CSSMin.php

Go to the documentation of this file.

<?php
class CSSMin {
 
    /* Constants */
 
    const PLACEHOLDER = "\x7fPLACEHOLDER\x7f";
 
    const DATA_URI_SIZE_LIMIT = 32768;
 
    const EMBED_REGEX = '\/\*\s*\@embed\s*\*\/';
    const COMMENT_REGEX = '\/\*.*?\*\/';
 
    /* Protected Static Members */
 
    protected static $mimeTypes = [
        'gif' => 'image/gif',
        'jpe' => 'image/jpeg',
        'jpeg' => 'image/jpeg',
        'jpg' => 'image/jpeg',
        'png' => 'image/png',
        'tif' => 'image/tiff',
        'tiff' => 'image/tiff',
        'xbm' => 'image/x-xbitmap',
        'svg' => 'image/svg+xml',
    ];
 
    /* Static Methods */
 
    public static function getLocalFileReferences( $source, $path ) {
        $stripped = preg_replace( '/' . self::COMMENT_REGEX . '/s', '', $source );
        $path = rtrim( $path, '/' ) . '/';
        $files = [];
 
        $rFlags = PREG_OFFSET_CAPTURE | PREG_SET_ORDER;
        if ( preg_match_all( '/' . self::getUrlRegex() . '/', $stripped, $matches, $rFlags ) ) {
            foreach ( $matches as $match ) {
                self::processUrlMatch( $match, $rFlags );
                $url = $match['file'][0];
 
                // Skip fully-qualified and protocol-relative URLs and data URIs
                // Also skips the rare `behavior` property specifying application's default behavior
                if (
                    substr( $url, 0, 2 ) === '//' ||
                    parse_url( $url, PHP_URL_SCHEME ) ||
                    substr( $url, 0, 9 ) === '#default#'
                ) {
                    break;
                }
 
                $files[] = $path . $url;
            }
        }
        return $files;
    }
 
    public static function encodeImageAsDataURI( $file, $type = null, $ie8Compat = true ) {
        // Fast-fail for files that definitely exceed the maximum data URI length
        if ( $ie8Compat && filesize( $file ) >= self::DATA_URI_SIZE_LIMIT ) {
            return false;
        }
 
        if ( $type === null ) {
            $type = self::getMimeType( $file );
        }
        if ( !$type ) {
            return false;
        }
 
        return self::encodeStringAsDataURI( file_get_contents( $file ), $type, $ie8Compat );
    }
 
    public static function encodeStringAsDataURI( $contents, $type, $ie8Compat = true ) {
        // Try #1: Non-encoded data URI
        // The regular expression matches ASCII whitespace and printable characters.
        if ( preg_match( '/^[\r\n\t\x20-\x7e]+$/', $contents ) ) {
            // Do not base64-encode non-binary files (sane SVGs).
            // (This often produces longer URLs, but they compress better, yielding a net smaller size.)
            $encoded = rawurlencode( $contents );
            // Unencode some things that don't need to be encoded, to make the encoding smaller
            $encoded = strtr( $encoded, [
                '%20' => ' ', // Unencode spaces
                '%2F' => '/', // Unencode slashes
                '%3A' => ':', // Unencode colons
                '%3D' => '=', // Unencode equals signs
            ] );
            $uri = 'data:' . $type . ',' . $encoded;
            if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
                return $uri;
            }
        }
 
        // Try #2: Encoded data URI
        $uri = 'data:' . $type . ';base64,' . base64_encode( $contents );
        if ( !$ie8Compat || strlen( $uri ) < self::DATA_URI_SIZE_LIMIT ) {
            return $uri;
        }
 
        // A data URI couldn't be produced
        return false;
    }
 
    public static function serializeStringValue( $value ) {
        if ( strstr( $value, "\0" ) ) {
            throw new Exception( "Invalid character in CSS string" );
        }
        $value = strtr( $value, [ '\\' => '\\\\', '"' => '\\"' ] );
        $value = preg_replace_callback( '/[\x01-\x1f\x7f-\x9f]/', function ( $match ) {
            return '\\' . base_convert( ord( $match[0] ), 10, 16 ) . ' ';
        }, $value );
        return '"' . $value . '"';
    }
 
    public static function getMimeType( $file ) {
        // Infer the MIME-type from the file extension
        $ext = strtolower( pathinfo( $file, PATHINFO_EXTENSION ) );
        if ( isset( self::$mimeTypes[$ext] ) ) {
            return self::$mimeTypes[$ext];
        }
 
        return mime_content_type( realpath( $file ) );
    }
 
    public static function buildUrlValue( $url ) {
        // The list below has been crafted to match URLs such as:
        //   scheme://user@domain:port/~user/fi%20le.png?query=yes&really=y+s
        //   data:image/png;base64,R0lGODlh/+==
        if ( preg_match( '!^[\w\d:@/~.%+;,?&=-]+$!', $url ) ) {
            return "url($url)";
        } else {
            return 'url("' . strtr( $url, [ '\\' => '\\\\', '"' => '\\"' ] ) . '")';
        }
    }
 
    public static function remap( $source, $local, $remote, $embedData = true ) {
        // High-level overview:
        // * For each CSS rule in $source that includes at least one url() value:
        //   * Check for an @embed comment at the start indicating that all URIs should be embedded
        //   * For each url() value:
        //     * Check for an @embed comment directly preceding the value
        //     * If either @embed comment exists:
        //       * Embedding the URL as data: URI, if it's possible / allowed
        //       * Otherwise remap the URL to work in generated stylesheets
 
        // Guard against trailing slashes, because "some/remote/../foo.png"
        // resolves to "some/remote/foo.png" on (some?) clients (T29052).
        if ( substr( $remote, -1 ) == '/' ) {
            $remote = substr( $remote, 0, -1 );
        }
 
        // Disallow U+007F DELETE, which is illegal anyway, and which
        // we use for comment placeholders.
        $source = str_replace( "\x7f", "?", $source );
 
        // Replace all comments by a placeholder so they will not interfere with the remapping.
        // Warning: This will also catch on anything looking like the start of a comment between
        // quotation marks (e.g. "foo /* bar").
        $comments = [];
 
        $pattern = '/(?!' . self::EMBED_REGEX . ')(' . self::COMMENT_REGEX . ')/s';
 
        $source = preg_replace_callback(
            $pattern,
            function ( $match ) use ( &$comments ) {
                $comments[] = $match[ 0 ];
                return CSSMin::PLACEHOLDER . ( count( $comments ) - 1 ) . 'x';
            },
            $source
        );
 
        // Note: This will not correctly handle cases where ';', '{' or '}'
        // appears in the rule itself, e.g. in a quoted string. You are advised
        // not to use such characters in file names. We also match start/end of
        // the string to be consistent in edge-cases ('@import url(…)').
        $pattern = '/(?:^|[;{])\K[^;{}]*' . self::getUrlRegex() . '[^;}]*(?=[;}]|$)/';
 
        $source = preg_replace_callback(
            $pattern,
            function ( $matchOuter ) use ( $local, $remote, $embedData ) {
                $rule = $matchOuter[0];
 
                // Check for global @embed comment and remove it. Allow other comments to be present
                // before @embed (they have been replaced with placeholders at this point).
                $embedAll = false;
                $rule = preg_replace(
                    '/^((?:\s+|' .
                        CSSMin::PLACEHOLDER .
                        '(\d+)x)*)' .
                        CSSMin::EMBED_REGEX .
                        '\s*/',
                    '$1',
                    $rule,
                    1,
                    $embedAll
                );
 
                // Build two versions of current rule: with remapped URLs
                // and with embedded data: URIs (where possible).
                $pattern = '/(?P<embed>' . CSSMin::EMBED_REGEX . '\s*|)' . self::getUrlRegex() . '/';
 
                $ruleWithRemapped = preg_replace_callback(
                    $pattern,
                    function ( $match ) use ( $local, $remote ) {
                        self::processUrlMatch( $match );
 
                        $remapped = CSSMin::remapOne( $match['file'], $match['query'], $local, $remote, false );
                        return CSSMin::buildUrlValue( $remapped );
                    },
                    $rule
                );
 
                if ( $embedData ) {
                    // Remember the occurring MIME types to avoid fallbacks when embedding some files.
                    $mimeTypes = [];
 
                    $ruleWithEmbedded = preg_replace_callback(
                        $pattern,
                        function ( $match ) use ( $embedAll, $local, $remote, &$mimeTypes ) {
                            self::processUrlMatch( $match );
 
                            $embed = $embedAll || $match['embed'];
                            $embedded = CSSMin::remapOne(
                                $match['file'],
                                $match['query'],
                                $local,
                                $remote,
                                $embed
                            );
 
                            $url = $match['file'] . $match['query'];
                            $file = "{$local}/{$match['file']}";
                            if (
                                !self::isRemoteUrl( $url ) && !self::isLocalUrl( $url )
                                && file_exists( $file )
                            ) {
                                $mimeTypes[ CSSMin::getMimeType( $file ) ] = true;
                            }
 
                            return CSSMin::buildUrlValue( $embedded );
                        },
                        $rule
                    );
 
                    // Are all referenced images SVGs?
                    $needsEmbedFallback = $mimeTypes !== [ 'image/svg+xml' => true ];
                }
 
                if ( !$embedData || $ruleWithEmbedded === $ruleWithRemapped ) {
                    // We're not embedding anything, or we tried to but the file is not embeddable
                    return $ruleWithRemapped;
                } elseif ( $embedData && $needsEmbedFallback ) {
                    // Build 2 CSS properties; one which uses a data URI in place of the @embed comment, and
                    // the other with a remapped and versioned URL with an Internet Explorer 6 and 7 hack
                    // making it ignored in all browsers that support data URIs
                    return "$ruleWithEmbedded;$ruleWithRemapped!ie";
                } else {
                    // Look ma, no fallbacks! This is for files which IE 6 and 7 don't support anyway: SVG.
                    return $ruleWithEmbedded;
                }
            }, $source );
 
        // Re-insert comments
        $pattern = '/' . self::PLACEHOLDER . '(\d+)x/';
        $source = preg_replace_callback( $pattern, function ( $match ) use ( &$comments ) {
            return $comments[ $match[1] ];
        }, $source );
 
        return $source;
    }
 
    protected static function isRemoteUrl( $maybeUrl ) {
        if ( substr( $maybeUrl, 0, 2 ) === '//' || parse_url( $maybeUrl, PHP_URL_SCHEME ) ) {
            return true;
        }
        return false;
    }
 
    protected static function isLocalUrl( $maybeUrl ) {
        if ( $maybeUrl !== '' && $maybeUrl[0] === '/' && !self::isRemoteUrl( $maybeUrl ) ) {
            return true;
        }
        return false;
    }
 
    private static function getUrlRegex() {
        static $urlRegex;
        if ( $urlRegex === null ) {
            // Match these three variants separately to avoid broken urls when
            // e.g. a double quoted url contains a parenthesis, or when a
            // single quoted url contains a double quote, etc.
            // Note: PCRE doesn't support multiple capture groups with the same name by default.
            // - PCRE 6.7 introduced the "J" modifier (PCRE_INFO_JCHANGED for PCRE_DUPNAMES).
            //   https://secure.php.net/manual/en/reference.pcre.pattern.modifiers.php
            //   However this isn't useful since it just ignores all but the first one.
            //   Also, while the modifier was introduced in PCRE 6.7 (PHP 5.2+) it was
            //   not exposed to public preg_* functions until PHP 5.6.0.
            // - PCRE 8.36 fixed this to work as expected (e.g. merge conceptually to
            //   only return the one matched in the part that actually matched).
            //   However MediaWiki supports 5.5.9, which has PCRE 8.32
            //   Per https://secure.php.net/manual/en/pcre.installation.php:
            //   - PCRE 8.32 (PHP 5.5.0)
            //   - PCRE 8.34 (PHP 5.5.10, PHP 5.6.0)
            //   - PCRE 8.37 (PHP 5.5.26, PHP 5.6.9, PHP 7.0.0)
            //   Workaround by using different groups and merge via processUrlMatch().
            // - Using string concatenation for class constant or member assignments
            //   is only supported in PHP 5.6. Use a getter method for now.
            $urlRegex = '(' .
                // Unquoted url
                'url\(\s*(?P<file0>[^\'"][^\?\)]*?)(?P<query0>\?[^\)]*?|)\s*\)' .
                // Single quoted url
                '|url\(\s*\'(?P<file1>[^\?\']*?)(?P<query1>\?[^\']*?|)\'\s*\)' .
                // Double quoted url
                '|url\(\s*"(?P<file2>[^\?"]*?)(?P<query2>\?[^"]*?|)"\s*\)' .
                ')';
        }
        return $urlRegex;
    }
 
    private static function processUrlMatch( array &$match, $flags = 0 ) {
        if ( $flags & PREG_SET_ORDER ) {
            // preg_match_all with PREG_SET_ORDER will return each group in each
            // match array, and if it didn't match, instead of the sub array
            // being an empty array it is `[ '', -1 ]`...
            if ( isset( $match['file0'] ) && $match['file0'][1] !== -1 ) {
                $match['file'] = $match['file0'];
                $match['query'] = $match['query0'];
            } elseif ( isset( $match['file1'] ) && $match['file1'][1] !== -1 ) {
                $match['file'] = $match['file1'];
                $match['query'] = $match['query1'];
            } else {
                $match['file'] = $match['file2'];
                $match['query'] = $match['query2'];
            }
        } else {
            if ( isset( $match['file0'] ) && $match['file0'] !== '' ) {
                $match['file'] = $match['file0'];
                $match['query'] = $match['query0'];
            } elseif ( isset( $match['file1'] ) && $match['file1'] !== '' ) {
                $match['file'] = $match['file1'];
                $match['query'] = $match['query1'];
            } else {
                $match['file'] = $match['file2'];
                $match['query'] = $match['query2'];
            }
        }
    }
 
    public static function remapOne( $file, $query, $local, $remote, $embed ) {
        // The full URL possibly with query, as passed to the 'url()' value in CSS
        $url = $file . $query;
 
        // Expand local URLs with absolute paths like /w/index.php to possibly protocol-relative URL, if
        // wfExpandUrl() is available. (This will not be the case if we're running outside of MW.)
        if ( self::isLocalUrl( $url ) && function_exists( 'wfExpandUrl' ) ) {
            return wfExpandUrl( $url, PROTO_RELATIVE );
        }
 
        // Pass thru fully-qualified and protocol-relative URLs and data URIs, as well as local URLs if
        // we can't expand them.
        // Also skips the rare `behavior` property specifying application's default behavior
        if (
            self::isRemoteUrl( $url ) ||
            self::isLocalUrl( $url ) ||
            substr( $url, 0, 9 ) === '#default#'
        ) {
            return $url;
        }
 
        if ( $local === false ) {
            // Assume that all paths are relative to $remote, and make them absolute
            $url = $remote . '/' . $url;
        } else {
            // We drop the query part here and instead make the path relative to $remote
            $url = "{$remote}/{$file}";
            // Path to the actual file on the filesystem
            $localFile = "{$local}/{$file}";
            if ( file_exists( $localFile ) ) {
                if ( $embed ) {
                    $data = self::encodeImageAsDataURI( $localFile );
                    if ( $data !== false ) {
                        return $data;
                    }
                }
                if ( method_exists( 'OutputPage', 'transformFilePath' ) ) {
                    $url = OutputPage::transformFilePath( $remote, $local, $file );
                } else {
                    // Add version parameter as the first five hex digits
                    // of the MD5 hash of the file's contents.
                    $url .= '?' . substr( md5_file( $localFile ), 0, 5 );
                }
            }
            // If any of these conditions failed (file missing, we don't want to embed it
            // or it's not embeddable), return the URL (possibly with ?timestamp part)
        }
        if ( function_exists( 'wfRemoveDotSegments' ) ) {
            $url = wfRemoveDotSegments( $url );
        }
        return $url;
    }
 
    public static function minify( $css ) {
        return trim(
            str_replace(
                [ '; ', ': ', ' {', '{ ', ', ', '} ', ';}' ],
                [ ';', ':', '{', '{', ',', '}', '}' ],
                preg_replace( [ '/\s+/', '/\/\*.*?\*\//s' ], [ ' ', '' ], $css )
            )
        );
    }
}