cURL使用Mime类型检查将文件保存到服务器

Question

I am creating an internal tool for my team that will allow trusted team members to save remote files to our server via php and curl. I have the open, write, and close working perfectly, but I would like to add a check to make sure the file is of a certain mime type before it creates and writes the local file.

How could I do this, based on an array of mime types?

$ch = curl_init();
$fp = fopen($local_file, 'w+');
$ch = curl_init($remote_file);
curl_setopt($ch, CURLOPT_TIMEOUT, 50);
curl_setopt($ch, CURLOPT_FILE, $fp);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1);
curl_setopt($ch, CURLOPT_ENCODING, "");
curl_exec($ch);
curl_close($ch);
fclose($fp);

Answer 1

I solved this by checking the mime via fileinfo after the file has been transfered. If it is not a valid mime type, then I remove it.

$ch = curl_init();
$fp = fopen($local_file, 'w+');
$ch = curl_init($remote_file);
curl_setopt($ch, CURLOPT_TIMEOUT, 50);
curl_setopt($ch, CURLOPT_FILE, $fp);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1);
curl_setopt($ch, CURLOPT_ENCODING, "");
curl_exec($ch);
curl_close($ch);
fclose($fp);

$finfo = new finfo(FILEINFO_MIME);
$mime_type = $finfo->file($local_file);

if (strpos($mime_type, 'application/xml') === false) {
    unlink($local_file);
}

Answer 2

For example:

$ch = curl_init('http://static.php.net/www.php.net/images/php.gif');
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_exec($ch);

$mime = curl_getinfo($ch, CURLINFO_CONTENT_TYPE);

in $mime has mime type of file.

Answer 3

Tricky! You could download the whole file first (into memory, or a temporary folder.) If you want to stream it, you may have to:

1. set CURLOPT_HEADER to include the HTTP response headers in the data you get back

2. set CURLOPT_WRITEFUNCTION instead of CURLOPT_FILE, read and parse the http headers to see the mime type, and then decide if you're going to create/write a file or not.

Obviously this is quite a bit of work, as you'll have to do some basic parsing of the HTTP headers, and possibly buffering to get the whole HTTP headers at once.

Hopefully someone will post an easier solution.

pseudocode:

state = headers
buf = ''
fd = null
func writefunc(ch, data)
   if state is headers
      buf .= data
      div = buf.strpos "\r\n\r\n"
      if div !== false
         mime = get_mime buf, div
         if mime_ok mime
            fd = fopen ...
            fd.write buf.substr div+4
            state = saving
         else
            # returns other than data.length() abort connection
            return 0
  else
     fd.write data
  return data.length()