From 6068a6a54f5695ecac400191d6237e3d255be6dd Mon Sep 17 00:00:00 2001
From: Vladimir Bryzgalov <vladimir.bryzgalov@gmail.com>
Date: Fri, 17 Apr 2026 22:03:33 +0500
Subject: [PATCH] feat(domain-formatter): decode punycode hosts to Cyrillic
 with fallbacks

---
 includes/class-domain-formatter.php | 43 +++++++++++++++++++++++
 tests/DomainFormatterTest.php       | 53 +++++++++++++++++++++++++++++
 2 files changed, 96 insertions(+)
 create mode 100644 includes/class-domain-formatter.php
 create mode 100644 tests/DomainFormatterTest.php

diff --git a/includes/class-domain-formatter.php b/includes/class-domain-formatter.php
new file mode 100644
index 0000000..42add0c
--- /dev/null
+++ b/includes/class-domain-formatter.php
@@ -0,0 +1,43 @@
+<?php
+declare(strict_types=1);
+
+namespace Cf7stg;
+
+final class DomainFormatter
+{
+    public static function humanize(string $input): string
+    {
+        $input = trim($input);
+        if ($input === '') {
+            return '';
+        }
+
+        // Normalize: strip scheme, userinfo, path, query, fragment
+        $host = $input;
+        if (preg_match('#^[a-z][a-z0-9+\-.]*://#i', $host)) {
+            $parsed = parse_url($host);
+            $host = $parsed['host'] ?? $host;
+        }
+        // Strip any remaining path/query
+        $host = preg_replace('~[/?#].*$~', '', $host) ?? $host;
+        // Strip leading www.
+        $host = preg_replace('/^www\./i', '', $host) ?? $host;
+
+        if ($host === '') {
+            return '';
+        }
+
+        // Punycode → Unicode per label (.xn-- pieces may appear on only one side)
+        if (function_exists('idn_to_utf8') && strpos($host, 'xn--') !== false) {
+            $variant = defined('INTL_IDNA_VARIANT_UTS46')
+                ? INTL_IDNA_VARIANT_UTS46
+                : 0;
+            $decoded = @idn_to_utf8($host, IDNA_DEFAULT, $variant);
+            if (is_string($decoded) && $decoded !== '') {
+                return $decoded;
+            }
+        }
+
+        return $host;
+    }
+}
diff --git a/tests/DomainFormatterTest.php b/tests/DomainFormatterTest.php
new file mode 100644
index 0000000..d940fab
--- /dev/null
+++ b/tests/DomainFormatterTest.php
@@ -0,0 +1,53 @@
+<?php
+declare(strict_types=1);
+
+namespace Cf7stg\Tests;
+
+use Cf7stg\DomainFormatter;
+use PHPUnit\Framework\TestCase;
+
+final class DomainFormatterTest extends TestCase
+{
+    public function test_decodes_punycode_to_cyrillic(): void
+    {
+        self::assertSame(
+            'вашаняня.рф',
+            DomainFormatter::humanize('xn--80aae0ca7d8bb.xn--p1ai')
+        );
+    }
+
+    public function test_latin_domain_unchanged(): void
+    {
+        self::assertSame(
+            'washanyanya.ru',
+            DomainFormatter::humanize('washanyanya.ru')
+        );
+    }
+
+    public function test_strips_scheme_and_path(): void
+    {
+        self::assertSame(
+            'washanyanya.ru',
+            DomainFormatter::humanize('https://washanyanya.ru/contact/?a=1')
+        );
+    }
+
+    public function test_strips_leading_www(): void
+    {
+        self::assertSame(
+            'washanyanya.ru',
+            DomainFormatter::humanize('www.washanyanya.ru')
+        );
+    }
+
+    public function test_invalid_input_returns_empty_string(): void
+    {
+        self::assertSame('', DomainFormatter::humanize(''));
+    }
+
+    public function test_broken_idn_returns_original(): void
+    {
+        // Looks like IDN but is garbage; idn_to_utf8 returns false → we fall back
+        self::assertSame('xn--', DomainFormatter::humanize('xn--'));
+    }
+}