mimemagic.php

php 开发的内容管理系统

		$sub4 =  substr( $head, 0, 4 );
		if ( $sub4 == "\x01\x00\x09\x00" || $sub4 == "\xd7\xcd\xc6\x9a" ) {
			// WMF kill kill kill
			// Note that WMF may have a bare header, no magic number.
			// The former of the above two checks is theoretically prone to false positives
			$mime = "application/x-msmetafile";
		}

		if (strpos($mime,"text/")===0 || $mime==="application/xml") {

			$xml_type= NULL;
			$script_type= NULL;

			/*
			* look for XML formats (XHTML and SVG)
			*/
			if ($mime==="text/sgml" ||
			    $mime==="text/plain" ||
			    $mime==="text/html" ||
			    $mime==="text/xml" ||
		            $mime==="application/xml") {

				if (substr($head,0,5)=="<?xml") $xml_type= "ASCII";
				elseif (substr($head,0,8)=="\xef\xbb\xbf<?xml") $xml_type= "UTF-8";
				elseif (substr($head,0,10)=="\xfe\xff\x00<\x00?\x00x\x00m\x00l") $xml_type= "UTF-16BE";
				elseif (substr($head,0,10)=="\xff\xfe<\x00?\x00x\x00m\x00l\x00") $xml_type= "UTF-16LE";

				if ($xml_type) {
					if ($xml_type!=="UTF-8" && $xml_type!=="ASCII") $head= iconv($xml_type,"ASCII//IGNORE",$head);

					$match= array();
					$doctype= "";
					$tag= "";

					if (preg_match('%<!DOCTYPE\s+[\w-]+\s+PUBLIC\s+["'."'".'"](.*?)["'."'".'"].*>%sim',$head,$match)) $doctype= $match[1];
					if (preg_match('%<(\w+).*>%sim',$head,$match)) $tag= $match[1];

					#print "<br>ANALYSING $file ($mime): doctype= $doctype; tag= $tag<br>";

					if (strpos($doctype,"-//W3C//DTD SVG")===0) $mime= "image/svg";
					elseif ($tag==="svg") $mime= "image/svg";
					elseif (strpos($doctype,"-//W3C//DTD XHTML")===0) $mime= "text/html";
					elseif ($tag==="html") $mime= "text/html";

					$test_more= false;
				}
			}

			/*
			* look for shell scripts
			*/
			if (!$xml_type) {
				$script_type= NULL;

				#detect by shebang
				if (substr($head,0,2)=="#!") $script_type= "ASCII";
				elseif (substr($head,0,5)=="\xef\xbb\xbf#!") $script_type= "UTF-8";
				elseif (substr($head,0,7)=="\xfe\xff\x00#\x00!") $script_type= "UTF-16BE";
				elseif (substr($head,0,7)=="\xff\xfe#\x00!") $script_type= "UTF-16LE";

				if ($script_type) {
					if ($script_type!=="UTF-8" && $script_type!=="ASCII") $head= iconv($script_type,"ASCII//IGNORE",$head);

					$match= array();
					$prog= "";

					if (preg_match('%/?([^\s]+/)(w+)%sim',$head,$match)) $script= $match[2];

					$mime= "application/x-$prog";
				}
			}

			/*
			* look for PHP
			*/
			if( !$xml_type && !$script_type ) {

				if( ( strpos( $head, '<?php' ) !== false ) ||
				    ( strpos( $head, '<? ' ) !== false ) ||
				    ( strpos( $head, "<?\n" ) !== false ) ||
				    ( strpos( $head, "<?\t" ) !== false ) ||
				    ( strpos( $head, "<?=" ) !== false ) ||

				    ( strpos( $head, "<\x00?\x00p\x00h\x00p" ) !== false ) ||
				    ( strpos( $head, "<\x00?\x00 " ) !== false ) ||
				    ( strpos( $head, "<\x00?\x00\n" ) !== false ) ||
				    ( strpos( $head, "<\x00?\x00\t" ) !== false ) ||
				    ( strpos( $head, "<\x00?\x00=" ) !== false ) ) {

					$mime= "application/x-php";
				}
			}

		}

		if (isset($this->mMimeTypeAliases[$mime])) $mime= $this->mMimeTypeAliases[$mime];

		wfDebug("$fname: final mime type of $file: $mime\n");
		return $mime;
	}

	/** Internal mime type detection, please use guessMimeType() for application code instead.
	* Detection is done using an external program, if $wgMimeDetectorCommand is set.
	* Otherwise, the fileinfo extension and mime_content_type are tried (in this order), if they are available.
	* If the dections fails and $useExt is true, the mime type is guessed from the file extension, using guessTypesForExtension.
	* If the mime type is still unknown, getimagesize is used to detect the mime type if the file is an image.
	* If no mime type can be determined, this function returns "unknown/unknown".
	*
	* @param string $file The file to check
	* @param bool $useExt switch for allowing to use the file extension to guess the mime type. true by default.
	*
	* @return string the mime type of $file
	* @access private
	*/
	function detectMimeType( $file, $useExt=true ) {
		$fname = 'MimeMagic::detectMimeType';

		global $wgMimeDetectorCommand;

		$m= NULL;
		if ($wgMimeDetectorCommand) {
			$fn= wfEscapeShellArg($file);
			$m= `$wgMimeDetectorCommand $fn`;
		}
		else if (function_exists("finfo_open") && function_exists("finfo_file")) {

			# This required the fileinfo extension by PECL,
			# see http://pecl.php.net/package/fileinfo
			# This must be compiled into PHP
			#
			# finfo is the official replacement for the deprecated
			# mime_content_type function, see below.
			#
			# If you may need to load the fileinfo extension at runtime, set
			# $wgLoadFileinfoExtension in LocalSettings.php

			$mime_magic_resource = finfo_open(FILEINFO_MIME); /* return mime type ala mimetype extension */

			if ($mime_magic_resource) {
				$m= finfo_file($mime_magic_resource, $file);

				finfo_close($mime_magic_resource);
			}
			else wfDebug("$fname: finfo_open failed on ".FILEINFO_MIME."!\n");
		}
		else if (function_exists("mime_content_type")) {

			# NOTE: this function is available since PHP 4.3.0, but only if
			# PHP was compiled with --with-mime-magic or, before 4.3.2, with --enable-mime-magic.
			#
			# On Winodws, you must set mime_magic.magicfile in php.ini to point to the mime.magic file bundeled with PHP;
			# sometimes, this may even be needed under linus/unix.
			#
			# Also note that this has been DEPRECATED in favor of the fileinfo extension by PECL, see above.
			# see http://www.php.net/manual/en/ref.mime-magic.php for details.

			$m= mime_content_type($file);
		}
		else wfDebug("$fname: no magic mime detector found!\n");

		if ($m) {
			#normalize
			$m= preg_replace('![;, ].*$!','',$m); #strip charset, etc
			$m= trim($m);
			$m= strtolower($m);

			if (strpos($m,'unknown')!==false) $m= NULL;
			else {
				wfDebug("$fname: magic mime type of $file: $m\n");
				return $m;
			}
		}

		#if still not known, use getimagesize to find out the type of image
		#TODO: skip things that do not have a well-known image extension? Would that be safe?
		wfSuppressWarnings();
		$gis = getimagesize( $file );
		wfRestoreWarnings();

		$notAnImage= false;

		if ($gis && is_array($gis) && $gis[2]) {
			switch ($gis[2]) {
				case IMAGETYPE_GIF: $m= "image/gif"; break;
				case IMAGETYPE_JPEG: $m= "image/jpeg"; break;
				case IMAGETYPE_PNG: $m= "image/png"; break;
				case IMAGETYPE_SWF: $m= "application/x-shockwave-flash"; break;
				case IMAGETYPE_PSD: $m= "application/photoshop"; break;
				case IMAGETYPE_BMP: $m= "image/bmp"; break;
				case IMAGETYPE_TIFF_II: $m= "image/tiff"; break;
				case IMAGETYPE_TIFF_MM: $m= "image/tiff"; break;
				case IMAGETYPE_JPC: $m= "image"; break;
				case IMAGETYPE_JP2: $m= "image/jpeg2000"; break;
				case IMAGETYPE_JPX: $m= "image/jpeg2000"; break;
				case IMAGETYPE_JB2: $m= "image"; break;
				case IMAGETYPE_SWC: $m= "application/x-shockwave-flash"; break;
				case IMAGETYPE_IFF: $m= "image/vnd.xiff"; break;
				case IMAGETYPE_WBMP: $m= "image/vnd.wap.wbmp"; break;
				case IMAGETYPE_XBM: $m= "image/x-xbitmap"; break;
			}

			if ($m) {
				wfDebug("$fname: image mime type of $file: $m\n");
				return $m;
			}
			else $notAnImage= true;
		} else {
			// Also test DjVu
			$deja = new DjVuImage( $file );
			if( $deja->isValid() ) {
				wfDebug("$fname: detected $file as image/vnd.djvu\n");
				return 'image/vnd.djvu';
			}
		}

		#if desired, look at extension as a fallback.
		if ($useExt) {
			$i = strrpos( $file, '.' );
			$e= strtolower( $i ? substr( $file, $i + 1 ) : '' );

			$m= $this->guessTypesForExtension($e);

			#TODO: if $notAnImage is set, do not trust the file extension if
			# the results is one of the image types that should have been recognized
			# by getimagesize

			if ($m) {
				wfDebug("$fname: extension mime type of $file: $m\n");
				return $m;
			}
		}

		#unknown type
		wfDebug("$fname: failed to guess mime type for $file!\n");
		return "unknown/unknown";
	}

	/**
	* Determine the media type code for a file, using its mime type, name and possibly
	* its contents.
	*
	* This function relies on the findMediaType(), mapping extensions and mime
	* types to media types.
	*
	* @todo analyse file if need be
	* @todo look at multiple extension, separately and together.
	*
	* @param string $path full path to the image file, in case we have to look at the contents
	*        (if null, only the mime type is used to determine the media type code).
	* @param string $mime mime type. If null it will be guessed using guessMimeType.
	*
	* @return (int?string?) a value to be used with the MEDIATYPE_xxx constants.
	*/
	function getMediaType($path=NULL,$mime=NULL) {
		if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;

		#if mime type is unknown, guess it
		if( !$mime ) $mime= $this->guessMimeType($path,false);

		#special code for ogg - detect if it's video (theora),
		#else label it as sound.
		if( $mime=="application/ogg" && file_exists($path) ) {

			// Read a chunk of the file
			$f = fopen( $path, "rt" );
			if( !$f ) return MEDIATYPE_UNKNOWN;
			$head = fread( $f, 256 );
			fclose( $f );

			$head= strtolower( $head );

			#This is an UGLY HACK, file should be parsed correctly
			if( strpos($head,'theora')!==false ) return MEDIATYPE_VIDEO;
			elseif( strpos($head,'vorbis')!==false ) return MEDIATYPE_AUDIO;
			elseif( strpos($head,'flac')!==false ) return MEDIATYPE_AUDIO;
			elseif( strpos($head,'speex')!==false ) return MEDIATYPE_AUDIO;
			else return MEDIATYPE_MULTIMEDIA;
		}

		#check for entry for full mime type
		if( $mime ) {
			$type= $this->findMediaType($mime);
			if( $type!==MEDIATYPE_UNKNOWN ) return $type;
		}

		#check for entry for file extension
		$e= NULL;
		if( $path ) {
			$i = strrpos( $path, '.' );
			$e= strtolower( $i ? substr( $path, $i + 1 ) : '' );

			#TODO: look at multi-extension if this fails, parse from full path

			$type= $this->findMediaType('.'.$e);
			if( $type!==MEDIATYPE_UNKNOWN ) return $type;
		}

		#check major mime type
		if( $mime ) {
			$i= strpos($mime,'/');
			if( $i !== false ) {
				$major= substr($mime,0,$i);
				$type= $this->findMediaType($major);
				if( $type!==MEDIATYPE_UNKNOWN ) return $type;
			}
		}

		if( !$type ) $type= MEDIATYPE_UNKNOWN;

		return $type;
	}

	/** returns a media code matching the given mime type or file extension.
	* File extensions are represented by a string starting with a dot (.) to
	* distinguish them from mime types.
	*
	* This funktion relies on the mapping defined by $this->mMediaTypes
	* @access private
	*/
	function findMediaType($extMime) {

		if (strpos($extMime,'.')===0) { #if it's an extension, look up the mime types
			$m= $this->getTypesForExtension(substr($extMime,1));
			if (!$m) return MEDIATYPE_UNKNOWN;

			$m= explode(' ',$m);
		}
		else { #normalize mime type
			if (isset($this->mMimeTypeAliases[$extMime])) {
				$extMime= $this->mMimeTypeAliases[$extMime];
			}

			$m= array($extMime);
		}

		foreach ($m as $mime) {
			foreach ($this->mMediaTypes as $type => $codes) {
				if (in_array($mime,$codes,true)) return $type;
			}
		}

		return MEDIATYPE_UNKNOWN;
	}
}

?>