Hi,
I want to parse the contents of a dynamic .csv file. However, this code (obviously with duff file names, etc.):
$socket = fsockopen("www.example.com", 443);
fwrite($socket, "GET /dynamicCsv.csv?param=value HTTP/1.1\r\n");
fwrite($socket, "Host: www.example.com\r\n");
fwrite($socket, "Connection: close\r\n");
fwrite($socket, "\r\n");
while(!feof($socket)) echo fgets($socket);
...just downloads the file.
Using openssl at command line I can type the exact same request and get the following response header sent back:
HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-disposition: attachment; filename=dynamicCsv.csv
Pragma: public
Cache-Control: max-age=0
Set-Cookie: SESSIONID=ASDFUHN023UIN0F; Path=/; Secure
Content-Type: text/csv;charset=UTF-8
Transfer-Encoding: chunked
Date: Thu, 04 Feb 2010 16:53:00 GMT
To avoid having the file download automatically I have tried to chop off the first character using substr
, in the hope that maybe the browser sees the response output by fgets
as a set of headers of the current document. That didn't work. Pursuing the same idea I also tried to insert a bunch of newlines before opening the socket, which didn't work either.
My two questions are:
- Why on earth does the browser think it should download the file?
- How can I stop it?
I should mention that I've only used Chrome so far, but I don't see why it would make a difference.
Many thanks, Andreas