From e12ac900e707706fe6213fe86ffd030bb5a60ad7 Mon Sep 17 00:00:00 2001
From: William Jacoby <bonelifer@gmail.com>
Date: Mon, 12 Oct 2026 16:14:05 -0500
Subject: [PATCH] Keep BBCode when articles are reparsed

The reparser guessed from the stored text whether an article used BBCode, and for
text stored as XML without a bbcode_uid it guessed no, so reparsing turned every
[b], [list] and [url] into plain text. The wiki always saves with BBCode,
smilies and links on, so the plugin now says so.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---
 tests/unit/article_reparser_test.php  | 54 +++++++++++++++++++++++++++
 textreparser/plugins/article_text.php | 21 +++++++++++
 2 files changed, 75 insertions(+)
 create mode 100644 tests/unit/article_reparser_test.php

diff --git a/tests/unit/article_reparser_test.php b/tests/unit/article_reparser_test.php
new file mode 100644
index 0000000..253b004
--- /dev/null
+++ b/tests/unit/article_reparser_test.php
@@ -0,0 +1,54 @@
+<?php
+/**
+ *
+ * Wiki extension for the phpBB Forum Software package
+ *
+ * @copyright (c) 2026, phpBB Modders, https://www.phpbbmodders.com/
+ * @license GNU General Public License, version 2 (GPL-2.0)
+ *
+ */
+
+namespace phpbbmodders\wiki\tests\unit;
+
+class article_reparser_test extends \PHPUnit\Framework\TestCase
+{
+	/**
+	 * What the reparser decides about an article row.
+	 *
+	 * @param array $record The row as the reparser reads it
+	 * @return array The row with its enable_* fields
+	 */
+	private function complete(array $record): array
+	{
+		$plugin = new \phpbbmodders\wiki\textreparser\plugins\article_text(
+			$this->createMock(\phpbb\db\driver\driver_interface::class),
+			'phpbb_wiki_article'
+		);
+		$method = new \ReflectionMethod($plugin, 'add_missing_fields');
+		$method->setAccessible(true);
+
+		return $method->invoke($plugin, $record);
+	}
+
+	public function test_an_article_stored_as_xml_keeps_its_bbcode()
+	{
+		// Stored as XML with no bbcode_uid, which is how the wiki saves a version: core would guess "no BBCode" here
+		$record = $this->complete(array(
+			'id'			=> 1,
+			'text'			=> '<r><B><s>[b]</s>bold<e>[/b]</e></B> <WIKILINK name="A">[[A]]</WIKILINK></r>',
+			'bbcode_uid'	=> '',
+		));
+
+		foreach (array('bbcode', 'smilies', 'magic_url', 'flash_bbcode', 'img_bbcode', 'quote_bbcode', 'url_bbcode') as $field)
+		{
+			$this->assertTrue($record['enable_' . $field], $field);
+		}
+	}
+
+	public function test_an_article_with_plain_text_is_parsed_the_same_way()
+	{
+		$record = $this->complete(array('id' => 2, 'text' => '<t>Plain [b]text[/b]</t>', 'bbcode_uid' => ''));
+
+		$this->assertTrue($record['enable_bbcode']);
+	}
+}
diff --git a/textreparser/plugins/article_text.php b/textreparser/plugins/article_text.php
index 4a1b2c4..3e1af0f 100644
--- a/textreparser/plugins/article_text.php
+++ b/textreparser/plugins/article_text.php
@@ -24,6 +24,27 @@ public function get_columns()
 		);
 	}
 
+	/**
+	 * Articles are always saved with BBCode, smilies and links on (see wiki/edit.php), and
+	 * the table has no columns to say so. Core would guess from the stored text, and for text
+	 * stored as XML without a bbcode_uid it guesses "no BBCode", which stripped every [b], [list]
+	 * and [url] from an article the first time it was reparsed.
+	 *
+	 * {@inheritdoc}
+	 */
+	protected function add_missing_fields(array $record)
+	{
+		return $record + array(
+			'enable_bbcode'		=> true,
+			'enable_smilies'	=> true,
+			'enable_magic_url'	=> true,
+			'enable_flash_bbcode'	=> true,
+			'enable_img_bbcode'	=> true,
+			'enable_quote_bbcode'	=> true,
+			'enable_url_bbcode'	=> true,
+		);
+	}
+
 	/**
 	 * Reparse with the wiki's own markup on, so headings and table of contents
 	 * markers in older articles are recognised.