UNPKG

chevrotain

Version:

Chevrotain is a high performance fault tolerant javascript parsing DSL for building recursive decent parsers

590 lines 24.4 kB
"use strict"; Object.defineProperty(exports, "__esModule", { value: true }); var tokens_public_1 = require("./tokens_public"); var lexer_public_1 = require("./lexer_public"); var utils_1 = require("../utils/utils"); var PATTERN = "PATTERN"; exports.DEFAULT_MODE = "defaultMode"; exports.MODES = "modes"; exports.SUPPORT_STICKY = typeof new RegExp("(?:)").sticky === "boolean"; function disableSticky() { exports.SUPPORT_STICKY = false; } exports.disableSticky = disableSticky; function enableSticky() { exports.SUPPORT_STICKY = true; } exports.enableSticky = enableSticky; function analyzeTokenTypes(tokenTypes, useSticky) { if (useSticky === void 0) { useSticky = exports.SUPPORT_STICKY; } var onlyRelevantTypes = utils_1.reject(tokenTypes, function (currType) { return currType[PATTERN] === lexer_public_1.Lexer.NA; }); var hasCustom = false; var allTransformedPatterns = utils_1.map(onlyRelevantTypes, function (currType) { var currPattern = currType[PATTERN]; if (utils_1.isRegExp(currPattern)) { var regExpSource = currPattern.source; if (regExpSource.length === 1 && // only these regExp meta characters which can appear in a length one regExp regExpSource !== "^" && regExpSource !== "$" && regExpSource !== ".") { return regExpSource; } else if (regExpSource.length === 2 && regExpSource[0] === "\\" && // not a meta character !utils_1.contains([ "d", "D", "s", "S", "t", "r", "n", "t", "0", "c", "b", "B", "f", "v", "w", "W" ], regExpSource[1])) { // escaped meta Characters: /\+/ /\[/ // or redundant escaping: /\a/ // without the escaping "\" return regExpSource[1]; } else { return useSticky ? addStickyFlag(currPattern) : addStartOfInput(currPattern); } } else if (utils_1.isFunction(currPattern)) { hasCustom = true; // CustomPatternMatcherFunc - custom patterns do not require any transformations, only wrapping in a RegExp Like object return { exec: currPattern }; } else if (utils_1.has(currPattern, "exec")) { hasCustom = true; // ICustomPattern return currPattern; } else if (typeof currPattern === "string") { // IGNORE ABOVE ELSE if (currPattern.length === 1) { return currPattern; } else { var escapedRegExpString = currPattern.replace(/[\\^$.*+?()[\]{}|]/g, "\\$&"); var wrappedRegExp = new RegExp(escapedRegExpString); // TODO: extract the "?" expression, it is duplicated return useSticky ? addStickyFlag(wrappedRegExp) : addStartOfInput(wrappedRegExp); } } else { /* istanbul ignore next */ throw Error("non exhaustive match"); } }); var patternIdxToType = utils_1.map(onlyRelevantTypes, function (currType) { return currType.tokenTypeIdx; }); var patternIdxToGroup = utils_1.map(onlyRelevantTypes, function (clazz) { var groupName = clazz.GROUP; if (groupName === lexer_public_1.Lexer.SKIPPED) { return undefined; } else if (utils_1.isString(groupName)) { return groupName; } else if (utils_1.isUndefined(groupName)) { return false; } else { /* istanbul ignore next */ throw Error("non exhaustive match"); } }); var patternIdxToLongerAltIdx = utils_1.map(onlyRelevantTypes, function (clazz) { var longerAltType = clazz.LONGER_ALT; if (longerAltType) { var longerAltIdx = utils_1.indexOf(onlyRelevantTypes, longerAltType); return longerAltIdx; } }); var patternIdxToPushMode = utils_1.map(onlyRelevantTypes, function (clazz) { return clazz.PUSH_MODE; }); var patternIdxToPopMode = utils_1.map(onlyRelevantTypes, function (clazz) { return utils_1.has(clazz, "POP_MODE"); }); var patternIdxToCanLineTerminator = utils_1.map(onlyRelevantTypes, function (clazz) { return clazz.LINE_BREAKS === true; }); var patternIdxToIsCustom = utils_1.map(onlyRelevantTypes, isCustomPattern); var patternIdxToShort = utils_1.map(allTransformedPatterns, isShortPattern); var emptyGroups = utils_1.reduce(onlyRelevantTypes, function (acc, clazz) { var groupName = clazz.GROUP; if (utils_1.isString(groupName) && !(groupName === lexer_public_1.Lexer.SKIPPED)) { acc[groupName] = []; } return acc; }, {}); var patternIdxToConfig = utils_1.map(allTransformedPatterns, function (x, idx) { return { pattern: allTransformedPatterns[idx], longerAlt: patternIdxToLongerAltIdx[idx], canLineTerminator: patternIdxToCanLineTerminator[idx], isCustom: patternIdxToIsCustom[idx], short: patternIdxToShort[idx], group: patternIdxToGroup[idx], push: patternIdxToPushMode[idx], pop: patternIdxToPopMode[idx], tokenTypeIdx: patternIdxToType[idx], tokenType: onlyRelevantTypes[idx] }; }); return { emptyGroups: emptyGroups, patternIdxToConfig: patternIdxToConfig, hasCustom: hasCustom }; } exports.analyzeTokenTypes = analyzeTokenTypes; function validatePatterns(tokenTypes, validModesNames) { var errors = []; var missingResult = findMissingPatterns(tokenTypes); errors = errors.concat(missingResult.errors); var invalidResult = findInvalidPatterns(missingResult.valid); var validTokenTypes = invalidResult.valid; errors = errors.concat(invalidResult.errors); errors = errors.concat(validateRegExpPattern(validTokenTypes)); errors = errors.concat(findInvalidGroupType(validTokenTypes)); errors = errors.concat(findModesThatDoNotExist(validTokenTypes, validModesNames)); errors = errors.concat(findUnreachablePatterns(validTokenTypes)); return errors; } exports.validatePatterns = validatePatterns; function validateRegExpPattern(tokenTypes) { var errors = []; var withRegExpPatterns = utils_1.filter(tokenTypes, function (currTokType) { return utils_1.isRegExp(currTokType[PATTERN]); }); errors = errors.concat(findEndOfInputAnchor(withRegExpPatterns)); errors = errors.concat(findStartOfInputAnchor(withRegExpPatterns)); errors = errors.concat(findUnsupportedFlags(withRegExpPatterns)); errors = errors.concat(findDuplicatePatterns(withRegExpPatterns)); errors = errors.concat(findEmptyMatchRegExps(withRegExpPatterns)); return errors; } function findMissingPatterns(tokenTypes) { var tokenTypesWithMissingPattern = utils_1.filter(tokenTypes, function (currType) { return !utils_1.has(currType, PATTERN); }); var errors = utils_1.map(tokenTypesWithMissingPattern, function (currType) { return { message: "Token Type: ->" + tokens_public_1.tokenName(currType) + "<- missing static 'PATTERN' property", type: lexer_public_1.LexerDefinitionErrorType.MISSING_PATTERN, tokenTypes: [currType] }; }); var valid = utils_1.difference(tokenTypes, tokenTypesWithMissingPattern); return { errors: errors, valid: valid }; } exports.findMissingPatterns = findMissingPatterns; function findInvalidPatterns(tokenTypes) { var tokenTypesWithInvalidPattern = utils_1.filter(tokenTypes, function (currType) { var pattern = currType[PATTERN]; return (!utils_1.isRegExp(pattern) && !utils_1.isFunction(pattern) && !utils_1.has(pattern, "exec") && !utils_1.isString(pattern)); }); var errors = utils_1.map(tokenTypesWithInvalidPattern, function (currType) { return { message: "Token Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'PATTERN' can only be a RegExp, a" + " Function matching the {CustomPatternMatcherFunc} type or an Object matching the {ICustomPattern} interface.", type: lexer_public_1.LexerDefinitionErrorType.INVALID_PATTERN, tokenTypes: [currType] }; }); var valid = utils_1.difference(tokenTypes, tokenTypesWithInvalidPattern); return { errors: errors, valid: valid }; } exports.findInvalidPatterns = findInvalidPatterns; var end_of_input = /[^\\][\$]/; function findEndOfInputAnchor(tokenTypes) { var invalidRegex = utils_1.filter(tokenTypes, function (currType) { var pattern = currType[PATTERN]; return end_of_input.test(pattern.source); }); var errors = utils_1.map(invalidRegex, function (currType) { return { message: "Unexpected RegExp Anchor Error:\n" + "\tToken Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'PATTERN' cannot contain end of input anchor '$'\n" + "\tSee sap.github.io/chevrotain/website/Building_Grammars/resolving_lexer_errors.html#ANCHORS" + "\tfor details.", type: lexer_public_1.LexerDefinitionErrorType.EOI_ANCHOR_FOUND, tokenTypes: [currType] }; }); return errors; } exports.findEndOfInputAnchor = findEndOfInputAnchor; function findEmptyMatchRegExps(tokenTypes) { var matchesEmptyString = utils_1.filter(tokenTypes, function (currType) { var pattern = currType[PATTERN]; return pattern.test(""); }); var errors = utils_1.map(matchesEmptyString, function (currType) { return { message: "Token Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'PATTERN' must not match an empty string", type: lexer_public_1.LexerDefinitionErrorType.EMPTY_MATCH_PATTERN, tokenTypes: [currType] }; }); return errors; } exports.findEmptyMatchRegExps = findEmptyMatchRegExps; var start_of_input = /[^\\[][\^]|^\^/; function findStartOfInputAnchor(tokenTypes) { var invalidRegex = utils_1.filter(tokenTypes, function (currType) { var pattern = currType[PATTERN]; return start_of_input.test(pattern.source); }); var errors = utils_1.map(invalidRegex, function (currType) { return { message: "Unexpected RegExp Anchor Error:\n" + "\tToken Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'PATTERN' cannot contain start of input anchor '^'\n" + "\tSee https://github.com/SAP/chevrotain/blob/master/docs/resolving_lexer_errors.md#ANCHORS\n" + "\tfor details.", type: lexer_public_1.LexerDefinitionErrorType.SOI_ANCHOR_FOUND, tokenTypes: [currType] }; }); return errors; } exports.findStartOfInputAnchor = findStartOfInputAnchor; function findUnsupportedFlags(tokenTypes) { var invalidFlags = utils_1.filter(tokenTypes, function (currType) { var pattern = currType[PATTERN]; return (pattern instanceof RegExp && (pattern.multiline || pattern.global)); }); var errors = utils_1.map(invalidFlags, function (currType) { return { message: "Token Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'PATTERN' may NOT contain global('g') or multiline('m')", type: lexer_public_1.LexerDefinitionErrorType.UNSUPPORTED_FLAGS_FOUND, tokenTypes: [currType] }; }); return errors; } exports.findUnsupportedFlags = findUnsupportedFlags; // This can only test for identical duplicate RegExps, not semantically equivalent ones. function findDuplicatePatterns(tokenTypes) { var found = []; var identicalPatterns = utils_1.map(tokenTypes, function (outerType) { return utils_1.reduce(tokenTypes, function (result, innerType) { if (outerType.PATTERN.source === innerType.PATTERN.source && !utils_1.contains(found, innerType) && innerType.PATTERN !== lexer_public_1.Lexer.NA) { // this avoids duplicates in the result, each Token Type may only appear in one "set" // in essence we are creating Equivalence classes on equality relation. found.push(innerType); result.push(innerType); return result; } return result; }, []); }); identicalPatterns = utils_1.compact(identicalPatterns); var duplicatePatterns = utils_1.filter(identicalPatterns, function (currIdenticalSet) { return currIdenticalSet.length > 1; }); var errors = utils_1.map(duplicatePatterns, function (setOfIdentical) { var tokenTypeNames = utils_1.map(setOfIdentical, function (currType) { return tokens_public_1.tokenName(currType); }); var dupPatternSrc = utils_1.first(setOfIdentical).PATTERN; return { message: "The same RegExp pattern ->" + dupPatternSrc + "<-" + ("has been used in all of the following Token Types: " + tokenTypeNames.join(", ") + " <-"), type: lexer_public_1.LexerDefinitionErrorType.DUPLICATE_PATTERNS_FOUND, tokenTypes: setOfIdentical }; }); return errors; } exports.findDuplicatePatterns = findDuplicatePatterns; function findInvalidGroupType(tokenTypes) { var invalidTypes = utils_1.filter(tokenTypes, function (clazz) { if (!utils_1.has(clazz, "GROUP")) { return false; } var group = clazz.GROUP; return group !== lexer_public_1.Lexer.SKIPPED && group !== lexer_public_1.Lexer.NA && !utils_1.isString(group); }); var errors = utils_1.map(invalidTypes, function (currType) { return { message: "Token Type: ->" + tokens_public_1.tokenName(currType) + "<- static 'GROUP' can only be Lexer.SKIPPED/Lexer.NA/A String", type: lexer_public_1.LexerDefinitionErrorType.INVALID_GROUP_TYPE_FOUND, tokenTypes: [currType] }; }); return errors; } exports.findInvalidGroupType = findInvalidGroupType; function findModesThatDoNotExist(tokenTypes, validModes) { var invalidModes = utils_1.filter(tokenTypes, function (clazz) { return (clazz.PUSH_MODE !== undefined && !utils_1.contains(validModes, clazz.PUSH_MODE)); }); var errors = utils_1.map(invalidModes, function (clazz) { var msg = "Token Type: ->" + tokens_public_1.tokenName(clazz) + "<- static 'PUSH_MODE' value cannot refer to a Lexer Mode ->" + clazz.PUSH_MODE + "<-" + "which does not exist"; return { message: msg, type: lexer_public_1.LexerDefinitionErrorType.PUSH_MODE_DOES_NOT_EXIST, tokenTypes: [clazz] }; }); return errors; } exports.findModesThatDoNotExist = findModesThatDoNotExist; function findUnreachablePatterns(tokenTypes) { var errors = []; var canBeTested = utils_1.reduce(tokenTypes, function (result, tokType, idx) { var pattern = tokType.PATTERN; if (pattern === lexer_public_1.Lexer.NA) { return result; } // a more comprehensive validation for all forms of regExps would require // deeper regExp analysis capabilities if (utils_1.isString(pattern)) { result.push({ str: pattern, idx: idx, tokenType: tokType }); } else if (utils_1.isRegExp(pattern) && noMetaChar(pattern)) { result.push({ str: pattern.source, idx: idx, tokenType: tokType }); } return result; }, []); utils_1.forEach(tokenTypes, function (tokType, testIdx) { utils_1.forEach(canBeTested, function (_a) { var str = _a.str, idx = _a.idx, tokenType = _a.tokenType; if (testIdx < idx && testTokenType(str, tokType.PATTERN)) { var msg = "Token: ->" + tokens_public_1.tokenName(tokenType) + "<- can never be matched.\n" + ("Because it appears AFTER the Token Type ->" + tokens_public_1.tokenName(tokType) + "<-") + "in the lexer's definition.\n" + "See http://sap.github.io/chevrotain/website/Building_Grammars/resolving_lexer_errors.html#UNREACHABLE"; errors.push({ message: msg, type: lexer_public_1.LexerDefinitionErrorType.UNREACHABLE_PATTERN, tokenTypes: [tokType, tokenType] }); } }); }); return errors; } exports.findUnreachablePatterns = findUnreachablePatterns; function testTokenType(str, pattern) { if (utils_1.isRegExp(pattern)) { var regExpArray = pattern.exec(str); return regExpArray !== null && regExpArray.index === 0; } else if (utils_1.isFunction(pattern)) { // maintain the API of custom patterns return pattern(str, 0, [], {}); } else if (utils_1.has(pattern, "exec")) { // maintain the API of custom patterns return pattern.exec(str, 0, [], {}); } else if (typeof pattern === "string") { return pattern === str; } else { /* istanbul ignore next */ throw Error("non exhaustive match"); } } function noMetaChar(regExp) { //https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp var metaChars = [ ".", "\\", "[", "]", "|", "^", "$", "(", ")", "?", "*", "+", "{" ]; return (utils_1.find(metaChars, function (char) { return regExp.source.indexOf(char) !== -1; }) === undefined); } function addStartOfInput(pattern) { var flags = pattern.ignoreCase ? "i" : ""; // always wrapping in a none capturing group preceded by '^' to make sure matching can only work on start of input. // duplicate/redundant start of input markers have no meaning (/^^^^A/ === /^A/) return new RegExp("^(?:" + pattern.source + ")", flags); } exports.addStartOfInput = addStartOfInput; function addStickyFlag(pattern) { var flags = pattern.ignoreCase ? "iy" : "y"; // always wrapping in a none capturing group preceded by '^' to make sure matching can only work on start of input. // duplicate/redundant start of input markers have no meaning (/^^^^A/ === /^A/) return new RegExp("" + pattern.source, flags); } exports.addStickyFlag = addStickyFlag; function performRuntimeChecks(lexerDefinition, trackLines) { var errors = []; // some run time checks to help the end users. if (!utils_1.has(lexerDefinition, exports.DEFAULT_MODE)) { errors.push({ message: "A MultiMode Lexer cannot be initialized without a <" + exports.DEFAULT_MODE + "> property in its definition\n", type: lexer_public_1.LexerDefinitionErrorType.MULTI_MODE_LEXER_WITHOUT_DEFAULT_MODE }); } if (!utils_1.has(lexerDefinition, exports.MODES)) { errors.push({ message: "A MultiMode Lexer cannot be initialized without a <" + exports.MODES + "> property in its definition\n", type: lexer_public_1.LexerDefinitionErrorType.MULTI_MODE_LEXER_WITHOUT_MODES_PROPERTY }); } if (utils_1.has(lexerDefinition, exports.MODES) && utils_1.has(lexerDefinition, exports.DEFAULT_MODE) && !utils_1.has(lexerDefinition.modes, lexerDefinition.defaultMode)) { errors.push({ message: "A MultiMode Lexer cannot be initialized with a " + exports.DEFAULT_MODE + ": <" + lexerDefinition.defaultMode + ">" + "which does not exist\n", type: lexer_public_1.LexerDefinitionErrorType.MULTI_MODE_LEXER_DEFAULT_MODE_VALUE_DOES_NOT_EXIST }); } if (utils_1.has(lexerDefinition, exports.MODES)) { utils_1.forEach(lexerDefinition.modes, function (currModeValue, currModeName) { utils_1.forEach(currModeValue, function (currTokType, currIdx) { if (utils_1.isUndefined(currTokType)) { errors.push({ message: "A Lexer cannot be initialized using an undefined Token Type. Mode:" + ("<" + currModeName + "> at index: <" + currIdx + ">\n"), type: lexer_public_1.LexerDefinitionErrorType.LEXER_DEFINITION_CANNOT_CONTAIN_UNDEFINED }); } }); }); } var allTokenTypes = utils_1.flatten(utils_1.mapValues(lexerDefinition.modes, function (tokTypes) { return tokTypes; })); if (trackLines && utils_1.find(allTokenTypes, function (currTokType) { return currTokType.LINE_BREAKS; }) === undefined) { errors.push({ message: "No LINE_BREAKS Error:\n" + "\tThis Lexer has been defined to track line and column information,\n" + "\tyet none of the Token definitions contain a LINE_BREAK flag.\n" + "\tSee http://sap.github.io/chevrotain/website/Building_Grammars/resolving_lexer_errors.html#LINE_BREAKS \n" + "\tfor details.", type: lexer_public_1.LexerDefinitionErrorType.NO_LINE_BREAKS_FLAGS }); } return errors; } exports.performRuntimeChecks = performRuntimeChecks; function cloneEmptyGroups(emptyGroups) { var clonedResult = {}; var groupKeys = utils_1.keys(emptyGroups); utils_1.forEach(groupKeys, function (currKey) { var currGroupValue = emptyGroups[currKey]; /* istanbul ignore else */ if (utils_1.isArray(currGroupValue)) { clonedResult[currKey] = []; } else { /* istanbul ignore next */ throw Error("non exhaustive match"); } }); return clonedResult; } exports.cloneEmptyGroups = cloneEmptyGroups; // TODO: refactor to avoid duplication function isCustomPattern(tokenType) { var pattern = tokenType.PATTERN; if (utils_1.isRegExp(pattern)) { return false; } else if (utils_1.isFunction(pattern)) { // CustomPatternMatcherFunc - custom patterns do not require any transformations, only wrapping in a RegExp Like object return true; } else if (utils_1.has(pattern, "exec")) { // ICustomPattern return true; } else if (utils_1.isString(pattern)) { return false; } else { /* istanbul ignore next */ throw Error("non exhaustive match"); } } exports.isCustomPattern = isCustomPattern; function isShortPattern(pattern) { if (utils_1.isString(pattern) && pattern.length === 1) { return pattern.charCodeAt(0); } else { return false; } } exports.isShortPattern = isShortPattern; /** * Faster than using a RegExp for default newline detection during lexing. */ exports.LineTerminatorOptimizedTester = { // implements /\n|\r\n?/g.test test: function (text) { var len = text.length; for (var i = this.lastIndex; i < len; i++) { var c = text.charCodeAt(i); if (c === 10) { this.lastIndex = i + 1; return true; } else if (c === 13) { if (text.charCodeAt(i + 1) === 10) { this.lastIndex = i + 2; } else { this.lastIndex = i + 1; } return true; } } return false; }, lastIndex: 0 }; //# sourceMappingURL=lexer.js.map