* v2.1.0.19
* enhanced error logging, fix seniority marker join blocker, fix sync message size limit defaults
* resolve signature failure
* additional error logging for merge-related signatures
* fix: one-shot sync message size, app shard TC signature size, collector/hotstuff race condition, expired joins blocking new joins due to pruning disable
* remove compat with old 2.0.0 blossomsub
* fix: resolve abandoned prover joins
* reload prover registry
* fix stale worker proposal edge
* add full sanity check on join before submitting to identify bug
* resolve non-fallthrough condition that should be fallthrough
* fix: resolve rare SIGFPE, fix orphan expired joins blocking workers from reallocating
* add reconnect fallback if no peers are found with variable reconnect time (#511)
Co-authored-by: Tyler Sturos <55340199+tjsturos@users.noreply.github.com>
* update base peer count to 1 (#513)
* fix: expired prover join frames, starting port ranges, proposer getting stuck, and seniority on joins
* fix: panic on shutdown, libp2p discovery picking inaccessible peers, coverage event check not in shutdown logic, amend app shard worker behavior to mirror global for prover root reconciliation
* fix: shutdown scenario quirks, reload hanging
* fix: do not bailout early on shutdown of coverage check
* fix: force registry refresh on worker waiting for registration
* add more logging to wait for prover
* fix: worker manager refreshes the filter on allocation, snapshots blocking close on shutdown
* tweak: force shutdown after five seconds for app worker
* fix: don't loop when shutting down
* fix: slight reordering, also added named workers to trace hanging shutdowns
* use deterministic key for peer id of workers to stop flagging workers as sybil attacks
* fix: remove pubsub stop from app consensus engine as it shouldn't manage pubsub lifecycle, integrate shutdown context to PerformSync to prevent stuck syncs from halting respawn
* fix: blossomsub pubsub interface does not properly track subscription status
* fix: subscribe order to avoid nil panic
* switch from dnsaddr to dns4
* add missing quic-v1
* additional logging to isolate respawn quirks
* fix: dnsaddr -> dns4 for blossomsub
* sort-of fix: apply sledgehammer to restart logic
* fix: restore proper respawn logic, fix frozen hypergraph post respawn, unsubscribe from bitmask previously missing
---------
Co-authored-by: winged-pegasus <55340199+winged-pegasus@users.noreply.github.com>
Co-authored-by: Tyler Sturos <55340199+tjsturos@users.noreply.github.com>
