bug #33807 [String] fix toAscii() (nicolas-grekas)

nicolas-grekas · nicolas-grekas · commit a176d1a92897 · 2019-10-02T14:13:38.000+02:00
This PR was merged into the 5.0-dev branch. Discussion ---------- [String] fix toAscii() | Q | A | ------------- | --- | Branch? | master | Bug fix? | yes | New feature? | no | Deprecations? | | Tickets | - | License | MIT | Doc PR | - Commits ------- 46a23f8 [String] fix toAscii()
diff --git a/src/Symfony/Component/String/AbstractUnicodeString.php b/src/Symfony/Component/String/AbstractUnicodeString.php
@@ -81,17 +81,20 @@ public function ascii(array $rules = []): self
         $s = $str->string;
         $str->string = '';
 
-        $rules[] = 'nfkd';
-        $rules[] = '[:nonspacing mark:] remove';
+        if (!$rules) {
+            $rules[] = '[:nonspacing mark:] remove';
+        }
+
+        array_unshift($rules, 'nfd');
 
         if (\function_exists('transliterator_transliterate')) {
             $rules[] = 'any-latin/bgn';
             $rules[] = 'nfkd';
             $rules[] = '[:nonspacing mark:] remove';
         }
 
-        while (\strlen($s) !== $i = strspn($s, self::ASCII)) {
-            if (0 !== $i) {
+        while (\strlen($s) - 1 > $i = strspn($s, self::ASCII)) {
+            if (0 < --$i) {
                 $str->string .= substr($s, 0, $i);
                 $s = substr($s, $i);
             }
@@ -106,10 +109,10 @@ public function ascii(array $rules = []): self
                     continue;
                 }
 
-                if ('nfkd' === $rule = strtolower($rule)) {
-                    if (!normalizer_is_normalized($s, self::NFKD)) {
-                        $s = normalizer_normalize($s, self::NFKD);
-                    }
+                if ('nfd' === $rule = strtolower($rule)) {
+                    normalizer_is_normalized($s, self::NFD) ?: $s = normalizer_normalize($s, self::NFD);
+                } elseif ('nfkd' === $rule) {
+                    normalizer_is_normalized($s, self::NFKD) ?: $s = normalizer_normalize($s, self::NFKD);
                 } elseif ('[:nonspacing mark:] remove' === $rule) {
                     $s = preg_replace('/\p{Mn}++/u', '', $s);
                 } elseif ('de-ascii' === $rule) {
diff --git a/src/Symfony/Component/String/Tests/AbstractUtf8TestCase.php b/src/Symfony/Component/String/Tests/AbstractUtf8TestCase.php
@@ -13,6 +13,13 @@ public function testCreateFromStringWithInvalidUtf8Input()
         static::createFromString("\xE9");
     }
 
+    public function testAscii()
+    {
+        $s = static::createFromString('Dieser Wert sollte größer oder gleich');
+        $this->assertSame('Dieser Wert sollte grosser oder gleich', (string) $s->ascii());
+        $this->assertSame('Dieser Wert sollte groesser oder gleich', (string) $s->ascii(['de-ASCII']));
+    }
+
     public function provideCreateFromCodePoint(): array
     {
         return [

Original file line number	Diff line number	Diff line change
`@@ -13,6 +13,13 @@ public function testCreateFromStringWithInvalidUtf8Input()`
`13`	`13`	`static::createFromString("\xE9");`
`14`	`14`	`}`
`15`	`15`
	`16`	`+ public function testAscii()`
	`17`	`+ {`
	`18`	`+ $s = static::createFromString('Dieser Wert sollte größer oder gleich');`
	`19`	`+ $this->assertSame('Dieser Wert sollte grosser oder gleich', (string) $s->ascii());`
	`20`	`+ $this->assertSame('Dieser Wert sollte groesser oder gleich', (string) $s->ascii(['de-ASCII']));`
	`21`	`+ }`
	`22`	`+`
`16`	`23`	`public function provideCreateFromCodePoint(): array`
`17`	`24`	`{`
`18`	`25`	`return [`