On Mon, 14 Dec 1998, Jason Haar wrote:
> If I have a Squid-2.1 server talking to a HTTP/1.1 server, and the user
> (running Netscape/IE 4+) breaks their connection while downloading a really
> large URL, what happens the next time they go and try to download it?
Depends on your quick_abort settings and browser behavior.
> Will Squid cache the half-finished download
No. Squid cannot cache partial objects.
> and then use HTTP/1.1
> byte-ranges to get the rest? (assuming timestamps/etc imply the URL hasn't
> changed of course).
The browser itself may use range requests to fetch the missing part. Whether
that part will come from Squid depends on your quick_abort settings,
complexity of the range request, and other factors.
Alex.
