X-Git-Url: http://source.bookstackapp.com/bookstack/blobdiff_plain/4ddbc9556bf02ecd31aff093ed4acb200c9fcfd6..refs/pull/4467/head:/app/Entities/Tools/PageContent.php diff --git a/app/Entities/Tools/PageContent.php b/app/Entities/Tools/PageContent.php index c60cf0311..949816eff 100644 --- a/app/Entities/Tools/PageContent.php +++ b/app/Entities/Tools/PageContent.php @@ -3,8 +3,7 @@ namespace BookStack\Entities\Tools; use BookStack\Entities\Models\Page; -use BookStack\Entities\Tools\Markdown\CustomListItemRenderer; -use BookStack\Entities\Tools\Markdown\CustomStrikeThroughExtension; +use BookStack\Entities\Tools\Markdown\MarkdownToHtml; use BookStack\Exceptions\ImageUploadException; use BookStack\Facades\Theme; use BookStack\Theming\ThemeEvents; @@ -12,31 +11,23 @@ use BookStack\Uploads\ImageRepo; use BookStack\Uploads\ImageService; use BookStack\Util\HtmlContentFilter; use DOMDocument; +use DOMElement; +use DOMNode; use DOMNodeList; use DOMXPath; use Illuminate\Support\Str; -use League\CommonMark\Block\Element\ListItem; -use League\CommonMark\CommonMarkConverter; -use League\CommonMark\Environment; -use League\CommonMark\Extension\Table\TableExtension; -use League\CommonMark\Extension\TaskList\TaskListExtension; class PageContent { - protected $page; - - /** - * PageContent constructor. - */ - public function __construct(Page $page) - { - $this->page = $page; + public function __construct( + protected Page $page + ) { } /** * Update the content of the page with new provided HTML. */ - public function setNewHTML(string $html) + public function setNewHTML(string $html): void { $html = $this->extractBase64ImagesFromHtml($html); $this->page->html = $this->formatHtml($html); @@ -47,38 +38,21 @@ class PageContent /** * Update the content of the page with new provided Markdown content. */ - public function setNewMarkdown(string $markdown) + public function setNewMarkdown(string $markdown): void { $markdown = $this->extractBase64ImagesFromMarkdown($markdown); $this->page->markdown = $markdown; - $html = $this->markdownToHtml($markdown); + $html = (new MarkdownToHtml($markdown))->convert(); $this->page->html = $this->formatHtml($html); $this->page->text = $this->toPlainText(); } - /** - * Convert the given Markdown content to a HTML string. - */ - protected function markdownToHtml(string $markdown): string - { - $environment = Environment::createCommonMarkEnvironment(); - $environment->addExtension(new TableExtension()); - $environment->addExtension(new TaskListExtension()); - $environment->addExtension(new CustomStrikeThroughExtension()); - $environment = Theme::dispatch(ThemeEvents::COMMONMARK_ENVIRONMENT_CONFIGURE, $environment) ?? $environment; - $converter = new CommonMarkConverter([], $environment); - - $environment->addBlockRenderer(ListItem::class, new CustomListItemRenderer(), 10); - - return $converter->convertToHtml($markdown); - } - /** * Convert all base64 image data to saved images. */ protected function extractBase64ImagesFromHtml(string $htmlText): string { - if (empty($htmlText) || strpos($htmlText, 'data:image') === false) { + if (empty($htmlText) || !str_contains($htmlText, 'data:image')) { return $htmlText; } @@ -107,15 +81,35 @@ class PageContent /** * Convert all inline base64 content to uploaded image files. + * Regex is used to locate the start of data-uri definitions then + * manual looping over content is done to parse the whole data uri. + * Attempting to capture the whole data uri using regex can cause PHP + * PCRE limits to be hit with larger, multi-MB, files. */ - protected function extractBase64ImagesFromMarkdown(string $markdown) + protected function extractBase64ImagesFromMarkdown(string $markdown): string { $matches = []; - preg_match_all('/!\[.*?]\(.*?(data:image\/.*?)[)"\s]/', $markdown, $matches); + $contentLength = strlen($markdown); + $replacements = []; + preg_match_all('/!\[.*?]\(.*?(data:image\/.{1,6};base64,)/', $markdown, $matches, PREG_OFFSET_CAPTURE); + + foreach ($matches[1] as $base64MatchPair) { + [$dataUri, $index] = $base64MatchPair; + + for ($i = strlen($dataUri) + $index; $i < $contentLength; $i++) { + $char = $markdown[$i]; + if ($char === ')' || $char === ' ' || $char === "\n" || $char === '"') { + break; + } + $dataUri .= $char; + } - foreach ($matches[1] as $base64Match) { - $newUrl = $this->base64ImageUriToUploadedImageUrl($base64Match); - $markdown = str_replace($base64Match, $newUrl, $markdown); + $newUrl = $this->base64ImageUriToUploadedImageUrl($dataUri); + $replacements[] = [$dataUri, $newUrl]; + } + + foreach ($replacements as [$dataUri, $newUrl]) { + $markdown = str_replace($dataUri, $newUrl, $markdown); } return $markdown; @@ -184,32 +178,13 @@ class PageContent $childNodes = $body->childNodes; $xPath = new DOMXPath($doc); - // Set ids on top-level nodes + // Map to hold used ID references $idMap = []; - foreach ($childNodes as $index => $childNode) { - [$oldId, $newId] = $this->setUniqueId($childNode, $idMap); - if ($newId && $newId !== $oldId) { - $this->updateLinks($xPath, '#' . $oldId, '#' . $newId); - } - } - - // Set ids on nested header nodes - $nestedHeaders = $xPath->query('//body//*//h1|//body//*//h2|//body//*//h3|//body//*//h4|//body//*//h5|//body//*//h6'); - foreach ($nestedHeaders as $nestedHeader) { - [$oldId, $newId] = $this->setUniqueId($nestedHeader, $idMap); - if ($newId && $newId !== $oldId) { - $this->updateLinks($xPath, '#' . $oldId, '#' . $newId); - } - } + // Map to hold changing ID references + $changeMap = []; - // Ensure no duplicate ids within child items - $idElems = $xPath->query('//body//*//*[@id]'); - foreach ($idElems as $domElem) { - [$oldId, $newId] = $this->setUniqueId($domElem, $idMap); - if ($newId && $newId !== $oldId) { - $this->updateLinks($xPath, '#' . $oldId, '#' . $newId); - } - } + $this->updateIdsRecursively($body, 0, $idMap, $changeMap); + $this->updateLinks($xPath, $changeMap); // Generate inner html as a string $html = ''; @@ -217,35 +192,71 @@ class PageContent $html .= $doc->saveHTML($childNode); } + // Perform required string-level tweaks + $html = str_replace(' ', ' ', $html); + return $html; } /** - * Update the all links to the $old location to instead point to $new. + * For the given DOMNode, traverse its children recursively and update IDs + * where required (Top-level, headers & elements with IDs). + * Will update the provided $changeMap array with changes made, where keys are the old + * ids and the corresponding values are the new ids. */ - protected function updateLinks(DOMXPath $xpath, string $old, string $new) + protected function updateIdsRecursively(DOMNode $element, int $depth, array &$idMap, array &$changeMap): void { - $old = str_replace('"', '', $old); - $matchingLinks = $xpath->query('//body//*//*[@href="' . $old . '"]'); - foreach ($matchingLinks as $domElem) { - $domElem->setAttribute('href', $new); + /* @var DOMNode $child */ + foreach ($element->childNodes as $child) { + if ($child instanceof DOMElement && ($depth === 0 || in_array($child->nodeName, ['h1', 'h2', 'h3', 'h4', 'h5', 'h6']) || $child->getAttribute('id'))) { + [$oldId, $newId] = $this->setUniqueId($child, $idMap); + if ($newId && $newId !== $oldId && !isset($idMap[$oldId])) { + $changeMap[$oldId] = $newId; + } + } + + if ($child->hasChildNodes()) { + $this->updateIdsRecursively($child, $depth + 1, $idMap, $changeMap); + } + } + } + + /** + * Update the all links in the given xpath to apply requires changes within the + * given $changeMap array. + */ + protected function updateLinks(DOMXPath $xpath, array $changeMap): void + { + if (empty($changeMap)) { + return; + } + + $links = $xpath->query('//body//*//*[@href]'); + /** @var DOMElement $domElem */ + foreach ($links as $domElem) { + $href = ltrim($domElem->getAttribute('href'), '#'); + $newHref = $changeMap[$href] ?? null; + if ($newHref) { + $domElem->setAttribute('href', '#' . $newHref); + } } } /** * Set a unique id on the given DOMElement. - * A map for existing ID's should be passed in to check for current existence. + * A map for existing ID's should be passed in to check for current existence, + * and this will be updated with any new IDs set upon elements. * Returns a pair of strings in the format [old_id, new_id]. */ - protected function setUniqueId(\DOMNode $element, array &$idMap): array + protected function setUniqueId(DOMNode $element, array &$idMap): array { - if (!$element instanceof \DOMElement) { + if (!$element instanceof DOMElement) { return ['', '']; } // Stop if there's an existing valid id that has not already been used. $existingId = $element->getAttribute('id'); - if (strpos($existingId, 'bkmrk') === 0 && !isset($idMap[$existingId])) { + if (str_starts_with($existingId, 'bkmrk') && !isset($idMap[$existingId])) { $idMap[$existingId] = true; return [$existingId, $existingId]; @@ -256,7 +267,7 @@ class PageContent // the same content is passed through. $contentId = 'bkmrk-' . mb_substr(strtolower(preg_replace('/\s+/', '-', trim($element->nodeValue))), 0, 20); $newId = urlencode($contentId); - $loopIndex = 0; + $loopIndex = 1; while (isset($idMap[$newId])) { $newId = urlencode($contentId . '-' . $loopIndex); @@ -293,7 +304,9 @@ class PageContent if ($blankIncludes) { $content = $this->blankPageIncludes($content); } else { - $content = $this->parsePageIncludes($content); + for ($includeDepth = 0; $includeDepth < 3; $includeDepth++) { + $content = $this->parsePageIncludes($content); + } } return $content; @@ -321,7 +334,7 @@ class PageContent */ protected function headerNodesToLevelList(DOMNodeList $nodeList): array { - $tree = collect($nodeList)->map(function ($header) { + $tree = collect($nodeList)->map(function (DOMElement $header) { $text = trim(str_replace("\xc2\xa0", '', $header->nodeValue)); $text = mb_substr($text, 0, 100); @@ -372,23 +385,30 @@ class PageContent continue; } - // Find page and skip this if page not found + // Find page to use, and default replacement to empty string for non-matches. /** @var ?Page $matchedPage */ $matchedPage = Page::visible()->find($pageId); - if ($matchedPage === null) { - $html = str_replace($fullMatch, '', $html); - continue; + $replacement = ''; + + if ($matchedPage && count($splitInclude) === 1) { + // If we only have page id, just insert all page html and continue. + $replacement = $matchedPage->html; + } elseif ($matchedPage && count($splitInclude) > 1) { + // Otherwise, if our include tag defines a section, load that specific content + $innerContent = $this->fetchSectionOfPage($matchedPage, $splitInclude[1]); + $replacement = trim($innerContent); } - // If we only have page id, just insert all page html and continue. - if (count($splitInclude) === 1) { - $html = str_replace($fullMatch, $matchedPage->html, $html); - continue; - } + $themeReplacement = Theme::dispatch( + ThemeEvents::PAGE_INCLUDE_PARSE, + $includeId, + $replacement, + clone $this->page, + $matchedPage ? (clone $matchedPage) : null, + ); - // Create and load HTML into a document - $innerContent = $this->fetchSectionOfPage($matchedPage, $splitInclude[1]); - $html = str_replace($fullMatch, trim($innerContent), $html); + // Perform the content replacement + $html = str_replace($fullMatch, $themeReplacement ?? $replacement, $html); } return $html; @@ -431,8 +451,8 @@ class PageContent { libxml_use_internal_errors(true); $doc = new DOMDocument(); - $html = '' . $html . ''; - $doc->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8')); + $html = '' . $html . ''; + $doc->loadHTML($html); return $doc; }