character sets support escaping of special chars

Author: ColinEberhardt

__tests__/character-sets.js

@@ -5,6 +5,17 @@ it("matches discrete characters", () => {
   expectNotMatch("[abce]", ["", "f", "h"]);
 });
 
+it("throws an error if no closing bracket is found", () => {
+  expect(() => new RegExp("[abce")).toThrow();
+});
+
+it("supports escaping of special characters", () => {
+  expectMatch("[a\\^b]", ["a", "b", "^"]);
+  expectMatch("[a\\-c]", ["a", "c", "-"]);
+  expectMatch("[a\\]]", ["a", "]"]);
+  expectMatch("[a\\\\b]", ["a", "\\"]);
+});
+
 it("matches character ranges", () => {
   expectMatch("[a-c]", ["a", "b", "c"]);
   expectNotMatch("[a-c]", ["d", "e", ""]);

__tests__/quantifiers.js

@@ -32,7 +32,6 @@ it("zero or more is greedy", () => {
 
 it("one or more is greedy", () => {
   let match = matches("a+", "aaaaa");
-  console.log(match);
   expect(match).not.toBeNull();
   expect(match.matches[0]).toEqual("aaaaa");
 });

assembly/parser/parser.ts

@@ -18,6 +18,16 @@ function isQuantifier(code: Char): bool {
   return code == Char.Question || code == Char.Plus || code == Char.Asterisk;
 }
 
+// characters which have special meaning within character sets
+function isCharacterSetSpecialChar(code: Char): bool {
+  return (
+    code == Char.Caret ||
+    code == Char.Minus ||
+    code == Char.RightSquareBracket ||
+    code == Char.Backslash
+  );
+}
+
 function isAssertion(code: u32): bool {
   return code == Char.Dollar || code == Char.Caret; // "$" or "^"
 }
@@ -228,16 +238,25 @@ export class Parser {
     while (this.currentToken != "]" || nodes.length == 0) {
       // lookahead for character range
       if (
-        this.cursor + 1 < u32(this.input.length) &&
+        this.cursor + 2 < u32(this.input.length) &&
+        this.currentToken != "\\" &&
         this.input.charCodeAt(this.cursor + 1) == Char.Minus &&
         this.input.charCodeAt(this.cursor + 2) != Char.RightSquareBracket
       ) {
         nodes.push(this.parseCharacterRange());
       } else {
-        nodes.push(this.parseCharacter());
+        if (
+          this.currentToken == "\\" &&
+          isCharacterSetSpecialChar(this.input.charCodeAt(this.cursor + 1))
+        ) {
+          this.eatToken(Char.Backslash);
+        }
+        nodes.push(new CharacterNode(this.eatToken()));
       }
 
-      // TODO error if we run out of chars?
+      if (this.cursor >= u32(this.input.length)) {
+        throw new SyntaxError("Unterminated character class");
+      }
     }
     this.eatToken(Char.RightSquareBracket);
     return new CharacterSetNode(nodes, negated);

package.json

@@ -6,6 +6,7 @@
   "scripts": {
     "test": "npm run asbuild:untouched && npm run prettier:check && jest __tests__",
     "test:suite": "npm run asbuild:untouched && jest __spec_tests__ --reporter=jest-summary-reporter",
+    "jest": "jest __tests__",
     "prettier:check": "prettier --check .",
     "prettier:write": "prettier --write .",
     "asbuild:untouched": "asc assembly/index.ts --target debug",

ts/index.ts

@@ -5,7 +5,7 @@ globalAny.log = console.log;
 
 import { RegExp } from "../assembly/regexp";
 
-const regexObj = new RegExp("[]a]");
-const match = regexObj.exec("]");
+const regexObj = new RegExp("[a\\\\c]");
+const match = regexObj.exec("\\");
 
 console.log(match);