Merge pull request #252 from dmichon-msft/on-inserted-semicolon

feat(parser): add onInsertedSemicolon option

README.md

@@ -85,6 +85,9 @@ This is the available options:
   // Allows comment extraction. Accepts either a function or array
   onComment: []
 
+  // Allows detection of automatic semicolon insertion. Accepts a callback function that will be passed the charater offset where the semicolon was inserted
+  onInsertedSemicolon: (pos) => {}
+
   // Allows token extraction. Accepts either a function or array
   onToken: []
 
@@ -115,13 +118,20 @@ If an array is supplied, comments/tokens will be pushed to the array, the item i
 If a function callback is supplied, the signature must be
 
 ```ts
-function onComment(type: string, value: string, start: number, end: number, loc: SourceLocation): void {}
+declare function onComment(type: string, value: string, start: number, end: number, loc: SourceLocation): void;
 
-function onToken(token: string, start: number, end: number, loc: SourceLocation): void {}
+declare function onToken(token: string, start: number, end: number, loc: SourceLocation): void;
 ```
 
 Note the `start/end/loc` information are provided to the function callback regardless of the settings on ranges and loc flags. onComment callback has one extra argument `value: string` for the body string of the comment.
 
+### onInsertedSemicolon
+If a function callback is supplied, the signature must be
+
+```ts
+declare function onInsertedSemicolon(position: number): void;
+```
+
 ## Example usage
 
 ```js
@@ -184,4 +194,4 @@ This will return when serialized in json:
         }
     ]
 }
-```
+```

src/common.ts

@@ -173,6 +173,11 @@ export const enum ScopeKind {
  */
 export type OnComment = void | Comment[] | ((type: string, value: string, start: number, end: number, loc: SourceLocation) => any);
 
+/**
+ * The type of the `onInsertedSemicolon` option.
+ */
+export type OnInsertedSemicolon = void | ((pos: number) => any);
+
 /**
  * The type of the `onToken` option.
  */
@@ -214,6 +219,7 @@ export interface ParserState {
   end: number;
   token: Token;
   onComment: any;
+  onInsertedSemicolon: any;
   onToken: any;
   tokenValue: any;
   tokenRaw: string;
@@ -247,7 +253,11 @@ export function matchOrInsertSemicolon(parser: ParserState, context: Context, sp
   ) {
     report(parser, Errors.UnexpectedToken, KeywordDescTable[parser.token & Token.Type]);
   }
-  consumeOpt(parser, context, Token.Semicolon);
+
+  if (!consumeOpt(parser, context, Token.Semicolon)) {
+    // Automatic semicolon insertion has occurred
+    parser.onInsertedSemicolon?.(parser.startPos);
+  }
 }
 
 export function isValidStrictMode(parser: ParserState, index: number, tokenPos: number, tokenValue: string): 0 | 1 {

src/parser.ts

@@ -13,6 +13,7 @@ import {
   consume,
   Flags,
   OnComment,
+  OnInsertedSemicolon,
   OnToken,
   pushComment,
   pushToken,
@@ -55,7 +56,8 @@ export function create(
   source: string,
   sourceFile: string | void,
   onComment: OnComment | void,
-  onToken: OnToken | void
+  onToken: OnToken | void,
+  onInsertedSemicolon: OnInsertedSemicolon | void
 ): ParserState {
   return {
     /**
@@ -180,6 +182,11 @@ export function create(
      */
     onToken,
 
+    /**
+     * Function invoked with the character offset when automatic semicolon insertion occurs
+     */
+    onInsertedSemicolon,
+
     /**
      * Holds leading decorators before "export" or "class" keywords
      */
@@ -221,9 +228,11 @@ export interface Options {
   jsx?: boolean;
   // Allow edge cases that deviate from the spec
   specDeviation?: boolean;
-  // Allows comment extraction. Accepts either a a callback function or an array
+  // Allows comment extraction. Accepts either a callback function or an array
   onComment?: OnComment;
-  // Allows token extraction. Accepts either a a callback function or an array
+  // Allows detection of automatic semicolon insertion. Accepts a callback function that will be passed the charater offset where the semicolon was inserted
+  onInsertedSemicolon?: OnInsertedSemicolon;
+  // Allows token extraction. Accepts either a callback function or an array
   onToken?: OnToken;
   // Creates unique key for in ObjectPattern when key value are same
   uniqueKeyInPattern?: boolean;
@@ -235,6 +244,7 @@ export interface Options {
 export function parseSource(source: string, options: Options | void, context: Context): ESTree.Program {
   let sourceFile = '';
   let onComment;
+  let onInsertedSemicolon;
   let onToken;
   if (options != null) {
     if (options.module) context |= Context.Module | Context.Strict;
@@ -257,14 +267,15 @@ export function parseSource(source: string, options: Options | void, context: Co
     if (options.onComment != null) {
       onComment = Array.isArray(options.onComment) ? pushComment(context, options.onComment) : options.onComment;
     }
+    if (options.onInsertedSemicolon != null) onInsertedSemicolon = options.onInsertedSemicolon;
     // Accepts either a callback function to be invoked or an array to collect tokens
     if (options.onToken != null) {
       onToken = Array.isArray(options.onToken) ? pushToken(context, options.onToken) : options.onToken;
     }
   }
 
   // Initialize parser state
-  const parser = create(source, sourceFile, onComment, onToken);
+  const parser = create(source, sourceFile, onComment, onToken, onInsertedSemicolon);
 
   // See: https://github.com/tc39/proposal-hashbang
   if (context & Context.OptionsNext) skipHashBang(parser);

test/parser/miscellaneous/onInsertedSemicolon.ts

@@ -0,0 +1,14 @@
+import * as t from 'assert';
+import { parseScript } from '../../../src/meriyah';
+
+describe('Miscellaneous - oninsertedSemicolon', () => {
+  it('invokes the callback when automatic semicolon insertion occurs', () => {
+    const semicolons: number[] = [];
+    const lines = ['"use strict"', 'self.a;', 'self.b'];
+    const input = lines.join('\n');
+    parseScript(input, {
+      onInsertedSemicolon: (pos) => semicolons.push(pos)
+    });
+    t.deepEqual(semicolons, [lines[0].length, input.length]);
+  });
+});