fix: media usage search never reports a capped scan as unused

Adds widgets and header/background theme mods, integer galleryIds, wider pre-filter, regex failures as possible usages, newest-first candidates, and usages_complete/usages_truncated in media-get. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
2026-09-26 02:02:21 +02:00
parent a2a2da1ae7
commit b197d8171f
4 changed files with 382 additions and 103 deletions
@@ -20,11 +20,15 @@ final class MediaGetAbilityTest extends TestCase
        Functions\when('get_post')->justReturn($attachment);
        Functions\when('current_user_can')->justReturn(true);
        $finder = Mockery::mock(MediaUsageFinder::class);
-        $finder->expects('usages')->with(33)->andReturn([]);
+        $usages = [['post_id' => 5, 'title' => 'Team', 'post_type' => 'page', 'where' => 'featured', 'certainty' => 'sicher']];
+        $finder->expects('scan')->with(33)->andReturn(['usages' => $usages, 'complete' => false, 'truncated' => true]);
        $presenter = Mockery::mock(MediaPresenter::class);
-        $presenter->expects('full')->with($attachment, [])->andReturn(['id' => 33]);
+        $presenter->expects('full')->with($attachment, $usages)->andReturn(['id' => 33, 'usages' => $usages]);

-        self::assertSame(['id' => 33], (new MediaGetAbility($presenter, $finder))->execute(['id' => 33]));
+        self::assertSame(
+            ['id' => 33, 'usages' => $usages, 'usages_complete' => false, 'usages_truncated' => true],
+            (new MediaGetAbility($presenter, $finder))->execute(['id' => 33])
+        );
    }

    public function testTrashedAttachmentsCanBeRead(): void
@@ -51,5 +55,7 @@ final class MediaGetAbilityTest extends TestCase
        self::assertSame('lit/media-get', $ability->name());
        self::assertFalse($ability->isDirect());
        self::assertTrue($ability->annotations()['readonly']);
+        self::assertContains('usages_complete', $ability->outputSchema()['required']);
+        self::assertContains('usages_truncated', $ability->outputSchema()['required']);
    }
 }
@@ -16,30 +16,32 @@ final class MediaUsageFinderTest extends TestCase
    private array $queries = [];

    /**
-     * @param array<string, mixed>          $options
-     * @param array<int, string>            $contents Candidate post ID => post_content.
-     * @param list<object>                  $featured
-     * @param list<object>                  $fields
+     * @param array<int, string>   $contents  Candidate post ID => post_content.
+     * @param list<object>         $featured
+     * @param list<object>         $fields
+     * @param array<string, mixed> $options   Option name => value.
+     * @param array<string, mixed> $themeMods Theme mod name => value.
     */
    private function finder(
        array $contents = [],
        array $featured = [],
        array $fields = [],
        array $options = [],
-        int $customLogo = 0,
-        array $mediaKeys = ['photo'],
+        array $themeMods = [],
+        array $mediaKeys = [],
        string $file = '2026/09/foto-scaled.jpg'
    ): MediaUsageFinder {
        Functions\when('get_post_meta')->justReturn($file);
        Functions\when('get_option')->alias(static fn (string $name) => $options[$name] ?? false);
-        Functions\when('get_theme_mod')->justReturn($customLogo);
+        Functions\when('get_theme_mod')->alias(static fn (string $name) => $themeMods[$name] ?? false);
+        Functions\when('post_type_exists')->alias(static fn (string $type): bool => $type !== 'ghost');
        Functions\when('current_user_can')->alias(static fn (string $cap, int $id): bool => $cap === 'read_post' && $id !== 13);

        /** @var MockInterface&\wpdb $db */
        $db = Mockery::mock(\wpdb::class)->makePartial();
        $db->allows('get_results')->andReturnUsing(function (string $sql) use ($contents, $featured, $fields): array {
            $this->queries[] = $sql;
-            if (str_contains($sql, "_thumbnail_id")) {
+            if (str_contains($sql, '_thumbnail_id')) {
                return $featured;
            }
            if (str_contains($sql, 'pm.meta_key IN')) {
@@ -52,10 +54,13 @@ final class MediaUsageFinderTest extends TestCase
                );
            }
            if (str_contains($sql, 'p.ID IN')) {
+                // No regex here: the PCRE limit test runs this with a backtrack limit of 1.
+                $list = substr($sql, (int) strpos($sql, 'p.ID IN (') + strlen('p.ID IN ('));
+                $ids = array_map('intval', explode(',', substr($list, 0, (int) strpos($list, ')'))));
                $rows = [];
-                foreach ($contents as $id => $content) {
-                    if (preg_match('/\b' . $id . '\b/', substr($sql, (int) strpos($sql, 'p.ID IN')))) {
-                        $rows[] = (object) ['ID' => $id, 'post_content' => $content];
+                foreach ($ids as $id) {
+                    if (isset($contents[$id])) {
+                        $rows[] = (object) ['ID' => $id, 'post_content' => $contents[$id]];
                    }
                }

@@ -68,25 +73,31 @@ final class MediaUsageFinderTest extends TestCase
        return new MediaUsageFinder(static fn (): array => $mediaKeys, $db);
    }

-    /** @return array<int, string> post_id => certainty */
+    /** @return array<int|string, string> post_id (or title for site usages) => certainty */
    private static function certainties(array $usages): array
    {
        $result = [];
        foreach ($usages as $usage) {
-            $result[$usage['post_id']] = $usage['certainty'];
+            $result[$usage['post_id'] === 0 ? $usage['title'] : $usage['post_id']] = $usage['certainty'];
        }

        return $result;
    }

-    public function testCombinesFeaturedFieldContentAndSiteUsages(): void
+    private static function post(int $id, string $title, string $type = 'page'): object
+    {
+        return (object) ['ID' => $id, 'post_title' => $title, 'post_type' => $type];
+    }
+
+    public function testCombinesFeaturedFieldSiteAndContentUsages(): void
    {
        $usages = $this->finder(
            [5 => '<img class="wp-image-33">', 9 => '<img class="wp-image-33 size-full">'],
-            [(object) ['ID' => 5, 'post_title' => 'Team', 'post_type' => 'page']],
-            [(object) ['ID' => 12, 'post_title' => 'Anna', 'post_type' => 'lit_person']],
+            [self::post(5, 'Team')],
+            [self::post(12, 'Anna', 'lit_person')],
            ['site_icon' => '33'],
-            33
+            ['custom_logo' => 33],
+            ['photo']
        )->usages(33);

        self::assertSame(
@@ -101,19 +112,24 @@ final class MediaUsageFinderTest extends TestCase
        );
    }

-    public function testQueriesArePreparedAndBounded(): void
+    public function testQueriesArePreparedBoundedAndNewestFirst(): void
    {
-        $this->finder([5 => 'x'])->usages(33, 7);
+        $this->finder([5 => 'x'], [], [], [], [], ['photo'])->usages(33, 7);

        [$featured, $fields, $candidates] = $this->queries;
        self::assertStringContainsString("pm.meta_value = '33'", $featured);
-        self::assertStringContainsString('LIMIT 7', $featured);
+        self::assertStringContainsString('ORDER BY p.ID DESC LIMIT 8', $featured);
        self::assertStringContainsString("pm.meta_key IN ('photo')", $fields);
-        self::assertStringContainsString('LIMIT 7', $fields);
-        foreach (['%wp-image-33%', '%"id":33%', '%"id":"33"%', '%"mediaId":33%', '%ids=%33%', '%"ids":[%33%', '%galleryIds":%33%', '%/2026/09/foto%', '%\\\\/2026\\\\/09\\\\/foto%'] as $like) {
+        self::assertStringContainsString('ORDER BY p.ID DESC LIMIT 8', $fields);
+        $likes = [
+            '%wp-image-33%', '%"id":33%', '%"id":"33"%', '%"id": 33%', '%"id": "33"%', '%"mediaId":33%',
+            '%"mediaId":"33"%', '%"mediaId": 33%', '%ids=%33%', '%"ids":%33%', '%galleryIds":%33%',
+            '%/2026/09/foto%', '%\\\\/2026\\\\/09\\\\/foto%',
+        ];
+        foreach ($likes as $like) {
            self::assertStringContainsString("'" . $like . "'", $candidates, $like);
        }
-        self::assertStringContainsString('LIMIT 200', $candidates);
+        self::assertStringContainsString('ORDER BY p.ID DESC LIMIT 200', $candidates);
        self::assertStringNotContainsString('post_content,', $candidates, 'candidates must not load content');
    }

@@ -124,16 +140,17 @@ final class MediaUsageFinderTest extends TestCase
            2 => '[et_pb_gallery gallery_ids="33,8"][/et_pb_gallery]',
            3 => '<!-- wp:image {"id":33,"sizeSlug":"large"} --><figure></figure><!-- /wp:image -->',
            4 => '<!-- wp:gallery {"ids":[2,33]} /-->',
-            5 => '<!-- wp:media-text {"mediaId":33,"mediaType":"image"} --><div></div><!-- /wp:media-text -->',
-            6 => '<!-- wp:cover {"url":"x","id":33} --><div></div><!-- /wp:cover -->',
+            5 => '<!-- wp:media-text {"mediaId":"33","mediaType":"image"} --><div></div><!-- /wp:media-text -->',
+            6 => '<!-- wp:cover {"url":"x", "id": 33} --><div></div><!-- /wp:cover -->',
            7 => '<!-- wp:divi/image {"image":{"innerContent":{"desktop":{"value":{"src":"https://k.de/a.jpg","id":"33"}}}}} /-->',
            8 => '<!-- wp:divi/gallery {"image":{"advanced":{"galleryIds":{"desktop":{"value":"12,33"}}}}} /-->',
            9 => '<!-- wp:divi/text {"content":{"innerContent":{"desktop":{"value":"[et_pb_gallery gallery_ids=\\u002233\\u0022]"}}}} /-->',
+            10 => '<!-- wp:divi/gallery {"image":{"advanced":{"galleryIds":{"desktop":{"value":[12,33]}}}}} /-->',
        ];

-        $usages = $this->finder($contents, [], [], [], 0, [])->usages(33);
+        $usages = $this->finder($contents)->usages(33);

-        self::assertSame(array_fill_keys(range(1, 9), 'sicher'), self::certainties($usages));
+        self::assertSame(array_fill_keys(range(1, 10), 'sicher'), self::certainties($usages));
    }

    public function testFileUrlMatchesAreOnlyPossibleUsages(): void
@@ -146,7 +163,7 @@ final class MediaUsageFinderTest extends TestCase

        self::assertSame(
            [1 => 'möglicherweise', 2 => 'möglicherweise', 3 => 'möglicherweise'],
-            self::certainties($this->finder($contents, [], [], [], 0, [])->usages(33))
+            self::certainties($this->finder($contents)->usages(33))
        );
    }

@@ -157,24 +174,113 @@ final class MediaUsageFinderTest extends TestCase
            2 => '<!-- wp:image {"id":330} /--> <!-- wp:gallery {"ids":[133,334]} /-->',
            3 => '[gallery ids="330,1334"] <!-- wp:paragraph {"id":33} /-->',
            4 => '<img src="https://k.de/wp-content/uploads/2026/09/foto-2.jpg"> /2025/09/foto.jpg',
+            5 => '<!-- wp:divi/gallery {"image":{"advanced":{"galleryIds":{"desktop":{"value":[330,3]}}}}} /-->',
        ];

-        self::assertSame([], $this->finder($contents, [], [], [], 0, [])->usages(33));
+        $finder = $this->finder($contents);
+
+        self::assertSame([], $finder->usages(33));
+        self::assertFalse($finder->isUsed(33));
+    }
+
+    public function testCappedContentSearchIsNeverReportedAsUnused(): void
+    {
+        $finder = $this->finder(array_fill_keys(range(1, 200), 'nichts'));
+
+        self::assertTrue($finder->isUsed(33));
+
+        $scan = $finder->scan(33);
+        self::assertFalse($scan['complete']);
+        self::assertFalse($scan['truncated']);
+        self::assertSame(
+            [[
+                'post_id' => 0,
+                'title' => 'Inhaltssuche unvollständig – weitere Verwendungen möglich',
+                'post_type' => 'content',
+                'where' => 'content',
+                'certainty' => 'möglicherweise',
+            ]],
+            $scan['usages']
+        );
+    }
+
+    public function testScanReportsTruncationBeyondMax(): void
+    {
+        $finder = $this->finder([], [self::post(1, 'A'), self::post(2, 'B'), self::post(3, 'C')]);
+
+        $scan = $finder->scan(33, 2);
+
+        self::assertCount(2, $scan['usages']);
+        self::assertTrue($scan['truncated']);
+        self::assertTrue($scan['complete']);
+        self::assertFalse($finder->scan(33, 3)['truncated']);
+    }
+
+    public function testRegexFailureCountsAsPossibleUsage(): void
+    {
+        $jit = ini_get('pcre.jit');
+        $backtrack = ini_get('pcre.backtrack_limit');
+        $finder = $this->finder([7 => str_repeat('wp-image-3 ', 200)]);
+        ini_set('pcre.jit', '0');
+        ini_set('pcre.backtrack_limit', '1');
+        try {
+            $usages = $finder->usages(33);
+        } finally {
+            ini_set('pcre.jit', (string) $jit);
+            ini_set('pcre.backtrack_limit', (string) $backtrack);
+        }
+
+        self::assertSame([7 => 'möglicherweise'], self::certainties($usages));
+    }
+
+    public function testWidgetsAndThemeMods(): void
+    {
+        $usages = $this->finder(
+            [],
+            [],
+            [],
+            [
+                'widget_media_image' => [3 => ['attachment_id' => 33, 'url' => 'x'], 4 => ['attachment_id' => 330], '_multiwidget' => 1],
+                'widget_media_gallery' => [2 => ['ids' => [5, 33]], 3 => ['ids' => [330]], '_multiwidget' => 1],
+                'widget_block' => [
+                    4 => ['content' => '<!-- wp:image {"id":33} --><figure></figure><!-- /wp:image -->'],
+                    5 => ['content' => '<img src="https://k.de/wp-content/uploads/2026/09/foto-150x150.jpg">'],
+                    6 => ['content' => '<p>nichts</p>'],
+                    '_multiwidget' => 1,
+                ],
+            ],
+            [
+                'header_image_data' => (object) ['attachment_id' => 33, 'url' => 'x'],
+                'background_image' => 'https://k.de/wp-content/uploads/2026/09/foto.jpg',
+            ]
+        )->usages(33);
+
+        self::assertSame(
+            [
+                'Header-Bild' => 'sicher',
+                'Hintergrundbild' => 'möglicherweise',
+                'Widget media_image-3' => 'sicher',
+                'Widget media_gallery-2' => 'sicher',
+                'Widget block-4' => 'sicher',
+                'Widget block-5' => 'möglicherweise',
+            ],
+            self::certainties($usages)
+        );
+        self::assertSame(['header_image', 'background_image', 'widget', 'widget', 'widget', 'widget'], array_column($usages, 'where'));
    }

    public function testUnreadableEntriesHideTheirTitleAndDiviLogoIsPossible(): void
    {
        $usages = $this->finder(
            [13 => '<img class="wp-image-33">'],
+            [self::post(20, 'Geist', 'ghost')],
            [],
-            [],
-            ['et_divi' => ['divi_logo' => 'https://k.de/wp-content/uploads/2026/09/foto.jpg']],
-            0,
-            []
+            ['et_divi' => ['divi_logo' => 'https://k.de/wp-content/uploads/2026/09/foto.jpg']]
        )->usages(33);

        self::assertSame(
            [
+                ['post_id' => 20, 'title' => '(nicht lesbar)', 'post_type' => 'ghost', 'where' => 'featured', 'certainty' => 'sicher'],
                ['post_id' => 0, 'title' => 'Divi-Logo', 'post_type' => 'option', 'where' => 'divi_logo', 'certainty' => 'möglicherweise'],
                ['post_id' => 13, 'title' => '(nicht lesbar)', 'post_type' => 'page', 'where' => 'content', 'certainty' => 'sicher'],
            ],
@@ -184,20 +290,20 @@ final class MediaUsageFinderTest extends TestCase

    public function testIdReferenceWinsOverAFileMatchInTheSameContent(): void
    {
-        $usages = $this->finder([5 => '/2026/09/foto.jpg <img class="wp-image-33">'], [], [], [], 0, [])->usages(33);
+        $usages = $this->finder([5 => '/2026/09/foto.jpg <img class="wp-image-33">'])->usages(33);

        self::assertSame([5 => 'sicher'], self::certainties($usages));
    }

    public function testNoFieldQueryWithoutMediaFieldsAndUnusedMedia(): void
    {
-        self::assertFalse($this->finder([], [], [], [], 0, [])->isUsed(33));
+        self::assertFalse($this->finder()->isUsed(33));
        self::assertCount(2, $this->queries);
    }

-    public function testIsUsedStopsBeforeTheContentScanWhenAlreadyUsed(): void
+    public function testIsUsedStopsAfterTheFirstHit(): void
    {
-        $finder = $this->finder([5 => 'wp-image-33'], [(object) ['ID' => 5, 'post_title' => 'Team', 'post_type' => 'page']]);
+        $finder = $this->finder([5 => 'wp-image-33'], [self::post(5, 'Team')], [], [], [], ['photo']);

        self::assertTrue($finder->isUsed(33));
        self::assertCount(1, $this->queries);
@@ -205,7 +311,7 @@ final class MediaUsageFinderTest extends TestCase

    public function testWithoutAttachedFileNoFilePatterns(): void
    {
-        $this->finder([], [], [], [], 0, [], '')->usages(33);
+        $this->finder([], [], [], [], [], [], '')->usages(33);

        self::assertStringNotContainsString('foto', $this->queries[1]);
        self::assertStringNotContainsString("'%/%'", $this->queries[1]);