I2P_Developers/i2p.itoopie
3
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

2006-02-15 jrandom

Browse Source 
* Merged in the i2p_0_6_1_10_PRE branch to the trunk, so CVS HEAD is no
      longer backwards compatible (and should not be used until 0.6.1.1 is
      out)
This commit is contained in:
jrandom
2006-02-15 05:33:17 +00:00
committed by zzz
parent 1374ea0ea1
commit 113fbc1df3
127 changed files with 2687 additions and 1309 deletions
Download Patch File Download Diff File Expand all files Collapse all files

124
history.txt
View File

@ -1,4 +1,126 @@
$Id: history.txt,v 1.395 2006/01/25 10:34:28 dust Exp $
$Id: history.txt,v 1.396.2.19 2006/02/15 00:16:31 jrandom Exp $
2006-02-15 jrandom
* Merged in the i2p_0_6_1_10_PRE branch to the trunk, so CVS HEAD is no
longer backwards compatible (and should not be used until 0.6.1.1 is
out)
2006-02-14 jrandom
* Syndie ui bugfixes (thanks all!)
2006-02-13 jrandom
* Use the current directory for some temporary I2PSnark files, rather than
the OS default temp dir (thanks anon!)
* Increase the base streaming lib window size (still shrinks to 1 on
retransmission though, of course)
* Fixed the I2PTunnel newlines to work with lighthttpd (thanks all!)
* Implement fast retransmit in the streaming lib (fires at most once per
packet), and increased the default ack delay to 2 seconds (from .5s)
* Don't ask for garlic level message acks for end to end messages unless
they're useful (e.g. to ack session tags)
2006-02-12 cervantes
* Use a different santisation method for some SML attributes
* Make router console update config save button actually save.
* Fix console bandwidth limiter burst rate dropdowns, so the display
relates to what is saved in the config.
2006-02-12 cervantes
* SML is now stricter in it's formatting (attributes should only use
double quotes instead of being allowed to mix with singles).
* Using apostrophes in SML attributes will no longer invalidate the tag.
* Some instances of [blog] tag description were not being displayed
correctly.
2006-02-12 jrandom
* Further SSU peer test throttling
* Put the most common router console features on the main index page too
2006-02-11 jrandom
* Be more careful about SSU peer test floods
2006-02-09 jrandom
* Adjusted one of the SSU timeouts so we don't drop peers as easily (duh)
2006-02-08 jrandom
* Added transparent support for VIA C3 CPUs to jbigi (thanks Nekow42), and
bundled a precompiled libjbigi.so in the jbigi.jar
* Cleaned up the synchronization for some SSU packet handling code
* Allow explicit rejection of more lagged tunnel build requests, rather
than dropping them outright
* Use lighter load testing
2006-02-07 jrandom
* Handle HTTP headers without any values (thanks Sugadude!)
* Don't show the option to make Syndie multiuser, since very few people
need it, and multiuser mode is a lot more complex to use. Geeks can
enable it by adding "syndie.singleUser=false" to syndie/syndie.config
(or in the router's advanced config, for the embedded Syndie)
* When a peer rejects participation in a tunnel, they mean it (duh)
* Decrease tunnel test timeout period to 20s (a 40s lag is insane)
* Remove a throttle on the size of the SSU active outbound pool, since
it was essentially arbitrary
* Use a more appropriate SSU bloom filter size
* Don't "proactively" drop SSU connections if we have partially received
inbound messages (duh)
* Migrate most of the message state across SSU connection reestablishment
2006-02-06 jrandom
* Reduce the SSU retransmit timeout range, and increase the number of ACKs
piggybacked
2006-02-05 jrandom
* Experiment with short exponents for DH/ElGamal, using a 226bit x instead
of a 2048bit x, as reports suggest that size is sufficient for 2048bit
DH/ElGamal when using safe primes (see KeyGenerator.java for references)
* Enable the messageHistory.txt by default, for debugging
2006-02-05 jrandom
* Substantial bugfix for the duplicate message detection in the transport
layer
* Handle tunnel build responses ASAP, rather than queueing them up to wait
in line (processing them is really fast - just a few AES loops)
* Don't bother handling build requests that we have queued up for a while
locally, as the requestor will have timed it out anyway (perhaps we
should reply regardless, but with a backoff instead?)
2006-02-04 jrandom
* Further tunnel test cleanup and disabling of the old tunnel creation
code
2006-02-04 jrandom
* Clean up and reenable the tunnel testing for the new tunnel system.
2006-02-04 jrandom
* Don't cache the archive.txt in syndie when fetching it through the web
interface.
* Logging updates
2006-02-03 jrandom
* Added further replay prevention on the tunnel build requests
* More aggressive streaming lib closing on reset
2006-02-03 jrandom
* More aggressive refusal of peers from the wrong network (oops)
2006-02-01 jrandom
* Instruct the router to reseed against a new URL, for migration purposes:
http://dev.i2p.net/i2pdb2/
* Aggressive error handling during UDP packet creation (thanks cervantes)
2006-02-01 jrandom
* Fix the new tunnel creation crypto, including the addition of a 4 byte
"next message ID" to the encrypted request structure in the spec.
* Backwards incompatible change, using the new tunnel creation crypto, the
fixed MD5 HMAC size, and a new network ID (to prevent cross pollination
with the old incompatible network).
* Reworked the leaseSet request process to handle a race condition
* Disable the TCP transport
* Run four separate threads on the job queue to cut down on job lag
2006-01-28 jrandom
* Removed a race that could show up in leaseSet requesting with the new
tunnel building process
2006-01-25 jrandom
* Run the peer profile coalescing/reorganization outside the job queue