logging mods
i really need to rewrite the tcp transport - the code is all functional, but the design sucks.
with the FIFO bandwidth limiter we could get away with a single 'send' thread rather than each TCPConnection having its own writer thread (but we'd still need the per-con reader thread, at least until nio is solid enough)
but maybe the rewrite can hold off until the AMOC implementation. we'll see
* add a new ClientWriterRunner thread (1 per I2CP connection) so that a client application that hangs or otherwise doesn't read from its i2cp socket quickly doesn't hang the whole router (since we've previously used the jobQueue for pushing I2CP messages). This may or may not clear the intermittent eepsite bug, but I'm not counting on it to (yet).
* update various points to deal with the client writer's operation (aka doSend won't throw IOException)
* logging
* lots and lots of metrics (yeah i know some of them vary based on the compiler)
- prepare the cache prior to use if you want to have the hash cache.
also fix the ejection policy to not clear the cache, but merely to remove sufficient values.
though maybe clearing the cache is the right thing to do so as to avoid ejection churn... hmm.
both of these fixes brought to you by the keen eyes of the one called mihi
close connections to peers who are so slow that they leave messages on the queue to expire
reduce the default max queue size per connection to 10 messages
(as always, this is a configurable param, via "i2np.tcp.maxQueuedMessages" in router.config)
2) implement an optimized 'should contain' algorithm, rather than being a wuss and building + comparing a BigInteger of the xor.
3) more unit tests
this stuff is called a *lot*, since we need to pick what bucket things go in all the time.
do i need to wrap the Input/Output streams we use to pipe data over the net with a verification wrapper for the messages?
e.g. prefix the serialization of all I2NPMessages sent on the wire with the SHA256 of that serialization and verify on read?
Ho hum, dunno. maybe its something else, but the ElG/AES+SessionTag already has integrity verification so the only thing I can
think of is a checksum error that got past TCP's checking and corrupted the AES stream.
see router.config.template mods and the new unit tests.
this implementation can cause starvation -
e.g. lots of 1KB writes will go through before a 32KB write if the queue is low and the bwlimiter only replenishes say, 16KBps
another impl would enforce a FIFO through thread wait/notify, etc, but would have the related overhead.
i dont know whether starvation situations will be the norm or the exception. i'm running this on a few routers so we'll see.
