feat: SOCKS/proxy detection + tunnel auto-reconnect, uptime, copy command#12
Closed
rekurt wants to merge 4 commits into
Closed
feat: SOCKS/proxy detection + tunnel auto-reconnect, uptime, copy command#12rekurt wants to merge 4 commits into
rekurt wants to merge 4 commits into
Conversation
Add a ProxyListening heuristic that flags processes listening on well-known SOCKS/proxy ports (1080, 1081, 3128, 9050, 9150). 8080/8443 are deliberately excluded to avoid false positives on http-alt servers. The new reason plugs into the existing enrichment pipeline, so proxy listeners get the magenta + [!] treatment and match the 'suspicious' filter automatically. Also adds tor-socks/socks names to known_ports.
- Add an Uptime column (compact 45s/12m/3h04m/2d05h formatting), reset on restart. - Layer a listener health check over the Alive status: an alive ssh child with no local listener on its port (broken -D/-L bind) now shows a yellow 'no listener' warning instead of a misleading green 'alive'. Reuses already-scanned LISTEN entries — opens no new connections. - Add 'c' key to copy the tunnel's full ssh command to the clipboard, reusing the existing copy_to_clipboard helper. - New i18n strings (uptime column, no-listener, copy hint) in en/ru/zh.
Tunnels that die on their own are now automatically restarted with exponential backoff (2s doubling up to 60s) so an unreachable host isn't hammered. reconnect_failed() runs each loop iteration right after cleanup(); gating is by next_retry_at rather than the tick, and a successful restart resets the backoff. Manual kill removes a tunnel entirely, so only self-inflicted failures reconnect.
…t, model health Auto-reconnect (#1,#2,#3,#8): - Backoff now actually grows for unreachable hosts. The reset was tied to a 'successful' 150ms spawn, which an unreachable host passes (it blocks on TCP timeout), so the delay never increased. Growth now happens on every attempt and is reset only after a tunnel stays Alive for STABLE_THRESHOLD (30s), via refresh_health. - reconnect uses a new non-blocking restart_async() (no 150ms sleep in the render thread); failures are caught on the next tick. Manual restart() stays blocking for immediate feedback. - After MAX_RETRIES attempts without recovery a tunnel is marked permanently failed (auto_reconnect=false); drop_failed() now prunes only those, so 'save' no longer discards a tunnel that's mid-reconnect. auto_reconnect is now a meaningful flag. Listener health in the model (#4,#5): - New TunnelStatus::Unhealthy (process alive, local port not listening). Decision extracted to a pure, unit-tested decide_status(); a HEALTH_GRACE window and a scan_usable flag suppress false 'no listener' right after start and while auto-refresh is paused. cleanup() now takes the listening port set; the view no longer reaches into the scan. Cleanups: - command_string() shell-quotes args via shlex (#6). - uptime column reuses prt-core format_duration (#7). - private from_parts() constructor removes spawn duplication (#9). - cross-reference comments between PROXY_PORTS and known_ports (#10). New unit tests: decide_status, next_backoff, command_string quoting.
Owner
Author
|
Closing as superseded by #13 ( That overlap is what made this branch conflict in every shared file. Generated by Claude Code |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Что и зачем
Три фичи вокруг SSH-туннелей / SOCKS5. Туннельная инфраструктура (
forward.rs,ssh_tunnel.rs,views/tunnels.rs, авто-старт из конфига) уже существовала — здесь закрыты конкретные пробелы.1. Детект локальных SOCKS/proxy-листенеров (
prt-core)Новая эвристика
ProxyListeningвsuspicious.rs: процесс, слушающий на well-known SOCKS/proxy-порту (1080, 1081, 3128, 9050, 9150).8080/8443намеренно исключены, чтобы не плодить ложные срабатывания на http-alt. Новая причина встраивается в существующий конвейер — листенер автоматически получает magenta-подсветку + бейдж[!]в Connections и попадает под фильтрsuspicious. Плюс именаsocks/tor-socksвknown_ports.2. Доработка UI туннелей (дёшевый набор, без новых соединений)
45s/12m/3h04m/2d05h), сбрасывается при рестарте.Alive: живой ssh-процесс без реального LISTEN на своём порту (типичный признак сломанного-D/-L) показывается жёлтымno listenerвместо обманчивого зелёногоalive. Переиспользует уже отсканированные данные — никаких новых соединений.c) — копирует полнуюssh-команду туннеля в буфер через существующийcopy_to_clipboard.3. Авто-reconnect упавших туннелей (
prt)Туннели, упавшие сами по себе, автоматически перезапускаются с экспоненциальным backoff (2s → ×2 → потолок 60s), чтобы не долбить недоступный хост.
reconnect_failed()вызывается каждый цикл послеcleanup(); гейтинг поnext_retry_at, успешный рестарт сбрасывает backoff. Ручной kill удаляет туннель целиком — реконнектятся только самопроизвольные падения.i18n
Новые строки (
tunnel_col_uptime,tunnel_health_no_listener,hint_copy_tunnel) добавлены во все три языка (en/ru/zh) — compile-time completeness-check проходит.Тесты
suspicious.rs: 1080/9050 LISTEN → flagged, 8080 → нет, ESTABLISHED → нет.fmt_uptime.cargo test --workspace— зелёный (197+ тестов),clippyиfmt --checkчисты.Ручная проверка
ssh -D 1080 localhost(илиnc -l 1080) → строка порта 1080 в Connections magenta +[!], ловится фильтромsuspicious.n) → растёт Uptime;c→ ssh-команда в буфере; health =aliveпри живом листенере.failed, затем повторы с растущим backoff; при восстановлении хоста сам переходит вalive.https://claude.ai/code/session_013YMo3DYYxbDAr6coo72M9W
Generated by Claude Code