| 1 | import { chevrotain } from '../../../lib.js'; |
| 2 | import { MacroLexer } from './MacroLexer.js'; |
| 3 | |
| 4 | const { CstParser } = chevrotain; |
| 5 | |
| 6 | /** @typedef {import('chevrotain').TokenType} TokenType */ |
| 7 | /** @typedef {import('chevrotain').CstNode} CstNode */ |
| 8 | /** @typedef {import('chevrotain').ILexingError} ILexingError */ |
| 9 | /** @typedef {import('chevrotain').IRecognitionException} IRecognitionException */ |
| 10 | |
| 11 | /** |
| 12 | * The singleton instance of the MacroParser. |
| 13 | * |
| 14 | * @type {MacroParser} |
| 15 | */ |
| 16 | let instance; |
| 17 | export { instance as MacroParser }; |
| 18 | |
| 19 | class MacroParser extends CstParser { |
| 20 | /** @type {MacroParser} */ static #instance; |
| 21 | /** @type {MacroParser} */ static get instance() { return MacroParser.#instance ?? (MacroParser.#instance = new MacroParser()); } |
| 22 | |
| 23 | /** @private */ |
| 24 | constructor() { |
| 25 | super(MacroLexer.def, { |
| 26 | traceInitPerf: false, |
| 27 | nodeLocationTracking: 'full', |
| 28 | recoveryEnabled: true, |
| 29 | }); |
| 30 | const Tokens = MacroLexer.tokens; |
| 31 | |
| 32 | const $ = this; |
| 33 | |
| 34 | // Top-level document rule that can handle both plaintext and macros |
| 35 | $.document = $.RULE('document', () => { |
| 36 | $.MANY(() => { |
| 37 | $.OR([ |
| 38 | { ALT: () => $.CONSUME(Tokens.Plaintext, { LABEL: 'plaintext' }) }, |
| 39 | { ALT: () => $.CONSUME(Tokens.PlaintextOpenBrace, { LABEL: 'plaintext' }) }, |
| 40 | { ALT: () => $.SUBRULE($.macro) }, |
| 41 | { ALT: () => $.CONSUME(Tokens.Macro.Start, { LABEL: 'plaintext' }) }, |
| 42 | ]); |
| 43 | }); |
| 44 | }); |
| 45 | |
| 46 | // Basic Macro Structure - can be either a regular macro or a variable expression |
| 47 | $.macro = $.RULE('macro', () => { |
| 48 | $.CONSUME(Tokens.Macro.Start); |
| 49 | |
| 50 | // Optional flags before the identifier (e.g., {{!user}}, {{?~macro}}, {{>filtered}}) |
| 51 | // Both regular flags and filter flag are captured under the 'flags' label |
| 52 | $.MANY(() => { |
| 53 | $.OR1([ |
| 54 | { ALT: () => $.CONSUME(Tokens.Macro.Flags, { LABEL: 'flags' }) }, |
| 55 | { ALT: () => $.CONSUME(Tokens.Macro.FilterFlag, { LABEL: 'flags' }) }, |
| 56 | ]); |
| 57 | }); |
| 58 | |
| 59 | // Branch: either a variable expression (starts with . or $) or a regular macro |
| 60 | $.OR([ |
| 61 | // Variable expression branch |
| 62 | { ALT: () => $.SUBRULE($.variableExpr) }, |
| 63 | // Regular macro branch |
| 64 | { ALT: () => $.SUBRULE($.macroBody) }, |
| 65 | ]); |
| 66 | |
| 67 | $.CONSUME(Tokens.Macro.End); |
| 68 | }); |
| 69 | |
| 70 | // Regular macro body (flags + identifier + optional arguments) |
| 71 | $.macroBody = $.RULE('macroBody', () => { |
| 72 | // Macro identifier (name) |
| 73 | $.OR2([ |
| 74 | { ALT: () => $.CONSUME(Tokens.Macro.DoubleSlash, { LABEL: 'Macro.identifier' }) }, |
| 75 | { ALT: () => $.CONSUME(Tokens.Macro.Identifier, { LABEL: 'Macro.identifier' }) }, |
| 76 | ]); |
| 77 | $.OPTION(() => $.SUBRULE($.arguments)); |
| 78 | }); |
| 79 | |
| 80 | // Variable expression: .varName or $varName with optional operator |
| 81 | $.variableExpr = $.RULE('variableExpr', () => { |
| 82 | // Variable scope prefix |
| 83 | $.OR3([ |
| 84 | { ALT: () => $.CONSUME(Tokens.Var.LocalPrefix, { LABEL: 'Var.scope' }) }, |
| 85 | { ALT: () => $.CONSUME(Tokens.Var.GlobalPrefix, { LABEL: 'Var.scope' }) }, |
| 86 | ]); |
| 87 | |
| 88 | // Variable identifier (name) |
| 89 | $.CONSUME(Tokens.Var.Identifier, { LABEL: 'Var.identifier' }); |
| 90 | |
| 91 | // Optional operator (and expression, if operator requires one) |
| 92 | $.OPTION2(() => $.SUBRULE($.variableOperator)); |
| 93 | }); |
| 94 | |
| 95 | // Variable operator: ++, --, = value, += value, -= value, ||, ??, ||=, ??=, ==, !=, >, >=, <, <= |
| 96 | $.variableOperator = $.RULE('variableOperator', () => { |
| 97 | $.OR4([ |
| 98 | { ALT: () => $.CONSUME(Tokens.Var.Operators.Increment, { LABEL: 'Var.operator' }) }, |
| 99 | { ALT: () => $.CONSUME(Tokens.Var.Operators.Decrement, { LABEL: 'Var.operator' }) }, |
| 100 | { |
| 101 | ALT: () => { |
| 102 | $.OR5([ |
| 103 | { ALT: () => $.CONSUME(Tokens.Var.Operators.NullishCoalescingEquals, { LABEL: 'Var.operator' }) }, |
| 104 | { ALT: () => $.CONSUME(Tokens.Var.Operators.NullishCoalescing, { LABEL: 'Var.operator' }) }, |
| 105 | { ALT: () => $.CONSUME(Tokens.Var.Operators.LogicalOrEquals, { LABEL: 'Var.operator' }) }, |
| 106 | { ALT: () => $.CONSUME(Tokens.Var.Operators.LogicalOr, { LABEL: 'Var.operator' }) }, |
| 107 | { ALT: () => $.CONSUME(Tokens.Var.Operators.MinusEquals, { LABEL: 'Var.operator' }) }, |
| 108 | { ALT: () => $.CONSUME(Tokens.Var.Operators.DoubleEquals, { LABEL: 'Var.operator' }) }, |
| 109 | { ALT: () => $.CONSUME(Tokens.Var.Operators.NotEquals, { LABEL: 'Var.operator' }) }, |
| 110 | { ALT: () => $.CONSUME(Tokens.Var.Operators.GreaterThanOrEqual, { LABEL: 'Var.operator' }) }, |
| 111 | { ALT: () => $.CONSUME(Tokens.Var.Operators.GreaterThan, { LABEL: 'Var.operator' }) }, |
| 112 | { ALT: () => $.CONSUME(Tokens.Var.Operators.LessThanOrEqual, { LABEL: 'Var.operator' }) }, |
| 113 | { ALT: () => $.CONSUME(Tokens.Var.Operators.LessThan, { LABEL: 'Var.operator' }) }, |
| 114 | { ALT: () => $.CONSUME(Tokens.Var.Operators.PlusEquals, { LABEL: 'Var.operator' }) }, |
| 115 | { ALT: () => $.CONSUME(Tokens.Var.Operators.Equals, { LABEL: 'Var.operator' }) }, |
| 116 | ]); |
| 117 | $.SUBRULE($.variableValue, { LABEL: 'Var.value' }); |
| 118 | }, |
| 119 | }, |
| 120 | ]); |
| 121 | }); |
| 122 | |
| 123 | // Variable value: everything after = or += until the end |
| 124 | // Can contain nested macros and any other tokens |
| 125 | $.variableValue = $.RULE('variableValue', () => { |
| 126 | $.MANY2(() => { |
| 127 | $.OR5([ |
| 128 | { ALT: () => $.SUBRULE($.macro) }, // Nested macros |
| 129 | { ALT: () => $.CONSUME(Tokens.Identifier) }, |
| 130 | { ALT: () => $.CONSUME(Tokens.Unknown) }, |
| 131 | ]); |
| 132 | }); |
| 133 | }); |
| 134 | |
| 135 | // Arguments Parsing |
| 136 | $.arguments = $.RULE('arguments', () => { |
| 137 | $.OR([ |
| 138 | { |
| 139 | ALT: () => { |
| 140 | $.CONSUME(Tokens.Args.DoubleColon, { LABEL: 'separator' }); |
| 141 | $.AT_LEAST_ONE_SEP({ |
| 142 | SEP: Tokens.Args.DoubleColon, |
| 143 | DEF: () => $.SUBRULE($.argument, { LABEL: 'argument' }), |
| 144 | }); |
| 145 | }, |
| 146 | }, |
| 147 | { |
| 148 | ALT: () => { |
| 149 | $.OPTION(() => { |
| 150 | $.CONSUME(Tokens.Args.Colon, { LABEL: 'separator' }); |
| 151 | }); |
| 152 | $.SUBRULE($.argumentAllowingColons, { LABEL: 'argument' }); |
| 153 | }, |
| 154 | // So, this is a bit hacky. But implemented below, the argument capture does explicitly exclude double colons |
| 155 | // from being captured as the first token. The potential ambiguity chevrotain claims here is not possible. |
| 156 | // It says stuff like <Args.DoubleColon, Identifier/Macro/Unknown> is possible in both branches, but it is not. |
| 157 | IGNORE_AMBIGUITIES: true, |
| 158 | }, |
| 159 | ]); |
| 160 | }); |
| 161 | |
| 162 | // List the argument tokens here, as we need two rules, one to be able to parse with double colons and one without |
| 163 | const validArgumentTokens = [ |
| 164 | { ALT: () => $.SUBRULE($.macro) }, // Nested Macros |
| 165 | { ALT: () => $.CONSUME(Tokens.Identifier) }, |
| 166 | { ALT: () => $.CONSUME(Tokens.Unknown) }, |
| 167 | { ALT: () => $.CONSUME(Tokens.Args.Colon) }, |
| 168 | { ALT: () => $.CONSUME(Tokens.Args.Equals) }, |
| 169 | { ALT: () => $.CONSUME(Tokens.Args.Quote) }, |
| 170 | ]; |
| 171 | |
| 172 | $.argument = $.RULE('argument', () => { |
| 173 | $.MANY(() => { |
| 174 | $.OR([...validArgumentTokens]); |
| 175 | }); |
| 176 | }); |
| 177 | $.argumentAllowingColons = $.RULE('argumentAllowingColons', () => { |
| 178 | $.AT_LEAST_ONE(() => { |
| 179 | $.OR([ |
| 180 | ...validArgumentTokens, |
| 181 | { ALT: () => $.CONSUME(Tokens.Args.DoubleColon) }, |
| 182 | ]); |
| 183 | }); |
| 184 | }); |
| 185 | |
| 186 | this.performSelfAnalysis(); |
| 187 | } |
| 188 | |
| 189 | /** |
| 190 | * Parses a document into a CST. |
| 191 | * |
| 192 | * @param {string} input |
| 193 | * @returns {{ cst: CstNode|null, errors: ({ message: string }|ILexingError|IRecognitionException)[] , lexingErrors: ILexingError[], parserErrors: IRecognitionException[] }} |
| 194 | */ |
| 195 | parseDocument(input) { |
| 196 | if (!input) { |
| 197 | return { cst: null, errors: [{ message: 'Input is empty' }], lexingErrors: [], parserErrors: [] }; |
| 198 | } |
| 199 | |
| 200 | const lexingResult = MacroLexer.tokenize(input); |
| 201 | |
| 202 | this.input = lexingResult.tokens; |
| 203 | const cst = this.document(); |
| 204 | |
| 205 | const errors = [ |
| 206 | ...lexingResult.errors, |
| 207 | ...this.errors, |
| 208 | ]; |
| 209 | |
| 210 | return { cst, errors, lexingErrors: lexingResult.errors, parserErrors: this.errors }; |
| 211 | } |
| 212 | |
| 213 | test(input) { |
| 214 | const lexingResult = MacroLexer.tokenize(input); |
| 215 | // "input" is a setter which will reset the parser's state. |
| 216 | this.input = lexingResult.tokens; |
| 217 | const cst = this.macro(); |
| 218 | |
| 219 | // For testing purposes we need to actually persist the error messages in the object, |
| 220 | // otherwise the test cases cannot read those, as they don't have access to the exception object type. |
| 221 | const errors = this.errors.map(x => ({ message: x.message, ...x, stack: x.stack })); |
| 222 | |
| 223 | return { cst, errors: errors }; |
| 224 | } |
| 225 | } |
| 226 | |
| 227 | instance = MacroParser.instance; |