markdown.php

   1 <?php
   2
   3 /**
   4  * Plugin Markdown.
   5  *
   6  * Shaare's descriptions are parsed with Markdown.
   7  */
   8
   9 /*
  10  * If this tag is used on a shaare, the description won't be processed by Parsedown.
  11  */
  12 define('NO_MD_TAG', 'nomarkdown');
  13
  14 /**
  15  * Parse linklist descriptions.
  16  *
  17  * @param array         $data linklist data.
  18  * @param ConfigManager $conf instance.
  19  *
  20  * @return mixed linklist data parsed in markdown (and converted to HTML).
  21  */
  22 function hook_markdown_render_linklist($data, $conf)
  23 {
  24     foreach ($data['links'] as &$value) {
  25         if (!empty($value['tags']) && noMarkdownTag($value['tags'])) {
  26             $value = stripNoMarkdownTag($value);
  27             continue;
  28         }
  29         $value['description'] = process_markdown(
  30             $value['description'],
  31             $conf->get('security.markdown_escape', true),
  32             $conf->get('security.allowed_protocols')
  33         );
  34     }
  35     return $data;
  36 }
  37
  38 /**
  39  * Parse feed linklist descriptions.
  40  *
  41  * @param array $data linklist data.
  42  * @param ConfigManager $conf instance.
  43  *
  44  * @return mixed linklist data parsed in markdown (and converted to HTML).
  45  */
  46 function hook_markdown_render_feed($data, $conf)
  47 {
  48     foreach ($data['links'] as &$value) {
  49         if (!empty($value['tags']) && noMarkdownTag($value['tags'])) {
  50             $value = stripNoMarkdownTag($value);
  51             continue;
  52         }
  53         $value['description'] = process_markdown(
  54             $value['description'],
  55             $conf->get('security.markdown_escape', true),
  56             $conf->get('security.allowed_protocols')
  57         );
  58     }
  59
  60     return $data;
  61 }
  62
  63 /**
  64  * Parse daily descriptions.
  65  *
  66  * @param array         $data daily data.
  67  * @param ConfigManager $conf instance.
  68  *
  69  * @return mixed daily data parsed in markdown (and converted to HTML).
  70  */
  71 function hook_markdown_render_daily($data, $conf)
  72 {
  73     //var_dump($data);die;
  74     // Manipulate columns data
  75     foreach ($data['linksToDisplay'] as &$value) {
  76         if (!empty($value['tags']) && noMarkdownTag($value['tags'])) {
  77             $value = stripNoMarkdownTag($value);
  78             continue;
  79         }
  80         $value['formatedDescription'] = process_markdown(
  81             $value['formatedDescription'],
  82             $conf->get('security.markdown_escape', true),
  83             $conf->get('security.allowed_protocols')
  84         );
  85     }
  86
  87     return $data;
  88 }
  89
  90 /**
  91  * Check if noMarkdown is set in tags.
  92  *
  93  * @param string $tags tag list
  94  *
  95  * @return bool true if markdown should be disabled on this link.
  96  */
  97 function noMarkdownTag($tags)
  98 {
  99     return preg_match('/(^|\s)'. NO_MD_TAG .'(\s|$)/', $tags);
 100 }
 101
 102 /**
 103  * Remove the no-markdown meta tag so it won't be displayed.
 104  *
 105  * @param array $link Link data.
 106  *
 107  * @return array Updated link without no markdown tag.
 108  */
 109 function stripNoMarkdownTag($link)
 110 {
 111     if (! empty($link['taglist'])) {
 112         $offset = array_search(NO_MD_TAG, $link['taglist']);
 113         if ($offset !== false) {
 114             unset($link['taglist'][$offset]);
 115         }
 116     }
 117
 118     if (!empty($link['tags'])) {
 119         str_replace(NO_MD_TAG, '', $link['tags']);
 120     }
 121
 122     return $link;
 123 }
 124
 125 /**
 126  * When link list is displayed, include markdown CSS.
 127  *
 128  * @param array $data includes data.
 129  *
 130  * @return mixed - includes data with markdown CSS file added.
 131  */
 132 function hook_markdown_render_includes($data)
 133 {
 134     if ($data['_PAGE_'] == Router::$PAGE_LINKLIST
 135         || $data['_PAGE_'] == Router::$PAGE_DAILY
 136         || $data['_PAGE_'] == Router::$PAGE_EDITLINK
 137     ) {
 138
 139         $data['css_files'][] = PluginManager::$PLUGINS_PATH . '/markdown/markdown.css';
 140     }
 141
 142     return $data;
 143 }
 144
 145 /**
 146  * Hook render_editlink.
 147  * Adds an help link to markdown syntax.
 148  *
 149  * @param array $data data passed to plugin
 150  *
 151  * @return array altered $data.
 152  */
 153 function hook_markdown_render_editlink($data)
 154 {
 155     // Load help HTML into a string
 156     $txt = file_get_contents(PluginManager::$PLUGINS_PATH .'/markdown/help.html');
 157     $translations = [
 158         t('Description will be rendered with'),
 159         t('Markdown syntax documentation'),
 160         t('Markdown syntax'),
 161     ];
 162     $data['edit_link_plugin'][] = vsprintf($txt, $translations);
 163     // Add no markdown 'meta-tag' in tag list if it was never used, for autocompletion.
 164     if (! in_array(NO_MD_TAG, $data['tags'])) {
 165         $data['tags'][NO_MD_TAG] = 0;
 166     }
 167
 168     return $data;
 169 }
 170
 171
 172 /**
 173  * Remove HTML links auto generated by Shaarli core system.
 174  * Keeps HREF attributes.
 175  *
 176  * @param string $description input description text.
 177  *
 178  * @return string $description without HTML links.
 179  */
 180 function reverse_text2clickable($description)
 181 {
 182     $descriptionLines = explode(PHP_EOL, $description);
 183     $descriptionOut = '';
 184     $codeBlockOn = false;
 185     $lineCount = 0;
 186
 187     foreach ($descriptionLines as $descriptionLine) {
 188         // Detect line of code: starting with 4 spaces,
 189         // except lists which can start with +/*/- or `2.` after spaces.
 190         $codeLineOn = preg_match('/^    +(?=[^\+\*\-])(?=(?!\d\.).)/', $descriptionLine) > 0;
 191         // Detect and toggle block of code
 192         if (!$codeBlockOn) {
 193             $codeBlockOn = preg_match('/^```/', $descriptionLine) > 0;
 194         }
 195         elseif (preg_match('/^```/', $descriptionLine) > 0) {
 196             $codeBlockOn = false;
 197         }
 198
 199         $hashtagTitle = ' title="Hashtag [^"]+"';
 200         // Reverse `inline code` hashtags.
 201         $descriptionLine = preg_replace(
 202             '!(`[^`\n]*)<a href="[^ ]*"'. $hashtagTitle .'>([^<]+)</a>([^`\n]*`)!m',
 203             '$1$2$3',
 204             $descriptionLine
 205         );
 206
 207         // Reverse all links in code blocks, only non hashtag elsewhere.
 208         $hashtagFilter = (!$codeBlockOn && !$codeLineOn) ? '(?!'. $hashtagTitle .')': '(?:'. $hashtagTitle .')?';
 209         $descriptionLine = preg_replace(
 210             '#<a href="[^ ]*"'. $hashtagFilter .'>([^<]+)</a>#m',
 211             '$1',
 212             $descriptionLine
 213         );
 214
 215         $descriptionOut .= $descriptionLine;
 216         if ($lineCount++ < count($descriptionLines) - 1) {
 217             $descriptionOut .= PHP_EOL;
 218         }
 219     }
 220     return $descriptionOut;
 221 }
 222
 223 /**
 224  * Remove <br> tag to let markdown handle it.
 225  *
 226  * @param string $description input description text.
 227  *
 228  * @return string $description without <br> tags.
 229  */
 230 function reverse_nl2br($description)
 231 {
 232     return preg_replace('!<br */?>!im', '', $description);
 233 }
 234
 235 /**
 236  * Remove HTML spaces '&nbsp;' auto generated by Shaarli core system.
 237  *
 238  * @param string $description input description text.
 239  *
 240  * @return string $description without HTML links.
 241  */
 242 function reverse_space2nbsp($description)
 243 {
 244     return preg_replace('/(^| )&nbsp;/m', '$1 ', $description);
 245 }
 246
 247 /**
 248  * Replace not whitelisted protocols with http:// in given description.
 249  *
 250  * @param string $description      input description text.
 251  * @param array  $allowedProtocols list of allowed protocols.
 252  *
 253  * @return string $description without malicious link.
 254  */
 255 function filter_protocols($description, $allowedProtocols)
 256 {
 257     return preg_replace_callback(
 258         '#]\((.*?)\)#is',
 259         function ($match) use ($allowedProtocols) {
 260             return ']('. whitelist_protocols($match[1], $allowedProtocols) .')';
 261         },
 262         $description
 263     );
 264 }
 265
 266 /**
 267  * Remove dangerous HTML tags (tags, iframe, etc.).
 268  * Doesn't affect <code> content (already escaped by Parsedown).
 269  *
 270  * @param string $description input description text.
 271  *
 272  * @return string given string escaped.
 273  */
 274 function sanitize_html($description)
 275 {
 276     $escapeTags = array(
 277         'script',
 278         'style',
 279         'link',
 280         'iframe',
 281         'frameset',
 282         'frame',
 283     );
 284     foreach ($escapeTags as $tag) {
 285         $description = preg_replace_callback(
 286             '#<\s*'. $tag .'[^>]*>(.*</\s*'. $tag .'[^>]*>)?#is',
 287             function ($match) { return escape($match[0]); },
 288             $description);
 289     }
 290     $description = preg_replace(
 291         '#(<[^>]+)on[a-z]*="?[^ "]*"?#is',
 292         '$1',
 293         $description);
 294     return $description;
 295 }
 296
 297 /**
 298  * Render shaare contents through Markdown parser.
 299  *   1. Remove HTML generated by Shaarli core.
 300  *   2. Reverse the escape function.
 301  *   3. Generate markdown descriptions.
 302  *   4. Sanitize sensible HTML tags for security.
 303  *   5. Wrap description in 'markdown' CSS class.
 304  *
 305  * @param string $description input description text.
 306  * @param bool   $escape      escape HTML entities
 307  *
 308  * @return string HTML processed $description.
 309  */
 310 function process_markdown($description, $escape = true, $allowedProtocols = [])
 311 {
 312     $parsedown = new Parsedown();
 313
 314     $processedDescription = $description;
 315     $processedDescription = reverse_nl2br($processedDescription);
 316     $processedDescription = reverse_space2nbsp($processedDescription);
 317     $processedDescription = reverse_text2clickable($processedDescription);
 318     $processedDescription = filter_protocols($processedDescription, $allowedProtocols);
 319     $processedDescription = unescape($processedDescription);
 320     $processedDescription = $parsedown
 321         ->setMarkupEscaped($escape)
 322         ->setBreaksEnabled(true)
 323         ->text($processedDescription);
 324     $processedDescription = sanitize_html($processedDescription);
 325
 326     if(!empty($processedDescription)){
 327         $processedDescription = '<div class="markdown">'. $processedDescription . '</div>';
 328     }
 329
 330     return $processedDescription;
 331 }
 332
 333 /**
 334  * This function is never called, but contains translation calls for GNU gettext extraction.
 335  */
 336 function markdown_dummy_translation()
 337 {
 338     // meta
 339     t('Render shaare description with Markdown syntax.<br><strong>Warning</strong>:
 340 If your shaared descriptions contained HTML tags before enabling the markdown plugin,
 341 enabling it might break your page.
 342 See the <a href="https://github.com/shaarli/Shaarli/tree/master/plugins/markdown#html-rendering">README</a>.');
 343 }