Forced download issue with IE6

[Simon Rigby]

Hi folks,

I appreciate that this topic of forced downloads has been discussed in
various threads but my situtation is a little different to those that I
have come across and am experiencing a wierd issue in IE6.

Essentially I am executing a select query against a MySql database and
taking the result of that query and building a csv stream on the fly to
send to the browser as a forced download.

I have a function which has been cobbled together from a few internet
sources. Code below:

function ForceDownload ($data, $name, $filesize=false ) {

if ($filesize == false OR !is_numeric($fi lesize)) {
$filesize = strlen($data);
}

$mimetype = 'application/force-download';

header("Pragma: public");
header("Expires : 0");
header("Cache-Control: must-revalidate, post-check=0,
pre-check=0");
header("Cache-Control: private",false) ;
header("Content-Transfer-Encoding: binary");
header("Content-Type: " . $mimetype);
header("Content-Length: " . $filesize);
header("Content-Disposition: attachment; filename=\"" . $name .
"\";" );

echo $data;
die();
}

I am calling it from my page like so:

$result = GetContacts($se archTerm, $searchType);
$data = '';
while($row = mysql_fetch_ass oc($result)) {
$data .= $row['ContactName'] . ',' . $row['Email'];
}
ForceDownload($ data, 'contactExport. csv');

This works fine in Firefox but IE6 refuses to do anything at all. It
doesn't even display the content of the stream inline (ie in the
browser window).

Any ideas greatly appreciated.

Simon

[Rik]

Simon Rigby wrote:

Hi folks,

I appreciate that this topic of forced downloads has been discussed in
various threads but my situtation is a little different to those that
I have come across and am experiencing a wierd issue in IE6.

Essentially I am executing a select query against a MySql database and
taking the result of that query and building a csv stream on the fly
to send to the browser as a forced download.

I have a function which has been cobbled together from a few internet
sources. Code below:

I don't know wether it's the problem but this:

$mimetype = 'application/force-download';
header("Content-Type: " . $mimetype);

Is something that's very bad practise IMO.

This is normally enough for me:

header("Content-Disposition: attachment; filename=\"" . $name .
"\";" );

Grtz,
--
Rik Wasmus

[Simon Rigby]

I'll try that Rik. But I'm interested to know why its bad practice?

Simon

[Simon Rigby]

Hmm. Actually thats still not working. Works in Firefox still but
nothing in IE. Just a browser flash as the form posts and then nothing.
Well not striclty true. The form posts back to itself and the
underlying page is rendering correctly but no download.

Simon Rigby wrote:

I'll try that Rik. But I'm interested to know why its bad practice?

Simon

[John Dunlop]

Simon Rigby:

[re 'application/force-download' for CSVs]

But I'm interested to know why its bad practice?

because there's already a MIME type registered for CSV, text/csv.

'application/force-download' isn't registered, and this subtype doesn't
mark its proprietariness with 'x-' - going against the MIME spec.

and content-type 'indicates the media type of the entity-body sent to
the recipient' (RFC2616: 14.17); it doesn't disposition the message.

more details in the archived discussions. HTH

--
Jock

[Simon Rigby]

Ah right, yes I understand. I was under the impression that using
anything that wouldnt be recognised by the browser was part of
"forcing" the download. I get it!

Does anyone know why this is nt working in IE6.

I have cut down the code to just apply the content disposition header.
Still workin Firefox but no joy in IE.

Simon

John Dunlop wrote:

Simon Rigby:

[re 'application/force-download' for CSVs]

But I'm interested to know why its bad practice?

because there's already a MIME type registered for CSV, text/csv.

'application/force-download' isn't registered, and this subtype doesn't
mark its proprietariness with 'x-' - going against the MIME spec.

and content-type 'indicates the media type of the entity-body sent to
the recipient' (RFC2616: 14.17); it doesn't disposition the message.

more details in the archived discussions. HTH

--
Jock

[Jerry Stuckle]

Simon Rigby wrote:

Ah right, yes I understand. I was under the impression that using
anything that wouldnt be recognised by the browser was part of
"forcing" the download. I get it!

Does anyone know why this is nt working in IE6.

I have cut down the code to just apply the content disposition header.
Still workin Firefox but no joy in IE.

Simon

Probably because IE is ignoring the headers and doing what it wants -
it's notorious for that.

At this point, you'll probably get better responses in alt.html. This
doesn't look like anything PHP is doing.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
js*******@attgl obal.net
=============== ===

[Simon Rigby]

Thanks for the pointer Jerry.

Simon

Jerry Stuckle wrote:

Simon Rigby wrote:

Ah right, yes I understand. I was under the impression that using
anything that wouldnt be recognised by the browser was part of
"forcing" the download. I get it!

Does anyone know why this is nt working in IE6.

I have cut down the code to just apply the content disposition header.
Still workin Firefox but no joy in IE.

Simon

Probably because IE is ignoring the headers and doing what it wants -
it's notorious for that.

At this point, you'll probably get better responses in alt.html. This
doesn't look like anything PHP is doing.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
js*******@attgl obal.net
=============== ===

[s a n j a y]

Simon Rigby wrote:

Thanks for the pointer Jerry.

Simon

Jerry Stuckle wrote:

>Simon Rigby wrote:

>>Ah right, yes I understand. I was under the impression that using
anything that wouldnt be recognised by the browser was part of
"forcing" the download. I get it!

Does anyone know why this is nt working in IE6.

I have cut down the code to just apply the content disposition header.
Still workin Firefox but no joy in IE.

Simon

Probably because IE is ignoring the headers and doing what it wants -
it's notorious for that.

At this point, you'll probably get better responses in alt.html. This
doesn't look like anything PHP is doing.

--
============== ====
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
js*******@attgl obal.net
============== ====

This happened to me once while developing a J2EE application, but I
think it will apply here as well.

Try changing value for "expires". You can try making it something like
30 minutes to start with and then if that works, lower the value to few
seconds.

What IE does, in my opinion, is that it downloads the file first to disk
cache before it opens it. But as soon as download is finished, it
realizes that the content is supposed to be expire in 0 seconds, so it
doesn't display anything.

If I am correct, it used to give file not found error in older versions.
Try it and let me know. I am curious as well.

[Richard Levasseur]

Simon Rigby wrote:

Thanks for the pointer Jerry.

Simon

Jerry Stuckle wrote:

Simon Rigby wrote:

Ah right, yes I understand. I was under the impression that using
anything that wouldnt be recognised by the browser was part of
"forcing" the download. I get it!
>
Does anyone know why this is nt working in IE6.
>
I have cut down the code to just apply the content disposition header.
Still workin Firefox but no joy in IE.
>
Simon
>

Probably because IE is ignoring the headers and doing what it wants -
it's notorious for that.

At this point, you'll probably get better responses in alt.html. This
doesn't look like anything PHP is doing.

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
js*******@attgl obal.net
=============== ===

By and large, i find force-download and saying the content type is some
sort of other format (x-octet-stream, iirc) works fairly well in IE. I
always have had the sneaking suspicion IE caches any file it recognizes
regardless of what any header says, though.

If you resolve this _please_ post how you did it - this is a problem i
always run into with IE. The solution always seems to be different,
too.