scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
/*
|
|
|
|
|
* net/transport.h — TCP control channel + UDP media channel.
|
|
|
|
|
*
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
* Design: docs/architecture.md (Net thread), docs/protocol.md §1 (framing).
|
|
|
|
|
* Implementation uses standalone Asio for sockets and timers.
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
*/
|
|
|
|
|
#ifndef VOICECAT_NET_TRANSPORT_H
|
|
|
|
|
#define VOICECAT_NET_TRANSPORT_H
|
|
|
|
|
|
|
|
|
|
#include <cstdint>
|
|
|
|
|
#include <string>
|
|
|
|
|
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
#define ASIO_STANDALONE 1
|
|
|
|
|
#include <asio.hpp>
|
|
|
|
|
|
|
|
|
|
#include <atomic>
|
|
|
|
|
#include <deque>
|
|
|
|
|
#include <functional>
|
|
|
|
|
#include <memory>
|
|
|
|
|
#include <mutex>
|
|
|
|
|
#include <thread>
|
|
|
|
|
#include <vector>
|
|
|
|
|
|
|
|
|
|
#include "protocol/protocol.h"
|
|
|
|
|
|
|
|
|
|
// Forward-declare TlsContext so transport.h does not pull in mbedTLS headers.
|
|
|
|
|
namespace voicecat::crypto { class TlsContext; }
|
|
|
|
|
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
namespace voicecat::net {
|
|
|
|
|
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
// Callbacks delivered on the net thread. Callers must not block inside them.
|
|
|
|
|
struct TcpChannelCallbacks {
|
|
|
|
|
std::function<void()> on_connected;
|
|
|
|
|
std::function<void(std::error_code)> on_connect_error;
|
|
|
|
|
std::function<void(std::vector<uint8_t>)> on_frame; // one decoded frame payload
|
|
|
|
|
std::function<void(std::error_code)> on_error;
|
|
|
|
|
std::function<void()> on_disconnected;
|
feat(M2): UDP voice/media plane -- SFU relay, Opus, AEAD, jitter buffer
Adds the full voice pipeline: 14-byte binary frame header, ChaCha20-Poly1305
AEAD keyed from the TLS exporter, libopus encode/decode with FEC/PLC/DTX,
an adaptive per-ssrc jitter buffer, a miniaudio capture/playback engine, an
APM passthrough stub, and the UdpBinding/StreamAnnounce signaling chain
wired through ConnSession/SessionRegistry into a new server-side SFU
(MediaRelay) that decrypts and re-encrypts frames per channel member.
Exit criterion verified: test_m2_voice — two headless clients relay 50
encrypted Opus frames through the server; ctest --preset m1-dev is 9/9
green. Also corrects protocol.md's UdpBinding diagram, which described the
UDP-side binding packet as AEAD-sealed when it is in fact a plaintext
bootstrap frame (separate from the TCP/TLS UdpBinding ack).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-16 01:31:14 +02:00
|
|
|
// Called (on the handshake thread) right after TLS succeeds, before reads begin.
|
|
|
|
|
// Use to export keying material while the handshake context is still fresh.
|
|
|
|
|
std::function<void(voicecat::crypto::TlsContext&)> on_tls_ready;
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
// ── Client-side: owns an io_context + dedicated net thread ──────────────────
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
class TcpControlChannel {
|
|
|
|
|
public:
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
explicit TcpControlChannel(TcpChannelCallbacks cbs);
|
|
|
|
|
~TcpControlChannel();
|
|
|
|
|
|
|
|
|
|
// Async connect; calls on_connected or on_connect_error on the net thread.
|
|
|
|
|
void async_connect(const std::string& host, uint16_t port);
|
|
|
|
|
|
|
|
|
|
// Queue a framed send (thread-safe; callable from any thread).
|
|
|
|
|
void send_frame(std::vector<uint8_t> payload);
|
|
|
|
|
|
|
|
|
|
// Graceful close; safe to call from any thread. Waits for the net thread to join.
|
|
|
|
|
void close();
|
|
|
|
|
|
|
|
|
|
bool connected() const { return connected_.load(std::memory_order_acquire); }
|
|
|
|
|
|
|
|
|
|
// Access the io_context so callers can post work back to the net thread.
|
|
|
|
|
asio::io_context& io() { return io_; }
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
|
|
|
|
|
private:
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
void run_loop();
|
|
|
|
|
void start_read();
|
|
|
|
|
void handle_length(std::error_code ec, std::size_t n);
|
|
|
|
|
void handle_body(uint32_t length, std::error_code ec, std::size_t n);
|
|
|
|
|
void do_send();
|
|
|
|
|
|
|
|
|
|
asio::io_context io_;
|
|
|
|
|
asio::executor_work_guard<asio::io_context::executor_type> work_guard_;
|
|
|
|
|
asio::ip::tcp::socket socket_;
|
|
|
|
|
asio::strand<asio::io_context::executor_type> strand_;
|
|
|
|
|
std::thread net_thread_;
|
|
|
|
|
|
|
|
|
|
TcpChannelCallbacks cbs_;
|
|
|
|
|
protocol::FrameCodec codec_;
|
|
|
|
|
|
|
|
|
|
uint8_t len_buf_[4]{};
|
|
|
|
|
std::vector<uint8_t> body_buf_;
|
|
|
|
|
std::deque<std::vector<uint8_t>> send_queue_;
|
|
|
|
|
bool sending_{false};
|
|
|
|
|
std::atomic<bool> connected_{false};
|
|
|
|
|
std::atomic<bool> closing_{false};
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
};
|
|
|
|
|
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
// ── Server-side: one per accepted socket, shares the server's io_context ────
|
|
|
|
|
class TcpServerConn : public std::enable_shared_from_this<TcpServerConn> {
|
|
|
|
|
public:
|
|
|
|
|
// Plain TCP constructor (no TLS — for tests or future plaintext paths).
|
|
|
|
|
TcpServerConn(asio::ip::tcp::socket socket, TcpChannelCallbacks cbs);
|
|
|
|
|
|
|
|
|
|
// TLS constructor: takes ownership of a TlsContext; start() will run the
|
|
|
|
|
// handshake on a temporary thread then switch to a TLS I/O thread.
|
|
|
|
|
TcpServerConn(asio::ip::tcp::socket socket, TcpChannelCallbacks cbs,
|
|
|
|
|
std::unique_ptr<voicecat::crypto::TlsContext> tls);
|
|
|
|
|
|
|
|
|
|
~TcpServerConn();
|
|
|
|
|
|
|
|
|
|
// Begin reading; must be called once after construction (on the io thread).
|
|
|
|
|
void start();
|
|
|
|
|
|
|
|
|
|
// Thread-safe send (safe to call from the server's io thread or another strand).
|
|
|
|
|
void send_frame(std::vector<uint8_t> payload);
|
|
|
|
|
|
|
|
|
|
// Close the connection (safe from any thread).
|
|
|
|
|
void close();
|
|
|
|
|
|
feat(macos): validate dev + apple-dev presets on macOS, fix 3 cross-platform bugs
macOS port groundwork — core, server, tools, and tests now build and run on
macOS 26.5 / Apple Silicon. ctest --preset dev green 21/21 (2 consecutive runs).
apple-dev produces valid arm64 libvoicecat.a + XCFramework for the Swift Package.
Three real cross-platform bugs found and fixed (all latent on Windows/Linux):
1. test_m2_voice.cpp POSIX branch missing <netdb.h> — Linux glibc transitively
includes it, macOS doesn't. Would fail on any strict POSIX system.
2. SIGPIPE killing processes on macOS — writing to a closed TCP socket raises
SIGPIPE by default (doesn't exist on Windows, benign on Linux). Fixed by
ignoring SIGPIPE in both core client init and server startup (POSIX-only,
#ifndef _WIN32). Production fix, not just tests.
3. Use-after-free of Asio's kqueue reactor on server shutdown — the
deterministic test_tofu_flow segfault. TcpServerConn's tls_read_loop runs on
a blocking-I/O thread; when Server::run() returned, io_context was destroyed
while those threads were still running. On macOS kqueue the reactor pointer
is null'd immediately -> segfault in socket.close(). Latent on Windows IOCP
and Linux epoll. Fix: TcpAcceptor now tracks connections; new shutdown()
closes all + joins threads before io is destroyed; Server::stop() now closes
acceptor + media_relay too (was just io.stop()).
Verified: dev + apple-dev presets build green, 21/21 tests pass, server starts
+ two vccli text chat over TLS (M1 on Mac), vccli --voice starts MIC stream via
CoreAudio (M2 protocol-level), vccli --list-devices enumerates CoreAudio
devices, xcodebuild -create-xcframework produces valid VoiceCatCore.xcframework.
No ABI or proto changes. Docs updated: building.md, clients/apple/README.md,
PROGRESS.md, CLAUDE.md status line.
2026-06-18 13:24:42 +02:00
|
|
|
// Block until the TLS I/O thread (if any) has finished. Must be called after close().
|
|
|
|
|
// Safe to call from any thread except the TLS I/O thread itself.
|
|
|
|
|
void wait_closed();
|
|
|
|
|
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
bool connected() const { return connected_.load(std::memory_order_acquire); }
|
|
|
|
|
|
|
|
|
|
private:
|
|
|
|
|
// ── Asio path (no TLS) ───────────────────────────────────────────────────
|
|
|
|
|
void start_read();
|
|
|
|
|
void handle_length(std::error_code ec, std::size_t n);
|
|
|
|
|
void handle_body(uint32_t length, std::error_code ec, std::size_t n);
|
|
|
|
|
void do_send();
|
|
|
|
|
|
|
|
|
|
// ── TLS path ─────────────────────────────────────────────────────────────
|
|
|
|
|
void tls_read_loop();
|
|
|
|
|
void tls_drain_sends();
|
|
|
|
|
|
|
|
|
|
asio::ip::tcp::socket socket_;
|
|
|
|
|
asio::strand<asio::any_io_executor> strand_;
|
|
|
|
|
TcpChannelCallbacks cbs_;
|
|
|
|
|
protocol::FrameCodec codec_;
|
|
|
|
|
|
|
|
|
|
uint8_t len_buf_[4]{};
|
|
|
|
|
std::vector<uint8_t> body_buf_;
|
|
|
|
|
std::deque<std::vector<uint8_t>> send_queue_;
|
|
|
|
|
bool sending_{false};
|
|
|
|
|
std::atomic<bool> connected_{false};
|
|
|
|
|
std::atomic<bool> closing_{false};
|
|
|
|
|
|
|
|
|
|
// TLS members (null in plain-TCP mode)
|
|
|
|
|
std::unique_ptr<voicecat::crypto::TlsContext> tls_;
|
|
|
|
|
std::thread tls_thread_;
|
|
|
|
|
std::mutex tls_send_mutex_;
|
|
|
|
|
std::deque<std::vector<uint8_t>> tls_send_queue_;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
// ── Server-side acceptor ─────────────────────────────────────────────────────
|
|
|
|
|
// Spawns a TcpServerConn (via factory) for each accepted TCP connection.
|
|
|
|
|
class TcpAcceptor {
|
|
|
|
|
public:
|
|
|
|
|
using ConnFactory = std::function<std::shared_ptr<TcpServerConn>(asio::ip::tcp::socket)>;
|
|
|
|
|
|
|
|
|
|
TcpAcceptor(asio::io_context& io, uint16_t port, ConnFactory factory);
|
|
|
|
|
|
|
|
|
|
// Start accepting. Call once; re-arms itself automatically.
|
|
|
|
|
void start();
|
|
|
|
|
|
feat(macos): validate dev + apple-dev presets on macOS, fix 3 cross-platform bugs
macOS port groundwork — core, server, tools, and tests now build and run on
macOS 26.5 / Apple Silicon. ctest --preset dev green 21/21 (2 consecutive runs).
apple-dev produces valid arm64 libvoicecat.a + XCFramework for the Swift Package.
Three real cross-platform bugs found and fixed (all latent on Windows/Linux):
1. test_m2_voice.cpp POSIX branch missing <netdb.h> — Linux glibc transitively
includes it, macOS doesn't. Would fail on any strict POSIX system.
2. SIGPIPE killing processes on macOS — writing to a closed TCP socket raises
SIGPIPE by default (doesn't exist on Windows, benign on Linux). Fixed by
ignoring SIGPIPE in both core client init and server startup (POSIX-only,
#ifndef _WIN32). Production fix, not just tests.
3. Use-after-free of Asio's kqueue reactor on server shutdown — the
deterministic test_tofu_flow segfault. TcpServerConn's tls_read_loop runs on
a blocking-I/O thread; when Server::run() returned, io_context was destroyed
while those threads were still running. On macOS kqueue the reactor pointer
is null'd immediately -> segfault in socket.close(). Latent on Windows IOCP
and Linux epoll. Fix: TcpAcceptor now tracks connections; new shutdown()
closes all + joins threads before io is destroyed; Server::stop() now closes
acceptor + media_relay too (was just io.stop()).
Verified: dev + apple-dev presets build green, 21/21 tests pass, server starts
+ two vccli text chat over TLS (M1 on Mac), vccli --voice starts MIC stream via
CoreAudio (M2 protocol-level), vccli --list-devices enumerates CoreAudio
devices, xcodebuild -create-xcframework produces valid VoiceCatCore.xcframework.
No ABI or proto changes. Docs updated: building.md, clients/apple/README.md,
PROGRESS.md, CLAUDE.md status line.
2026-06-18 13:24:42 +02:00
|
|
|
// Stop accepting and close all tracked connections (safe while io_context is alive).
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
void stop();
|
|
|
|
|
|
feat(macos): validate dev + apple-dev presets on macOS, fix 3 cross-platform bugs
macOS port groundwork — core, server, tools, and tests now build and run on
macOS 26.5 / Apple Silicon. ctest --preset dev green 21/21 (2 consecutive runs).
apple-dev produces valid arm64 libvoicecat.a + XCFramework for the Swift Package.
Three real cross-platform bugs found and fixed (all latent on Windows/Linux):
1. test_m2_voice.cpp POSIX branch missing <netdb.h> — Linux glibc transitively
includes it, macOS doesn't. Would fail on any strict POSIX system.
2. SIGPIPE killing processes on macOS — writing to a closed TCP socket raises
SIGPIPE by default (doesn't exist on Windows, benign on Linux). Fixed by
ignoring SIGPIPE in both core client init and server startup (POSIX-only,
#ifndef _WIN32). Production fix, not just tests.
3. Use-after-free of Asio's kqueue reactor on server shutdown — the
deterministic test_tofu_flow segfault. TcpServerConn's tls_read_loop runs on
a blocking-I/O thread; when Server::run() returned, io_context was destroyed
while those threads were still running. On macOS kqueue the reactor pointer
is null'd immediately -> segfault in socket.close(). Latent on Windows IOCP
and Linux epoll. Fix: TcpAcceptor now tracks connections; new shutdown()
closes all + joins threads before io is destroyed; Server::stop() now closes
acceptor + media_relay too (was just io.stop()).
Verified: dev + apple-dev presets build green, 21/21 tests pass, server starts
+ two vccli text chat over TLS (M1 on Mac), vccli --voice starts MIC stream via
CoreAudio (M2 protocol-level), vccli --list-devices enumerates CoreAudio
devices, xcodebuild -create-xcframework produces valid VoiceCatCore.xcframework.
No ABI or proto changes. Docs updated: building.md, clients/apple/README.md,
PROGRESS.md, CLAUDE.md status line.
2026-06-18 13:24:42 +02:00
|
|
|
// Stop accepting, close all connections, and block until every connection's
|
|
|
|
|
// I/O thread has finished. Call BEFORE the io_context is destroyed — the TLS read
|
|
|
|
|
// threads do blocking I/O (not async on the io_context) and will touch the reactor
|
|
|
|
|
// on socket close if the io_context is already gone (manifests as a null-kqueue-reactor
|
|
|
|
|
// segfault on macOS; latent on Windows IOCP / Linux epoll where timing is more forgiving).
|
|
|
|
|
void shutdown();
|
|
|
|
|
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
// Actual bound port (useful when bind_port=0 lets the OS pick).
|
|
|
|
|
uint16_t local_port() const { return static_cast<uint16_t>(acceptor_.local_endpoint().port()); }
|
|
|
|
|
|
|
|
|
|
private:
|
|
|
|
|
void do_accept();
|
|
|
|
|
|
feat(macos): validate dev + apple-dev presets on macOS, fix 3 cross-platform bugs
macOS port groundwork — core, server, tools, and tests now build and run on
macOS 26.5 / Apple Silicon. ctest --preset dev green 21/21 (2 consecutive runs).
apple-dev produces valid arm64 libvoicecat.a + XCFramework for the Swift Package.
Three real cross-platform bugs found and fixed (all latent on Windows/Linux):
1. test_m2_voice.cpp POSIX branch missing <netdb.h> — Linux glibc transitively
includes it, macOS doesn't. Would fail on any strict POSIX system.
2. SIGPIPE killing processes on macOS — writing to a closed TCP socket raises
SIGPIPE by default (doesn't exist on Windows, benign on Linux). Fixed by
ignoring SIGPIPE in both core client init and server startup (POSIX-only,
#ifndef _WIN32). Production fix, not just tests.
3. Use-after-free of Asio's kqueue reactor on server shutdown — the
deterministic test_tofu_flow segfault. TcpServerConn's tls_read_loop runs on
a blocking-I/O thread; when Server::run() returned, io_context was destroyed
while those threads were still running. On macOS kqueue the reactor pointer
is null'd immediately -> segfault in socket.close(). Latent on Windows IOCP
and Linux epoll. Fix: TcpAcceptor now tracks connections; new shutdown()
closes all + joins threads before io is destroyed; Server::stop() now closes
acceptor + media_relay too (was just io.stop()).
Verified: dev + apple-dev presets build green, 21/21 tests pass, server starts
+ two vccli text chat over TLS (M1 on Mac), vccli --voice starts MIC stream via
CoreAudio (M2 protocol-level), vccli --list-devices enumerates CoreAudio
devices, xcodebuild -create-xcframework produces valid VoiceCatCore.xcframework.
No ABI or proto changes. Docs updated: building.md, clients/apple/README.md,
PROGRESS.md, CLAUDE.md status line.
2026-06-18 13:24:42 +02:00
|
|
|
asio::ip::tcp::acceptor acceptor_;
|
|
|
|
|
ConnFactory factory_;
|
|
|
|
|
bool stopped_{false};
|
|
|
|
|
|
|
|
|
|
std::mutex conns_mu_;
|
|
|
|
|
std::vector<std::shared_ptr<TcpServerConn>> conns_;
|
feat(M1): TCP/TLS control plane -- auth, channels, ephemeral text
Implements the full M1 milestone. Two clients authenticate over TLS 1.3
(guest + Argon2id password) and exchange channel + private text messages
through a real server. All five ctest --preset m1-dev tests pass in ~1 s.
Key components added:
- vcpkg baseline + m1-dev preset (protobuf/mbedTLS/libsodium/asio/sqlite3)
- FrameCodec feed+emit, encode/decode_envelope, protobuf codegen
- TcpServerConn with blocking TLS handshake thread + tls_read_loop
- TlsContext (mbedTLS 1.3, ECDSA-P256 self-signed cert, TOFU on client)
- WorkerPool (3 threads, used for Argon2id)
- Database: SQLite + libsodium Argon2id, account lifecycle, bootstrap admin
- ServerIdentityManager: Ed25519 key + cert generate/persist/fingerprint
- ConnSession state machine: WaitingHello -> WaitingAuth -> Authenticated
- SessionRegistry: channel tree, user map, text routing, broadcast
- vc_client full M1 C ABI: connect/TLS/handshake/auth/text/disconnect
- voicecat-admin CLI: account add/reset/del/list
- test_m1_integration: M1 exit criterion, verified green
Bug fixed: double-framing in ConnSession::send_envelope -- encode_envelope
was adding the [4-byte len] prefix, then TcpServerConn::send_frame added
a second one, causing the client to parse [len][proto] as protobuf (silent
failure). Fixed by serializing raw protobuf bytes in send_envelope and
letting send_frame apply the single length prefix.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-15 23:48:44 +02:00
|
|
|
};
|
|
|
|
|
|
chore: comment cleanup pass ahead of open-sourcing
Removes leftover debug scaffolding (stray Console.WriteLine/NSLog traces,
dead nick_buf_ptr, a no-op --print-config flag now implemented for real),
fixes stale/misleading comments (channel passwords are no longer a "future
M5+" feature, a wrong cross-reference, a stale TlsContext::close() mention,
an incomplete BanRecord::subject_type doc, and a smoke test pointing at a
build/m1-dev preset that no longer exists), strips internal M1-M5 milestone
jargon from comments now that the roadmap is done, trims comments that just
restated the following line, and consolidates a few "why" explanations that
were duplicated 2-3 times in the same file.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-03 10:20:18 +01:00
|
|
|
// ── UDP media channel ────────────────────────────────────────────────────────
|
feat(M2): UDP voice/media plane -- SFU relay, Opus, AEAD, jitter buffer
Adds the full voice pipeline: 14-byte binary frame header, ChaCha20-Poly1305
AEAD keyed from the TLS exporter, libopus encode/decode with FEC/PLC/DTX,
an adaptive per-ssrc jitter buffer, a miniaudio capture/playback engine, an
APM passthrough stub, and the UdpBinding/StreamAnnounce signaling chain
wired through ConnSession/SessionRegistry into a new server-side SFU
(MediaRelay) that decrypts and re-encrypts frames per channel member.
Exit criterion verified: test_m2_voice — two headless clients relay 50
encrypted Opus frames through the server; ctest --preset m1-dev is 9/9
green. Also corrects protocol.md's UdpBinding diagram, which described the
UDP-side binding packet as AEAD-sealed when it is in fact a plaintext
bootstrap frame (separate from the TCP/TLS UdpBinding ack).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-16 01:31:14 +02:00
|
|
|
// Thin async UDP socket. send_to() is thread-safe. Recv callbacks fire on the
|
|
|
|
|
// io_context's thread (same thread that runs the io_context::run() loop).
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
class UdpMediaChannel {
|
|
|
|
|
public:
|
feat(M2): UDP voice/media plane -- SFU relay, Opus, AEAD, jitter buffer
Adds the full voice pipeline: 14-byte binary frame header, ChaCha20-Poly1305
AEAD keyed from the TLS exporter, libopus encode/decode with FEC/PLC/DTX,
an adaptive per-ssrc jitter buffer, a miniaudio capture/playback engine, an
APM passthrough stub, and the UdpBinding/StreamAnnounce signaling chain
wired through ConnSession/SessionRegistry into a new server-side SFU
(MediaRelay) that decrypts and re-encrypts frames per channel member.
Exit criterion verified: test_m2_voice — two headless clients relay 50
encrypted Opus frames through the server; ctest --preset m1-dev is 9/9
green. Also corrects protocol.md's UdpBinding diagram, which described the
UDP-side binding packet as AEAD-sealed when it is in fact a plaintext
bootstrap frame (separate from the TCP/TLS UdpBinding ack).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-16 01:31:14 +02:00
|
|
|
using FrameCallback =
|
|
|
|
|
std::function<void(const uint8_t*, size_t, asio::ip::udp::endpoint)>;
|
|
|
|
|
|
|
|
|
|
UdpMediaChannel() = default;
|
|
|
|
|
~UdpMediaChannel() { close(); }
|
|
|
|
|
|
|
|
|
|
UdpMediaChannel(const UdpMediaChannel&) = delete;
|
|
|
|
|
UdpMediaChannel& operator=(const UdpMediaChannel&) = delete;
|
|
|
|
|
|
|
|
|
|
// Bind to 0.0.0.0:port (0 = OS-assigned). Must be called before start_recv/send_to.
|
|
|
|
|
bool bind(asio::io_context& io, uint16_t port = 0);
|
|
|
|
|
|
|
|
|
|
// Begin the async recv loop. cb is called on the io_context thread.
|
|
|
|
|
void start_recv(FrameCallback cb);
|
|
|
|
|
|
|
|
|
|
// Thread-safe fire-and-forget send. Copies data into a heap buffer.
|
|
|
|
|
void send_to(const uint8_t* data, size_t len, asio::ip::udp::endpoint dst);
|
|
|
|
|
|
|
|
|
|
// Cancel all async ops and close the socket. Safe to call from any thread.
|
|
|
|
|
void close();
|
|
|
|
|
|
|
|
|
|
asio::ip::udp::endpoint local_endpoint() const;
|
|
|
|
|
bool bound() const { return bound_.load(std::memory_order_acquire); }
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
|
|
|
|
|
private:
|
feat(M2): UDP voice/media plane -- SFU relay, Opus, AEAD, jitter buffer
Adds the full voice pipeline: 14-byte binary frame header, ChaCha20-Poly1305
AEAD keyed from the TLS exporter, libopus encode/decode with FEC/PLC/DTX,
an adaptive per-ssrc jitter buffer, a miniaudio capture/playback engine, an
APM passthrough stub, and the UdpBinding/StreamAnnounce signaling chain
wired through ConnSession/SessionRegistry into a new server-side SFU
(MediaRelay) that decrypts and re-encrypts frames per channel member.
Exit criterion verified: test_m2_voice — two headless clients relay 50
encrypted Opus frames through the server; ctest --preset m1-dev is 9/9
green. Also corrects protocol.md's UdpBinding diagram, which described the
UDP-side binding packet as AEAD-sealed when it is in fact a plaintext
bootstrap frame (separate from the TCP/TLS UdpBinding ack).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-16 01:31:14 +02:00
|
|
|
void do_recv();
|
|
|
|
|
|
|
|
|
|
// Socket + recv state live here; only accessed from the io_context thread after bind().
|
|
|
|
|
std::unique_ptr<asio::ip::udp::socket> socket_;
|
|
|
|
|
asio::ip::udp::endpoint sender_ep_;
|
|
|
|
|
std::array<uint8_t, 1500> recv_buf_{};
|
|
|
|
|
FrameCallback frame_cb_;
|
|
|
|
|
std::atomic<bool> bound_{false};
|
|
|
|
|
std::atomic<bool> closed_{false};
|
scaffold: M0 skeleton + agent onboarding (build, architecture, progress)
Turn the design into a buildable, dependency-free M0 skeleton plus the
onboarding layer so a new agent can pick up instantly.
Build system:
- CMake + CMakePresets (dev = no deps; server-release = vcpkg) + vcpkg.json
- Skeleton builds with just a C++20 compiler; deps stay off until needed
- .gitattributes (LF), .gitignore, .clang-format
Core (libvoicecat):
- core/include/voicecat.h: full C ABI (the client/server contract), stubbed
- core/proto/voicecat.proto: control-plane wire format, matches docs/protocol.md
- src/{net,crypto,codec,protocol,session,audio,core}: subsystem stubs that
return VC_ERR_NOT_IMPLEMENTED, each pointing to its design doc
- server/ (voicecat-server) and tools/vccli/ link the core
- tests/: CTest smoke test asserting the C ABI contract (behavior, not just build)
- clients/{apple,windows}: M4 placeholders
Onboarding for agents:
- CLAUDE.md: hub — build/test commands, architecture at a glance, doc map, rules
- AGENTS.md: working method (behavior-driven; clean compile is the floor not the goal)
- PROGRESS.md: living tracker — M0 done, M1 task checklist, "where we left off"
Verified: cmake --preset dev && cmake --build --preset dev && ctest --preset dev → green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-15 21:09:09 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
} // namespace voicecat::net
|
|
|
|
|
|
|
|
|
|
#endif // VOICECAT_NET_TRANSPORT_H
|