Refactor UTF-8 seeking

lib/Encoding/UTF8.php

@@ -102,62 +102,31 @@ class UTF8 implements StatelessEncoding {
         return $bytes;
     }
 
-    /** Advance $distance characters through the string
-     *
-     * If $distance is negative, the operation will be performed in reverse
-     *
-     * If the end (or beginning) of the string was reached before the end of the operation, the remaining number of requested characters is returned
-     */
-    public function seek(int $distance): int {
-        if ($distance > 0) {
-            if ($this->posByte == strlen($this->string)) {
-                // if we're already at the end of the string, we can't go further
-                return $distance;
-            }
-            do {
-                // get the next code point; this automatically increments the character position
-                $p = $this->nextCode();
-            } while (--$distance && $p !== false); // stop after we have skipped the desired number of characters, or reached EOF
-            return $distance;
-        } elseif ($distance < 0) {
-            $distance = abs($distance);
-            if (!$this->posByte) {
-                // if we're already at the start of the string, we can't go further back
-                return $distance;
-            }
-            $mode = $this->errMode;
-            $this->errMode = self::MODE_NULL;
-            do {
-                $this->sync($this->posByte - 1);
-                // manually decrement the character position
-                $this->posChar--;
-            } while (--$distance && $this->posByte);
-            $this->errMode = $mode;
-            return $distance;
-        } else {
-            return 0;
-        }
-    }
-
-    /** Synchronize to the byte offset of the start of the nearest character at or before byte offset $pos */
-    protected function sync(int $pos) {
-        $b = ord(@$this->string[$pos]);
-        if ($b < 0x80) {
-            // if the byte is an ASCII byte or the end of input, then this is already a synchronized position
-            $this->posByte = $pos;
-        } else {
-            $s = $pos;
-            while ($b >= 0x80 && $b <= 0xBF && $pos > 0 && ($s - $pos) < 3) { // go back at most three bytes, no further than the start of the string, and only as long as the byte remains a continuation byte
-                $b = ord(@$this->string[--$pos]);
-            }
-            $this->posByte = $pos;
-            // decrement the character position because nextCode() increments it
+    /** Implements backward seeking $distance characters */
+    protected function seekBack(int $distance): int {
+        while ($distance > 0 && $this->posByte > 0) {
+            $distance--;
             $this->posChar--;
-            if (is_null($this->nextCode())) {
-                $this->posByte = $s;
+            $pos = $this->posByte - 1;
+            $b = ord(@$this->string[$pos]);
+            if ($b < 0x80) {
+                // if the byte is an ASCII byte or the end of input, then this is already a synchronized position
+                $this->posByte = $pos;
             } else {
-                $this->posByte = ($this->posByte > $s) ? $pos : $s;
+                $s = $pos;
+                while ($b >= 0x80 && $b <= 0xBF && $pos > 0 && ($s - $pos) < 3) { // go back at most three bytes, no further than the start of the string, and only as long as the byte remains a continuation byte
+                    $b = ord(@$this->string[--$pos]);
+                }
+                $this->posByte = $pos;
+                // decrement the character position because nextCode() increments it
+                $this->posChar--;
+                if (is_null($this->nextCode())) {
+                    $this->posByte = $s;
+                } else {
+                    $this->posByte = ($this->posByte > $s) ? $pos : $s;
+                }
             }
         }
+        return $distance;
     }
 }

tests/cases/Encoding/TestUTF8.php

@@ -57,7 +57,7 @@ class TestUTF8 extends \MensBeam\Intl\Test\EncodingTest {
 
     /**
      * @dataProvider provideStrings
-     * @covers MensBeam\Intl\Encoding\UTF8::sync
+     * @covers MensBeam\Intl\Encoding\UTF8::seekBack
     */
     public function testSTepBackThroughAString(string $input, array $exp) {
         return parent::testSTepBackThroughAString($input, $exp);