jrandom authored 20 years ago and zzz committed 20 years ago

parseParams throws exception on bad formatting, and its perfectly valid to have params with 0 values (e.g. DEST GENERATE\n)

jrandom authored 20 years ago and zzz committed 20 years ago

* don't create excess I2PAppContexts (if any old context will do, use the global)
keep track of keys per spec (when DESTINATION=blah, create (or reuse) the destination private
keys).  we still need to persist this data though.
* the DESTINATION in the SESSION STATUS is now the same as the one sent in the
SESSION CREATE, /not/ the base64 of the private key, per spec
* enum is a reserved word in 1.5, so s/enum/names/ for future compatability
* logging

jrandom authored 20 years ago and zzz committed 20 years ago

initial tests for HELLO and create session (style=stream).  covers the basics, but doesn't cover a single normal scenario yet

jrandom authored 20 years ago and zzz committed 20 years ago

removed nested synchronization (which had been causing undetected deadlocks)
made sync blocks smaller, though this may have opened holes related to
resent ACK/SYN/CLOSE packets that are delivered in a race.  I'm not as
fluent in the ministreaming lib code as i should be (yet), but duck's thread
dumps were showing hundreds of threads waiting on a lock that'll never get
released (since the only way to release it would be to receive another
packet, and no more packets can be received until the lock is released, etc)
also, I2PSession is threadsafe - i can see no reason to synchronize on it
(and it was being synchronized on only part of the time?)
also, refactored the charset encoding stuff and minor log tweaking
i've been testing this for the last hour or so, on eepsites and squid (large
and small files), as well as irc, and there haven't been any glitches.  but
it needs more testing before it can be released, obviously.

jrandom authored 20 years ago and zzz committed 20 years ago

include the timestamper (and fire the client on startup, using a pair of arbitrary NTP hosts - to be configured later)

jrandom authored 20 years ago and zzz committed 20 years ago

fire the LoadClientAppsJob right after the admin listener is booted, which now includes support for the onBoot property (which causes the client to run immediately, instead of waiting 2+ minutes)
(yeah, it'd suck if all routers started up, tried to connect to people, got shitlisted, then 2 minutes later got the right NTP time, 'eh?)

jrandom authored 20 years ago and zzz committed 20 years ago

yes, we'll want to filter the access to the admin manager ;)

jrandom authored 20 years ago and zzz committed 20 years ago

this stat is published in the netDb, but the quantity fields (how many acks the stat is averaged over) is h4x0red
(it always reads "666", since otherwise it'd be fairly easy to identify what routers run servers, and i can live without knowing the quantity)

jrandom authored 20 years ago and zzz committed 20 years ago

logging

jrandom authored 20 years ago and zzz committed 20 years ago

dont dieFatal() if the first leaseSet lookup fails (this was the cause of those "fast fail" errors [bugId=69])
try up to 6 times to search (or until the expiration, etc)
logging

jrandom authored 20 years ago and zzz committed 20 years ago

(min/max frequency=1m/30m, doubling every time we don't find something new)
reduce the bredth and duration of explorations
republish less often (every 60s send out one random one [or however many explicit ones])
logging

jrandom authored 20 years ago and zzz committed 20 years ago

mark the ping datapoint before actually sending it, for those instances where a reply can be sent back faster than the ACK of the send

jrandom authored 20 years ago and zzz committed 20 years ago

include the (redundant) rtt in the output file to let tools that dont know how to combine columns do so
*cough*

jrandom authored 20 years ago and zzz committed 20 years ago

(and set the file modification date to the 'published on' date in the netDb object)