fix(ui): only toast "Blockchain rescan complete" for user-initiated rescans

The daemon-output parser treats autonomous background witness rebuilds as a rescan (they set
state_.sync.rescanning via the foundWitness branch), so one completing fired "Blockchain rescan
complete" even though the user never started a rescan — most visibly after a minimize, where a
whole rebuild's start+finish arrives in one batch.

Add user_initiated_rescan_ (atomic — some triggers run on worker threads), set it at the wallet's
real rescan triggers (the Rescan button, a -rescan/salvage/zap/reindex restart, key import, seed
migration), and gate the three "rescan complete" toasts on it, clearing it when shown. The
rescan/witness progress state machine is untouched — only the toast is gated — so background
rebuilds no longer announce a completed rescan while genuine user rescans still do (and can't
double-toast).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/app.h

@@ -1302,6 +1302,11 @@ private:
     // the per-second mining/rescan-status pollers are suppressed (the daemon holds cs_main for
     // the whole scan and would block them); completion is signalled by the rescan RPC callback.
     bool runtime_rescan_active_ = false;
+    // True only for a rescan the WALLET/USER initiated (the Rescan button, a -rescan/salvage/zap/reindex
+    // restart, key import, seed migration) — not an autonomous background witness rebuild the daemon does
+    // on its own. Gates the "Blockchain rescan complete" toast so background rebuilds don't fire it;
+    // cleared when the toast is shown.
+    std::atomic<bool> user_initiated_rescan_{false};   // atomic: some rescan triggers run on worker threads
     // Set when a bootstrap completes; consumed once the daemon is connected to auto-run a rescan
     // that reconciles the preserved wallet.dat against the freshly-imported chain.
     bool post_bootstrap_rescan_pending_ = false;