Ðavîd Låndïs wrote:
...
Incidentally, does anyone know what exception is supposed to be thrown
when a user cancels the attachment download?
I am bit out of my depth here, but presumably when the user does that,
the browser closes its (reading) side of the socket. Whatever process is
at the other end writing to it should then get a "attempt to write to a
closed socket" kind of exception.
But if there is anything in-between, which does not really reflect the
condition up the chain, then it might be something else..
I mean if for instance there was a firewall, or a load-balancer, or an
Apache http + mod_jk (which I know you all don't have right now, but may
have in the future), then what you'd get at the Tomcat level is whatever
results from the socket Tomcat is really writing to, not necessarily the
browser directly.
I seem to remember a discussion 3-4 months ago on this forum about a
topic just like this one, ending finally in the consensus that it is the
JVM itself which does not really make available to the Java application
all the innards of the underlying OS-level socket.
It may do that for potability reasons, or maybe because the JVM
developers thought that Java programmers should not be concerned by such
low-level considerations.
Maybe Chuck remembers the topic of that thread, as I am sure he was part
of it.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
