Upload Probelms MIME type errors

Filmo

I've got some code that is designed to upload Microsoft Word and PDF documents to our server.

I'm using a MIME-type check to make sure that the file is a MS-Word document.

$MIME = 'application/msword|application/pdf';	// correct MIME type for uploaded files.
$file_type = $_FILES[$formname]['type'];

if (!preg_match("~($MIME)~",$file_type)) {	// $MIME contains '/'s, thus the use of '~'
echo "Wrong File Type....";
}

The problem is that it seems like some of the files that are being uploaded as "my_word.doc" are being recognized as 'application/octet-stream' instead of 'application/msword' and thus not failing the MIME type test.

Is the problem with the code and server or with the client's Browser/computer? (We had a perl script which used to upload the Word Files, so I know that the client files are actual word files). In fact the same client can 'retry' the upload and the same file will upload 'correctly' the second or third time around.

Any reason correctly formed Word documents are being recongized as /octet-stream instead of /msword???

stolzyboy

$file_type = $_FILES[$formname]['type'];

if ($file_type == "application/msword" || $file_type == "application/pdf") //take out the second condition if only want word doc
{
echo "allow upload for word and pdf";
}
else
{
echo "wrong file type";
}

Filmo

I'm not worried about my code. Your code and mine do the same thing.

I guess I should have expanded it out a bit further. When the file_type check function (your's or mine) 'fails', I am currently having it dump out an error message.

echo "File Type of attempted upload was: $file_type";

In order to see what the script was reading the file as since it is not seeing it as 'application/msword' or 'application/pdf'.

In this case, consistently, the results of the dump are:

File Type of attempted upload was: application/octet-stream

I know that the clients are uploading correctly named documents, (e.g. 'document.doc'). In fact, if they try to re-upload the file again, it often times succeeds.

Why would PHP think that a file with extension ".doc" is an octet-stream??

macosxguy

For what it's worth, I'm having the same problem. When a .doc or .xls is uploaded by a user on Safari/Mac, the file-type is shown correctly (application/msword or application/vdn.ms-excel) but when an IE/Win user uploads, it is always application/octet-stream.

Whassup with that??

Weedpacket

The only way to really be sure what type a file is is by direct examination: anything else (file extension, MIME type) can be wrong or a flat-out lie. PDF files start with the string "%PDF", while Microsoft Word documents - um, I'm not too sure, 'cos Microsoft likes to keep the file format "secret", but looking at a couple they have the string "Microsoft Word Document" toward the end. They also start with the character sequence D0 CF 11 E0 A1 B1 A1 E1, but I don't have enough MS Office-generated files around to know if that's common to all Word documents (of some version or other) or if Excel spreadsheets or Powerpoint presentations have it as well.