Say what would close a shortfall, and where the free days are

A year that cannot afford the whole week was already funded as far as it
goes rather than dropped; there is now a test that says so, across four
years of a repeating wish where only three and a half are affordable.

The advice to move a window onto the red days was looking three days
either side, which can never find Easter. It now searches four weeks
either way — enough for Easter's drift plus the weekday snap — and ranks
placements with the same cost model flexible periods are already placed
with, naming the dates and the days gained.

None of this reached the screen before. The overview now lists each short
window with its gap and the ranked remedies underneath, one line each.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ei7gCZzdd751uZtPzb1Pep

backend/scripts/demo_plan.py

@@ -206,7 +206,9 @@ _SV_TEMPLATES: dict[str, str] = {
     "planRemedies.lower_priority_wish": (
         "Sänk {wish_name} från {from_weeks} till {to_weeks} veckor frigör {freed_days} dagar"
     ),
-    "planRemedies.shift_window_into_free_days": "Flytta {wish_name} närmare en helg frigör {shift_gain} dagar",
+    "planRemedies.shift_window_into_free_days": (
+        "Flytta {wish_name} till {new_start_date} – {new_end_date} frigör {shift_gain} dagar"
+    ),
     "planWarnings.days_lost_at_age_4": (
         "riskerar att förlora {expiring_days} dagar vid barnets 4-årsdag "
         "({fourth_birthday}) om inte fler dagar tas ut innan dess"