The welist host publishes both A and AAAA records. On networks where the
IPv4 route is broken (connection stalls after the TCP handshake) while
IPv6 works, the default connector may pick the broken family and the
report request hangs.
When BAL_PUSHER_PREFER_IPV6 is truthy, the pusher now resolves the welist
host itself and pins the reqwest client to its first IPv6 address; the
original hostname is still used for the Host header and TLS SNI. When the
variable is unset (default) or no AAAA record exists, behavior is
completely unchanged.
Includes unit tests for the URL host/port parsing and documentation in
docs/07_deployment_and_ops.md.
- Add multi-stage Dockerfile with tini, non-root user, healthcheck
- Fix SQLite WAL mode race between bal-server and bal-pusher (busy_timeout + retry)
- Fix tbl_stats missing UNIQUE index for ON CONFLICT clause
- Replace unwrap() panics in bal-pusher with graceful error handling
- Add docker/entrypoint.sh with BAL_PUSHER_NETWORK support
- cargo fmt across all files
- xpub.rs: Replace convert_xpub() unwrap() with Result propagation
- xpub.rs: Replace calculate_fingerprint() unwrap() with Result propagation
- xpub.rs: Replace get_bitcoincore_descriptor() unwrap() with Result/match
- xpub.rs: Fix new_address_from_xpub() call in bal-server.rs to handle Result
- xpub.rs: Add prefix validation in convert_xpub (xpub/ypub/zpub/tpub/vpub/upub)
- bal-pusher.rs: Remove parse().unwrap() on env var bool, use unwrap_or(false)
- bal-pusher.rs: Replace port try_into().unwrap() with safe u16::try_from match
- bal-pusher.rs: Add error logging for invalid port values in env vars
- All tests pass: cargo test (5 tests: 3 SQL + 2 panic regression)
- Build verified: cargo check --bin=bal-server --bin=bal-pusher (0 errors)
- Replace all sqlite::open().unwrap() per-request with shared Arc<Mutex<Connection>>
- Add Mutex poisoning recovery in all cfg.lock() and db.lock() calls
- Fix std::str::from_utf8().unwrap() with safe match → 400 Bad Request
- Fix timestamp_nanos_opt().unwrap() with safe match
- Fix panic on RPC client failure (bal-pusher): error log + sleep + retry
- Fix ZMQ socket connect with retry loop (bal-pusher)
- Add ZMQ_RCVTIMEO=5000 and match recv for graceful timeout (bal-pusher)
- Fix ZMQ subscribe error with match instead of unwrap (bal-pusher)
- Add unwrap_or for all tbl_stats row fields in echo_stats (prevent NULL panic)
- Add panic_regression_tests.rs: test mutex poisoning recovery and NULL unwrap_or
All tests pass: cargo test --test panic_regression_tests + sql_injection_tests
Build verified: cargo check --bin=bal-server --bin=bal-pusher
- Add docs/INDEX.md with navigable index and quick reference guides
- Add 9 knowledge base files covering project overview, Bitcoin domain,
architecture, modules, API reference, database schema, deployment/security
- Update AGENTS.md with knowledge base reference and update policy
- Add tests/sql_injection_tests.rs with regression tests for SQL injection
- Fix SQL injection vulnerabilities in bal-pusher.rs:
* Replace string-formatted UPDATE IN with loop + parameterized queries
* Replace string-formatted UPDATE push_err with parameterized query
* Add chain name validation in calculate_stats to prevent env var tampering
- Update .gitignore to exclude bal-pusher.env and bal-pusher.sh
