remove autoload section in composer.json

[github/wallabag/wallabag.git] / inc / 3rdparty / htmlpurifier / HTMLPurifier / Generator.php

<?php\r
\r
/**\r
 * Generates HTML from tokens.\r
 * @todo Refactor interface so that configuration/context is determined\r
 *       upon instantiation, no need for messy generateFromTokens() calls\r
 * @todo Make some of the more internal functions protected, and have\r
 *       unit tests work around that\r
 */\r
class HTMLPurifier_Generator\r
{\r
\r
    /**\r
     * Whether or not generator should produce XML output.\r
     * @type bool\r
     */\r
    private $_xhtml = true;\r
\r
    /**\r
     * :HACK: Whether or not generator should comment the insides of <script> tags.\r
     * @type bool\r
     */\r
    private $_scriptFix = false;\r
\r
    /**\r
     * Cache of HTMLDefinition during HTML output to determine whether or\r
     * not attributes should be minimized.\r
     * @type HTMLPurifier_HTMLDefinition\r
     */\r
    private $_def;\r
\r
    /**\r
     * Cache of %Output.SortAttr.\r
     * @type bool\r
     */\r
    private $_sortAttr;\r
\r
    /**\r
     * Cache of %Output.FlashCompat.\r
     * @type bool\r
     */\r
    private $_flashCompat;\r
\r
    /**\r
     * Cache of %Output.FixInnerHTML.\r
     * @type bool\r
     */\r
    private $_innerHTMLFix;\r
\r
    /**\r
     * Stack for keeping track of object information when outputting IE\r
     * compatibility code.\r
     * @type array\r
     */\r
    private $_flashStack = array();\r
\r
    /**\r
     * Configuration for the generator\r
     * @type HTMLPurifier_Config\r
     */\r
    protected $config;\r
\r
    /**\r
     * @param HTMLPurifier_Config $config\r
     * @param HTMLPurifier_Context $context\r
     */\r
    public function __construct($config, $context)\r
    {\r
        $this->config = $config;\r
        $this->_scriptFix = $config->get('Output.CommentScriptContents');\r
        $this->_innerHTMLFix = $config->get('Output.FixInnerHTML');\r
        $this->_sortAttr = $config->get('Output.SortAttr');\r
        $this->_flashCompat = $config->get('Output.FlashCompat');\r
        $this->_def = $config->getHTMLDefinition();\r
        $this->_xhtml = $this->_def->doctype->xml;\r
    }\r
\r
    /**\r
     * Generates HTML from an array of tokens.\r
     * @param HTMLPurifier_Token[] $tokens Array of HTMLPurifier_Token\r
     * @return string Generated HTML\r
     */\r
    public function generateFromTokens($tokens)\r
    {\r
        if (!$tokens) {\r
            return '';\r
        }\r
\r
        // Basic algorithm\r
        $html = '';\r
        for ($i = 0, $size = count($tokens); $i < $size; $i++) {\r
            if ($this->_scriptFix && $tokens[$i]->name === 'script'\r
                && $i + 2 < $size && $tokens[$i+2] instanceof HTMLPurifier_Token_End) {\r
                // script special case\r
                // the contents of the script block must be ONE token\r
                // for this to work.\r
                $html .= $this->generateFromToken($tokens[$i++]);\r
                $html .= $this->generateScriptFromToken($tokens[$i++]);\r
            }\r
            $html .= $this->generateFromToken($tokens[$i]);\r
        }\r
\r
        // Tidy cleanup\r
        if (extension_loaded('tidy') && $this->config->get('Output.TidyFormat')) {\r
            $tidy = new Tidy;\r
            $tidy->parseString(\r
                $html,\r
                array(\r
                   'indent'=> true,\r
                   'output-xhtml' => $this->_xhtml,\r
                   'show-body-only' => true,\r
                   'indent-spaces' => 2,\r
                   'wrap' => 68,\r
                ),\r
                'utf8'\r
            );\r
            $tidy->cleanRepair();\r
            $html = (string) $tidy; // explicit cast necessary\r
        }\r
\r
        // Normalize newlines to system defined value\r
        if ($this->config->get('Core.NormalizeNewlines')) {\r
            $nl = $this->config->get('Output.Newline');\r
            if ($nl === null) {\r
                $nl = PHP_EOL;\r
            }\r
            if ($nl !== "\n") {\r
                $html = str_replace("\n", $nl, $html);\r
            }\r
        }\r
        return $html;\r
    }\r
\r
    /**\r
     * Generates HTML from a single token.\r
     * @param HTMLPurifier_Token $token HTMLPurifier_Token object.\r
     * @return string Generated HTML\r
     */\r
    public function generateFromToken($token)\r
    {\r
        if (!$token instanceof HTMLPurifier_Token) {\r
            trigger_error('Cannot generate HTML from non-HTMLPurifier_Token object', E_USER_WARNING);\r
            return '';\r
\r
        } elseif ($token instanceof HTMLPurifier_Token_Start) {\r
            $attr = $this->generateAttributes($token->attr, $token->name);\r
            if ($this->_flashCompat) {\r
                if ($token->name == "object") {\r
                    $flash = new stdclass();\r
                    $flash->attr = $token->attr;\r
                    $flash->param = array();\r
                    $this->_flashStack[] = $flash;\r
                }\r
            }\r
            return '<' . $token->name . ($attr ? ' ' : '') . $attr . '>';\r
\r
        } elseif ($token instanceof HTMLPurifier_Token_End) {\r
            $_extra = '';\r
            if ($this->_flashCompat) {\r
                if ($token->name == "object" && !empty($this->_flashStack)) {\r
                    // doesn't do anything for now\r
                }\r
            }\r
            return $_extra . '</' . $token->name . '>';\r
\r
        } elseif ($token instanceof HTMLPurifier_Token_Empty) {\r
            if ($this->_flashCompat && $token->name == "param" && !empty($this->_flashStack)) {\r
                $this->_flashStack[count($this->_flashStack)-1]->param[$token->attr['name']] = $token->attr['value'];\r
            }\r
            $attr = $this->generateAttributes($token->attr, $token->name);\r
             return '<' . $token->name . ($attr ? ' ' : '') . $attr .\r
                ( $this->_xhtml ? ' /': '' ) // <br /> v. <br>\r
                . '>';\r
\r
        } elseif ($token instanceof HTMLPurifier_Token_Text) {\r
            return $this->escape($token->data, ENT_NOQUOTES);\r
\r
        } elseif ($token instanceof HTMLPurifier_Token_Comment) {\r
            return '<!--' . $token->data . '-->';\r
        } else {\r
            return '';\r
\r
        }\r
    }\r
\r
    /**\r
     * Special case processor for the contents of script tags\r
     * @param HTMLPurifier_Token $token HTMLPurifier_Token object.\r
     * @return string\r
     * @warning This runs into problems if there's already a literal\r
     *          --> somewhere inside the script contents.\r
     */\r
    public function generateScriptFromToken($token)\r
    {\r
        if (!$token instanceof HTMLPurifier_Token_Text) {\r
            return $this->generateFromToken($token);\r
        }\r
        // Thanks <http://lachy.id.au/log/2005/05/script-comments>\r
        $data = preg_replace('#//\s*$#', '', $token->data);\r
        return '<!--//--><![CDATA[//><!--' . "\n" . trim($data) . "\n" . '//--><!]]>';\r
    }\r
\r
    /**\r
     * Generates attribute declarations from attribute array.\r
     * @note This does not include the leading or trailing space.\r
     * @param array $assoc_array_of_attributes Attribute array\r
     * @param string $element Name of element attributes are for, used to check\r
     *        attribute minimization.\r
     * @return string Generated HTML fragment for insertion.\r
     */\r
    public function generateAttributes($assoc_array_of_attributes, $element = '')\r
    {\r
        $html = '';\r
        if ($this->_sortAttr) {\r
            ksort($assoc_array_of_attributes);\r
        }\r
        foreach ($assoc_array_of_attributes as $key => $value) {\r
            if (!$this->_xhtml) {\r
                // Remove namespaced attributes\r
                if (strpos($key, ':') !== false) {\r
                    continue;\r
                }\r
                // Check if we should minimize the attribute: val="val" -> val\r
                if ($element && !empty($this->_def->info[$element]->attr[$key]->minimized)) {\r
                    $html .= $key . ' ';\r
                    continue;\r
                }\r
            }\r
            // Workaround for Internet Explorer innerHTML bug.\r
            // Essentially, Internet Explorer, when calculating\r
            // innerHTML, omits quotes if there are no instances of\r
            // angled brackets, quotes or spaces.  However, when parsing\r
            // HTML (for example, when you assign to innerHTML), it\r
            // treats backticks as quotes.  Thus,\r
            //      <img alt="``" />\r
            // becomes\r
            //      <img alt=`` />\r
            // becomes\r
            //      <img alt='' />\r
            // Fortunately, all we need to do is trigger an appropriate\r
            // quoting style, which we do by adding an extra space.\r
            // This also is consistent with the W3C spec, which states\r
            // that user agents may ignore leading or trailing\r
            // whitespace (in fact, most don't, at least for attributes\r
            // like alt, but an extra space at the end is barely\r
            // noticeable).  Still, we have a configuration knob for\r
            // this, since this transformation is not necesary if you\r
            // don't process user input with innerHTML or you don't plan\r
            // on supporting Internet Explorer.\r
            if ($this->_innerHTMLFix) {\r
                if (strpos($value, '`') !== false) {\r
                    // check if correct quoting style would not already be\r
                    // triggered\r
                    if (strcspn($value, '"\' <>') === strlen($value)) {\r
                        // protect!\r
                        $value .= ' ';\r
                    }\r
                }\r
            }\r
            $html .= $key.'="'.$this->escape($value).'" ';\r
        }\r
        return rtrim($html);\r
    }\r
\r
    /**\r
     * Escapes raw text data.\r
     * @todo This really ought to be protected, but until we have a facility\r
     *       for properly generating HTML here w/o using tokens, it stays\r
     *       public.\r
     * @param string $string String data to escape for HTML.\r
     * @param int $quote Quoting style, like htmlspecialchars. ENT_NOQUOTES is\r
     *               permissible for non-attribute output.\r
     * @return string escaped data.\r
     */\r
    public function escape($string, $quote = null)\r
    {\r
        // Workaround for APC bug on Mac Leopard reported by sidepodcast\r
        // http://htmlpurifier.org/phorum/read.php?3,4823,4846\r
        if ($quote === null) {\r
            $quote = ENT_COMPAT;\r
        }\r
        return htmlspecialchars($string, $quote, 'UTF-8');\r
    }\r
}\r
\r
// vim: et sw=4 sts=4\r