fix(rpc): drop a stale connect result after teardown; don't let open() erase a pending abort

Concurrency audit, connect-lifecycle races (no fund-path call touched — connect-result gating + abort-flag
lifetime only):
- The connect task's MainCb ran onConnected() on any successful probe without re-checking that a wallet
  switch / adopt / restore / shutdown had already torn the connection down — so a probe that succeeded
  against the old daemon could start a refresh against one about to be stopped. Capture rpc_session_ when
  posting (main thread) and, in the MainCb, drop the result (requestAbort + disconnect, not a blocking bare
  disconnect) when the session changed or daemon_restarting_/shutting_down_ is set. Same guard on the
  cookie-auth retry post.
- RPCClient::open() called clearAbort(), so a requestAbort() issued while a connect was still queued was
  erased, stalling teardown/shutdown. open() no longer clears it; the two reconnect entry points
  (tryConnect for rpc_, onConnected for fast_rpc_) clear it on the main thread first, skipped while
  shutting_down_ so a shutdown abort survives. A fresh client defaults to not-aborted, so first connects
  are unaffected.

HELD (not in this commit): making abortMaintenanceCalls() abort unconditionally (the UI-freeze-on-disconnect
finding) — it would also abort an in-flight user send during teardown, a send-lifecycle behaviour change
that wants explicit sign-off (it is fund-safe via the lost-reply "outcome unknown" handling, but still a
behaviour change).

Threading fix — clean build + ctest; the races are runtime-only.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/app_network.cpp

@@ -563,7 +563,12 @@ void App::tryConnect()
                    attempt, isUsingEmbeddedDaemon() ? "yes" : "no");
     }
     
-    worker_->post([this, config, daemonStarting, externalDetected, attempt]() -> rpc::RPCWorker::MainCb {
+    // A previous teardown leaves rpc_'s abort flag set (open() no longer clears it, so an abort issued
+    // while this connect is still queued is not erased). This is the reconnect entry point: clear it here
+    // on the main thread, after the daemon_restarting_/shutting_down_ gates above.
+    if (rpc_ && !shutting_down_.load()) rpc_->clearAbort();
+    const std::uint64_t connectSession = rpc_session_;   // main thread; a teardown/onConnected bumps it
+    worker_->post([this, config, daemonStarting, externalDetected, attempt, connectSession]() -> rpc::RPCWorker::MainCb {
         bool connected = rpc_->connect(config.host, config.port, config.rpcuser, config.rpcpassword, config.use_tls);
         std::string connectErr = rpc_->getLastConnectError();
         const rpc::RpcFailureKind connectKind = rpc_->getLastConnectFailureKind();
@@ -571,7 +576,18 @@ void App::tryConnect()
         std::string warmupStatus = rpc_->getWarmupStatus();
 
         return [this, config, connected, warmingUp, warmupStatus, daemonStarting, externalDetected, attempt, connectErr,
-                connectKind]() {
+                connectKind, connectSession]() {
+            // A probe that succeeded after a wallet switch / adopt / restore / shutdown already tore the
+            // connection down must not run onConnected() against a daemon that is about to be stopped.
+            if (connected && (connectSession != rpc_session_ || daemon_restarting_.load() || shutting_down_.load())) {
+                VERBOSE_LOGF("[connect #%d] Dropping stale connect result (teardown raced the probe)\n", attempt);
+                if (rpc_) {
+                    rpc_->requestAbort();   // a worker call may hold curl_mutex_; don't block the UI on disconnect()
+                    rpc_->disconnect();
+                }
+                connection_in_progress_ = false;
+                return;
+            }
             noteDaemonRpcResult(connected, connectKind);   // a busy node (timeouts / queue full) feeds the quit guard
             if (connected) connect_backoff_.reset();
             if (connected) {
@@ -600,12 +616,19 @@ void App::tryConnect()
                     if (rpc::Connection::buildCookieAuthConfig(config, cookieConfig)) {
                         VERBOSE_LOGF("[connect #%d] HTTP 401 — retrying with .cookie auth from %s\n",
                                    attempt, cookieConfig.hush_dir.c_str());
-                        worker_->post([this, cookieConfig, attempt]() -> rpc::RPCWorker::MainCb {
+                        worker_->post([this, cookieConfig, attempt, connectSession]() -> rpc::RPCWorker::MainCb {
                             bool ok = rpc_->connect(cookieConfig.host, cookieConfig.port,
                                                     cookieConfig.rpcuser, cookieConfig.rpcpassword,
                                                     cookieConfig.use_tls);
-                            return [this, cookieConfig, ok, attempt]() {
+                            return [this, cookieConfig, ok, attempt, connectSession]() {
                                 connection_in_progress_ = false;
+                                if (ok && (connectSession != rpc_session_ || daemon_restarting_.load() || shutting_down_.load())) {
+                                    if (rpc_) {
+                                        rpc_->requestAbort();
+                                        rpc_->disconnect();
+                                    }
+                                    return;   // stale: a teardown raced the probe
+                                }
                                 if (ok) {
                                     VERBOSE_LOGF("[connect #%d] Connected via .cookie auth\n", attempt);
                                     saved_config_ = cookieConfig;
@@ -836,6 +859,7 @@ void App::onConnected()
         fast_worker_->start();
     }
     // Connect on the fast worker's own thread (non-blocking to main)
+    if (fast_rpc_ && !shutting_down_.load()) fast_rpc_->clearAbort();   // open() no longer clears a prior teardown's abort
     fast_worker_->post([this]() -> rpc::RPCWorker::MainCb {
         bool ok = fast_rpc_->connect(saved_config_.host, saved_config_.port,
                                      saved_config_.rpcuser, saved_config_.rpcpassword,

src/rpc/rpc_client.cpp

@@ -216,7 +216,8 @@ bool RPCClient::open(const std::string& host, const std::string& port,
     curl_easy_setopt(impl_->curl, CURLOPT_HTTPHEADER, impl_->headers);
     curl_easy_setopt(impl_->curl, CURLOPT_WRITEFUNCTION, WriteCallback);
     // Progress callback so requestAbort() can unblock an in-flight curl_easy_perform.
-    clearAbort();  // a fresh connection must not start in the aborted state
+    // NOTE: abort is deliberately NOT cleared here: a requestAbort() issued while this connect was queued
+    // must still cut it short. Reconnect callers (App::tryConnect / onConnected) call clearAbort() first.
     curl_easy_setopt(impl_->curl, CURLOPT_NOPROGRESS, 0L);
     curl_easy_setopt(impl_->curl, CURLOPT_XFERINFOFUNCTION, xferInfoCallback);
     curl_easy_setopt(impl_->curl, CURLOPT_XFERINFODATA, this);