From e2c41360013e28d024395428c95389493c08a109 Mon Sep 17 00:00:00 2001
From: "J. King" <jking@jkingweb.ca>
Date: Fri, 10 Aug 2018 12:24:10 -0400
Subject: [PATCH] Change iterator to a set of generators

Not only is the faster than a classical iterator (though still not as
fast as a while loop), but it also offers the choice of characters
or code points.
---
 lib/Encoding/UTF8.php             | 22 +++++++----------
 perf/perf.php                     |  9 +++++--
 tests/cases/Encoding/TestUTF8.php | 39 +++++++++++++++++++++++++++----
 3 files changed, 50 insertions(+), 20 deletions(-)

diff --git a/lib/Encoding/UTF8.php b/lib/Encoding/UTF8.php
index 946fc21..0c2955e 100644
--- a/lib/Encoding/UTF8.php
+++ b/lib/Encoding/UTF8.php
@@ -6,7 +6,7 @@
 declare(strict_types=1);
 namespace MensBeam\Intl\Encoding;
 
-class UTF8 implements \Iterator {
+class UTF8 {
     const MODE_NULL = 0;
     const MODE_REPLACE = 1;
     const MODE_HTML = 2;
@@ -31,20 +31,16 @@ class UTF8 implements \Iterator {
         $this->current = null;
     }
 
-    public function valid() {
-        return $this->posByte < $this->lenByte;
-    }
-
-    public function current() {
-        return $this->current ?? ($this->current = $this->nextCode());
-    }
-
-    public function key() {
-        return isset($this->current) ? $this->posChar - 1 : $this->posChar;
+    public function chars(): \Generator {
+        while (($c = $this->nextChar()) !== "") {
+            yield ($this->posChar - 1) => $c;
+        }
     }
 
-    public function next() {
-        $this->current = null;
+    public function codes(): \Generator {
+        while (($c = $this->nextCode()) !== false) {
+            yield ($this->posChar - 1) => $c;
+        }
     }
 
     public function __construct(string $string, bool $fatal = false) {
diff --git a/perf/perf.php b/perf/perf.php
index 49e3c1e..47ba8b6 100644
--- a/perf/perf.php
+++ b/perf/perf.php
@@ -35,6 +35,11 @@ $tests = [
             $c = $i->nextChar();
         }
     }],
+    'Character generator' => ["", function(string $text) {
+        $c = null;
+        $i = new UTF8($text);
+        foreach ($i->chars() as $c);
+    }],
     'Intl code points' => ["intl", function(string $text) {
         $i = (function($text) {
             $i = \IntlBreakIterator::createCodePointInstance();
@@ -54,10 +59,10 @@ $tests = [
             $p = $i->nextCode();
         }
     }],
-    'Code point iterator' => ["", function(string $text) {
+    'Code point generator' => ["", function(string $text) {
         $c = null;
         $i = new UTF8($text);
-        foreach ($i as $c);
+        foreach ($i->codes() as $c);
     }],
 ];
 
diff --git a/tests/cases/Encoding/TestUTF8.php b/tests/cases/Encoding/TestUTF8.php
index c8cfa3a..63b06c5 100644
--- a/tests/cases/Encoding/TestUTF8.php
+++ b/tests/cases/Encoding/TestUTF8.php
@@ -59,21 +59,50 @@ class TestUTF8 extends \PHPUnit\Framework\TestCase {
     /**
      * @dataProvider provideStrings
      * @covers MensBeam\Intl\Encoding\UTF8::rewind
-     * @covers MensBeam\Intl\Encoding\UTF8::valid
-     * @covers MensBeam\Intl\Encoding\UTF8::current
-     * @covers MensBeam\Intl\Encoding\UTF8::key
-     * @covers MensBeam\Intl\Encoding\UTF8::next
+     * @covers MensBeam\Intl\Encoding\UTF8::chars
+     * @covers MensBeam\Intl\Encoding\UTF8::codes
     */
     public function testIterateThroughAString(string $input, array $exp) {
         $out = [];
         $s = new UTF8($input);
         $a = 0;
         $this->assertTrue(true); // prevent risky test of empty string
-        foreach ($s as $index => $p) {
+        foreach ($s->codes() as $index => $p) {
             $this->assertSame($a, $index, "Character key at index $a reported incorrectly");
             $this->assertSame($exp[$a], $p, "Character at index $a decoded incorrectly");
             $a++;
         }
+        $a = 0;
+        foreach ($s->codes() as $p) {
+            $a++;
+        }
+        $this->assertSame(0, $a);
+        $s->rewind();
+        foreach ($s->codes() as $p) {
+            $a++;
+        }
+        $this->assertSame(sizeof($exp), $a);
+
+        $exp = array_map(function ($v) {
+            return \IntlChar::chr($v);
+        }, $exp);
+
+        foreach ($s->chars() as $index => $p) {
+            $this->assertSame($a, $index, "Character key at index $a reported incorrectly");
+            $this->assertSame(bin2hex($exp[$a]), bin2hex($p), "Character at index $a decoded incorrectly");
+            $a++;
+        }
+        $a = 0;
+        foreach ($s->chars() as $p) {
+            $a++;
+        }
+        $this->assertSame(0, $a);
+        $s->rewind();
+        foreach ($s->chars() as $p) {
+            $a++;
+        }
+        $this->assertSame(sizeof($exp), $a);
+
     }
 
     /**