The root cause of issue #4760 was that the sniffing configuration and inbound tag
were being stored during Process() instead of at initialization time. This caused
a race condition where concurrent connections would overwrite each other's session
information, breaking domain-based routing.
The fix follows the same pattern as the TUN handler:
- Save tag and sniffingRequest during NewServer (at creation time)
- Use these saved values in forwardConnection for each new connection
This ensures that each connection uses the correct sniffing configuration that was
set for the inbound, regardless of concurrent connections from multiple WireGuard peers.
Co-authored-by: RPRX <63339210+RPRX@users.noreply.github.com>
* Refactor log
* Add new log methods
* Fix logger test
* Change all logging code
* Clean up pathObj
* Rebase to latest main
* Remove invoking method name after the dot
* Add session context outbounds as slice
slice is needed for dialer proxy where two outbounds work on top of each other
There are two sets of target addr for example
It also enable Xtls to correctly do splice copy by checking both outbounds are ready to do direct copy
* Fill outbound tag info
* Splice now checks capalibility from all outbounds
* Fix unit tests
* feat: wireguard inbound
* feat(command): generate wireguard compatible keypair
* feat(wireguard): connection idle timeout
* fix(wireguard): close endpoint after connection closed
* fix(wireguard): resolve conflicts
* feat(wireguard): set cubic as default cc algorithm in gVisor TUN
* chore(wireguard): resolve conflict
* chore(wireguard): remove redurant code
* chore(wireguard): remove redurant code
* feat: rework server for gvisor tun
* feat: keep user-space tun as an option
* fix: exclude android from native tun build
* feat: auto kernel tun
* fix: build
* fix: regulate function name & fix test
copilot-swe-agent[bot]