X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FMimeMagic.php;h=cde30bc009978117c0550ec0d7558a7737f1de69;hb=92ec8a2bec1c34474e5450eec34203c19f253908;hp=77a306214a679070911008cf697a1fb741b60efd;hpb=d3ddd6efdbfefa8065a2abcdbcb019b3effd209b;p=lhc%2Fweb%2Fwiklou.git
diff --git a/includes/MimeMagic.php b/includes/MimeMagic.php
index 77a306214a..cde30bc009 100644
--- a/includes/MimeMagic.php
+++ b/includes/MimeMagic.php
@@ -1,23 +1,42 @@
getTypesForExtension( $ext );
- if ( is_null( $m ) ) return NULL;
+ if ( is_null( $m ) ) return null;
$m = trim( $m );
$m = preg_replace( '/\s.*$/', '', $m );
@@ -309,7 +349,7 @@ class MimeMagic {
$ext = $this->getExtensionsForType( $mime );
if ( !$ext ) {
- return NULL; //unknown
+ return null; //unknown
}
$ext = explode( ' ', $ext );
@@ -358,28 +398,84 @@ class MimeMagic {
'bmp', 'tiff', 'tif', 'jpc', 'jp2',
'jpx', 'jb2', 'swc', 'iff', 'wbmp',
'xbm',
-
+
// Formats we recognize magic numbers for
- 'djvu', 'ogg', 'mid', 'pdf', 'wmf', 'xcf',
-
+ 'djvu', 'ogx', 'ogg', 'ogv', 'oga', 'spx',
+ 'mid', 'pdf', 'wmf', 'xcf', 'webm', 'mkv', 'mka',
+
// XML formats we sure hope we recognize reliably
'svg',
);
return in_array( strtolower( $extension ), $types );
}
+ /** improves a mime type using the file extension. Some file formats are very generic,
+ * so their mime type is not very meaningful. A more useful mime type can be derived
+ * by looking at the file extension. Typically, this method would be called on the
+ * result of guessMimeType().
+ *
+ * Currently, this method does the following:
+ *
+ * If $mime is "unknown/unknown" and isRecognizableExtension( $ext ) returns false,
+ * return the result of guessTypesForExtension($ext).
+ *
+ * If $mime is "application/x-opc+zip" and isMatchingExtension( $ext, $mime )
+ * gives true, return the result of guessTypesForExtension($ext).
+ *
+ * @param $mime String: the mime type, typically guessed from a file's content.
+ * @param $ext String: the file extension, as taken from the file name
+ *
+ * @return string the mime type
+ */
+ function improveTypeFromExtension( $mime, $ext ) {
+ if ( $mime === "unknown/unknown" ) {
+ if( $this->isRecognizableExtension( $ext ) ) {
+ wfDebug( __METHOD__. ": refusing to guess mime type for .$ext file, " .
+ "we should have recognized it\n" );
+ } else {
+ /* Not something we can detect, so simply
+ * trust the file extension */
+ $mime = $this->guessTypesForExtension( $ext );
+ }
+ }
+ else if ( $mime === "application/x-opc+zip" ) {
+ if ( $this->isMatchingExtension( $ext, $mime ) ) {
+ /* A known file extension for an OPC file,
+ * find the proper mime type for that file extension */
+ $mime = $this->guessTypesForExtension( $ext );
+ } else {
+ wfDebug( __METHOD__. ": refusing to guess better type for $mime file, " .
+ ".$ext is not a known OPC extension.\n" );
+ $mime = "application/zip";
+ }
+ }
+
+ if ( isset( $this->mMimeTypeAliases[$mime] ) ) {
+ $mime = $this->mMimeTypeAliases[$mime];
+ }
+
+ wfDebug(__METHOD__.": improved mime type for .$ext: $mime\n");
+ return $mime;
+ }
/** mime type detection. This uses detectMimeType to detect the mime type of the file,
* but applies additional checks to determine some well known file formats that may be missed
- * or misinterpreter by the default mime detection (namely xml based formats like XHTML or SVG).
+ * or misinterpreter by the default mime detection (namely XML based formats like XHTML or SVG,
+ * as well as ZIP based formats like OPC/ODF files).
*
- * @param string $file The file to check
- * @param mixed $ext The file extension, or true to extract it from the filename.
- * Set it to false to ignore the extension.
+ * @param $file String: the file to check
+ * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
+ * Set it to false to ignore the extension. DEPRECATED! Set to false, use
+ * improveTypeFromExtension($mime, $ext) later to improve mime type.
*
* @return string the mime type of $file
*/
function guessMimeType( $file, $ext = true ) {
+ if( $ext ) { # TODO: make $ext default to false. Or better, remove it.
+ wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
+ "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
+ }
+
$mime = $this->doGuessMimeType( $file, $ext );
if( !$mime ) {
@@ -391,38 +487,42 @@ class MimeMagic {
$mime = $this->mMimeTypeAliases[$mime];
}
- wfDebug(__METHOD__.": final mime type of $file: $mime\n");
+ wfDebug(__METHOD__.": guessed mime type of $file: $mime\n");
return $mime;
}
-
- function doGuessMimeType( $file, $ext = true ) {
+
+ private function doGuessMimeType( $file, $ext ) { # TODO: remove $ext param
// Read a chunk of the file
wfSuppressWarnings();
$f = fopen( $file, "rt" );
wfRestoreWarnings();
if( !$f ) return "unknown/unknown";
$head = fread( $f, 1024 );
+ fseek( $f, -65558, SEEK_END );
+ $tail = fread( $f, 65558 ); // 65558 = maximum size of a zip EOCDR
fclose( $f );
+ wfDebug( __METHOD__ . ": analyzing head and tail of $file for magic numbers.\n" );
+
// Hardcode a few magic number checks...
$headers = array(
// Multimedia...
'MThd' => 'audio/midi',
'OggS' => 'application/ogg',
-
+
// Image formats...
// Note that WMF may have a bare header, no magic number.
"\x01\x00\x09\x00" => 'application/x-msmetafile', // Possibly prone to false positives?
"\xd7\xcd\xc6\x9a" => 'application/x-msmetafile',
- 'PDF%' => 'application/pdf',
+ '%PDF' => 'application/pdf',
'gimp xcf' => 'image/x-xcf',
-
+
// Some forbidden fruit...
'MZ' => 'application/octet-stream', // DOS/Windows executable
"\xca\xfe\xba\xbe" => 'application/octet-stream', // Mach-O binary
"\x7fELF" => 'application/octet-stream', // ELF binary
);
-
+
foreach( $headers as $magic => $candidate ) {
if( strncmp( $head, $magic, strlen( $magic ) ) == 0 ) {
wfDebug( __METHOD__ . ": magic header in $file recognized as $candidate\n" );
@@ -430,17 +530,37 @@ class MimeMagic {
}
}
+ /* Look for WebM and Matroska files */
+ if( strncmp( $head, pack( "C4", 0x1a, 0x45, 0xdf, 0xa3 ), 4 ) == 0 ) {
+ $doctype = strpos( $head, "\x42\x82" );
+ if( $doctype ) {
+ // Next byte is datasize, then data (sizes larger than 1 byte are very stupid muxers)
+ $data = substr($head, $doctype+3, 8);
+ if( strncmp( $data, "matroska", 8 ) == 0 ) {
+ wfDebug( __METHOD__ . ": recognized file as video/x-matroska\n" );
+ return "video/x-matroska";
+ } else if ( strncmp( $data, "webm", 4 ) == 0 ) {
+ wfDebug( __METHOD__ . ": recognized file as video/webm\n" );
+ return "video/webm";
+ }
+ }
+ wfDebug( __METHOD__ . ": unknown EBML file\n" );
+ return "unknown/unknown";
+ }
+
/*
- * look for PHP
- * Check for this before HTML/XML...
- * Warning: this is a heuristic, and won't match a file with a lot of non-PHP before.
- * It will also match text files which could be PHP. :)
+ * Look for PHP. Check for this before HTML/XML... Warning: this is a
+ * heuristic, and won't match a file with a lot of non-PHP before. It
+ * will also match text files which could be PHP. :)
+ *
+ * FIXME: For this reason, the check is probably useless -- an attacker
+ * could almost certainly just pad the file with a lot of nonsense to
+ * circumvent the check in any case where it would be a security
+ * problem. On the other hand, it causes harmful false positives (bug
+ * 16583). The heuristic has been cut down to exclude three-character
+ * strings like " 126 ) {
- printf( "\\x%02x", $c );
- } else {
- print $head{$i};
- }
- }
- echo "\n";
- */
- }
-
- if( $xml_type == 'UTF-16BE' || $xml_type == 'UTF-16LE' ) {
- // Quick and dirty fold down to ASCII!
- $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
- $chars = unpack( $pack[$xml_type], substr( $head, 2 ) );
- $head = '';
- foreach( $chars as $codepoint ) {
- if( $codepoint < 128 ) {
- $head .= chr( $codepoint );
- } else {
- $head .= '?';
- }
- }
- }
-
- $match = array();
- $doctype = "";
- $tag = "";
-
- if ( preg_match( '%%siD',
- $head, $match ) ) {
- $doctype = $match[1];
- }
-
- if( $xml_type || $doctype ) {
- if ( preg_match( '%<(\w+)\b%si', $head, $match ) ) {
- $tag = $match[1];
- }
-
- #print "
ANALYSING $file: doctype= $doctype; tag= $tag
";
-
- if ( strpos( $doctype, "-//W3C//DTD SVG" ) === 0 ) {
- return "image/svg+xml";
- } elseif ( $tag === "svg" ) {
- return "image/svg+xml";
- } elseif ( strpos( $doctype, "-//W3C//DTD XHTML" ) === 0 ) {
- return "text/html";
- } elseif ( $tag === "html" ) {
- return "text/html";
+ $xml = new XmlTypeCheck( $file );
+ if( $xml->wellFormed ) {
+ global $wgXMLMimeTypes;
+ if( isset( $wgXMLMimeTypes[$xml->getRootElement()] ) ) {
+ return $wgXMLMimeTypes[$xml->getRootElement()];
} else {
- /// Fixme -- this would be the place to allow additional XML type checks
- return "application/xml";
+ return 'application/xml';
}
}
/*
* look for shell scripts
*/
- $script_type = NULL;
+ $script_type = null;
# detect by shebang
if ( substr( $head, 0, 2) == "#!" ) {
@@ -541,7 +603,17 @@ class MimeMagic {
if ( $script_type ) {
if ( $script_type !== "UTF-8" && $script_type !== "ASCII") {
- $head = iconv( $script_type, "ASCII//IGNORE", $head);
+ // Quick and dirty fold down to ASCII!
+ $pack = array( 'UTF-16BE' => 'n*', 'UTF-16LE' => 'v*' );
+ $chars = unpack( $pack[$script_type], substr( $head, 2 ) );
+ $head = '';
+ foreach( $chars as $codepoint ) {
+ if( $codepoint < 128 ) {
+ $head .= chr( $codepoint );
+ } else {
+ $head .= '?';
+ }
+ }
}
$match = array();
@@ -552,17 +624,21 @@ class MimeMagic {
return $mime;
}
}
-
+
+ // Check for ZIP variants (before getimagesize)
+ if ( strpos( $tail, "PK\x05\x06" ) !== false ) {
+ wfDebug( __METHOD__.": ZIP header present in $file\n" );
+ return $this->detectZipType( $head, $tail, $ext );
+ }
+
wfSuppressWarnings();
$gis = getimagesize( $file );
wfRestoreWarnings();
-
+
if( $gis && isset( $gis['mime'] ) ) {
$mime = $gis['mime'];
wfDebug( __METHOD__.": getimagesize detected $file as $mime\n" );
return $mime;
- } else {
- return false;
}
// Also test DjVu
@@ -571,27 +647,133 @@ class MimeMagic {
wfDebug( __METHOD__.": detected $file as image/vnd.djvu\n" );
return 'image/vnd.djvu';
}
+
+ return false;
+ }
+
+ /**
+ * Detect application-specific file type of a given ZIP file from its
+ * header data. Currently works for OpenDocument and OpenXML types...
+ * If can't tell, returns 'application/zip'.
+ *
+ * @param $header String: some reasonably-sized chunk of file header
+ * @param $tail String: the tail of the file
+ * @param $ext Mixed: the file extension, or true to extract it from the filename.
+ * Set it to false (default) to ignore the extension. DEPRECATED! Set to false,
+ * use improveTypeFromExtension($mime, $ext) later to improve mime type.
+ *
+ * @return string
+ */
+ function detectZipType( $header, $tail = null, $ext = false ) {
+ if( $ext ) { # TODO: remove $ext param
+ wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. " .
+ "Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
+ }
+
+ $mime = 'application/zip';
+ $opendocTypes = array(
+ 'chart-template',
+ 'chart',
+ 'formula-template',
+ 'formula',
+ 'graphics-template',
+ 'graphics',
+ 'image-template',
+ 'image',
+ 'presentation-template',
+ 'presentation',
+ 'spreadsheet-template',
+ 'spreadsheet',
+ 'text-template',
+ 'text-master',
+ 'text-web',
+ 'text' );
+
+ // http://lists.oasis-open.org/archives/office/200505/msg00006.html
+ $types = '(?:' . implode( '|', $opendocTypes ) . ')';
+ $opendocRegex = "/^mimetype(application\/vnd\.oasis\.opendocument\.$types)/";
+
+ $openxmlRegex = "/^\[Content_Types\].xml/";
+
+ if( preg_match( $opendocRegex, substr( $header, 30 ), $matches ) ) {
+ $mime = $matches[1];
+ wfDebug( __METHOD__.": detected $mime from ZIP archive\n" );
+ } elseif( preg_match( $openxmlRegex, substr( $header, 30 ) ) ) {
+ $mime = "application/x-opc+zip";
+ # TODO: remove the block below, as soon as improveTypeFromExtension is used everywhere
+ if( $ext !== true && $ext !== false ) {
+ /** This is the mode used by getPropsFromPath
+ * These mime's are stored in the database, where we don't really want
+ * x-opc+zip, because we use it only for internal purposes
+ */
+ if( $this->isMatchingExtension( $ext, $mime) ) {
+ /* A known file extension for an OPC file,
+ * find the proper mime type for that file extension */
+ $mime = $this->guessTypesForExtension( $ext );
+ } else {
+ $mime = "application/zip";
+ }
+ }
+ wfDebug( __METHOD__.": detected an Open Packaging Conventions archive: $mime\n" );
+ } else if( substr( $header, 0, 8 ) == "\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1" &&
+ ($headerpos = strpos( $tail, "PK\x03\x04" ) ) !== false &&
+ preg_match( $openxmlRegex, substr( $tail, $headerpos + 30 ) ) ) {
+ if( substr( $header, 512, 4) == "\xEC\xA5\xC1\x00" ) {
+ $mime = "application/msword";
+ }
+ switch( substr( $header, 512, 6) ) {
+ case "\xEC\xA5\xC1\x00\x0E\x00":
+ case "\xEC\xA5\xC1\x00\x1C\x00":
+ case "\xEC\xA5\xC1\x00\x43\x00":
+ $mime = "application/vnd.ms-powerpoint";
+ break;
+ case "\xFD\xFF\xFF\xFF\x10\x00":
+ case "\xFD\xFF\xFF\xFF\x1F\x00":
+ case "\xFD\xFF\xFF\xFF\x22\x00":
+ case "\xFD\xFF\xFF\xFF\x23\x00":
+ case "\xFD\xFF\xFF\xFF\x28\x00":
+ case "\xFD\xFF\xFF\xFF\x29\x00":
+ case "\xFD\xFF\xFF\xFF\x10\x02":
+ case "\xFD\xFF\xFF\xFF\x1F\x02":
+ case "\xFD\xFF\xFF\xFF\x22\x02":
+ case "\xFD\xFF\xFF\xFF\x23\x02":
+ case "\xFD\xFF\xFF\xFF\x28\x02":
+ case "\xFD\xFF\xFF\xFF\x29\x02":
+ $mime = "application/vnd.msexcel";
+ break;
+ }
+
+ wfDebug( __METHOD__.": detected a MS Office document with OPC trailer\n");
+ } else {
+ wfDebug( __METHOD__.": unable to identify type of ZIP archive\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 $ext is not false, the mime type is guessed from the file extension, using
+ * If the dections fails and $ext is not false, 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 mixed $ext The file extension, or true to extract it from the filename.
- * Set it to false to ignore the extension.
+ * @param $file String: the file to check
+ * @param $ext Mixed: the file extension, or true (default) to extract it from the filename.
+ * Set it to false to ignore the extension. DEPRECATED! Set to false, use
+ * improveTypeFromExtension($mime, $ext) later to improve mime type.
*
* @return string the mime type of $file
* @access private
*/
- function detectMimeType( $file, $ext = true ) {
+ private function detectMimeType( $file, $ext = true ) {
global $wgMimeDetectorCommand;
- $m = NULL;
+ if( $ext ) { # TODO: make $ext default to false. Or better, remove it.
+ wfDebug( __METHOD__.": WARNING: use of the \$ext parameter is deprecated. Use improveTypeFromExtension(\$mime, \$ext) instead.\n" );
+ }
+
+ $m = null;
if ( $wgMimeDetectorCommand ) {
$fn = wfEscapeShellArg( $file );
$m = `$wgMimeDetectorCommand $fn`;
@@ -638,7 +820,7 @@ class MimeMagic {
$m = strtolower( $m );
if ( strpos( $m, 'unknown' ) !== false ) {
- $m = NULL;
+ $m = null;
} else {
wfDebug( __METHOD__.": magic mime type of $file: $m\n" );
return $m;
@@ -677,13 +859,13 @@ class MimeMagic {
* @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
+ * @param $path String: 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.
+ * @param $mime String: 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 ) {
+ function getMediaType( $path = null, $mime = null ) {
if( !$mime && !$path ) return MEDIATYPE_UNKNOWN;
# If mime type is unknown, guess it
@@ -716,7 +898,7 @@ class MimeMagic {
}
# Check for entry for file extension
- $e = NULL;
+ $e = null;
if ( $path ) {
$i = strrpos( $path, '.' );
$e = strtolower( $i ? substr( $path, $i + 1 ) : '' );
@@ -755,7 +937,7 @@ class MimeMagic {
if ( !$m ) return MEDIATYPE_UNKNOWN;
$m = explode( ' ', $m );
- } else {
+ } else {
# Normalize mime type
if ( isset( $this->mMimeTypeAliases[$extMime] ) ) {
$extMime = $this->mMimeTypeAliases[$extMime];
@@ -774,6 +956,27 @@ class MimeMagic {
return MEDIATYPE_UNKNOWN;
}
-}
+ /**
+ * Get the MIME types that various versions of Internet Explorer would
+ * detect from a chunk of the content.
+ *
+ * @param $fileName String: the file name (unused at present)
+ * @param $chunk String: the first 256 bytes of the file
+ * @param $proposed String: the MIME type proposed by the server
+ */
+ public function getIEMimeTypes( $fileName, $chunk, $proposed ) {
+ $ca = $this->getIEContentAnalyzer();
+ return $ca->getRealMimesFromData( $fileName, $chunk, $proposed );
+ }
+ /**
+ * Get a cached instance of IEContentAnalyzer
+ */
+ protected function getIEContentAnalyzer() {
+ if ( is_null( $this->mIEAnalyzer ) ) {
+ $this->mIEAnalyzer = new IEContentAnalyzer;
+ }
+ return $this->mIEAnalyzer;
+ }
+}