Merge pull request #1291 from noinkling/normalize

Add string.normalize for unicode normalization
hapijs · Sep 20, 2017 · 2adcc09 · 2adcc09
2 parents 9a8eb5a + a79b59a
commit 2adcc09
Show file tree

Hide file tree

Showing 5 changed files with 409 additions and 1 deletion.
diff --git a/API.md b/API.md
@@ -125,6 +125,7 @@
     - [`string.hex()`](#stringhex)
     - [`string.base64([options])`](#stringbase64options)
     - [`string.hostname()`](#stringhostname)
+    - [`string.normalize([form])`](#stringnormalizeform)
     - [`string.lowercase()`](#stringlowercase)
     - [`string.uppercase()`](#stringuppercase)
     - [`string.trim()`](#stringtrim)
@@ -1910,6 +1911,21 @@ Requires the string value to be a valid hostname as per [RFC1123](http://tools.i
 const schema = Joi.string().hostname();
 ```
 
+#### `string.normalize([form])`
+
+Requires the string value to be in a [unicode normalized](https://developer.mozilla.org/en/docs/Web/JavaScript/Reference/Global_Objects/String/normalize)
+form. If the validation `convert` option is on (enabled by default), the string will be normalized.
+
+- `form` - The unicode normalization form to use. Valid values: `NFC` [default], `NFD`, `NFKC`, `NFKD`
+
+```js
+const schema = Joi.string().normalize(); // defaults to NFC
+const schema = Joi.string().normalize('NFC'); // canonical composition
+const schema = Joi.string().normalize('NFD'); // canonical decomposition
+const schema = Joi.string().normalize('NFKC'); // compatibility composition
+const schema = Joi.string().normalize('NFKD'); // compatibility decomposition
+```
+
 #### `string.lowercase()`
 
 Requires the string value to be all lowercase. If the validation `convert` option is on (enabled by default), the string

diff --git a/lib/language.js b/lib/language.js
@@ -139,6 +139,7 @@ exports.errors = {
         hex: 'must only contain hexadecimal characters',
         base64: 'must be a valid base64 string',
         hostname: 'must be a valid hostname',
+        normalize: 'must be unicode normalized in the {{form}} form',
         lowercase: 'must only contain lowercase characters',
         uppercase: 'must only contain uppercase characters',
         trim: 'must not have leading or trailing whitespace',

diff --git a/lib/types/any/index.js b/lib/types/any/index.js
@@ -54,6 +54,7 @@ module.exports = internals.Any = class {
              encoding: undefined,
              insensitive: false,
              trim: false,
+             normalize: undefined,                   // NFC, NFD, NFKC, NFKD
              case: undefined,                        // upper, lower
              empty: undefined,
              func: false,

diff --git a/lib/types/string/index.js b/lib/types/string/index.js
@@ -26,7 +26,8 @@ const internals = {
         uuidv4: '4',
         uuidv5: '5'
     },
-    cidrPresences: ['required', 'optional', 'forbidden']
+    cidrPresences: ['required', 'optional', 'forbidden'],
+    normalizationForms: ['NFC', 'NFD', 'NFKC', 'NFKD']
 };
 
 internals.String = class extends Any {
@@ -43,6 +44,10 @@ internals.String = class extends Any {
         if (typeof value === 'string' &&
             options.convert) {
 
+            if (this._flags.normalize) {
+                value = value.normalize(this._flags.normalize);
+            }
+
             if (this._flags.case) {
                 value = (this._flags.case === 'upper' ? value.toLocaleUpperCase() : value.toLocaleLowerCase());
             }
@@ -456,6 +461,26 @@ internals.String = class extends Any {
         });
     }
 
+    normalize(form) {
+
+        form = form || 'NFC';
+        Hoek.assert(Hoek.contain(internals.normalizationForms, form), 'normalization form must be one of ' + internals.normalizationForms.join(', '));
+
+        const obj = this._test('normalize', form, function (value, state, options) {
+
+            if (options.convert ||
+                value === value.normalize(form)) {
+
+                return value;
+            }
+
+            return this.createError('string.normalize', { value, form }, state, options);
+        });
+
+        obj._flags.normalize = form;
+        return obj;
+    }
+
     lowercase() {
 
         const obj = this._test('lowercase', undefined, function (value, state, options) {