X-Git-Url: http://source.bookstackapp.com/bookstack/blobdiff_plain/c7fea8fe08e8d26840da7a0d353c05f7a84d3ff1..refs/pull/3333/head:/app/Entities/Tools/PageContent.php

diff --git a/app/Entities/Tools/PageContent.php b/app/Entities/Tools/PageContent.php
index c5b17ddef..b1c750adb 100644
--- a/app/Entities/Tools/PageContent.php
+++ b/app/Entities/Tools/PageContent.php
@@ -12,6 +12,8 @@ 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;
@@ -107,15 +109,35 @@ 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)
     {
         $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;
+            }
+
+            $newUrl = $this->base64ImageUriToUploadedImageUrl($dataUri);
+            $replacements[] = [$dataUri, $newUrl];
+        }
 
-        foreach ($matches[1] as $base64Match) {
-            $newUrl = $this->base64ImageUriToUploadedImageUrl($base64Match);
-            $markdown = str_replace($base64Match, $newUrl, $markdown);
+        foreach ($replacements as [$dataUri, $newUrl]) {
+            $markdown = str_replace($dataUri, $newUrl, $markdown);
         }
 
         return $markdown;
@@ -135,6 +157,12 @@ class PageContent
             return '';
         }
 
+        // Validate that the content is not over our upload limit
+        $uploadLimitBytes = (config('app.upload_limit') * 1000000);
+        if (strlen($imageInfo['data']) > $uploadLimitBytes) {
+            return '';
+        }
+
         // Save image from data with a random name
         $imageName = 'embedded-image-' . Str::random(8) . '.' . $imageInfo['extension'];
 
@@ -149,15 +177,17 @@ class PageContent
 
     /**
      * Parse a base64 image URI into the data and extension.
-     * @return array{extension: array, data: string}
+     *
+     * @return array{extension: string, data: string}
      */
     protected function parseBase64ImageUri(string $uri): array
     {
         [$dataDefinition, $base64ImageData] = explode(',', $uri, 2);
         $extension = strtolower(preg_split('/[\/;]/', $dataDefinition)[1] ?? '');
+
         return [
             'extension' => $extension,
-            'data' => base64_decode($base64ImageData) ?: '',
+            'data'      => base64_decode($base64ImageData) ?: '',
         ];
     }
 
@@ -185,6 +215,15 @@ class PageContent
             }
         }
 
+        // 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);
+            }
+        }
+
         // Ensure no duplicate ids within child items
         $idElems = $xPath->query('//body//*//*[@id]');
         foreach ($idElems as $domElem) {
@@ -200,6 +239,9 @@ class PageContent
             $html .= $doc->saveHTML($childNode);
         }
 
+        // Perform required string-level tweaks
+        $html = str_replace(' ', '&nbsp;', $html);
+
         return $html;
     }
 
@@ -220,9 +262,9 @@ class PageContent
      * A map for existing ID's should be passed in to check for current existence.
      * 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 (get_class($element) !== 'DOMElement') {
+        if (!$element instanceof DOMElement) {
             return ['', ''];
         }
 
@@ -234,7 +276,7 @@ class PageContent
             return [$existingId, $existingId];
         }
 
-        // Create an unique id for the element
+        // Create a unique id for the element
         // Uses the content as a basis to ensure output is the same every time
         // the same content is passed through.
         $contentId = 'bkmrk-' . mb_substr(strtolower(preg_replace('/\s+/', '-', trim($element->nodeValue))), 0, 20);
@@ -304,7 +346,7 @@ 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);
 
@@ -382,7 +424,7 @@ class PageContent
      */
     protected function fetchSectionOfPage(Page $page, string $sectionId): string
     {
-        $topLevelTags = ['table', 'ul', 'ol'];
+        $topLevelTags = ['table', 'ul', 'ol', 'pre'];
         $doc = $this->loadDocumentFromHtml($page->html);
 
         // Search included content for the id given and blank out if not exists.