I'm having a hard time figuring out the solution to a problem that I thought would be very common or straight-forward to solve, but apparently I was wrong.
I have to re-write some web code (that I didn't write) that's causing a problem. When a user clicks a link, a request is sent to the web server, which in turn fetches or creates a PDF document from somewhere. The PDF data is returned with the Content-Disposition
header set to attachment
, and the browser shows the save-as dialog.
The reason the save-as dialog appears is because when the user clicks the link, the Javascript sets window.location.href
to the server URL (with some parameters).
There's no loading animation other than the one the browser shows in the tab etc. while the request is being processed.
The problem is that if a request hangs or takes a while, users tend to click the link again (possibly multiple times) which means requests for that same resource just keep building up on the server (even accidental double clicks on a link, which are common, cause two requests to be processed).
How can I prevent this from happening? If I do something like this (with window.location.href
replaced by window.open
:
var REQUEST_PENDING = false;
function getPDF(param1, param2) {
if (REQUEST_PENDING) return;
REQUEST_PENDING = true;
var w = window.open("/GetPdf.servlet?param1="+param1+"¶m2="+param2);
w.onload = function() {
DOC_REQUEST_PENDING = false;
}
}
...then only one request will be processed at any one time, but the onload callback only works if the return content is HTML. When it's an attachment, which is what I have, the DOC_REQUEST_PENDING variable is never set back to false, so no further requests can be made.
I know that the ultimate solution should probably be implemented server-side, but is it not possible to achieve what I'm trying to do client-side? (I can use jQuery).