* Fix a long standing leak in I2PTunnel (hanging on to i2psocket objects)
* Fix a leak injected into the SimpleTimer
* Fix a race condition in the tunnel message handling
* Added throttles on how many I2PTunnel client connections we open at once
* Replaced some buffered streams in I2PTunnel with unbuffered streams, as
the streaming library used should take care of any buffering.
* Added a cache for some objects used in I2PTunnel, especially useful when
there are many short lived connections.
* Trimmed the SimpleTimer's processing a bit
* Added throttles on how many I2PTunnel client connections we open at once
* Replaced some buffered streams in I2PTunnel with unbuffered streams, as
the streaming library used should take care of any buffering.
* Added a cache for some objects used in I2PTunnel, especially useful when
there are many short lived connections.
* Trimmed the SimpleTimer's processing a bit
* Remove spurious flush calls from I2PTunnel, and work with the
I2PSocket's output stream directly (as the various implementations
do their own buffering).
* Another pass at a long standing JobQueue bug - dramatically simplify
the job management synchronization since we dont need to deal with
high contention (unlike last year when we had dozens of queue runners
going at once).
* Logging
* Strip the Referer, Via, and From headers completely, rather than
inserting a bogus value ("i2p"). This should help with the use of
SnipSnap and Geeklog (thanks nickster and DrWoo!)
* Don't go into a fast busy if an I2PTunnel 'server' is explicitly killed
(thanks mule!)
* Handle some more error conditions regarding abruptly closing sockets
(thanks Jonva!)
2004-09-30 jrandom
* Bundle the configuration necessary to run an eepsite out of the box
with Jetty - simply edit ./eepsite/docroot/index.html and give people
the key listed on the I2PTunnel configuration page, and its up.
* Router console cleanup, and some (off by default) tunnels -
smtp.postman.i2p (port 7659), pop.postman.i2p (port 7660), and
irc.baffled.i2p (port 7661)
* Disable the timestamper by default for all applications except the router
(enable via -Dtime.disabled=false)
* Simplify the retrieval of the full destination with text based browsers.
* Bundle the updated wrapper.config and hosts.txt in the i2pupdate.tar.bz2
* Write the native libraries to the current directory when they are loaded
from a resource, and load them from that file on subsequent runs (in
turn, we no longer *cough* delete the running libraries...)
* Added support for a graceful restart.
* Added new pseudo-shutdown hook specific to the router, allowing
applications to request tasks to be run when the router shuts down. We
use this for integration with the service manager, since otherwise a
graceful shutdown would cause a timeout, followed by a forced hard
shutdown.
* Handle a bug in the SimpleTimer with requeued tasks.
* Made the capacity calculator a bit more dynamic by not outright ignoring
the otherwise valid capacity data for a period with a single rejected
tunnel (except for the 10 minute period). In addition, peers with an
equal capacity are ordered by speed rather than by their hashes.
* Cleaned up the SimpleTimer, addressing some threading and synchronization
issues.
* When an I2PTunnel client or httpclient is explicitly closed, destroy the
associated session (unless there are other clients using it), and deal
with a closed session when starting a new I2PTunnel instance.
* Refactoring and logging.
* Address a race condition in the key management code that would manifest
itself as a corrupt router identity.
* Properly clear old transport addresses from being displayed on the old
console after soft restarts.
* Properly refuse to load the client applications more than once in the
same JVM.
* Added support for a graceful restart (a graceful shutdown followed by a
full JVM restart - useful for restarting client apps).
* More defensive programming, HTML cleanup, logging
* wrapper.config cleanup of duplicate lines
* Added some filters to the HTTP request, replacing the User-Agent,
Referrer, Via, and From headers, which helps until we have a more
comprehensive filtering system.
2004-09-03 jrandom
* Disabled the old listener on port 7655.
the /i2ptunnel/ page so they can easily be shared.
* By default, try to create any needed private key files (for server tunnels only)
* keep track of the I2PSession objects used by the I2PTunnel instances (only needed
for exposing the associated Destination)
* new config property to have a tunnel start on load (default=true), so tunnels, er, start on load
* use i2ptunnel.config instead of i2ptunnel.cfg (for consistency)
* minor refactoring
instead of the standard 'httpclient 4444' or 'httpclient 4444 squid.i2p', you can now specify a comma delimited list of outproxies:
'httpclient 4444 squid.i2p,www1.squid.i2p,www2.squid.i2p' and each individual http request goes through a randomly selected proxy
there are a few general issues with this, such as a lack of affinity (web applications that require a session to always come from a single IP address will break)
but it should work most of the time.
(this is what caused the runtime errors on sun jvms but not on kaffe)
((aka i slacked and didn't test sufficiently. off with my head))
this now builds and runs fine in sun 1.3-1.5 jvms, as well as kaffe
e.g.: -e "clientoptions tunnels.depthInbound=0" -e "httpclient 6666"
this updates so many files because they all need a reference to an I2PTunnel object on construction so they query tunnel.getClientOptions() instead of System.getProperties
ERR_REQUEST_DENIED now returns a 403 status code. Note that I'm assuming
this error occurs because the proxy is configured to not allow connection
to non-i2p destinations. If it's issued for other reasons we might want
to consider using 503 instead.
instead of 404 for everything.
ERR_DESTINATION_UNKNOWN now returns 503 instead of 404 with a slightly
more clear error message.
ERR_TIMEOUT now returns 504 gateway timeout.
* rather than have all jobs created hooked into the clock for offset updates, have the jobQueue stay hooked up and update any active jobs accordingly (killing a memory leak of a JobTiming objects - one per job)
* dont go totally insane during shutdown and log like mad (though the clientApp things still log like mad, since they don't know the router is going down)
* adjust memory buffer sizes based on real world values so we don't have to expand/contract a lot
* dont display things that are completely useless (who cares what the first 32 bytes of a public key are?)
* reduce temporary object creation
* use more efficient collections at times
* on shutdown, log some state information (ready/timed jobs, pending messages, etc)
* explicit GC every 10 jobs. yeah, not efficient, but just for now we'll keep 'er in there
* only reread the router config file if it changes (duh)
