Support Unicode international characters

lib/index.js

@@ -2,7 +2,10 @@
 
 // Load modules
 
-const Dns = require('dns');
+/* eslint-disable */
+let Dns = require('dns');
+/* eslint-enable */
+const Punycode = require('punycode');
 
 
 // Declare internals
@@ -132,7 +135,19 @@ internals.specials = function () {
     }
 
     for (let i = 0; i < specials.length; ++i) {
-        lookup[specials.charCodeAt(i)] = true;
+        lookup[specials.codePointAt(i)] = true;
+    }
+
+    // add C0 control characters
+
+    for (let i = 0; i < 33; ++i) {
+        lookup[i] = true;
+    }
+
+    // add C1 control characters
+
+    for (let i = 127; i < 160; ++i) {
+        lookup[i] = true;
     }
 
     return function (code) {
@@ -173,7 +188,7 @@ internals.validDomain = function (tldAtom, options) {
 
 
 /**
- * Check that an email address conforms to RFCs 5321, 5322 and others
+ * Check that an email address conforms to RFCs 5321, 5322, 6530 and others
  *
  * We distinguish clearly between a Mailbox as defined by RFC 5321 and an
  * addr-spec as defined by RFC 5322. Depending on the context, either can be
@@ -462,10 +477,10 @@ exports.validate = internals.validate = function (email, options, callback) {
                         }
                         else {
                             context.prev = context.now;
-                            charCode = token.charCodeAt(0);
+                            charCode = token.codePointAt(0);
 
                             // Especially if charCode == 10
-                            if (charCode < 33 || charCode > 126 || internals.specials(charCode)) {
+                            if (internals.specials(charCode)) {
 
                                 // Fatal error
                                 updateResult(internals.diagnoses.errExpectingATEXT);
@@ -660,11 +675,11 @@ exports.validate = internals.validate = function (email, options, callback) {
                             }
                         }
 
-                        charCode = token.charCodeAt(0);
+                        charCode = token.codePointAt(0);
                         // Assume this token isn't a hyphen unless we discover it is
                         hyphenFlag = false;
 
-                        if (charCode < 33 || charCode > 126 || internals.specials(charCode)) {
+                        if (internals.specials(charCode)) {
                             // Fatal error
                             updateResult(internals.diagnoses.errExpectingATEXT);
                         }
@@ -676,8 +691,8 @@ exports.validate = internals.validate = function (email, options, callback) {
 
                             hyphenFlag = true;
                         }
-                            // Check if it's a neither a number nor a latin letter
-                        else if (charCode < 48 || charCode > 122 || (charCode > 57 && charCode < 65) || (charCode > 90 && charCode < 97)) {
+                            // Check if it's a neither a number nor a latin/unicode letter
+                        else if (charCode < 48 || (charCode > 122 && charCode < 192) || (charCode > 57 && charCode < 65) || (charCode > 90 && charCode < 97)) {
                             // This is not an RFC 5321 subdomain, but still OK by RFC 5322
                             updateResult(internals.diagnoses.rfc5322Domain);
                         }
@@ -864,7 +879,7 @@ exports.validate = internals.validate = function (email, options, callback) {
                         //                     %d12 /     ;  include the carriage
                         //                     %d14-31 /  ;  return, line feed, and
                         //                     %d127      ;  white space characters
-                        charCode = token.charCodeAt(0);
+                        charCode = token.codePointAt(0);
 
                         // '\r', '\n', ' ', and '\t' have already been parsed above
                         if (charCode > 127 || charCode === 0 || token === '[') {
@@ -954,7 +969,7 @@ exports.validate = internals.validate = function (email, options, callback) {
                         //                      %d12 /             ;  include the carriage
                         //                      %d14-31 /          ;  return, line feed, and
                         //                      %d127              ;  white space characters
-                        charCode = token.charCodeAt(0);
+                        charCode = token.codePointAt(0);
 
                         if (charCode > 127 || charCode === 0 || charCode === 10) {
                             updateResult(internals.diagnoses.errExpectingQTEXT);
@@ -992,7 +1007,7 @@ exports.validate = internals.validate = function (email, options, callback) {
                 //                       %d127     ;  white space characters
                 //
                 // i.e. obs-qp       =  "\" (%d0-8, %d10-31 / %d127)
-                charCode = token.charCodeAt(0);
+                charCode = token.codePointAt(0);
 
                 if (charCode > 127) {
                     // Fatal error
@@ -1099,7 +1114,7 @@ exports.validate = internals.validate = function (email, options, callback) {
                         //                   %d12 /     ;  include the carriage
                         //                   %d14-31 /  ;  return, line feed, and
                         //                   %d127      ;  white space characters
-                        charCode = token.charCodeAt(0);
+                        charCode = token.codePointAt(0);
 
                         if (charCode > 127 || charCode === 0 || charCode === 10) {
                             // Fatal error
@@ -1266,7 +1281,7 @@ exports.validate = internals.validate = function (email, options, callback) {
 
         if (!dnsPositive && maxResult < internals.categories.dnsWarn) {
             // Per RFC 5321, domain atoms are limited to letter-digit-hyphen, so we only need to check code <= 57 to check for a digit
-            const code = atomData.domains[elementCount].charCodeAt(0);
+            const code = atomData.domains[elementCount].codePointAt(0);
             if (code <= 57) {
                 updateResult(internals.diagnoses.rfc5321TLDNumeric);
             }
@@ -1311,7 +1326,7 @@ exports.validate = internals.validate = function (email, options, callback) {
             parseData.domain += '.';
         }
 
-        const dnsDomain = parseData.domain;
+        const dnsDomain = Punycode.toASCII(parseData.domain);
         Dns.resolveMx(dnsDomain, (err, mxRecords) => {
 
             // If we have a fatal error, then we must assume that there are no records

package.json

@@ -18,10 +18,12 @@
     "node": ">=4.0.0"
   },
   "dependencies": {
+    "punycode": "^2.1.0"
   },
   "devDependencies": {
     "code": "3.x.x",
-    "lab": "10.x.x"
+    "lab": "10.x.x",
+    "rewire": "^2.5.2"
   },
   "scripts": {
     "test": "lab -a code -t 100 -L -m 5000",

test/index.js

@@ -48,6 +48,10 @@ const tldExpectations = [
     ['shouldbe@example.com', diag.valid]
 ];
 
+const noDNSExpectations = [
+    ['伊昭傑@郵件.商務', diag.valid],
+    ['ñoñó1234@ñomething.com', diag.valid]
+];
 
 describe('validate()', () => {
 
@@ -207,6 +211,23 @@ describe('validate()', () => {
         });
     });
 
+    noDNSExpectations.forEach((obj, i) => {
+
+        const email = obj[0];
+        const result = obj[1];
+        it('should handle noDNS test ' + (i + 1), (done) => {
+
+            Isemail.validate(email, {
+                errorLevel: 0,
+                checkDNS: false
+            }, (res) => {
+
+                expect(res).to.equal(result);
+                done();
+            });
+        });
+    });
+
     it('should handle domain atom test 1', (done) => {
 
         expect(Isemail.validate('shouldbe@invalid', {

test/tests.json

@@ -11,6 +11,9 @@
     ["test@nominet.org.uk", "valid"],
     ["test@about.museum", "valid"],
     ["a@iana.org", "valid"],
+    ["êjness@iana.org", "valid"],
+    ["ñoñó1234@iana.org", "valid"],
+    ["ñoñó1234@something.com", "valid"],
     ["test@e.com", "dnsWarnNoRecord"],
     ["test@iana.a", "dnsWarnNoRecord"],
     ["test.test@iana.org", "valid"],

test/tmp.js

@@ -0,0 +1,47 @@
+'use strict';
+
+// Load modules
+
+const Lab = require('lab');
+const Code = require('code');
+const Rewire = require('rewire');
+const Isemail = Rewire('..');
+
+// Test shortcuts
+
+const lab = exports.lab = Lab.script();
+const describe = lab.describe;
+const it = lab.it;
+const expect = Code.expect;
+
+// declare internals
+
+const internals = {};
+
+// Dns rewire stub
+
+const dns_stub = {
+    resolveMx: (domainName, cb) => {
+
+        internals.punyDomain = domainName;
+
+        return cb(null, [domainName]);
+    }
+};
+
+Isemail.__set__('Dns', dns_stub);
+
+describe('validate() international domains', () => {
+
+    it('should punycode domains', (done) => {
+
+        Isemail.validate('伊昭傑@郵件.商務', {
+            errorLevel: 0,
+            checkDNS: true
+        }, () => {
+
+            expect(internals.punyDomain).to.equal('xn--5nqv22n.xn--lhr59c');
+            done();
+        });
+    });
+});