@scinorandex/sparse 0.0.8 → 0.1.2

package/dist/utils/loadFiles.js

@@ -0,0 +1,47 @@
+"use strict";
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.loadTable = exports.loadGrammar = void 0;
+const promises_1 = __importDefault(require("fs/promises"));
+const common_1 = require("../meta/common");
+const selfhosted_1 = require("../table/selfhosted");
+const selfhosted_2 = require("../meta/selfhosted");
+const loadGrammar = async (grammarPath, options = {}) => {
+    let source;
+    try {
+        source = await promises_1.default.readFile(grammarPath, "utf8");
+    }
+    catch (err) {
+        return {
+            success: false,
+            reason: `Could not read the grammar file: ${err instanceof Error ? err.message : String(err)}`,
+            token: (0, common_1.syntheticGrammarToken)(),
+        };
+    }
+    const result = (0, selfhosted_2.tryBuildProductions)(source, options);
+    if (result.success === false)
+        return Object.assign(Object.assign({}, result), { reason: `${result.reason} (in ${grammarPath})` });
+    return result;
+};
+exports.loadGrammar = loadGrammar;
+const loadTable = async (tablePath, options = {}) => {
+    let source;
+    try {
+        source = await promises_1.default.readFile(tablePath, "utf8");
+    }
+    catch (err) {
+        return {
+            success: false,
+            reason: `Could not read the parsing table file: ${err instanceof Error ? err.message : String(err)}`,
+            token: (0, common_1.syntheticGrammarToken)(),
+        };
+    }
+    const result = (0, selfhosted_1.tryBuildStates)(source, Object.assign(Object.assign({}, options), { source: tablePath }));
+    if (result.success === false)
+        return Object.assign(Object.assign({}, result), { reason: `${result.reason} (in ${tablePath})` });
+    return result;
+};
+exports.loadTable = loadTable;
+//# sourceMappingURL=loadFiles.js.map

package/dist/utils/loadFiles.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"loadFiles.js","sourceRoot":"","sources":["../../src/utils/loadFiles.ts"],"names":[],"mappings":";;;;;;AAAA,2DAA6B;AAC7B,2CAA+F;AAE/F,oDAAyE;AACzE,mDAAyD;AAMlD,MAAM,WAAW,GAAG,KAAK,EAC9B,WAAmB,EACnB,UAA8B,EAAE,EACD,EAAE;IACjC,IAAI,MAAc,CAAC;IACnB,IAAI,CAAC;QACH,MAAM,GAAG,MAAM,kBAAE,CAAC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;IAClD,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO;YACL,OAAO,EAAE,KAAK;YACd,MAAM,EAAE,oCAAoC,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE;YAC9F,KAAK,EAAE,IAAA,8BAAqB,GAAE;SAC/B,CAAC;IACJ,CAAC;IAED,MAAM,MAAM,GAAG,IAAA,gCAAmB,EAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACpD,IAAI,MAAM,CAAC,OAAO,KAAK,KAAK;QAC1B,uCAAY,MAAM,KAAE,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,QAAQ,WAAW,GAAG,IAAG;IACvE,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAnBW,QAAA,WAAW,eAmBtB;AAKK,MAAM,SAAS,GAAG,KAAK,EAC5B,SAAiB,EACjB,UAA4B,EAAE,EACC,EAAE;IACjC,IAAI,MAAc,CAAC;IACnB,IAAI,CAAC;QACH,MAAM,GAAG,MAAM,kBAAE,CAAC,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;IAChD,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO;YACL,OAAO,EAAE,KAAK;YACd,MAAM,EAAE,0CAA0C,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE;YACpG,KAAK,EAAE,IAAA,8BAAqB,GAAE;SAC/B,CAAC;IACJ,CAAC;IAED,MAAM,MAAM,GAAG,IAAA,2BAAc,EAAC,MAAM,kCAAO,OAAO,KAAE,MAAM,EAAE,SAAS,IAAG,CAAC;IACzE,IAAI,MAAM,CAAC,OAAO,KAAK,KAAK;QAC1B,uCAAY,MAAM,KAAE,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,QAAQ,SAAS,GAAG,IAAG;IACrE,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAnBW,QAAA,SAAS,aAmBpB"}

package/dist/utils/reducers.d.ts

@@ -0,0 +1,9 @@
+import { Production } from "../meta/common";
+export declare const KLEENE_REDUCER_NAME = "autogenerated-kleene";
+export declare const namedProductions: (productions: Production[]) => string[];
+export declare const missingReducerNames: (productions: Production[], reducers: Record<string, unknown>) => string[];
+export declare const missingReducerMessage: (productions: Production[], reducers: Record<string, unknown>) => string;
+export type ReducerMap<Node> = Record<string, (newInput: any, oldInput: any) => Node>;
+export declare const defineReducers: <Reducers extends Record<string, (...args: never[]) => unknown>>(reducers: Reducers) => (newInput: any, oldInput: any) => any;
+export declare const assertReducersCoverGrammar: (productions: Production[], reducers: ReducerMap<unknown>) => void;
+export declare const enumToString: <TokenType = unknown>(tokenTypes: object) => (tokenType: TokenType) => string;

package/dist/utils/reducers.js

@@ -0,0 +1,56 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.enumToString = exports.assertReducersCoverGrammar = exports.defineReducers = exports.missingReducerMessage = exports.missingReducerNames = exports.namedProductions = exports.KLEENE_REDUCER_NAME = void 0;
+exports.KLEENE_REDUCER_NAME = "autogenerated-kleene";
+const namedProductions = (productions) => {
+    const names = new Set();
+    for (const production of productions)
+        if (production.name != null)
+            names.add(production.name);
+    return [...names].sort();
+};
+exports.namedProductions = namedProductions;
+const missingReducerNames = (productions, reducers) => (0, exports.namedProductions)(productions).filter((name) => reducers[name] == null);
+exports.missingReducerNames = missingReducerNames;
+const describeKleene = (productions) => {
+    const needed = productions.filter((production) => production.name === exports.KLEENE_REDUCER_NAME ||
+        production.identifier.startsWith("<autogen-"));
+    if (needed.length === 0)
+        return "";
+    return `\nNote: this grammar uses "*" or "+", so it has ${needed.length} autogenerated list production(s) named "${exports.KLEENE_REDUCER_NAME}". That reducer has to flatten the list it is handed (see README, "Repetition with * and +").`;
+};
+const missingReducerMessage = (productions, reducers) => {
+    const missing = (0, exports.missingReducerNames)(productions, reducers);
+    if (missing.length === 0)
+        return "";
+    const has = (0, exports.namedProductions)(productions).filter((name) => reducers[name] != null);
+    const kleene = missing.includes(exports.KLEENE_REDUCER_NAME) ? describeKleene(productions) : "";
+    return (`No reducer for ${missing.length === 1 ? "production" : "productions"} ${missing
+        .map((name) => `"${name}"`)
+        .join(", ")}.` +
+        `\nReducers are keyed by the name given in the grammar, e.g. "<PROGRAM: program>" is reduced by reducers.program.` +
+        (has.length === 0 ? "\nYour reducer map is empty." : `\nYou have reducers for: ${has.join(", ")}.`) +
+        kleene);
+};
+exports.missingReducerMessage = missingReducerMessage;
+const defineReducers = (reducers) => function sparseReducer(newInput, oldInput) {
+    const name = newInput === null || newInput === void 0 ? void 0 : newInput.name;
+    const lookup = reducers;
+    const reducer = name == null ? undefined : lookup[name];
+    if (reducer == null)
+        throw new Error(`No reducer for production "${name !== null && name !== void 0 ? name : ""}" (production ${oldInput === null || oldInput === void 0 ? void 0 : oldInput.index}). Reducers are keyed by the name given in the grammar.`);
+    return reducer(newInput, oldInput);
+};
+exports.defineReducers = defineReducers;
+const assertReducersCoverGrammar = (productions, reducers) => {
+    const message = (0, exports.missingReducerMessage)(productions, reducers);
+    if (message !== "")
+        throw new Error(message);
+};
+exports.assertReducersCoverGrammar = assertReducersCoverGrammar;
+const enumToString = (tokenTypes) => (tokenType) => {
+    const name = tokenTypes[tokenType];
+    return `${name !== null && name !== void 0 ? name : tokenType}`;
+};
+exports.enumToString = enumToString;
+//# sourceMappingURL=reducers.js.map

package/dist/utils/reducers.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reducers.js","sourceRoot":"","sources":["../../src/utils/reducers.ts"],"names":[],"mappings":";;;AAEa,QAAA,mBAAmB,GAAG,sBAAsB,CAAC;AAGnD,MAAM,gBAAgB,GAAG,CAAC,WAAyB,EAAY,EAAE;IACtE,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;IAChC,KAAK,MAAM,UAAU,IAAI,WAAW;QAClC,IAAI,UAAU,CAAC,IAAI,IAAI,IAAI;YAAE,KAAK,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;IAC1D,OAAO,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;AAC3B,CAAC,CAAC;AALW,QAAA,gBAAgB,oBAK3B;AAGK,MAAM,mBAAmB,GAAG,CACjC,WAAyB,EACzB,QAAiC,EACvB,EAAE,CAAC,IAAA,wBAAgB,EAAC,WAAW,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,CAAC;AAHzE,QAAA,mBAAmB,uBAGsD;AAEtF,MAAM,cAAc,GAAG,CAAC,WAAyB,EAAE,EAAE;IACnD,MAAM,MAAM,GAAG,WAAW,CAAC,MAAM,CAC/B,CAAC,UAAU,EAAE,EAAE,CACb,UAAU,CAAC,IAAI,KAAK,2BAAmB;QACvC,UAAU,CAAC,UAAU,CAAC,UAAU,CAAC,WAAW,CAAC,CAChD,CAAC;IACF,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEnC,OAAO,mDAAmD,MAAM,CAAC,MAAM,4CAA4C,2BAAmB,+FAA+F,CAAC;AACxO,CAAC,CAAC;AAEK,MAAM,qBAAqB,GAAG,CACnC,WAAyB,EACzB,QAAiC,EACzB,EAAE;IACV,MAAM,OAAO,GAAG,IAAA,2BAAmB,EAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;IAC3D,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEpC,MAAM,GAAG,GAAG,IAAA,wBAAgB,EAAC,WAAW,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,CAAC;IACnF,MAAM,MAAM,GAAG,OAAO,CAAC,QAAQ,CAAC,2BAAmB,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAExF,OAAO,CACL,kBAAkB,OAAO,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,aAAa,IAAI,OAAO;SAC7E,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,IAAI,GAAG,CAAC;SAC1B,IAAI,CAAC,IAAI,CAAC,GAAG;QAChB,kHAAkH;QAClH,CAAC,GAAG,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,8BAA8B,CAAC,CAAC,CAAC,4BAA4B,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;QACnG,MAAM,CACP,CAAC;AACJ,CAAC,CAAC;AAlBW,QAAA,qBAAqB,yBAkBhC;AAWK,MAAM,cAAc,GAAG,CAAiE,QAAkB,EAAE,EAAE,CACnH,SAAS,aAAa,CAAC,QAAa,EAAE,QAAa;IACjD,MAAM,IAAI,GAA8B,QAAQ,aAAR,QAAQ,uBAAR,QAAQ,CAAE,IAAI,CAAC;IACvD,MAAM,MAAM,GAAG,QAA8D,CAAC;IAC9E,MAAM,OAAO,GAAG,IAAI,IAAI,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IACxD,IAAI,OAAO,IAAI,IAAI;QACjB,MAAM,IAAI,KAAK,CACb,8BAA8B,IAAI,aAAJ,IAAI,cAAJ,IAAI,GAAI,EAAE,iBAAiB,QAAQ,aAAR,QAAQ,uBAAR,QAAQ,CAAE,KAAK,yDAAyD,CAClI,CAAC;IACJ,OAAO,OAAO,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;AACrC,CAAC,CAAC;AAVS,QAAA,cAAc,kBAUvB;AAMG,MAAM,0BAA0B,GAAG,CAAC,WAAyB,EAAE,QAA6B,EAAE,EAAE;IACrG,MAAM,OAAO,GAAG,IAAA,6BAAqB,EAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;IAC7D,IAAI,OAAO,KAAK,EAAE;QAAE,MAAM,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC;AAC/C,CAAC,CAAC;AAHW,QAAA,0BAA0B,8BAGrC;AAUK,MAAM,YAAY,GACvB,CAAsB,UAAkB,EAAE,EAAE,CAC5C,CAAC,SAAoB,EAAU,EAAE;IAC/B,MAAM,IAAI,GAAI,UAAsC,CAAC,SAA8B,CAAC,CAAC;IACrF,OAAO,GAAG,IAAI,aAAJ,IAAI,cAAJ,IAAI,GAAI,SAAS,EAAE,CAAC;AAChC,CAAC,CAAC;AALS,QAAA,YAAY,gBAKrB"}

package/example/LoLang/example.ts

@@ -1,9 +1,9 @@
 import fs from "fs/promises";
-import { buildProductions, LR1StackSymbol, Sparse } from "../../src/index";
+import { LR1StackSymbol, Sparse, buildProductions, buildStates } from "../../src/index";
 import { ColumnAndRow, Slex, Token } from "@scinorandex/slex";
 
 // prettier-ignore
-enum LoLangTokenType {
+export enum LoLangTokenType {
   EOF,
 
   IMPORT,
@@ -42,7 +42,7 @@ enum LoLangTokenType {
   SINGLE_LINE_COMMENT, MULTI_LINE_COMMENT
 }
 
-const lexerGenerator = new Slex<LoLangTokenType, {}>({
+export const lexerGenerator = new Slex<LoLangTokenType, {}>({
   EOF_TYPE: LoLangTokenType.EOF,
   isHigherPrecedence: ({ current, next }) => current === LoLangTokenType.IDENTIFIER,
   // Specifies that SINGLE_LINE_COMMENT and MULTI_LINE_COMMENT tokens should be ignored.
@@ -165,17 +165,18 @@ class Node {
   }
 }
 
-async function main() {
+async function parse(tablePath: string) {
   const exampleSourceCode = await fs.readFile("./example/LoLang/Features_Array_Methods.lol", "utf8");
   const lexer = lexerGenerator.generate(exampleSourceCode, () => ({}));
 
   const toStringifiedTokenType = (type: LoLangTokenType) => LoLangTokenType[type];
   const productions = buildProductions(await fs.readFile("./example/LoLang/grammar.txt", "utf8"));
-  const parserGenerator = Sparse.fromProductions<LoLangTokenType, {}, Node>({ productions, toStringifiedTokenType });
+  const states = buildStates(await fs.readFile(tablePath, "utf8"));
+  const parserGenerator = new Sparse<LoLangTokenType, {}, Node>({ productions, toStringifiedTokenType, states });
 
   const parser = parserGenerator.generate(lexer, {
     reducer: (_, { input }) => new Node(input),
-    recover({ lexer, statesStack, symbolsStack, isSafe, addError, crash, finish, states }) {
+    recover({ lexer, statesStack, symbolsStack, isSafe, addError, crash, finish, insertToken, states }) {
       let token = lexer.peekNextToken();
       const currentState = statesStack.peek();
 
@@ -188,25 +189,26 @@ async function main() {
       token = lexer.peekNextToken();
 
       // check for lacking semicolons
-      if (states[currentState].actions.has("[SEMICOLON]")) {
-        const action = states[currentState].getTerminalAction("SEMICOLON");
+      const semicolonAction = states[currentState].getTerminalAction("SEMICOLON");
+      if (semicolonAction != null) {
+        // A semicolon the author forgot. It gets the same position as the token we choked on.
         const newToken = new Token<LoLangTokenType, {}>(
           LoLangTokenType.SEMICOLON,
           ";",
-          new ColumnAndRow(token.column, token.line),
+          new ColumnAndRow(token.line, token.column),
           token.metadata,
         );
 
-        if (action) {
-          if (action.type === "reduce") return finish({ newToken });
+        // If the parser only wanted a semicolon to reduce, hand it back as the next token.
+        if (semicolonAction.type === "reduce") return finish({ newToken });
 
-          statesStack.push(action.value);
-          symbolsStack.push({ type: "token", token: newToken });
-          addError(
-            `Expected SEMICOLON but received ${LoLangTokenType[token.type]}. SEMICOLON was automatically inserted.`,
-          );
-        }
+        // Otherwise shift the synthesized semicolon onto the stacks and carry on from there.
+        const inserted = insertToken(newToken);
+        if (inserted != null) return crash(inserted.reason);
 
+        addError(
+          `Expected SEMICOLON but received ${LoLangTokenType[token.type]}. SEMICOLON was automatically inserted.`,
+        );
         if (isSafe()) return finish();
       }
 
@@ -215,7 +217,29 @@ async function main() {
     },
   });
 
-  console.log(parser.parse());
+  const parsingResult = parser.parse();
+  return JSON.stringify(parsingResult.result?.toObject());
+}
+
+async function main() {
+  let nsTime = Date.now();
+  const lr1Result = await parse("./example/LoLang/table.txt");
+  const lr1Time = Date.now() - nsTime;
+
+  nsTime = Date.now();
+  const lalrResult = await parse("./example/LoLang/tablelalr.txt");
+  const lalrTime = Date.now() - nsTime;
+
+  console.log(`LR(1) time: ${lr1Time}ms`);
+  console.log(`LALR(1) time: ${lalrTime}ms`);
+  const delta = Math.abs(lalrTime - lr1Time);
+  console.log("LR(1) is " + (lr1Time < lalrTime ? "faster" : "slower") + " than LALR(1)" + ` (${delta}ms)`);
+
+  console.log("==============================");
+  console.log(`LR(1) result: ${lr1Result}`);
+  console.log(`LALR(1) result: ${lalrResult}`);
+  console.log("==============================");
+  console.log(lr1Result === lalrResult ? "Equal output" : "Not equal output");
 }
 
 main();