[RESOLVED] File Content Type

[dclamp]

is there an easy way to get the content type of a file?

I am working on a file sharing script and i need to get the file content type for download.

the only way i can think of is like having a database with all the different content types and getting it by the file extension...

[penagate]

You could use the MIMEType extension to sniff for a content type. It is more reliable however to store the content type when the file is uploaded if possible.

[k1ll3rdr4g0n]

If you are on a *nix system, you can tell the script to execute:

file -bi file

And that will return the content type

:~$ file -bi hosts.allow
text/plain; charset=us-ascii

In fact this is what mediawiki does. Or can do I should say.

If you are on a windows box, there probably is a port of that program. And if you really wanted to get the content type, that would be the best way imo.

[dclamp]

You could use the MIMEType extension to sniff for a content type. It is more reliable however to store the content type when the file is uploaded if possible.

there are so many though. it would take forever to get them all.

[visualAd]

is there an easy way to get the content type of a file?

I am working on a file sharing script and i need to get the file content type for download.

the only way i can think of is like having a database with all the different content types and getting it by the file extension...

It is very difficult determine the exact file type of an external file; even when that file is just a text file. If the user uploading the file wants others to know the true file type then they will use the appropriate extension from which the browser can determine and send the correct MIME type for.

If the user wants to disguise the file type they can change the extension. To determine the true type you may need to open the file and check that it conforms to the appropriate format (e.g: XML,CSV). This is more difficult with binary files that have undocumented file formats.

Checking for executable files may involve looking for specific binary instruction combinations; but is by no means reliable and is even more difficult if the user has encoded or encrypted the file. In short; if you are doing this from a security perspective; the controls you can put in place a limited.

[penagate]

there are so many though. it would take forever to get them all.

Many what? A file can only have one content type. I'm not sure what you mean.


When a file is uploaded, the MIME type is sent by the client and available in the $_FILES array.

Maybe if you clarify what you are doing we can provide better assistance.

[dclamp]

oh. the $_FILES tells you what MIME type it is? i thought i would have to write code to figure out what MIME type it is

[penagate]

$_FILES[$index]['type']

[k1ll3rdr4g0n]

It is only available if its sent by the browser.
I don't know under what circumstances it doesn't, but I wouldn't touch it with a 6ft pole. I would use file -bi, that way you know for sure what the mime type is.

See http://www.php.net/manual/en/feature...load.php#53133
Also: http://www.php.net/manual/en/feature...load.php#52989
And: http://www.php.net/manual/en/feature...load.php#51256

[penagate]

You can't know for sure. All file with the -i flag does is use a 'magic' behaviour which is the same as the MIMEType and PECL/fileinfo extension. This is fallible; all type sniffing is. visualAd is correct — if this is for security purposes, there is not a lot you can realistically do.

Relying on the MIME type sent by the client is fine if you just need it to send back when streaming the file for download to another client. If the type is missing, you can simply use application/octet-stream to force a download. This is more than adequate for a file sharing service.

[dclamp]

i used application/octet-stream for a .doc file, and Word wasnt able to open it after i downloaded it. it said it was corrupt

[penagate]

That suggests you are not streaming it correctly...

[dclamp]

should i show you my code then?

[penagate]

At your leisure.

In fact, don't post it in code tags. Attach the .php file (if possible).

[dclamp]

thanks all. the $_FILES['file']['type'] solved it!

[visualAd]

Dylan, the first four bytes of the file you upload can also be used to verify the file type. A JPEG file for example should always start with "ff d8 ff e0", if it doesn't then you can be absolutely sure that the file is not a JPEG. That doesn't prevent other data from being hidden within it but is is another step you can use:

// example
$jpeg = chr(0xff) . chr(0xd8) . chr(0xff) . chr(0xe0);

$fhwnd = fopen('file.jpg', 'rb');

$signature = fread($fhwnd, 4);

if ($signature != $jpeg) {
    // wrong file type
} 

You can see some other common signatures here: http://www.garykessler.net/library/file_sigs.html. Watch out though, some file types may have several signatures.

[dclamp]

thanks for that vis. i will look into it!