Merge branch 'MDL-67547' of https://github.com/paulholden/moodle

dataformat/html/classes/writer.php

@@ -92,13 +92,24 @@ public function start_sheet($columns) {
         echo \html_writer::end_tag('tr');
     }
 
+    /**
+     * Method to define whether the dataformat supports export of HTML
+     *
+     * @return bool
+     */
+    public function supports_html(): bool {
+        return true;
+    }
+
     /**
      * Write a single record
      *
      * @param array $record
      * @param int $rownum
      */
     public function write_record($record, $rownum) {
+        $record = $this->format_record($record);
+
         echo \html_writer::start_tag('tr');
         foreach ($record as $cell) {
             echo \html_writer::tag('td', $cell);

dataformat/json/classes/writer.php

@@ -80,7 +80,7 @@ public function write_record($record, $rownum) {
             echo ",";
         }
 
-        echo json_encode($record, JSON_UNESCAPED_UNICODE | JSON_PRETTY_PRINT);
+        echo json_encode($this->format_record($record), JSON_UNESCAPED_UNICODE | JSON_PRETTY_PRINT);
 
         $this->sheetdatadded = true;
     }

dataformat/pdf/classes/writer.php

@@ -93,16 +93,52 @@ public function start_sheet($columns) {
         $this->print_heading();
     }
 
+    /**
+     * Method to define whether the dataformat supports export of HTML
+     *
+     * @return bool
+     */
+    public function supports_html(): bool {
+        return true;
+    }
+
+    /**
+     * When exporting images, we need to return their Base64 encoded content. Otherwise TCPDF will create a HTTP
+     * request for them, which will lead to the login page (i.e. not the image it expects) and throw an exception
+     *
+     * Note: ideally we would copy the file to a temp location and return it's path, but a bug in TCPDF currently
+     * prevents that
+     *
+     * @param \stored_file $file
+     * @return string|null
+     */
+    protected function export_html_image_source(\stored_file $file): ?string {
+        // Set upper dimensions for embedded images.
+        $resizedimage = $file->resize_image(400, 300);
+
+        return '@' . base64_encode($resizedimage);
+    }
+
+    /**
+     * Write a single record
+     *
+     * @param array $record
+     * @param int $rownum
+     */
     public function write_record($record, $rownum) {
         $rowheight = 0;
 
-        // If $record is an object convert it to an array.
-        if (is_object($record)) {
-            $record = (array)$record;
-        }
-
+        $record = $this->format_record($record);
         foreach ($record as $cell) {
-            $rowheight = max($rowheight, $this->pdf->getStringHeight($this->colwidth, $cell, false, true, '', 1));
+            // We need to calculate the row height (accounting for any content). Unfortunately TCPDF doesn't provide an easy
+            // method to do that, so we create a second PDF inside a transaction, add cell content and use the largest cell by
+            // height. Solution similar to that at https://stackoverflow.com/a/1943096.
+            $pdf2 = clone $this->pdf;
+            $pdf2->startTransaction();
+            $pdf2->AddPage('L');
+            $pdf2->writeHTMLCell($this->colwidth, 0, '', '', $cell, 1, 1, false, true, 'L');
+            $rowheight = max($rowheight, $pdf2->getY() - $pdf2->getMargins()['top']);
+            $pdf2->rollbackTransaction();
         }
 
         $margins = $this->pdf->getMargins();
@@ -123,7 +159,7 @@ public function write_record($record, $rownum) {
             // Determine whether we're at the last element of the record.
             $nextposition = ($lastkey === $key) ? 1 : 0;
             // Write the element.
-            $this->pdf->Multicell($this->colwidth, $rowheight, $cell, 1, 'L', false, $nextposition);
+            $this->pdf->writeHTMLCell($this->colwidth, $rowheight, '', '', $cell, 1, $nextposition, false, true, 'L');
         }
     }
 

dataformat/pdf/tests/writer_test.php

@@ -0,0 +1,74 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle is free software: you can redistribute it and/or modify
+// it under the terms of the GNU General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU General Public License for more details.
+//
+// You should have received a copy of the GNU General Public License
+// along with Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Tests for the dataformat_pdf writer
+ *
+ * @package    dataformat_pdf
+ * @copyright  2020 Paul Holden <[email protected]>
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+namespace dataformat_pdf;
+
+use core\dataformat;
+use context_system;
+use html_writer;
+use moodle_url;
+
+/**
+ * Writer tests
+ *
+ * @package    dataformat_pdf
+ * @copyright  2020 Paul Holden <[email protected]>
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+class writer_testcase extends \advanced_testcase {
+
+    /**
+     * Test writing data whose content contains an image with pluginfile.php source
+     */
+    public function test_write_data_with_pluginfile_image(): void {
+        global $CFG;
+
+        $this->resetAfterTest(true);
+
+        $imagefixture = "{$CFG->dirroot}/lib/filestorage/tests/fixtures/testimage.jpg";
+        $image = get_file_storage()->create_file_from_pathname([
+            'contextid' => context_system::instance()->id,
+            'component' => 'dataformat_pdf',
+            'filearea'  => 'test',
+            'itemid'    => 0,
+            'filepath'  => '/',
+            'filename'  => basename($imagefixture),
+
+        ], $imagefixture);
+
+        $imageurl = moodle_url::make_pluginfile_url($image->get_contextid(), $image->get_component(), $image->get_filearea(),
+            $image->get_itemid(), $image->get_filepath(), $image->get_filename());
+
+        // Insert out test image into the data so it is exported.
+        $columns = ['animal', 'image'];
+        $row = ['cat', html_writer::img($imageurl->out(), 'My image')];
+
+        // Export to file. Assert that the exported file exists.
+        $exportfile = dataformat::write_data('My export', 'pdf', $columns, [$row]);
+        $this->assertFileExists($exportfile);
+
+        // The exported file should be a reasonable size (~275kb).
+        $this->assertGreaterThan(270000, filesize($exportfile));
+    }
+}

dataformat/upgrade.txt

@@ -9,6 +9,13 @@ information provided here is intended especially for developers.
 * Calls to the following dataformat plugin methods have been removed:
   - write_header()
   - write_footer()
+* The following methods have been added to the base class to allow instances to define support for exporting
+  HTML content, with additional support for defining how images should be embedded:
+  - supports_html()
+  - export_html_image_source()
+* Dataformat writers should also call the following method to ensure data is properly formatted before being
+  written, which takes into account prior methods defining support for HTML:
+  - format_record()
 
 === 3.4 ===
 * In order to allow multiple sheets in an exported file the functions write_header() and write_footer() have

lib/classes/dataformat/base.php

@@ -145,6 +145,82 @@ public function start_sheet($columns) {
         // Override me if needed.
     }
 
+    /**
+     * Method to define whether the dataformat supports export of HTML
+     *
+     * @return bool
+     */
+    public function supports_html(): bool {
+        return false;
+    }
+
+    /**
+     * Apply formatting to the cells of a given record
+     *
+     * @param array|\stdClass $record
+     * @return array
+     */
+    protected function format_record($record): array {
+        $record = (array)$record;
+
+        // If the dataformat supports export of HTML, we need to allow them to manage embedded images.
+        if ($this->supports_html()) {
+            $record = array_map([$this, 'replace_pluginfile_images'], $record);
+        }
+
+        return $record;
+    }
+
+    /**
+     * Given a stored_file, return a suitable source attribute for an img element in the export (or null to use the original)
+     *
+     * @param \stored_file $file
+     * @return string|null
+     */
+    protected function export_html_image_source(\stored_file $file): ?string {
+        return null;
+    }
+
+    /**
+     * We need to locate all img tags within a given cell that match pluginfile URL's. Partly so the exported file will show
+     * the image without requiring the user is logged in; and also to prevent some of the dataformats requesting the file
+     * themselves, which is likely to fail due to them not having an active session
+     *
+     * @param string|null $content
+     * @return string
+     */
+    protected function replace_pluginfile_images(?string $content): string {
+        $content = (string)$content;
+
+        // Examine content to see if it contains any HTML image tags.
+        return preg_replace_callback('/(?<pre><img[^>]+src=")(?<source>[^"]*)(?<post>".*>)/i', function(array $matches) {
+            $source = $matches['source'];
+
+            // Now check if the image source looks like a pluginfile URL.
+            if (preg_match('/pluginfile.php\/(?<context>\d+)\/(?<component>[^\/]+)\/(?<filearea>[^\/]+)\/(?:(?<itemid>\d+)\/)?' .
+                    '(?<path>.*)/u', $source, $args)) {
+
+                $context = $args['context'];
+                $component = clean_param($args['component'], PARAM_COMPONENT);
+                $filearea = clean_param($args['filearea'], PARAM_AREA);
+                $itemid = $args['itemid'] ?: 0;
+                $path = clean_param(urldecode($args['path']), PARAM_PATH);
+
+                // Try and get the matching file from storage, allow the dataformat to define the replacement source.
+                $fullpath = "/{$context}/{$component}/{$filearea}/{$itemid}/{$path}";
+                if ($file = get_file_storage()->get_file_by_hash(sha1($fullpath))) {
+                    $exportsource = $this->export_html_image_source($file);
+
+                    if ($exportsource) {
+                        $source = $exportsource;
+                    }
+                }
+            }
+
+            return $matches['pre'] . $source . $matches['post'];
+        }, $content);
+    }
+
     /**
      * Write a single record
      *

lib/classes/dataformat/spout_base.php

@@ -116,11 +116,11 @@ public function start_sheet($columns) {
     /**
      * Write a single record
      *
-     * @param object $record
+     * @param array $record
      * @param int $rownum
      */
     public function write_record($record, $rownum) {
-        $row = \Box\Spout\Writer\Common\Creator\WriterEntityFactory::createRowFromArray((array)$record);
+        $row = \Box\Spout\Writer\Common\Creator\WriterEntityFactory::createRowFromArray($this->format_record($record));
         $this->writer->addRow($row);
     }
 

lib/tablelib.php

@@ -2103,7 +2103,7 @@ function format_text($text, $format=FORMAT_MOODLE, $options=NULL, $courseid=NULL
  */
 class table_dataformat_export_format extends table_default_export_format_parent {
 
-    /** @var $dataformat */
+    /** @var \core\dataformat\base $dataformat */
     protected $dataformat;
 
     /** @var $rownum */
@@ -2138,6 +2138,15 @@ public function __construct(&$table, $dataformat) {
         \core\session\manager::write_close();
     }
 
+    /**
+     * Whether the current dataformat supports export of HTML
+     *
+     * @return bool
+     */
+    public function supports_html(): bool {
+        return $this->dataformat->supports_html();
+    }
+
     /**
      * Start document
      *

mod/forum/export.php

@@ -120,7 +120,7 @@
                 'messageformat', 'messagetrust', 'attachment', 'totalscore', 'mailnow', 'deleted', 'privatereplyto',
                 'privatereplytofullname', 'wordcount', 'charcount'];
 
-    $canviewfullname = has_capability('moodle/site:viewfullnames', $forum->get_context());
+    $canviewfullname = has_capability('moodle/site:viewfullnames', $context);
 
     $datamapper = $legacydatamapperfactory->get_post_data_mapper();
     $exportdata = new ArrayObject($datamapper->to_legacy_objects($posts));
@@ -132,7 +132,7 @@
         $dataformat,
         $fields,
         $iterator,
-        function($exportdata) use ($fields, $striphtml, $humandates, $canviewfullname) {
+        function($exportdata) use ($fields, $striphtml, $humandates, $canviewfullname, $context) {
             $data = new stdClass();
 
             foreach ($fields as $field) {
@@ -149,18 +149,18 @@ function($exportdata) use ($fields, $striphtml, $humandates, $canviewfullname) {
                     $data->privatereplytofullname = fullname($user, $canviewfullname);
                 }
 
+                if ($field == 'message') {
+                    $data->message = file_rewrite_pluginfile_urls($data->message, 'pluginfile.php', $context->id, 'mod_forum',
+                        'post', $data->id);
+                }
+
                 // Convert any boolean fields to their integer equivalent for output.
                 if (is_bool($data->$field)) {
                     $data->$field = (int) $data->$field;
                 }
             }
 
             if ($striphtml) {
-                // The following call to html_to_text uses the option that strips out
-                // all URLs, but format_text complains if it finds @@PLUGINFILE@@ tokens.
-                // So, we need to replace @@PLUGINFILE@@ with a real URL, but it doesn't
-                // matter what. We use http://example.com/.
-                $data->message = str_replace('@@PLUGINFILE@@/', 'http://example.com/', $data->message);
                 $data->message = html_to_text(format_text($data->message, $data->messageformat), 0, false);
                 $data->messageformat = FORMAT_PLAIN;
             }