Quoting "[EMAIL PROTECTED]" from ml.softs.gtk-gnutella.devel:
:#5 0x8118bae in node_shutdown_mode (n=0x8b7c6a8, delay=120) at nodes.c:1714
:1714 mq_shutdown(n->outq);
:(gdb) p *n
:$1 = {node_handle = 350, peermode = NODE_P_LEAF, start_peermode = :NODE_P_ULTRA,
: remove_msg = 0x8b7c6b4 "Compression flush failed: Stream error", ip =
:3575912895, port = 2255, guid = 0x0, proxy_ip = 0, proxy_port = 0, outq =
:0x0, searchq = 0x0, rx = 0x8ab95e0, routing_data = 0x0,
Thanks. It looks like the nagle timeout gets triggered for a node which
has already been removed, hence the assertion failure.
I looked briefly (I don't have much time to spend on this now) at
tx_deflate.c and mq.c and could not find a place where we would
not call tx_free() when freeing the outq, and tx_free() invoked
tx_deflate_destroy(), in which we take care of cancelling the nagle
timeout if it is instantiated. I'm clueless, and maybe my initial
hypothesis is wrong (but the null socket, searcha and outq pointers in
the node structure lead me to think that it's correct, only I haven't
nailed down the exact troublesome place).
So I have no idea what the problem can be, and I never personally encountered
it. I'll leave it to someone else to debug... anyone who can spend the
time required to track it down.
I'm sorry to not be able to help you out quickly here...
Raphael
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click
_______________________________________________
Gtk-gnutella-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/gtk-gnutella-devel
