Html.php

Go to the documentation of this file.

<?php
class Html {
    // List of void elements from HTML5, section 8.1.2 as of 2011-08-12
    private static $voidElements = array(
        'area',
        'base',
        'br',
        'col',
        'command',
        'embed',
        'hr',
        'img',
        'input',
        'keygen',
        'link',
        'meta',
        'param',
        'source',
        'track',
        'wbr',
    );
 
    // Boolean attributes, which may have the value omitted entirely.  Manually
    // collected from the HTML5 spec as of 2011-08-12.
    private static $boolAttribs = array(
        'async',
        'autofocus',
        'autoplay',
        'checked',
        'controls',
        'default',
        'defer',
        'disabled',
        'formnovalidate',
        'hidden',
        'ismap',
        'itemscope',
        'loop',
        'multiple',
        'muted',
        'novalidate',
        'open',
        'pubdate',
        'readonly',
        'required',
        'reversed',
        'scoped',
        'seamless',
        'selected',
        'truespeed',
        'typemustmatch',
        // HTML5 Microdata
        'itemscope',
    );
 
    public static function rawElement( $element, $attribs = array(), $contents = '' ) {
        $start = self::openElement( $element, $attribs );
        if ( in_array( $element, self::$voidElements ) ) {
            // Silly XML.
            return substr( $start, 0, -1 ) . '/>';
        } else {
            return "$start$contents" . self::closeElement( $element );
        }
    }
 
    public static function element( $element, $attribs = array(), $contents = '' ) {
        return self::rawElement( $element, $attribs, strtr( $contents, array(
            // There's no point in escaping quotes, >, etc. in the contents of
            // elements.
            '&' => '&amp;',
            '<' => '&lt;'
        ) ) );
    }
 
    public static function openElement( $element, $attribs = array() ) {
        $attribs = (array)$attribs;
        // This is not required in HTML5, but let's do it anyway, for
        // consistency and better compression.
        $element = strtolower( $element );
 
        // Remove invalid input types
        if ( $element == 'input' ) {
            $validTypes = array(
                'hidden',
                'text',
                'password',
                'checkbox',
                'radio',
                'file',
                'submit',
                'image',
                'reset',
                'button',
 
                // HTML input types
                'datetime',
                'datetime-local',
                'date',
                'month',
                'time',
                'week',
                'number',
                'range',
                'email',
                'url',
                'search',
                'tel',
                'color',
            );
            if ( isset( $attribs['type'] )
            && !in_array( $attribs['type'], $validTypes ) ) {
                unset( $attribs['type'] );
            }
        }
 
        // According to standard the default type for <button> elements is "submit".
        // Depending on compatibility mode IE might use "button", instead.
        // We enforce the standard "submit".
        if ( $element == 'button' && !isset( $attribs['type'] ) ) {
            $attribs['type'] = 'submit';
        }
 
        return "<$element" . self::expandAttributes(
            self::dropDefaults( $element, $attribs ) ) . '>';
    }
 
    public static function closeElement( $element ) {
        $element = strtolower( $element );
 
        return "</$element>";
    }
 
    private static function dropDefaults( $element, $attribs ) {
 
        // Whenever altering this array, please provide a covering test case
        // in HtmlTest::provideElementsWithAttributesHavingDefaultValues
        static $attribDefaults = array(
            'area' => array( 'shape' => 'rect' ),
            'button' => array(
                'formaction' => 'GET',
                'formenctype' => 'application/x-www-form-urlencoded',
            ),
            'canvas' => array(
                'height' => '150',
                'width' => '300',
            ),
            'command' => array( 'type' => 'command' ),
            'form' => array(
                'action' => 'GET',
                'autocomplete' => 'on',
                'enctype' => 'application/x-www-form-urlencoded',
            ),
            'input' => array(
                'formaction' => 'GET',
                'type' => 'text',
            ),
            'keygen' => array( 'keytype' => 'rsa' ),
            'link' => array( 'media' => 'all' ),
            'menu' => array( 'type' => 'list' ),
            // Note: the use of text/javascript here instead of other JavaScript
            // MIME types follows the HTML5 spec.
            'script' => array( 'type' => 'text/javascript' ),
            'style' => array(
                'media' => 'all',
                'type' => 'text/css',
            ),
            'textarea' => array( 'wrap' => 'soft' ),
        );
 
        $element = strtolower( $element );
 
        foreach ( $attribs as $attrib => $value ) {
            $lcattrib = strtolower( $attrib );
            if ( is_array( $value ) ) {
                $value = implode( ' ', $value );
            } else {
                $value = strval( $value );
            }
 
            // Simple checks using $attribDefaults
            if ( isset( $attribDefaults[$element][$lcattrib] ) &&
            $attribDefaults[$element][$lcattrib] == $value ) {
                unset( $attribs[$attrib] );
            }
 
            if ( $lcattrib == 'class' && $value == '' ) {
                unset( $attribs[$attrib] );
            }
        }
 
        // More subtle checks
        if ( $element === 'link' && isset( $attribs['type'] )
        && strval( $attribs['type'] ) == 'text/css' ) {
            unset( $attribs['type'] );
        }
        if ( $element === 'input' ) {
            $type = isset( $attribs['type'] ) ? $attribs['type'] : null;
            $value = isset( $attribs['value'] ) ? $attribs['value'] : null;
            if ( $type === 'checkbox' || $type === 'radio' ) {
                // The default value for checkboxes and radio buttons is 'on'
                // not ''. By stripping value="" we break radio boxes that
                // actually wants empty values.
                if ( $value === 'on' ) {
                    unset( $attribs['value'] );
                }
            } elseif ( $type === 'submit' ) {
                // The default value for submit appears to be "Submit" but
                // let's not bother stripping out localized text that matches
                // that.
            } else {
                // The default value for nearly every other field type is ''
                // The 'range' and 'color' types use different defaults but
                // stripping a value="" does not hurt them.
                if ( $value === '' ) {
                    unset( $attribs['value'] );
                }
            }
        }
        if ( $element === 'select' && isset( $attribs['size'] ) ) {
            if ( in_array( 'multiple', $attribs )
                || ( isset( $attribs['multiple'] ) && $attribs['multiple'] !== false )
            ) {
                // A multi-select
                if ( strval( $attribs['size'] ) == '4' ) {
                    unset( $attribs['size'] );
                }
            } else {
                // Single select
                if ( strval( $attribs['size'] ) == '1' ) {
                    unset( $attribs['size'] );
                }
            }
        }
 
        return $attribs;
    }
 
    public static function expandAttributes( $attribs ) {
        $ret = '';
        $attribs = (array)$attribs;
        foreach ( $attribs as $key => $value ) {
            // Support intuitive array( 'checked' => true/false ) form
            if ( $value === false || is_null( $value ) ) {
                continue;
            }
 
            // For boolean attributes, support array( 'foo' ) instead of
            // requiring array( 'foo' => 'meaningless' ).
            if ( is_int( $key )
            && in_array( strtolower( $value ), self::$boolAttribs ) ) {
                $key = $value;
            }
 
            // Not technically required in HTML5 but we'd like consistency
            // and better compression anyway.
            $key = strtolower( $key );
 
            // Bug 23769: Blacklist all form validation attributes for now.  Current
            // (June 2010) WebKit has no UI, so the form just refuses to submit
            // without telling the user why, which is much worse than failing
            // server-side validation.  Opera is the only other implementation at
            // this time, and has ugly UI, so just kill the feature entirely until
            // we have at least one good implementation.
 
            // As the default value of "1" for "step" rejects decimal
            // numbers to be entered in 'type="number"' fields, allow
            // the special case 'step="any"'.
 
            if ( in_array( $key, array( 'max', 'min', 'pattern', 'required' ) )
                || $key === 'step' && $value !== 'any' ) {
                continue;
            }
 
            // http://www.w3.org/TR/html401/index/attributes.html ("space-separated")
            // http://www.w3.org/TR/html5/index.html#attributes-1 ("space-separated")
            $spaceSeparatedListAttributes = array(
                'class', // html4, html5
                'accesskey', // as of html5, multiple space-separated values allowed
                // html4-spec doesn't document rel= as space-separated
                // but has been used like that and is now documented as such
                // in the html5-spec.
                'rel',
            );
 
            // Specific features for attributes that allow a list of space-separated values
            if ( in_array( $key, $spaceSeparatedListAttributes ) ) {
                // Apply some normalization and remove duplicates
 
                // Convert into correct array. Array can contain space-separated
                // values. Implode/explode to get those into the main array as well.
                if ( is_array( $value ) ) {
                    // If input wasn't an array, we can skip this step
                    $newValue = array();
                    foreach ( $value as $k => $v ) {
                        if ( is_string( $v ) ) {
                            // String values should be normal `array( 'foo' )`
                            // Just append them
                            if ( !isset( $value[$v] ) ) {
                                // As a special case don't set 'foo' if a
                                // separate 'foo' => true/false exists in the array
                                // keys should be authoritative
                                $newValue[] = $v;
                            }
                        } elseif ( $v ) {
                            // If the value is truthy but not a string this is likely
                            // an array( 'foo' => true ), falsy values don't add strings
                            $newValue[] = $k;
                        }
                    }
                    $value = implode( ' ', $newValue );
                }
                $value = explode( ' ', $value );
 
                // Normalize spacing by fixing up cases where people used
                // more than 1 space and/or a trailing/leading space
                $value = array_diff( $value, array( '', ' ' ) );
 
                // Remove duplicates and create the string
                $value = implode( ' ', array_unique( $value ) );
            }
 
            $quote = '"';
 
            if ( in_array( $key, self::$boolAttribs ) ) {
                $ret .= " $key=\"\"";
            } else {
                // Apparently we need to entity-encode \n, \r, \t, although the
                // spec doesn't mention that.  Since we're doing strtr() anyway,
                // we may as well not call htmlspecialchars().
                // @todo FIXME: Verify that we actually need to
                // escape \n\r\t here, and explain why, exactly.
                #
                // We could call Sanitizer::encodeAttribute() for this, but we
                // don't because we're stubborn and like our marginal savings on
                // byte size from not having to encode unnecessary quotes.
                // The only difference between this transform and the one by
                // Sanitizer::encodeAttribute() is ' is not encoded.
                $map = array(
                    '&' => '&amp;',
                    '"' => '&quot;',
                    '>' => '&gt;',
                    // '<' allegedly allowed per spec
                    // but breaks some tools if not escaped.
                    "<" => '&lt;',
                    "\n" => '&#10;',
                    "\r" => '&#13;',
                    "\t" => '&#9;'
                );
                $ret .= " $key=$quote" . strtr( $value, $map ) . $quote;
            }
        }
        return $ret;
    }
 
    public static function inlineScript( $contents ) {
        $attrs = array();
 
        if ( preg_match( '/[<&]/', $contents ) ) {
            $contents = "/*<![CDATA[*/$contents/*]]>*/";
        }
 
        return self::rawElement( 'script', $attrs, $contents );
    }
 
    public static function linkedScript( $url ) {
        $attrs = array( 'src' => $url );
 
        return self::element( 'script', $attrs );
    }
 
    public static function inlineStyle( $contents, $media = 'all' ) {
        // Don't escape '>' since that is used
        // as direct child selector.
        // Remember, in css, there is no "x" for hexadecimal escapes, and
        // the space immediately after an escape sequence is swallowed.
        $contents = strtr( $contents, array(
            '<' => '\3C ',
            // CDATA end tag for good measure, but the main security
            // is from escaping the '<'.
            ']]>' => '\5D\5D\3E '
        ) );
 
        if ( preg_match( '/[<&]/', $contents ) ) {
            $contents = "/*<![CDATA[*/$contents/*]]>*/";
        }
 
        return self::rawElement( 'style', array(
            'type' => 'text/css',
            'media' => $media,
        ), $contents );
    }
 
    public static function linkedStyle( $url, $media = 'all' ) {
        return self::element( 'link', array(
            'rel' => 'stylesheet',
            'href' => $url,
            'type' => 'text/css',
            'media' => $media,
        ) );
    }
 
    public static function input( $name, $value = '', $type = 'text', $attribs = array() ) {
        $attribs['type'] = $type;
        $attribs['value'] = $value;
        $attribs['name'] = $name;
 
        return self::element( 'input', $attribs );
    }
 
    public static function hidden( $name, $value, $attribs = array() ) {
        return self::input( $name, $value, 'hidden', $attribs );
    }
 
    public static function textarea( $name, $value = '', $attribs = array() ) {
        $attribs['name'] = $name;
 
        if ( substr( $value, 0, 1 ) == "\n" ) {
            // Workaround for bug 12130: browsers eat the initial newline
            // assuming that it's just for show, but they do keep the later
            // newlines, which we may want to preserve during editing.
            // Prepending a single newline
            $spacedValue = "\n" . $value;
        } else {
            $spacedValue = $value;
        }
        return self::element( 'textarea', $attribs, $spacedValue );
    }
 
    public static function namespaceSelector( array $params = array(), array $selectAttribs = array() ) {
        global $wgContLang;
 
        ksort( $selectAttribs );
 
        // Is a namespace selected?
        if ( isset( $params['selected'] ) ) {
            // If string only contains digits, convert to clean int. Selected could also
            // be "all" or "" etc. which needs to be left untouched.
            // PHP is_numeric() has issues with large strings, PHP ctype_digit has other issues
            // and returns false for already clean ints. Use regex instead..
            if ( preg_match( '/^\d+$/', $params['selected'] ) ) {
                $params['selected'] = intval( $params['selected'] );
            }
            // else: leaves it untouched for later processing
        } else {
            $params['selected'] = '';
        }
 
        if ( !isset( $params['exclude'] ) || !is_array( $params['exclude'] ) ) {
            $params['exclude'] = array();
        }
        if ( !isset( $params['disable'] ) || !is_array( $params['disable'] ) ) {
            $params['disable'] = array();
        }
 
        // Associative array between option-values and option-labels
        $options = array();
 
        if ( isset( $params['all'] ) ) {
            // add an option that would let the user select all namespaces.
            // Value is provided by user, the name shown is localized for the user.
            $options[$params['all']] = wfMessage( 'namespacesall' )->text();
        }
        // Add all namespaces as options (in the content language)
        $options += $wgContLang->getFormattedNamespaces();
 
        // Convert $options to HTML and filter out namespaces below 0
        $optionsHtml = array();
        foreach ( $options as $nsId => $nsName ) {
            if ( $nsId < NS_MAIN || in_array( $nsId, $params['exclude'] ) ) {
                continue;
            }
            if ( $nsId === NS_MAIN ) {
                // For other namespaces use use the namespace prefix as label, but for
                // main we don't use "" but the user message describing it (e.g. "(Main)" or "(Article)")
                $nsName = wfMessage( 'blanknamespace' )->text();
            } elseif ( is_int( $nsId ) ) {
                $nsName = $wgContLang->convertNamespace( $nsId );
            }
            $optionsHtml[] = Html::element(
                'option', array(
                    'disabled' => in_array( $nsId, $params['disable'] ),
                    'value' => $nsId,
                    'selected' => $nsId === $params['selected'],
                ), $nsName
            );
        }
 
        if ( !array_key_exists( 'id', $selectAttribs ) ) {
            $selectAttribs['id'] = 'namespace';
        }
 
        if ( !array_key_exists( 'name', $selectAttribs ) ) {
            $selectAttribs['name'] = 'namespace';
        }
 
        $ret = '';
        if ( isset( $params['label'] ) ) {
            $ret .= Html::element(
                'label', array(
                    'for' => isset( $selectAttribs['id'] ) ? $selectAttribs['id'] : null,
                ), $params['label']
            ) . '&#160;';
        }
 
        // Wrap options in a <select>
        $ret .= Html::openElement( 'select', $selectAttribs )
            . "\n"
            . implode( "\n", $optionsHtml )
            . "\n"
            . Html::closeElement( 'select' );
 
        return $ret;
    }
 
    public static function htmlHeader( $attribs = array() ) {
        $ret = '';
 
        global $wgHtml5Version, $wgMimeType, $wgXhtmlNamespaces;
 
        $isXHTML = self::isXmlMimeType( $wgMimeType );
 
        if ( $isXHTML ) { // XHTML5
            // XML mimetyped markup should have an xml header.
            // However a DOCTYPE is not needed.
            $ret .= "<?xml version=\"1.0\" encoding=\"UTF-8\" ?" . ">\n";
 
            // Add the standard xmlns
            $attribs['xmlns'] = 'http://www.w3.org/1999/xhtml';
 
            // And support custom namespaces
            foreach ( $wgXhtmlNamespaces as $tag => $ns ) {
                $attribs["xmlns:$tag"] = $ns;
            }
        } else { // HTML5
            // DOCTYPE
            $ret .= "<!DOCTYPE html>\n";
        }
 
        if ( $wgHtml5Version ) {
            $attribs['version'] = $wgHtml5Version;
        }
 
        $html = Html::openElement( 'html', $attribs );
 
        if ( $html ) {
            $html .= "\n";
        }
 
        $ret .= $html;
 
        return $ret;
    }
 
    public static function isXmlMimeType( $mimetype ) {
        # http://www.whatwg.org/html/infrastructure.html#xml-mime-type
        # * text/xml
        # * application/xml
        # * Any mimetype with a subtype ending in +xml (this implicitly includes application/xhtml+xml)
        return (bool)preg_match( '!^(text|application)/xml$|^.+/.+\+xml$!', $mimetype );
    }
 
    static function infoBox( $text, $icon, $alt, $class = false, $useStylePath = true ) {
        global $wgStylePath;
 
        if ( $useStylePath ) {
            $icon = $wgStylePath . '/common/images/' . $icon;
        }
 
        $s = Html::openElement( 'div', array( 'class' => "mw-infobox $class" ) );
 
        $s .= Html::openElement( 'div', array( 'class' => 'mw-infobox-left' ) ) .
                Html::element( 'img',
                    array(
                        'src' => $icon,
                        'alt' => $alt,
                    )
                ) .
                Html::closeElement( 'div' );
 
        $s .= Html::openElement( 'div', array( 'class' => 'mw-infobox-right' ) ) .
                $text .
                Html::closeElement( 'div' );
        $s .= Html::element( 'div', array( 'style' => 'clear: left;' ), ' ' );
 
        $s .= Html::closeElement( 'div' );
 
        $s .= Html::element( 'div', array( 'style' => 'clear: left;' ), ' ' );
 
        return $s;
    }
 
    static function srcSet( $urls ) {
        $candidates = array();
        foreach ( $urls as $density => $url ) {
            // Image candidate syntax per current whatwg live spec, 2012-09-23:
            // http://www.whatwg.org/html/embedded-content-1.html#attr-img-srcset
            $candidates[] = "{$url} {$density}x";
        }
        return implode( ", ", $candidates );
    }
}