Implement unique work names to prevent duplicate WorkManager tasks
from being enqueued when multiple notifications are scheduled for
the same time or when the receiver is triggered multiple times.
Changes:
- DailyNotificationReceiver: Use enqueueUniqueWork with unique names
("display_{id}", "dismiss_{id}") and ExistingWorkPolicy.KEEP/REPLACE
- DailyNotificationFetcher: Use unique work names based on scheduled
time rounded to minutes ("fetch_{minutes}") with ExistingWorkPolicy.REPLACE
This resolves the issue where ~25+ concurrent workers were being
enqueued for the same notification, leading to race conditions and
resource waste. Now only one worker processes each notification/fetch
at a time.
Verified in logcat: Worker count reduced from 25+ to 1 per notification.