Add implementation of Myers differ

Intended to be used in the format-preserving pretty printer for
handling changes in node lists.

lib/PhpParser/Internal/DiffElem.php

@@ -0,0 +1,26 @@
+<?php
+
+namespace PhpParser\Internal;
+
+/**
+ * @internal
+ */
+class DiffElem {
+    const TYPE_KEEP = 0;
+    const TYPE_REMOVE = 1;
+    const TYPE_ADD = 2;
+    const TYPE_REPLACE = 3;
+
+    /** @var int One of the TYPE_* constants */
+    public $type;
+    /** @var mixed Is null for add operations */
+    public $old;
+    /** @var mixed Is null for remove operations */
+    public $new;
+
+    public function __construct(int $type, $old, $new) {
+        $this->type = $type;
+        $this->old = $old;
+        $this->new = $new;
+    }
+}

lib/PhpParser/Internal/Differ.php

@@ -0,0 +1,163 @@
+<?php
+
+namespace PhpParser\Internal;
+
+/**
+ * Implements the Myers diff algorithm.
+ *
+ * Myers, Eugene W. "An O (ND) difference algorithm and its variations."
+ * Algorithmica 1.1 (1986): 251-266.
+ *
+ * @internal
+ */
+class Differ {
+    private $isEqual;
+
+    /**
+     * Create differ over the given equality relation.
+     *
+     * @param callable $isEqual Equality relation with signature function($a, $b): bool
+     */
+    public function __construct(callable $isEqual) {
+        $this->isEqual = $isEqual;
+    }
+
+    /**
+     * Calculate diff (edit script) from $old to $new.
+     *
+     * @param array $old Original array
+     * @param array $new New array
+     *
+     * @return DiffElem[] Diff (edit script)
+     */
+    public function diff(array $old, array $new) {
+        list($trace, $x, $y) = $this->calculateTrace($old, $new);
+        return $this->extractDiff($trace, $x, $y, $old, $new);
+    }
+
+    /**
+     * Calculate diff, including "replace" operations.
+     *
+     * If a sequence of remove operations is followed by the same number of add operations, these
+     * will be coalesced into replace operations.
+     *
+     * @param array $old Original array
+     * @param array $new New array
+     *
+     * @return DiffElem[] Diff (edit script), including replace operations
+     */
+    public function diffWithReplacements(array $old, array $new) {
+        return $this->coalesceReplacements($this->diff($old, $new));
+    }
+
+    private function calculateTrace(array $a, array $b) {
+        $n = \count($a);
+        $m = \count($b);
+        $max = $n + $m;
+        $v = [1 => 0];
+        $trace = [];
+        for ($d = 0; $d <= $max; $d++) {
+            $trace[] = $v;
+            for ($k = -$d; $k <= $d; $k += 2) {
+                if ($k === -$d || ($k !== $d && $v[$k-1] < $v[$k+1])) {
+                    $x = $v[$k+1];
+                } else {
+                    $x = $v[$k-1] + 1;
+                }
+
+                $y = $x - $k;
+                while ($x < $n && $y < $m && ($this->isEqual)($a[$x], $b[$y])) {
+                    $x++;
+                    $y++;
+                }
+
+                $v[$k] = $x;
+                if ($x >= $n && $y >= $m) {
+                    return [$trace, $x, $y];
+                }
+            }
+        }
+        throw new \Exception('Should not happen');
+    }
+
+    private function extractDiff(array $trace, int $x, int $y, array $a, array $b) {
+        $result = [];
+        for ($d = \count($trace) - 1; $d >= 0; $d--) {
+            $v = $trace[$d];
+            $k = $x - $y;
+
+            if ($k === -$d || ($k !== $d && $v[$k-1] < $v[$k+1])) {
+                $prevK = $k + 1;
+            } else {
+                $prevK = $k - 1;
+            }
+
+            $prevX = $v[$prevK];
+            $prevY = $prevX - $prevK;
+
+            while ($x > $prevX && $y > $prevY) {
+                $result[] = new DiffElem(DiffElem::TYPE_KEEP, $a[$x-1], $b[$y-1]);
+                $x--;
+                $y--;
+            }
+
+            if ($d === 0) {
+                break;
+            }
+
+            while ($x > $prevX) {
+                $result[] = new DiffElem(DiffElem::TYPE_REMOVE, $a[$x-1], null);
+                $x--;
+            }
+
+            while ($y > $prevY) {
+                $result[] = new DiffElem(DiffElem::TYPE_ADD, null, $b[$y-1]);
+                $y--;
+            }
+        }
+        return array_reverse($result);
+    }
+
+    /**
+     * Coalesce equal-length sequences of remove+add into a replace operation.
+     *
+     * @param DiffElem[] $diff
+     * @return DiffElem[]
+     */
+    private function coalesceReplacements(array $diff) {
+        $newDiff = [];
+        $c = \count($diff);
+        for ($i = 0; $i < $c; $i++) {
+            $diffType = $diff[$i]->type;
+            if ($diffType !== DiffElem::TYPE_REMOVE) {
+                $newDiff[] = $diff[$i];
+                continue;
+            }
+
+            $j = $i;
+            while ($j < $c && $diff[$j]->type === DiffElem::TYPE_REMOVE) {
+                $j++;
+            }
+
+            $k = $j;
+            while ($k < $c && $diff[$k]->type === DiffElem::TYPE_ADD) {
+                $k++;
+            }
+
+            if ($j - $i === $k - $j) {
+                $len = $j - $i;
+                for ($n = 0; $n < $len; $n++) {
+                    $newDiff[] = new DiffElem(
+                        DiffElem::TYPE_REPLACE, $diff[$i + $n]->old, $diff[$j + $n]->new
+                    );
+                }
+            } else {
+                for (; $i < $k; $i++) {
+                    $newDiff[] = $diff[$i];
+                }
+            }
+            $i = $k - 1;
+        }
+        return $newDiff;
+    }
+}

test/PhpParser/Internal/DifferTest.php

@@ -0,0 +1,66 @@
+<?php
+
+namespace PhpParser\Internal;
+
+use PHPUnit\Framework\TestCase;
+
+class DifferTest extends TestCase {
+    private function formatDiffString(array $diff) {
+        $diffStr = '';
+        foreach ($diff as $diffElem) {
+            switch ($diffElem->type) {
+                case DiffElem::TYPE_KEEP:
+                    $diffStr .= $diffElem->old;
+                    break;
+                case DiffElem::TYPE_REMOVE:
+                    $diffStr .= '-' . $diffElem->old;
+                    break;
+                case DiffElem::TYPE_ADD:
+                    $diffStr .= '+' . $diffElem->new;
+                    break;
+                case DiffElem::TYPE_REPLACE:
+                    $diffStr .= '/' . $diffElem->old . $diffElem->new;
+                    break;
+                default:
+                    assert(false);
+                    break;
+            }
+        }
+        return $diffStr;
+    }
+
+    /** @dataProvider provideTestDiff */
+    public function testDiff($oldStr, $newStr, $expectedDiffStr) {
+        $differ = new Differ(function($a, $b) { return $a === $b; });
+        $diff = $differ->diff(str_split($oldStr), str_split($newStr));
+        $this->assertSame($expectedDiffStr, $this->formatDiffString($diff));
+    }
+
+    public function provideTestDiff() {
+        return [
+            ['abc', 'abc', 'abc'],
+            ['abc', 'abcdef', 'abc+d+e+f'],
+            ['abcdef', 'abc', 'abc-d-e-f'],
+            ['abcdef', 'abcxyzdef', 'abc+x+y+zdef'],
+            ['axyzb', 'ab', 'a-x-y-zb'],
+            ['abcdef', 'abxyef', 'ab-c-d+x+yef'],
+            ['abcdef', 'cdefab', '-a-bcdef+a+b'],
+        ];
+    }
+
+    /** @dataProvider provideTestDiffWithReplacements */
+    public function testDiffWithReplacements($oldStr, $newStr, $expectedDiffStr) {
+        $differ = new Differ(function($a, $b) { return $a === $b; });
+        $diff = $differ->diffWithReplacements(str_split($oldStr), str_split($newStr));
+        $this->assertSame($expectedDiffStr, $this->formatDiffString($diff));
+    }
+
+    public function provideTestDiffWithReplacements() {
+        return [
+            ['abcde', 'axyze', 'a/bx/cy/dze'],
+            ['abcde', 'xbcdy', '/axbcd/ey'],
+            ['abcde', 'axye', 'a-b-c-d+x+ye'],
+            ['abcde', 'axyzue', 'a-b-c-d+x+y+z+ue'],
+        ];
+    }
+}