mirror of
https://github.com/farion1231/cc-switch.git
synced 2026-10-26 01:34:51 +08:00
YUZHEthefool
157ff7e8d1
The routing takeover previously re-armed itself unconditionally on every launch: when the process was killed without cleanup (direct shutdown, power loss), the proxy address stayed in client configs and the next cc-switch launch re-took-over even if the proxy never came back, leaving CLI tools pointed at a dead 127.0.0.1:<port> address indefinitely. Add a proxy_restore_on_startup setting (default off) modeled after Clash-style proxy autostart options: - Off (default): any exit restores the original client configs and clears the per-app takeover flags. After an abnormal exit, startup recovery restores the live configs and clears the flags instead of re-taking over, so no stale routing address survives a relaunch. - On: previous behavior kept - exit preserves proxy_config.enabled and the next launch re-arms takeover automatically (requires Launch on Startup). cleanup_before_exit now branches on the setting (stop_with_restore vs stop_with_restore_keep_state) with a half-takeover fallback that clears stray enabled flags even when no backups or placeholders remain. The settings page gains a toggle in the Local Routing section with a hint when auto-resume is on but Launch on Startup is off, plus zh/zh-TW/en/ja translations and user-manual updates.