1 | /**
|
2 | * @fileoverview Abstraction of JavaScript source code.
|
3 | * @author Nicholas C. Zakas
|
4 | */
|
5 | ;
|
6 |
|
7 | //------------------------------------------------------------------------------
|
8 | // Requirements
|
9 | //------------------------------------------------------------------------------
|
10 |
|
11 | const
|
12 | { isCommentToken } = require("eslint-utils"),
|
13 | TokenStore = require("./token-store"),
|
14 | astUtils = require("../shared/ast-utils"),
|
15 | Traverser = require("../shared/traverser"),
|
16 | lodash = require("lodash");
|
17 |
|
18 | //------------------------------------------------------------------------------
|
19 | // Private
|
20 | //------------------------------------------------------------------------------
|
21 |
|
22 | /**
|
23 | * Validates that the given AST has the required information.
|
24 | * @param {ASTNode} ast The Program node of the AST to check.
|
25 | * @throws {Error} If the AST doesn't contain the correct information.
|
26 | * @returns {void}
|
27 | * @private
|
28 | */
|
29 | function validate(ast) {
|
30 | if (!ast.tokens) {
|
31 | throw new Error("AST is missing the tokens array.");
|
32 | }
|
33 |
|
34 | if (!ast.comments) {
|
35 | throw new Error("AST is missing the comments array.");
|
36 | }
|
37 |
|
38 | if (!ast.loc) {
|
39 | throw new Error("AST is missing location information.");
|
40 | }
|
41 |
|
42 | if (!ast.range) {
|
43 | throw new Error("AST is missing range information");
|
44 | }
|
45 | }
|
46 |
|
47 | /**
|
48 | * Check to see if its a ES6 export declaration.
|
49 | * @param {ASTNode} astNode An AST node.
|
50 | * @returns {boolean} whether the given node represents an export declaration.
|
51 | * @private
|
52 | */
|
53 | function looksLikeExport(astNode) {
|
54 | return astNode.type === "ExportDefaultDeclaration" || astNode.type === "ExportNamedDeclaration" ||
|
55 | astNode.type === "ExportAllDeclaration" || astNode.type === "ExportSpecifier";
|
56 | }
|
57 |
|
58 | /**
|
59 | * Merges two sorted lists into a larger sorted list in O(n) time.
|
60 | * @param {Token[]} tokens The list of tokens.
|
61 | * @param {Token[]} comments The list of comments.
|
62 | * @returns {Token[]} A sorted list of tokens and comments.
|
63 | * @private
|
64 | */
|
65 | function sortedMerge(tokens, comments) {
|
66 | const result = [];
|
67 | let tokenIndex = 0;
|
68 | let commentIndex = 0;
|
69 |
|
70 | while (tokenIndex < tokens.length || commentIndex < comments.length) {
|
71 | if (commentIndex >= comments.length || tokenIndex < tokens.length && tokens[tokenIndex].range[0] < comments[commentIndex].range[0]) {
|
72 | result.push(tokens[tokenIndex++]);
|
73 | } else {
|
74 | result.push(comments[commentIndex++]);
|
75 | }
|
76 | }
|
77 |
|
78 | return result;
|
79 | }
|
80 |
|
81 | /**
|
82 | * Determines if two nodes or tokens overlap.
|
83 | * @param {ASTNode|Token} first The first node or token to check.
|
84 | * @param {ASTNode|Token} second The second node or token to check.
|
85 | * @returns {boolean} True if the two nodes or tokens overlap.
|
86 | * @private
|
87 | */
|
88 | function nodesOrTokensOverlap(first, second) {
|
89 | return (first.range[0] <= second.range[0] && first.range[1] >= second.range[0]) ||
|
90 | (second.range[0] <= first.range[0] && second.range[1] >= first.range[0]);
|
91 | }
|
92 |
|
93 | //------------------------------------------------------------------------------
|
94 | // Public Interface
|
95 | //------------------------------------------------------------------------------
|
96 |
|
97 | class SourceCode extends TokenStore {
|
98 |
|
99 | /**
|
100 | * Represents parsed source code.
|
101 | * @param {string|Object} textOrConfig The source code text or config object.
|
102 | * @param {string} textOrConfig.text The source code text.
|
103 | * @param {ASTNode} textOrConfig.ast The Program node of the AST representing the code. This AST should be created from the text that BOM was stripped.
|
104 | * @param {Object|null} textOrConfig.parserServices The parser services.
|
105 | * @param {ScopeManager|null} textOrConfig.scopeManager The scope of this source code.
|
106 | * @param {Object|null} textOrConfig.visitorKeys The visitor keys to traverse AST.
|
107 | * @param {ASTNode} [astIfNoConfig] The Program node of the AST representing the code. This AST should be created from the text that BOM was stripped.
|
108 | */
|
109 | constructor(textOrConfig, astIfNoConfig) {
|
110 | let text, ast, parserServices, scopeManager, visitorKeys;
|
111 |
|
112 | // Process overloading.
|
113 | if (typeof textOrConfig === "string") {
|
114 | text = textOrConfig;
|
115 | ast = astIfNoConfig;
|
116 | } else if (typeof textOrConfig === "object" && textOrConfig !== null) {
|
117 | text = textOrConfig.text;
|
118 | ast = textOrConfig.ast;
|
119 | parserServices = textOrConfig.parserServices;
|
120 | scopeManager = textOrConfig.scopeManager;
|
121 | visitorKeys = textOrConfig.visitorKeys;
|
122 | }
|
123 |
|
124 | validate(ast);
|
125 | super(ast.tokens, ast.comments);
|
126 |
|
127 | /**
|
128 | * The flag to indicate that the source code has Unicode BOM.
|
129 | * @type boolean
|
130 | */
|
131 | this.hasBOM = (text.charCodeAt(0) === 0xFEFF);
|
132 |
|
133 | /**
|
134 | * The original text source code.
|
135 | * BOM was stripped from this text.
|
136 | * @type string
|
137 | */
|
138 | this.text = (this.hasBOM ? text.slice(1) : text);
|
139 |
|
140 | /**
|
141 | * The parsed AST for the source code.
|
142 | * @type ASTNode
|
143 | */
|
144 | this.ast = ast;
|
145 |
|
146 | /**
|
147 | * The parser services of this source code.
|
148 | * @type {Object}
|
149 | */
|
150 | this.parserServices = parserServices || {};
|
151 |
|
152 | /**
|
153 | * The scope of this source code.
|
154 | * @type {ScopeManager|null}
|
155 | */
|
156 | this.scopeManager = scopeManager || null;
|
157 |
|
158 | /**
|
159 | * The visitor keys to traverse AST.
|
160 | * @type {Object}
|
161 | */
|
162 | this.visitorKeys = visitorKeys || Traverser.DEFAULT_VISITOR_KEYS;
|
163 |
|
164 | // Check the source text for the presence of a shebang since it is parsed as a standard line comment.
|
165 | const shebangMatched = this.text.match(astUtils.shebangPattern);
|
166 | const hasShebang = shebangMatched && ast.comments.length && ast.comments[0].value === shebangMatched[1];
|
167 |
|
168 | if (hasShebang) {
|
169 | ast.comments[0].type = "Shebang";
|
170 | }
|
171 |
|
172 | this.tokensAndComments = sortedMerge(ast.tokens, ast.comments);
|
173 |
|
174 | /**
|
175 | * The source code split into lines according to ECMA-262 specification.
|
176 | * This is done to avoid each rule needing to do so separately.
|
177 | * @type string[]
|
178 | */
|
179 | this.lines = [];
|
180 | this.lineStartIndices = [0];
|
181 |
|
182 | const lineEndingPattern = astUtils.createGlobalLinebreakMatcher();
|
183 | let match;
|
184 |
|
185 | /*
|
186 | * Previously, this was implemented using a regex that
|
187 | * matched a sequence of non-linebreak characters followed by a
|
188 | * linebreak, then adding the lengths of the matches. However,
|
189 | * this caused a catastrophic backtracking issue when the end
|
190 | * of a file contained a large number of non-newline characters.
|
191 | * To avoid this, the current implementation just matches newlines
|
192 | * and uses match.index to get the correct line start indices.
|
193 | */
|
194 | while ((match = lineEndingPattern.exec(this.text))) {
|
195 | this.lines.push(this.text.slice(this.lineStartIndices[this.lineStartIndices.length - 1], match.index));
|
196 | this.lineStartIndices.push(match.index + match[0].length);
|
197 | }
|
198 | this.lines.push(this.text.slice(this.lineStartIndices[this.lineStartIndices.length - 1]));
|
199 |
|
200 | // Cache for comments found using getComments().
|
201 | this._commentCache = new WeakMap();
|
202 |
|
203 | // don't allow modification of this object
|
204 | Object.freeze(this);
|
205 | Object.freeze(this.lines);
|
206 | }
|
207 |
|
208 | /**
|
209 | * Split the source code into multiple lines based on the line delimiters.
|
210 | * @param {string} text Source code as a string.
|
211 | * @returns {string[]} Array of source code lines.
|
212 | * @public
|
213 | */
|
214 | static splitLines(text) {
|
215 | return text.split(astUtils.createGlobalLinebreakMatcher());
|
216 | }
|
217 |
|
218 | /**
|
219 | * Gets the source code for the given node.
|
220 | * @param {ASTNode} [node] The AST node to get the text for.
|
221 | * @param {int} [beforeCount] The number of characters before the node to retrieve.
|
222 | * @param {int} [afterCount] The number of characters after the node to retrieve.
|
223 | * @returns {string} The text representing the AST node.
|
224 | * @public
|
225 | */
|
226 | getText(node, beforeCount, afterCount) {
|
227 | if (node) {
|
228 | return this.text.slice(Math.max(node.range[0] - (beforeCount || 0), 0),
|
229 | node.range[1] + (afterCount || 0));
|
230 | }
|
231 | return this.text;
|
232 | }
|
233 |
|
234 | /**
|
235 | * Gets the entire source text split into an array of lines.
|
236 | * @returns {Array} The source text as an array of lines.
|
237 | * @public
|
238 | */
|
239 | getLines() {
|
240 | return this.lines;
|
241 | }
|
242 |
|
243 | /**
|
244 | * Retrieves an array containing all comments in the source code.
|
245 | * @returns {ASTNode[]} An array of comment nodes.
|
246 | * @public
|
247 | */
|
248 | getAllComments() {
|
249 | return this.ast.comments;
|
250 | }
|
251 |
|
252 | /**
|
253 | * Gets all comments for the given node.
|
254 | * @param {ASTNode} node The AST node to get the comments for.
|
255 | * @returns {Object} An object containing a leading and trailing array
|
256 | * of comments indexed by their position.
|
257 | * @public
|
258 | */
|
259 | getComments(node) {
|
260 | if (this._commentCache.has(node)) {
|
261 | return this._commentCache.get(node);
|
262 | }
|
263 |
|
264 | const comments = {
|
265 | leading: [],
|
266 | trailing: []
|
267 | };
|
268 |
|
269 | /*
|
270 | * Return all comments as leading comments of the Program node when
|
271 | * there is no executable code.
|
272 | */
|
273 | if (node.type === "Program") {
|
274 | if (node.body.length === 0) {
|
275 | comments.leading = node.comments;
|
276 | }
|
277 | } else {
|
278 |
|
279 | /*
|
280 | * Return comments as trailing comments of nodes that only contain
|
281 | * comments (to mimic the comment attachment behavior present in Espree).
|
282 | */
|
283 | if ((node.type === "BlockStatement" || node.type === "ClassBody") && node.body.length === 0 ||
|
284 | node.type === "ObjectExpression" && node.properties.length === 0 ||
|
285 | node.type === "ArrayExpression" && node.elements.length === 0 ||
|
286 | node.type === "SwitchStatement" && node.cases.length === 0
|
287 | ) {
|
288 | comments.trailing = this.getTokens(node, {
|
289 | includeComments: true,
|
290 | filter: isCommentToken
|
291 | });
|
292 | }
|
293 |
|
294 | /*
|
295 | * Iterate over tokens before and after node and collect comment tokens.
|
296 | * Do not include comments that exist outside of the parent node
|
297 | * to avoid duplication.
|
298 | */
|
299 | let currentToken = this.getTokenBefore(node, { includeComments: true });
|
300 |
|
301 | while (currentToken && isCommentToken(currentToken)) {
|
302 | if (node.parent && (currentToken.start < node.parent.start)) {
|
303 | break;
|
304 | }
|
305 | comments.leading.push(currentToken);
|
306 | currentToken = this.getTokenBefore(currentToken, { includeComments: true });
|
307 | }
|
308 |
|
309 | comments.leading.reverse();
|
310 |
|
311 | currentToken = this.getTokenAfter(node, { includeComments: true });
|
312 |
|
313 | while (currentToken && isCommentToken(currentToken)) {
|
314 | if (node.parent && (currentToken.end > node.parent.end)) {
|
315 | break;
|
316 | }
|
317 | comments.trailing.push(currentToken);
|
318 | currentToken = this.getTokenAfter(currentToken, { includeComments: true });
|
319 | }
|
320 | }
|
321 |
|
322 | this._commentCache.set(node, comments);
|
323 | return comments;
|
324 | }
|
325 |
|
326 | /**
|
327 | * Retrieves the JSDoc comment for a given node.
|
328 | * @param {ASTNode} node The AST node to get the comment for.
|
329 | * @returns {Token|null} The Block comment token containing the JSDoc comment
|
330 | * for the given node or null if not found.
|
331 | * @public
|
332 | * @deprecated
|
333 | */
|
334 | getJSDocComment(node) {
|
335 |
|
336 | /**
|
337 | * Checks for the presence of a JSDoc comment for the given node and returns it.
|
338 | * @param {ASTNode} astNode The AST node to get the comment for.
|
339 | * @returns {Token|null} The Block comment token containing the JSDoc comment
|
340 | * for the given node or null if not found.
|
341 | * @private
|
342 | */
|
343 | const findJSDocComment = astNode => {
|
344 | const tokenBefore = this.getTokenBefore(astNode, { includeComments: true });
|
345 |
|
346 | if (
|
347 | tokenBefore &&
|
348 | isCommentToken(tokenBefore) &&
|
349 | tokenBefore.type === "Block" &&
|
350 | tokenBefore.value.charAt(0) === "*" &&
|
351 | astNode.loc.start.line - tokenBefore.loc.end.line <= 1
|
352 | ) {
|
353 | return tokenBefore;
|
354 | }
|
355 |
|
356 | return null;
|
357 | };
|
358 | let parent = node.parent;
|
359 |
|
360 | switch (node.type) {
|
361 | case "ClassDeclaration":
|
362 | case "FunctionDeclaration":
|
363 | return findJSDocComment(looksLikeExport(parent) ? parent : node);
|
364 |
|
365 | case "ClassExpression":
|
366 | return findJSDocComment(parent.parent);
|
367 |
|
368 | case "ArrowFunctionExpression":
|
369 | case "FunctionExpression":
|
370 | if (parent.type !== "CallExpression" && parent.type !== "NewExpression") {
|
371 | while (
|
372 | !this.getCommentsBefore(parent).length &&
|
373 | !/Function/u.test(parent.type) &&
|
374 | parent.type !== "MethodDefinition" &&
|
375 | parent.type !== "Property"
|
376 | ) {
|
377 | parent = parent.parent;
|
378 |
|
379 | if (!parent) {
|
380 | break;
|
381 | }
|
382 | }
|
383 |
|
384 | if (parent && parent.type !== "FunctionDeclaration" && parent.type !== "Program") {
|
385 | return findJSDocComment(parent);
|
386 | }
|
387 | }
|
388 |
|
389 | return findJSDocComment(node);
|
390 |
|
391 | // falls through
|
392 | default:
|
393 | return null;
|
394 | }
|
395 | }
|
396 |
|
397 | /**
|
398 | * Gets the deepest node containing a range index.
|
399 | * @param {int} index Range index of the desired node.
|
400 | * @returns {ASTNode} The node if found or null if not found.
|
401 | * @public
|
402 | */
|
403 | getNodeByRangeIndex(index) {
|
404 | let result = null;
|
405 |
|
406 | Traverser.traverse(this.ast, {
|
407 | visitorKeys: this.visitorKeys,
|
408 | enter(node) {
|
409 | if (node.range[0] <= index && index < node.range[1]) {
|
410 | result = node;
|
411 | } else {
|
412 | this.skip();
|
413 | }
|
414 | },
|
415 | leave(node) {
|
416 | if (node === result) {
|
417 | this.break();
|
418 | }
|
419 | }
|
420 | });
|
421 |
|
422 | return result;
|
423 | }
|
424 |
|
425 | /**
|
426 | * Determines if two nodes or tokens have at least one whitespace character
|
427 | * between them. Order does not matter. Returns false if the given nodes or
|
428 | * tokens overlap.
|
429 | * @param {ASTNode|Token} first The first node or token to check between.
|
430 | * @param {ASTNode|Token} second The second node or token to check between.
|
431 | * @returns {boolean} True if there is a whitespace character between
|
432 | * any of the tokens found between the two given nodes or tokens.
|
433 | * @public
|
434 | */
|
435 | isSpaceBetween(first, second) {
|
436 | if (nodesOrTokensOverlap(first, second)) {
|
437 | return false;
|
438 | }
|
439 |
|
440 | const [startingNodeOrToken, endingNodeOrToken] = first.range[1] <= second.range[0]
|
441 | ? [first, second]
|
442 | : [second, first];
|
443 | const firstToken = this.getLastToken(startingNodeOrToken) || startingNodeOrToken;
|
444 | const finalToken = this.getFirstToken(endingNodeOrToken) || endingNodeOrToken;
|
445 | let currentToken = firstToken;
|
446 |
|
447 | while (currentToken !== finalToken) {
|
448 | const nextToken = this.getTokenAfter(currentToken, { includeComments: true });
|
449 |
|
450 | if (currentToken.range[1] !== nextToken.range[0]) {
|
451 | return true;
|
452 | }
|
453 |
|
454 | currentToken = nextToken;
|
455 | }
|
456 |
|
457 | return false;
|
458 | }
|
459 |
|
460 | /**
|
461 | * Determines if two nodes or tokens have at least one whitespace character
|
462 | * between them. Order does not matter. Returns false if the given nodes or
|
463 | * tokens overlap.
|
464 | * @param {...ASTNode|Token} args The nodes or tokens to check between.
|
465 | * @returns {boolean} True if there is a whitespace character between
|
466 | * any of the tokens found between the two given nodes or tokens.
|
467 | * @deprecated in favor of isSpaceBetween().
|
468 | * @public
|
469 | */
|
470 | isSpaceBetweenTokens(...args) {
|
471 | return this.isSpaceBetween(...args);
|
472 | }
|
473 |
|
474 | /**
|
475 | * Converts a source text index into a (line, column) pair.
|
476 | * @param {number} index The index of a character in a file
|
477 | * @returns {Object} A {line, column} location object with a 0-indexed column
|
478 | * @public
|
479 | */
|
480 | getLocFromIndex(index) {
|
481 | if (typeof index !== "number") {
|
482 | throw new TypeError("Expected `index` to be a number.");
|
483 | }
|
484 |
|
485 | if (index < 0 || index > this.text.length) {
|
486 | throw new RangeError(`Index out of range (requested index ${index}, but source text has length ${this.text.length}).`);
|
487 | }
|
488 |
|
489 | /*
|
490 | * For an argument of this.text.length, return the location one "spot" past the last character
|
491 | * of the file. If the last character is a linebreak, the location will be column 0 of the next
|
492 | * line; otherwise, the location will be in the next column on the same line.
|
493 | *
|
494 | * See getIndexFromLoc for the motivation for this special case.
|
495 | */
|
496 | if (index === this.text.length) {
|
497 | return { line: this.lines.length, column: this.lines[this.lines.length - 1].length };
|
498 | }
|
499 |
|
500 | /*
|
501 | * To figure out which line rangeIndex is on, determine the last index at which rangeIndex could
|
502 | * be inserted into lineIndices to keep the list sorted.
|
503 | */
|
504 | const lineNumber = lodash.sortedLastIndex(this.lineStartIndices, index);
|
505 |
|
506 | return { line: lineNumber, column: index - this.lineStartIndices[lineNumber - 1] };
|
507 | }
|
508 |
|
509 | /**
|
510 | * Converts a (line, column) pair into a range index.
|
511 | * @param {Object} loc A line/column location
|
512 | * @param {number} loc.line The line number of the location (1-indexed)
|
513 | * @param {number} loc.column The column number of the location (0-indexed)
|
514 | * @returns {number} The range index of the location in the file.
|
515 | * @public
|
516 | */
|
517 | getIndexFromLoc(loc) {
|
518 | if (typeof loc !== "object" || typeof loc.line !== "number" || typeof loc.column !== "number") {
|
519 | throw new TypeError("Expected `loc` to be an object with numeric `line` and `column` properties.");
|
520 | }
|
521 |
|
522 | if (loc.line <= 0) {
|
523 | throw new RangeError(`Line number out of range (line ${loc.line} requested). Line numbers should be 1-based.`);
|
524 | }
|
525 |
|
526 | if (loc.line > this.lineStartIndices.length) {
|
527 | throw new RangeError(`Line number out of range (line ${loc.line} requested, but only ${this.lineStartIndices.length} lines present).`);
|
528 | }
|
529 |
|
530 | const lineStartIndex = this.lineStartIndices[loc.line - 1];
|
531 | const lineEndIndex = loc.line === this.lineStartIndices.length ? this.text.length : this.lineStartIndices[loc.line];
|
532 | const positionIndex = lineStartIndex + loc.column;
|
533 |
|
534 | /*
|
535 | * By design, getIndexFromLoc({ line: lineNum, column: 0 }) should return the start index of
|
536 | * the given line, provided that the line number is valid element of this.lines. Since the
|
537 | * last element of this.lines is an empty string for files with trailing newlines, add a
|
538 | * special case where getting the index for the first location after the end of the file
|
539 | * will return the length of the file, rather than throwing an error. This allows rules to
|
540 | * use getIndexFromLoc consistently without worrying about edge cases at the end of a file.
|
541 | */
|
542 | if (
|
543 | loc.line === this.lineStartIndices.length && positionIndex > lineEndIndex ||
|
544 | loc.line < this.lineStartIndices.length && positionIndex >= lineEndIndex
|
545 | ) {
|
546 | throw new RangeError(`Column number out of range (column ${loc.column} requested, but the length of line ${loc.line} is ${lineEndIndex - lineStartIndex}).`);
|
547 | }
|
548 |
|
549 | return positionIndex;
|
550 | }
|
551 | }
|
552 |
|
553 | module.exports = SourceCode;
|