Fix dashboard 500 error and stale activeChild token after regen

- timer.py: MySQL returns timezone-naive datetimes; wrap occurred_at
  with _utc() before arithmetic against datetime.now(timezone.utc) to
  avoid TypeError on compute_block_elapsed / compute_break_elapsed.
  This was causing a 500 on GET /api/dashboard/{token} for any child
  with an active session, preventing schedule templates from loading.

- children.js: after regenerateToken, also update activeChild so it
  doesn't hold a stale tv_token reference. Also sync activeChild on
  fetchChildren re-fetch so navigation after token regen picks up the
  new token immediately.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

frontend/src/stores/children.js

@@ -11,6 +11,10 @@ export const useChildrenStore = defineStore('children', () => {
     children.value = res.data
     if (!activeChild.value && children.value.length > 0) {
       activeChild.value = children.value[0]
+    } else if (activeChild.value) {
+      // Sync activeChild to the fresh data so stale tokens don't persist
+      const updated = children.value.find((c) => c.id === activeChild.value.id)
+      if (updated) activeChild.value = updated
     }
   }
 
@@ -36,6 +40,9 @@ export const useChildrenStore = defineStore('children', () => {
     const res = await api.post(`/api/children/${id}/regenerate-token`)
     const idx = children.value.findIndex((c) => c.id === id)
     if (idx !== -1) children.value[idx] = res.data
+    if (activeChild.value?.id === id) {
+      activeChild.value = res.data
+    }
     return res.data
   }