Fix flaky regression test for Play->Back session persistence

The actual fix (main_menu.gd's _ready() checking ApiClient.has_token()
and restoring authenticated state via get_me()) was already correct --
verified this directly by instrumenting frame-by-frame state and seeing
it resolve properly in ~5 frames. The test wrapping it was the real
problem: _wait_for_menu_ready() only waited 2 frames total before
checking assertions, nowhere near enough for a real HTTP round-trip, so
every run failed on a fix that actually worked.

Root fix: added a real auth_state_ready signal to main_menu.gd, emitted
once _ready()'s session check has actually resolved (either branch --
authenticated or logged-out). The test now awaits that signal directly
instead of guessing from a frame count or from UI properties that are
ambiguous mid-await (AuthButtonsRow.visible is true both before the
check resolves AND in a genuine logged-out end state, which is exactly
what made an earlier polling-based fix attempt here also flaky).

Also fixed a second real bug in the test itself: unique_id used
Time.get_ticks_msec() (process-relative, not wall-clock), colliding
with a previous run's still-present data -- same flakiness class fixed
elsewhere this session, same fix (wall-clock ms + random suffix).

Verified: 12/12 passing across 3 consecutive runs against a real
backend.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

scripts/smoke_test_main_menu_back.gd

@@ -45,7 +45,8 @@ func _run() -> void:
 		return
 
 	# 1. Establish a real, valid session exactly the way a user would.
-	var unique_id := Time.get_ticks_msec()
+	randomize()
+	var unique_id: int = int(Time.get_unix_time_from_system() * 1000) + randi() % 100000
 	var username := USERNAME % unique_id
 	var email := EMAIL % unique_id
 	print("\n=== Setup: register ", username, " ===")
@@ -59,13 +60,21 @@ func _run() -> void:
 	# 2. First visit to the main menu: should already show the authenticated
 	#    state (no login form, Play enabled).
 	change_scene_to_file("res://scenes/main_menu.tscn")
-	await _wait_for_menu_ready()
+	# change_scene_to_file defers the actual swap -- a couple frames is
+	# plenty for the scene object itself to exist (unlike the real
+	# get_me() network round-trip afterward, which is what auth_state_ready
+	# actually waits for below).
+	for i in 5:
+		await process_frame
+		if current_scene is Control:
+			break
 
 	var menu: Control = current_scene
 	if menu == null or not menu is Control:
 		printerr("FAIL: main_menu.tscn did not become current_scene")
 		_finish()
 		return
+	await menu.auth_state_ready
 	_check("first main menu visit restores authenticated state", _is_authenticated(menu))
 
 	# 3. Play -> world_list.tscn (real button press).
@@ -90,6 +99,12 @@ func _run() -> void:
 		_finish()
 		return
 
+	# The new main_menu instance's own _ready() is *also* awaiting a real
+	# get_me() call -- _wait_for_scene_change only confirms the scene
+	# object exists, not that its authenticated-state fields have been
+	# set yet. Wait for the real signal, not a guess.
+	await menu.auth_state_ready
+
 	# 5. THE regression assertion: still logged in after returning.
 	_check("main menu stays authenticated after Play -> Back", _is_authenticated(menu))
 
@@ -113,18 +128,6 @@ func _is_authenticated(menu: Control) -> bool:
 		and not auth_panel.visible
 
 
-## Wait for the main menu's _ready() to finish, including its awaited
-## ApiClient.get_me() call (which is why we wait several frames).
-func _wait_for_menu_ready() -> void:
-	for i in 20:
-		await process_frame
-		if current_scene is Control:
-			# The menu's _ready() awaits get_me(); give it time to resolve.
-			if i >= 1:
-				return
-	await process_frame
-
-
 func _wait_for_scene_change(previous_scene: Node) -> void:
 	for i in 20:
 		await process_frame