Unearth a random old message
forwards a random message_id in [1, current], fun-gated with v1's fun_off reply. v1 wrote a 100-attempt retry and then returned inside its own except, so it tried once and answered nothing whenever that id was deleted; the retry is real here, bounded at 8. QA authored, not ported
/desenterrar/unearth- v1 scenarios
- —
- Ported
- 5
- Green
- 6
- Failing
- —
forwards a random message_id in [1, current], fun-gated with v1's fun_off reply. v1 wrote a 100-attempt retry and then returned inside its own except, so it tried once and answered nothing whenever that id was deleted; the retry is real here, bounded at 8. QA authored, not ported
Miscellaneous.py:325-333What it does
Behaviour that must not change
How to verify it
Sticker DB auto-reply
passive: pools an sfw group's alphanumeric-set-name, non-banned-emoji stickers into a new GLOBAL sticker_pool reference table (migration 0009, not per-group like fun_random - full reasoning in that migration's docstring), then replies with one at random to any sticker/document/animation sent in reply to the bot. Deviations: (1) 'reply is from the bot' now checks reply_to_message.from_user.id == bot.id, not v1's literal first_name == 'Cookiebot' (wrong for every other persona this codebase ships); (2) pooling has no funfunctions gate, matching v1's real asymmetry exactly (only sfw + sender-has-username) - only the reply side is fun-gated; (3) write is a Valkey-fronted ON CONFLICT DO NOTHING upsert, since a reference-table write is 2PC replicated to every node and most sends repeat a pack already pooled. Also unblocks the stickerdatabase importer collection (map_stickerdatabase mapped it to skip for want of a destination table; now maps every row - see platform_migration_etl). QA authored, not ported
Citus schema, colocation, UUIDv7
topology asserted by qa/integration/test_citus_topology.py