by processing DSM inline in InNetMessagePool.
Caused random lookup failures of RI and LS.
Made worse by previous change removing duplicate store from reply job.
thanks obscuratus and drzed

Use a single ack timer for each session rather than
creating a new one for every ack
Log tweaks

as the console will do the lookup if not

i2pd will send this after retry with zero token and termination

mark session dead after receiving destroy
log tweaks

ref: http://git.idk.i2p/mesh/i2p.i2p/-/issues/6

(i2pd bug, fixed today)

Add RTTdev to logging

(more prep for connection migration)

As fallback for failed lookup by RemoteHostId.
This will eventually be the primary lookup after most traffic is SSU2.
For now, only used to receive traffic from a peer that changed IP/port.
SSU 1/2 send destroy when replacing old session with new.
Prep for full SSU2 connection migration.

Comment out unused contact property

Retransmit handshake messages faster and expire faster (SSU 1/2)
log tweaks

Use minimum MTU for IPv6 session confirmed to avoid PMTU issues
Set default MTU for known IPv6 tunnel brokers
Check for s mismatch in session confirmed RI
Don't put DateTime block after Termination block
Fix first message failed check
Log sent data packet size
Log tweaks

Set max MTU for range48.org brokered addresses
Limit MTU for brokered addresses in UDPAddress also

Reduce max consecutive message failures to fail session (SSU 1/2)
Set peer as unreachable for these failures
Reduce max consectutive session confirmed retransmissions
Fix session confirmed retransmission timing
Respond to relay tag request in data phase
Log tweaks

was treating them as acks

not simply when a new file is downloaded

Move establisher and packet handler initialization
to after the address initialization