How Confluence implement downloads

[Lawrence Robinson]

I am trying to create a 'hit-counter' of sorts and would like to know how confluence handles client downloads (server-side) when a link on a page is clicked. The counter is intended as a way of tracking how many time a piece of in-house software (a tarball linked to a page) is downloaded by customers.

More specifically my question is:

How does the confluence application know the following when a link in a page is clicked:

1. Whether the download is cancelled when the File Download dialog box is presented by the client browser
2. Whether there is some interruption in the connection before the download completes

[Matthew J. Horn · Accepted solution]

What you want to do is very complex because of the nature of HTTP and the cross-browser incompatiblities. There might be a JavaScript library that supports such features, so I would start there.

[Lawrence Robinson]

So I am not opposed in anyway to implementing a solution with JavaScript however, at this point I am more so interested specifically in what is the current implementation of initiating the download once a link is clicked.

For instance, is one of the servlet processes that is always waiting for such a request to come from any link in the running tomcat instance?

Or something more like a download server routine running?

[Lawrence Robinson]

I would like to further clarify that I am asking this:

Does confluence implement anything like a download manager to handle serving files that maybe embedded by way of a link?

or

Does clicking on an embedded link simply load the URL into the clients browser using default browser behavior?