Section titles are stacked single-cell rows: the "TipiLAN 2026 CS2" title
sits above notice rows ("If a team withdraws..."), and parse_team_sections
kept only the LAST one, so the notice clobbered the title and resolve_divider
saw no game/year keywords -> None -> teams never positioned.
- parse_team_sections now accumulates all single-cell rows above a header, so
the game/year title survives alongside the notices.
- _refresh_teams_sync resolves the divider from "<tab name> <section title>",
so the game is taken reliably from the CS2/LoL worksheet name while the year
still comes from the title, keeping year-scoped TEAM_DIVIDER_*_2026 vars.
- regression test: title survives notice rows and still resolves.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XPRsW4tazVtYi2jNzVQkre
Divider placement matched the divider role by its exact Discord name, so
renaming the role in Discord silently broke positioning. Switch the
TEAM_DIVIDER_<SUFFIX> config to hold a role ID; resolve the ID to the
role's current name in apply_team_role_positions and keep the existing
name-based ordering maths downstream unchanged.
- config._parse_team_dividers now parses values as ints (rejects non-ints)
- resolve_divider / _team_divider cache / get_team_dividers return IDs
- apply_team_role_positions resolves each ID via guild.get_role once
- .env.example documents IDs and ships the CS2/LoL divider role IDs
- resolve_divider tests updated to assert IDs
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XPRsW4tazVtYi2jNzVQkre
The team-role feature was aimed at the wrong bot. Tournament participants
live in the economy/community guild, but team-role sync had been bolted onto
the dev bot's roster sync (sync_member), which bails out for anyone missing
from the internal member sheet - so it could never reach its actual audience.
Decouple it: keep the (roster-independent) team-sheet parsing, pull the team
wiring off the dev/member-sync path, and re-home it on the economy bot.
- core/member_sync: revert sync_member to add-only (drop team block +
SyncResult.roles_removed); add roster-independent sync_team_role and a
whole-guild sync_all_team_roles returning a reporting summary. Still only
ever touches role NAMES present in the team sheet.
- commands/economy_team_commands: new admin-only /teamsync command.
- bot.py: hourly team_sync_hourly task (economy-only, no-op unless
TEAM_SHEET_ID is set; first tick at boot covers startup load); register
/teamsync under the economy profile; drop the dev-side startup load.
- commands/dev_member_commands: /check no longer refreshes teams or reports
removed roles.
- strings + .env.example: TEAMSYNC_UI, CMD[teamsync], document TEAM_SHEET_ID.
- tests: retarget sync tests to sync_team_role; add sync_all_team_roles case.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R6JZkyszyDFuFtk25WBbcR
Match members by Discord username against the lineup nicknames in the
separate registration spreadsheet (TEAM_SHEET_ID) and give each their
team's role. One team per person: switching teams removes the old team
role, and a team with no Discord role yet is auto-created. Only role
names present in the sheet are ever touched, so organisation/field/base
roles are never at risk; the feature is a no-op when TEAM_SHEET_ID is
unset.
The sheet is not a single table (merged rows, stacked CS2/LoL sections
with different layouts), so it is parsed via raw-row scanning rather than
get_all_records. Citizenship markers like "(EST)" are used only as
player delimiters and discarded.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The economy had many coin faucets but almost no sinks: the /shop is
one-time ownership, and gambling/rob fines route to the house (which
players drain back via jackpots and heists), so they recirculate rather
than destroy coins. Result: steady inflation.
Add a consumables shop as a true recurring sink - buying destroys the
coins and grants a temporary boost, so there's always something to spend
on after gear is maxed:
- Energiajook XL (500) - 1h of 2x earnings on /work, /beg, /crime
- XP jook (500) - 1h of 2x EXP
- Kohv (300) - instantly clears all cooldowns
Timed buffs live in a new active_buffs field ({kind: expiry_iso}), pruned
on read; rebuying extends the timer. Effects hook where they belong:
earn_mult in income.do_work/do_beg/do_crime, exp_buff_mult in
levels.award_exp; kohv is self-contained. New /consumables command browses
the menu (with active buffs) or buys a boost. Covered by 9 tests.
Note: active_buffs is a new PocketBase field - run
scripts/sync_pb_schema.py before deploying or buffs won't persist.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
strings/ is split into domain submodules whose names are re-exported from
strings/__init__.py. Adding a constant to a submodule and forgetting to
re-export it - or shadowing a name across two submodules - would only
surface as a runtime crash in a command. This test asserts every submodule
__all__ entry is reachable as strings.NAME, that no name is defined twice,
and that strings.__all__ matches the union of the submodules. Submodules
are auto-discovered, so a new one is covered without editing the test.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
core/economy has long been a package (store/income/gambling/shop/levels/
jail/heist/prestige/fishing/quests/leaderboards/house/admin) rather than a
single core/economy.py, but README.md and docs/DEV_NOTES.md still described
it as one file in ~27 places. Point every reference at the real submodule,
add a "Defined in" column to the constants quick-reference and a Module
column to the strings table, and fix the stale FISH -> FISH_CATALOGUE name.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A follow-up reentrancy sweep of the five remaining money-moving views found
one more instance of the same bug class. The other four (quests, prestige,
fish, shop) verified clean - their underlying do_* functions are idempotent.
RequestView / FundModal (commands/economy_support_commands.py):
- on_submit read self._view.remaining, then awaited do_give, then decremented
remaining. Because discord.py dispatches each modal submit as its own task
and do_give is a plain non-idempotent transfer, a funder could open two
modals and submit both before the first resolved: both read the same
pre-decrement remaining, both passed the range check, and both transferred
`amount` - over-funding the request (remaining goes negative) and moving up
to the funder's whole balance.
- Fix: reserve the amount synchronously (decrement remaining BEFORE the do_give
await, with no await in between - atomic under asyncio), and roll the
reservation back if the transfer fails. The second concurrent submit now
sees the reduced remaining and is rejected. Added a cheap _fund guard
(remaining<=0 / is_finished) so a click on a funded request doesn't open a
dead modal.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Multi-agent audit of the money-moving economy modules surfaced three
confirmed correctness bugs. All three are fixed here with regression tests.
heist (core/economy/heist.py):
- `house = await get_user(house.HOUSE_ID)` shadowed the imported `house`
module for the whole function, so `house.HOUSE_ID` raised UnboundLocalError
on every successful heist (win payout was entirely dead) and on the
fail-path compensation branch. Rename the local to `house_rec`.
- Un-shadowing exposed a latent mint: the pot was floored at 300 but the
house was debited only min(total, balance), so a poor house paid out more
than it lost. Cap the pot at the balance and debit exactly what is paid
(house debit == sum of payouts). No mint, no leak.
- Add the missing `_is_jailed` import (do_heist_check referenced it unimported).
blackjack (commands/economy_games_commands.py):
- Button callbacks had no reentrancy guard; discord.py dispatches each click
as its own task, so double-clicking Stand within the dealer-reveal window
paid out twice (mint), and double-clicking Double/Split deducted the extra
bet twice. Add a synchronous `_busy` guard (matching the existing RpsGame
idiom) on all four callbacks plus a `_resolved` idempotency flag on
settlement, so a game can only pay out once.
bail (core/economy/jail.py, commands/economy_extra_commands.py):
- do_bail only checked balance, never jail state; a double-click or a stale
BailView from a re-run /jailbreak charged bail twice, destroying coins (bail
is a pure sink). Make do_bail a no-op when the user is not jailed, and add a
UI reentrancy guard + "already free" message.
Tests: 47 passed (4 new regression tests covering heist coin-conservation and
bail idempotency).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
