts

@@ -559,7 +559,7 @@ export class LangiumGrammarValidator {
 
     checkTerminalRuleReturnType(rule: ast.TerminalRule, accept: ValidationAcceptor): void {
         if (rule.type?.name && !isPrimitiveType(rule.type.name)) {
-            accept('error', "Terminal rules can only return primitive types like 'string', 'boolean', 'number' or 'date'.", { node: rule.type, property: 'name' });
+            accept('error', "Terminal rules can only return primitive types like 'string', 'boolean', 'number', 'Date' or 'bigint'.", { node: rule.type, property: 'name' });
         }
     }
 
@@ -627,7 +627,7 @@ export class LangiumGrammarValidator {
     }
 }
 
-const primitiveTypes = ['string', 'number', 'boolean', 'Date'];
+const primitiveTypes = ['string', 'number', 'boolean', 'Date', 'bigint'];
 
 function isPrimitiveType(type: string): boolean {
     return primitiveTypes.includes(type);

packages/langium/src/grammar/langium-grammar.langium

@@ -31,7 +31,7 @@ AtomType:
     (primitiveType=PrimitiveType | isRef?='@'? refType=[AbstractType]) isArray?='[]'? | keywordType=Keyword;
 
 PrimitiveType returns string:
-    'string' | 'number' | 'boolean' | 'date';
+    'string' | 'number' | 'boolean' | 'Date' | 'bigint';
 
 type AbstractType = Interface | Type | Action | ParserRule;
 

packages/langium/src/parser/value-converter.ts

@@ -49,6 +49,8 @@ export class DefaultValueConverter implements ValueConverter {
         switch (getRuleType(rule)?.toLowerCase()) {
             case 'number': return convertNumber(input);
             case 'boolean': return convertBoolean(input);
+            case 'bigint': return convertBigint(input);
+            case 'date': return convertDate(input);
             default: return input;
         }
     }
@@ -70,6 +72,14 @@ export function convertInt(input: string): number {
     return parseInt(input);
 }
 
+export function convertBigint(input: string): bigint {
+    return BigInt(input);
+}
+
+export function convertDate(input: string): Date {
+    return new Date(input);
+}
+
 export function convertNumber(input: string): number {
     return Number(input);
 }

packages/langium/test/parser/langium-parser-builder.test.ts

@@ -207,17 +207,112 @@ describe('One name for terminal and non-terminal rules', () => {
 
 });
 
+describe('Boolean value converter', () => {
+    let parser: LangiumParser;
+    const content = `
+    grammar G
+    entry M: value?='true';
+    hidden terminal WS: /\\s+/;
+    `;
+
+    beforeAll(async () => {
+        const grammar = (await helper(content)).parseResult.value;
+        parser = parserFromGrammar(grammar);
+    });
+
+    function expectValue(input: string, value: unknown): void {
+        const main = parser.parse(input).value as unknown as { value: unknown };
+        expect(main.value).toBe(value);
+    }
+
+    test('Should have no definition errors', () => {
+        expect(parser.definitionErrors).toHaveLength(0);
+    });
+
+    test('Parsed Boolean is correct', () => {
+        expectValue('true', true);
+        // normal behavior when a property type can be resolved to only boolean
+        // gives us true/false values representing the parse result
+        expectValue('false', false);
+        expectValue('something-else-entirely', false);
+    });
+});
+
+describe('BigInt Parser value converter', () => {
+    let parser: LangiumParser;
+    const content = `
+    grammar G
+    entry M: value=BIGINT;
+    terminal BIGINT returns bigint: /[0-9]+/;
+    hidden terminal WS: /\\s+/;
+    `;
+
+    beforeAll(async () => {
+        const grammar = (await helper(content)).parseResult.value;
+        parser = parserFromGrammar(grammar);
+    });
+
+    function expectValue(input: string, value: unknown): void {
+        const main = parser.parse(input).value as unknown as { value: unknown };
+        expect(main.value).toBe(value);
+    }
+
+    test('Should have no definition errors', () => {
+        expect(parser.definitionErrors).toHaveLength(0);
+    });
+
+    test('Parsed BigInt is correct', () => {
+        expectValue('149587349587234971', BigInt('149587349587234971'));
+        expectValue('9007199254740991', BigInt('9007199254740991')); // === 0x1fffffffffffff
+    });
+});
+
+describe('Date Parser value converter', () => {
+    let parser: LangiumParser;
+    const content = `
+    grammar G
+    entry M: value=DATE;
+    terminal DATE returns Date: /[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}/;
+    hidden terminal WS: /\\s+/;
+    `;
+
+    beforeAll(async () => {
+        const grammar = (await helper(content)).parseResult.value;
+        parser = parserFromGrammar(grammar);
+    });
+
+    test('Should have no definition errors', () => {
+        expect(parser.definitionErrors).toHaveLength(0);
+    });
+
+    test('Parsed Date is correct Date object', () => {
+        const parseResult = parser.parse('2022-10-12T00:00').value as unknown as { value: unknown };
+        expect(parseResult.value).toEqual(new Date('2022-10-12T00:00'));
+    });
+});
+
 describe('Parser calls value converter', () => {
 
     let parser: LangiumParser;
     const content = `
     grammar TestGrammar
-    entry Main: value=(QFN|Number);
+    entry Main:
+        'big' BigVal |
+        'b' value?='true' |
+        'q' value=QFN |
+        'n' value=Number |
+        'd' value=DATE;
 
     QFN returns string: ID ('.' QFN)?;
     terminal ID: /\\^?[_a-zA-Z][\\w_]*/;
 
-    Number returns number: INT ('.' INT)?;
+    fragment BigVal: value=BINT 'n';
+    terminal BINT returns bigint: INT /(?=n)/;
+
+    terminal DATE returns Date: /[0-9]{4}-[0-9]{2}-[0-9]{2}(T[0-9]{2}:[0-9]{2})?/;
+
+    Number returns number:
+        INT ('.' INT)?;
     terminal INT returns number: /[0-9]+/;
 
     hidden terminal WS: /\\s+/;
@@ -230,32 +325,64 @@ describe('Parser calls value converter', () => {
     });
 
     function expectValue(input: string, value: unknown): void {
-        const main = parser.parse(input).value as unknown as { value: string };
+        const main = parser.parse(input).value as unknown as { value: unknown };
         expect(main.value).toBe(value);
     }
 
+    function expectEqual(input: string, value: unknown): void {
+        const main = parser.parse(input).value as unknown as { value: unknown };
+        expect(main.value).toEqual(value);
+    }
+
+    test('Should have no definition errors', () => {
+        expect(parser.definitionErrors).toHaveLength(0);
+    });
+
     test('Should parse ID inside of data type rule correctly', () => {
-        expectValue('^x', 'x');
+        expectValue('q ^x', 'x');
     });
 
     test('Should parse FQN correctly', () => {
-        expectValue('^x.y.^z', 'x.y.z');
+        expectValue('q ^x.y.^z', 'x.y.z');
     });
 
     test('Should parse FQN with whitespace correctly', () => {
-        expectValue('^x. y . ^z', 'x.y.z');
+        expectValue('q ^x. y . ^z', 'x.y.z');
     });
 
     test('Should parse FQN with comment correctly', () => {
-        expectValue('^x./* test */y.^z', 'x.y.z');
+        expectValue('q ^x./* test */y.^z', 'x.y.z');
     });
 
     test('Should parse integer correctly', () => {
-        expectValue('123', 123);
+        expectValue('n 123', 123);
     });
 
     test('Should parse float correctly', () => {
-        expectValue('123.5', 123.5);
+        expectValue('n 123.5', 123.5);
+    });
+
+    test('Should parse bool correctly', () => {
+        expectValue('b true', true);
+        // this is the current 'boolean' behavior when a prop type can't be resolved to just a boolean
+        // either true/undefined, no false in this case
+        expectValue('b false', undefined);
+        // ...then no distinguishing between the bad parse case when the type is unclear
+        expectValue('b asdfg', undefined);
+    });
+
+    test('Should parse BigInt correctly', () => {
+        expectValue('big 9007199254740991n', BigInt('9007199254740991'));
+        expectValue('big 9007199254740991', undefined);
+        expectValue('big 1.1', undefined);
+        expectValue('big -19458438592374', undefined);
+    });
+
+    test('Should parse Date correctly', () => {
+        expectEqual('d 2020-01-01', new Date('2020-01-01'));
+        expectEqual('d 2020-01-01T00:00', new Date('2020-01-01T00:00'));
+        expectEqual('d 2022-10-04T12:13', new Date('2022-10-04T12:13'));
+        expectEqual('d 2022-Peach', undefined);
     });
 });
 

packages/langium/test/validation/grammar-validation.test.ts

@@ -86,6 +86,39 @@ describe('Checked Named CrossRefs', () => {
     });
 });
 
+describe('Check grammar with primitives', () => {
+    const grammar = `
+    grammar PrimGrammar
+    entry Expr:
+        (String | Bool | Num | BigInt | DateObj)*;
+    String:
+        'String' val=STR;
+    Bool:
+        'Bool' val?='true';
+    Num:
+        'Num' val=NUM;
+    BigInt:
+        'BigInt' val=BIG 'n';
+    DateObj:
+        'Date' val=DATE;
+    terminal STR: /[_a-zA-Z][\\w_]*/;
+    terminal BIG returns bigint: /[0-9]+(?=n)/;
+    terminal NUM returns number: /[0-9]+(\\.[0-9])?/;
+    terminal DATE returns Date: /[0-9]{4}-{0-9}2-{0-9}2/+;
+    `.trim();
+
+    let validationData: ValidatorData;
+
+    // 1. build a parser from this grammar, verify it works
+    beforeAll(async () => {
+        validationData = await parseAndValidate(grammar);
+    });
+
+    test('No validation errors in grammar', () => {
+        expect(validationData.diagnostics.filter(d => d.severity === DiagnosticSeverity.Error)).toHaveLength(0);
+    });
+});
+
 interface ValidatorData {
     document: LangiumDocument;
     diagnostics: Diagnostic[];