[T3350] FIX: birthday push cron stalled all scheduled actions

[Danielgergely]

Port of the 14.0 fix: drive the cron from device tokens instead of scanning 10'000+ children.

[greptile-apps]

Confidence Score: 4/5

T-Rex Logs

What T-Rex did

• T-Rex produced a proof for a posted P1 finding and attached three artifacts documenting the focused birthday push shared-partner harness source and its execution outputs.
• T-Rex produced a second proof for another posted P1 finding.
• T-Rex validated the contract by observing the CURRENT_SOURCE_SUCCESSFUL_SEND_USER_IDS output and a matching in-memory comparison, confirmed the relevant code in compassion_child.py (lines 92–112) and that the per-user loop exits after the first send, with no repository changes; only evidence artifacts were authored.

_{Ran code and verified through T-Rex}

Comments Outside Diff (1)

1. General comment

   Birthday push cron skips later active users who share a partner

   • Bug
     • For a target partnership where multiple active users have the same partner_id and each user has a sendable device token, only the first user whose _send_push_notification returns true receives the birthday push. The focused execution observed user 201 receiving the push while user 202 did not.
   • Cause
     • break on line 112 terminates for user_id in user_ids_by_partner[partner_id] immediately after the first successful per-user send. The partner-level already_sent marker and interaction log are also recorded at that point.
   • Fix
     • Remove or redesign the line-112 break so every user-specific token recordset is sent to; if only one partner-level interaction should be created, record it once after processing all users rather than using loop termination.

   _{Ran code and verified through T-Rex}

_{Reviews (1): Last reviewed commit: "[T3350] FIX: birthday push cron stalled ..." | Re-trigger Greptile}