ci / rust (pull_request) Failing after 5m3s
android / android (pull_request) Canceled after 6m35s
apple / swift (pull_request) Canceled after 1m41s
apple / distribute (pull_request) Canceled after 0s
apple / screenshots (pull_request) Canceled after 0s
ci / rust-arm64 (pull_request) Canceled after 5m46s
ci / web (pull_request) Canceled after 1m39s
ci / docs-site (pull_request) Canceled after 1m29s
ci / bun-nix (pull_request) Canceled after 0s
nix / flake (pull_request) Canceled after 2s
windows-client / client (arm64, --no-default-features, aarch64-pc-windows-msvc, C:\t-a64) (pull_request) Canceled after 40s
windows-client / client (x64, , x86_64-pc-windows-msvc, C:\t) (pull_request) Canceled after 0s
The core logs through tracing; the desktop and Android shells install a subscriber/logger and see those lines, the Swift client installed none and saw nothing — every transport warning, quinn connection event and rustls handshake note vanished, and a client log bundle sent to the host carried the shell's half only. - punktfunk_set_log_callback(max_level, cb, user): a log::Log backend behind a C callback (level, target, message, user), level-gated by log::set_max_level so anything above the ceiling costs no formatting; NULL detaches; Unsupported when another log backend already owns the process (android_logger). tracing's `log` feature declared explicitly by the core — it was on transitively via quinn, which an ABI promise must not rest on. ABI_VERSION 24 → 25, header regenerated, embedding doc §2.6. - Apple: CoreLog.install() at app start routes the lines into ClientLog under core.<crate> (os_log + the send-to-host ring), info ceiling by default (PUNKTFUNK_CORE_LOG_LEVEL raises it). - Test: a log record and a tracing event both reach the callback with level/target/message, interior NUL dropped not truncated, ceiling honoured, detach silences.