A file download system

[oOSephOo]

I've created a file download system in (somewhat) OOP. It's just short of a thousand lines so I've stripped everything away but the part I feel is the slowest (I actually did profiling, even though it's not needed because it's clearly slow)

What I want critique on is if there's a better(read faster) way to do it, I've already tried to optimize it as much as possible but it's still running slow on my local comp. (a real server is faster of course)

Here goes: http://pastebin.dk/index.php?show=2027

I included everything relevant, supporting resuming etc, but the timekiller is the actual fread statement, I did it this way to avoid the overhead of calling a var or a function to check if the file was done being read and it helped a lot.

[jazz_snob]

There is a problem I've seen first hand when having PHP force download of large files that a bunch of memory is eaten by apache and never released; if I find the link discussing it I'll post it. I've given up on php downloads altogether for now and instead, if necessary, stick in a php script to do whatever voodoo (like counter increment, etc) then do a header() to the location of the file and let apache do the download. Using .htaccess you can prevent direct access to the downloadables if necessary.

[oOSephOo]

Hmm indeed, I'd imagine ob_flush(); and flush(); will deal with that though. Will have to test this on larger files.