feat: single-user architecture with first-time setup flow#11
Open
nathanialhenniges wants to merge 3 commits intomainfrom
Open
feat: single-user architecture with first-time setup flow#11nathanialhenniges wants to merge 3 commits intomainfrom
nathanialhenniges wants to merge 3 commits intomainfrom
Conversation
Convert Dirework from multi-user to single-owner-per-instance model. Removes all per-user FK columns from config/state tables (true singletons), adds instanceConfig table for overlay tokens, gates signup to one owner, and introduces a /setup route that self-disables once claimed. Key changes: - Schema: drop userId FK from all singleton tables (timerState, timerConfig, timerStyle, taskStyle, botConfig, botAccount); add instanceConfig singleton with overlayTimerToken/overlayTasksToken; add isOwner flag to user table; drop ownerId from task (keep authorTwitchId/authorDisplayName); fresh init migration - Auth: replace allowlist with owner-gate hook (count-based, first signup wins, sets isOwner=true); session hook provisions all singletons on first login; export hasOwner() helper; add drizzle-orm as explicit dependency - Event bus: drop :userId suffix from event constants (TIMER_STATE_CHANGE, TASK_LIST_CHANGE, BOT_CONFIG_CHANGE) - tRPC routers: remove all userId/ownerId scoping from queries; overlay token lookup via instanceConfig; rename ensureUserConfig→ensureSingletons - Bot service/commands: remove userId parameter; load config unscoped; broadcaster check via unscoped user.findFirst() - Setup route: /setup server component calls notFound() if owner exists; client UI with Twitch claim button; home page shows toast on instance_claimed error - Remove ALLOWED_TWITCH_IDS env var entirely - Tests: update fixtures to singleton schema; fix vi.hoisted() for mock ordering Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- README: fix pnpm → bun commands in Getting Started; update step 7 to explain /setup claim flow on first run - CLAUDE.md: remove ALLOWED_TWITCH_IDS from env var docs (removed in refactor) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
Warning Rate limit exceeded
Your organization is not enrolled in usage-based pricing. Contact your admin to enable usage-based pricing to continue reviews beyond the rate limit, or try again in 43 minutes and 8 seconds. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (7)
WalkthroughArchitectural migration from multi-user to single-owner instance model: removes per-user scoped configurations and overlay tokens, consolidates config tables to singleton pattern with shared Changes
Sequence Diagram(s)sequenceDiagram
participant User as User/Browser
participant Server as Next.js Server
participant Auth as Auth Service
participant DB as Database
participant SetupUI as Setup UI
User->>Server: GET /
Server->>Auth: getSession(headers)
Auth->>DB: Query user (check existence)
DB-->>Auth: user row or null
alt No users exist (instance unclaimed)
Auth-->>Server: session = null
Server->>DB: hasOwner()
DB-->>Server: false
Server-->>User: Render HomeContent (Sign in link)
User->>Auth: signIn.social('twitch')
Auth-->>User: Redirect to /setup
else Users exist (instance claimed)
alt User is authenticated
Auth-->>Server: session with user
Server-->>User: Redirect to /dashboard
else User not authenticated
Auth-->>Server: session = null
Server->>DB: hasOwner()
DB-->>Server: true
Server-->>User: Render HomeContent
end
end
User->>Server: GET /setup
Server->>Auth: getSession(headers)
Server->>DB: hasOwner()
alt Instance already owned
DB-->>Server: true
Server-->>User: 404 Not Found
else Instance not owned
alt User authenticated
Auth-->>Server: session with user
Server->>DB: Create user (isOwner=true)
DB-->>Server: user created
Server->>DB: Provision instanceConfig
DB-->>Server: config created
Server-->>User: Redirect to /dashboard
else User not authenticated
Auth-->>Server: session = null
Server-->>SetupUI: Render SetupContent
SetupUI-->>User: Display "Claim with Twitch" button
end
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~75 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🧹 Nitpick comments (10)
apps/web/src/app/(app)/setup/page.tsx (1)
1-3: Consolidatenext/navigationimports.
redirectandnotFoundcome from the same module; merge into one import.♻️ Proposed fix
-import { redirect } from "next/navigation"; import { headers } from "next/headers"; -import { notFound } from "next/navigation"; +import { notFound, redirect } from "next/navigation";🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@apps/web/src/app/`(app)/setup/page.tsx around lines 1 - 3, Consolidate the duplicate import from "next/navigation" by combining the two separate imports into a single import statement that includes both redirect and notFound (leave headers imported from "next/headers" as-is); update the file-level imports so only one import from "next/navigation" exports redirect and notFound to remove redundancy and keep imports tidy.packages/api/src/routers/timer.ts (1)
89-91: Consider an explicit.where()on singleton updates.
ctx.db.update(schema.timerState).set(data).returning()has no predicate; it updates every row. That is correct today because the table is a singleton, but an explicit.where(eq(schema.timerState.id, "singleton"))documents intent and guards against accidental drift if the invariant changes. Same applies to the other update sites in this file (lines 113, 126, 144, 186, 194).🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/api/src/routers/timer.ts` around lines 89 - 91, The update calls like ctx.db.update(schema.timerState).set(data).returning() currently lack a predicate and will update all rows; add an explicit WHERE predicate such as .where(eq(schema.timerState.id, "singleton")) to each update call to document intent and guard the singleton invariant. Locate the update expressions (e.g., ctx.db.update(schema.timerState).set(data).returning()) and add .where(eq(schema.timerState.id, "singleton")) (import eq if necessary), and apply the same change to the other update sites in this file (the similar update calls referenced in the review).packages/db/src/schema/auth.ts (1)
14-14:isOwnercolumn looks good; single-owner invariant relies on app-level enforcement.The column is notNull with a safe default. Enforcement that only one user can exist lives in the Better Auth
user.create.beforehook inpackages/auth/src/index.ts(counts rows and throwsFORBIDDEN). If you want defense-in-depth at the DB level, consider adding a partial unique index (e.g.CREATE UNIQUE INDEX user_single_owner_idx ON "user" ((is_owner)) WHERE is_owner = true) in a future migration — not required for this PR.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/db/src/schema/auth.ts` at line 14, isOwner column is correctly not-null with default false and the single-owner invariant is enforced in the Better Auth hook user.create.before in packages/auth/src/index.ts; if you want DB-level defense-in-depth add a migration that creates a partial unique index (e.g. name it user_single_owner_idx) on the "user" table for rows WHERE is_owner = true so the database will prevent more than one owner in addition to the application-level check.apps/web/src/app/(app)/home-content.tsx (1)
3-5: Nit: merge the tworeactimports.
SuspenseanduseEffectare imported from"react"on two separate lines. Consolidate to a single import for readability.♻️ Proposed fix
-import { Suspense } from "react"; +import { Suspense, useEffect } from "react"; import { useSearchParams } from "next/navigation"; -import { useEffect } from "react"; import {🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@apps/web/src/app/`(app)/home-content.tsx around lines 3 - 5, Merge the two separate React imports into a single import statement by combining Suspense and useEffect into one import from "react" (replace the separate imports of Suspense and useEffect with a single line importing both) and remove the now-unused duplicate import; ensure the import includes both Suspense and useEffect in the same curly-braced import.packages/api/src/bot/index.ts (1)
174-186: Minor: log when reloadConfig fails to load a config row.If the insert succeeds but
findFirstreturnsundefined(e.g., due to an unexpected schema issue or a transaction anomaly),configCachestaysnulland every chat message silently short-circuits atonMessage(line 78). That's very hard to diagnose from outside. A singlelogger.warn/logger.errorwhenbotConfigRowis falsy would make the failure mode observable.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/api/src/bot/index.ts` around lines 174 - 186, In reloadConfig, add a log when the inserted-but-not-found case occurs: after awaiting this.db.query.botConfig.findFirst() check if botConfigRow is falsy and call a logger (e.g., this.logger.warn or this.logger.error) with a descriptive message including context like "reloadConfig: botConfig row missing after insert" and any available diagnostics (e.g., this.db transaction id, schema version, or the inserted result) before returning so that the unexpected state where configCache remains null is observable; update references in reloadConfig, botConfigRow, and configCache accordingly.packages/auth/src/__tests__/allowlist.test.ts (1)
1-71: Nit: rename file now that allowlist logic is gone.All
parseAllowedTwitchIdscoverage was removed and this file now exclusively testshasOwner(). Consider renaming tohas-owner.test.ts(orowner.test.ts) so the filename reflects the suite and future grep/navigation matches intent.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/auth/src/__tests__/allowlist.test.ts` around lines 1 - 71, Test file name no longer matches its contents: rename the test file from allowlist.test.ts to has-owner.test.ts (or owner.test.ts) so it reflects the single exported function under test (hasOwner) and improves discoverability; ensure any CI/test patterns or docs referencing the old filename are updated and keep the existing describe/it blocks as-is since they already target hasOwner.packages/api/src/bot/commands.ts (1)
170-171: Hot-path: avoid a DB lookup for the owner's twitchId on every!task/!next.
handleTaskAdd(line 170) andhandleTaskNext(line 430) each executedb.query.user.findFirst(...)per chat message purely to determineisBroadcaster. In a single-owner instance the owner'stwitchIdis effectively immutable for the lifetime of the bot; fetching it on every message is wasted I/O on the chat hot path.Consider caching it once at bot startup (e.g. store
ownerTwitchIdonTwitchBotServicenext tochannelName, loaded instart()from the samedb.query.user.findFirstthat already runs there) and passing it throughMessageContextso these handlers can do a plain string compare. Same semantics, no per-message query.Also applies to: 430-431
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/api/src/bot/commands.ts` around lines 170 - 171, The handlers handleTaskAdd and handleTaskNext are doing a DB lookup per message to compute isBroadcaster; instead, add a cached ownerTwitchId property on TwitchBotService (loaded once in start() using the existing db.query.user.findFirst call) and propagate that value into MessageContext so the handlers can compare strings (ownerTwitchId === userInfo.twitchId) instead of calling db.query.user.findFirst each time; update TwitchBotService.start to set ownerTwitchId, adjust MessageContext shape to include ownerTwitchId, and change handleTaskAdd and handleTaskNext to read from context and perform a plain string compare.packages/api/src/routers/config.ts (1)
378-380: Scope updates to the singleton row defensively.All six
db.update(schema.X).set(...).returning()calls here omit awhereclause, so they update every row in the table. That's the intended behavior today because these tables hold a singleid = 'singleton'row, but the invariant is implicit and easy to violate (a bad migration, a manual DB insert, or a future feature that adds a second row will silently corrupt every other record on the next save).Consider adding an explicit
.where(eq(schema.X.id, "singleton"))to every mutation. It encodes the assumption in the code and costs nothing at runtime.Also applies to: 391-393, 404-406, 457-494, 513-523, 537-539
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/api/src/routers/config.ts` around lines 378 - 380, The update calls like ctx.db.update(schema.timerConfig).set(input).returning() must be scoped to the singleton row; add an explicit .where(eq(schema.timerConfig.id, "singleton")) to that call and to the other five similar mutation calls (e.g., where you call ctx.db.update(schema.<Table>).set(...).returning()). Also ensure the eq symbol is imported/available in the file so you can reference eq(schema.<Table>.id, "singleton") for each update. This encodes the singleton invariant and prevents accidental multi-row updates.packages/db/drizzle/meta/0000_snapshot.json (1)
1-1695: Clean-slate migration: confirm no existing deployments need an upgrade path.This snapshot (and the replaced
0000_init.sqlper the PR description) represents a reset of the migration history — any existing database from the previous multi-user schema will not migrate cleanly withdrizzle-kit migrateand would need to be dropped/recreated. That's consistent with the pre-release single-owner pivot, but worth confirming:
- No production/staging DB holds data that must be preserved.
- Contributors with local dev databases will need to drop them (mention in README/CLAUDE.md if not already covered).
If confirmed, the snapshot contents themselves (singleton defaults, new
instance_config,is_owneron user, updated task indexes with removedowner_id) match the authored schema inpackages/db/src/schema/app.ts.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/db/drizzle/meta/0000_snapshot.json` around lines 1 - 1695, This change replaces migration history with a clean-slate snapshot (0000_snapshot.json + replaced 0000_init.sql) so confirm there are no production/staging databases or any contributor local DBs that must be preserved, and update documentation (README or CLAUDE.md) with explicit instructions to drop/recreate local dev databases before running drizzle-kit migrate; also verify the snapshot schema (instance_config table, user.is_owner column, task index changes) exactly matches packages/db/src/schema/app.ts and mention the required manual upgrade/drop steps in the PR description or docs so users are not surprised.packages/api/src/routers/overlay.ts (1)
53-115: Optional: deduplicate the initial-yield and change-loop read blocks.The
Promise.all([timerState, timerConfig, timerStyle])fetch and its mapping to the yielded shape appears three times across this file (initial + loop inonTimerState, plusgetTimerState), and the task equivalent three times as well. Extracting small helpers likereadTimerSnapshot(ctx)/readTaskSnapshot(ctx)would keep all yields in lockstep and reduce drift risk if the shape changes later.♻️ Sketch
async function readTimerSnapshot(ctx: Ctx) { const [timerState, timerConfigRow, timerStyleRow] = await Promise.all([ ctx.db.query.timerState.findFirst(), ctx.db.query.timerConfig.findFirst(), ctx.db.query.timerStyle.findFirst(), ]); return { timerState: timerState ?? null, timerConfig: timerConfigRow ? buildTimerConfig(timerConfigRow) : null, timerStyles: timerStyleRow ? buildTimerStylesConfig(timerStyleRow) : null, }; }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@packages/api/src/routers/overlay.ts` around lines 53 - 115, Duplicate DB-read + mapping logic in onTimerState, onTaskList (and the separate getTimerState path) should be extracted into small helpers to avoid drift; add async functions like readTimerSnapshot(ctx) and readTaskSnapshot(ctx) that perform the Promise.all queries (timerState/timerConfig/timerStyle and task/findMany/taskStyle respectively) and return the mapped shape (using buildTimerConfig/buildTimerStylesConfig and buildTaskStylesConfig). Replace the initial-yield blocks and the for-await loop bodies in onTimerState and onTaskList (and call sites such as getTimerState) to invoke these helpers and yield their result, ensuring all yields share the same mapping logic.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@apps/web/src/app/`(app)/setup/setup-content.tsx:
- Around line 57-63: The OAuth errorCallbackURL currently hardcodes
error=instance_claimed for all Twitch sign-in failures, which mislabels
unrelated failures; update the authClient.signIn.social call (provider:
"twitch") to use a generic error code such as error=signin_failed (or a distinct
error_type param) instead of instance_claimed, or add a more specific parameter
(e.g., error_type=instance_claimed) only when you can deterministically detect
the FORBIDDEN/owner-claimed case; adjust any home page toast logic to read the
new generic key and display appropriate copy for generic sign-in failures vs.
the specific instance-claimed case.
In `@packages/api/src/routers/user.ts`:
- Around line 30-41: The mutate path regenerateOverlayToken can cause an
uninitialized instanceConfig row to be created with the "other" token due to the
schema $defaultFn when using insert(...).onConflictDoUpdate; to prevent that,
call ensureSingletons() first (the routine that provisions instanceConfig) or
replace the insert-onConflict pattern with an explicit update targeting the
singleton row (use
ctx.db.update(schema.instanceConfig).set({...}).where(eq(schema.instanceConfig.id,
SINGLETON_ID))) so rotating one token does not implicitly mint the other; update
references: regenerateOverlayToken, schema.instanceConfig, ensureSingletons,
eq(...), and SINGLETON_ID.
In `@packages/auth/src/index.ts`:
- Around line 36-44: The current create.before hook in
packages/auth/src/index.ts (the async "before" handler that checks db.select
count on schema.user and throws APIError) is racy; add a DB-level partial unique
index to enforce a single owner (e.g. CREATE UNIQUE INDEX user_single_owner_idx
ON "user"(is_owner) WHERE is_owner = true) via a migration, and also catch the
DB unique-constraint failure when inserting a second owner and translate it into
the same APIError ("FORBIDDEN" / "This instance is already claimed. Single-user
only.") so concurrent requests get a friendly error; keep the existing before
hook for the common-path short-circuit but rely on the DB constraint plus the
insert error handling to guarantee correctness.
---
Nitpick comments:
In `@apps/web/src/app/`(app)/home-content.tsx:
- Around line 3-5: Merge the two separate React imports into a single import
statement by combining Suspense and useEffect into one import from "react"
(replace the separate imports of Suspense and useEffect with a single line
importing both) and remove the now-unused duplicate import; ensure the import
includes both Suspense and useEffect in the same curly-braced import.
In `@apps/web/src/app/`(app)/setup/page.tsx:
- Around line 1-3: Consolidate the duplicate import from "next/navigation" by
combining the two separate imports into a single import statement that includes
both redirect and notFound (leave headers imported from "next/headers" as-is);
update the file-level imports so only one import from "next/navigation" exports
redirect and notFound to remove redundancy and keep imports tidy.
In `@packages/api/src/bot/commands.ts`:
- Around line 170-171: The handlers handleTaskAdd and handleTaskNext are doing a
DB lookup per message to compute isBroadcaster; instead, add a cached
ownerTwitchId property on TwitchBotService (loaded once in start() using the
existing db.query.user.findFirst call) and propagate that value into
MessageContext so the handlers can compare strings (ownerTwitchId ===
userInfo.twitchId) instead of calling db.query.user.findFirst each time; update
TwitchBotService.start to set ownerTwitchId, adjust MessageContext shape to
include ownerTwitchId, and change handleTaskAdd and handleTaskNext to read from
context and perform a plain string compare.
In `@packages/api/src/bot/index.ts`:
- Around line 174-186: In reloadConfig, add a log when the
inserted-but-not-found case occurs: after awaiting
this.db.query.botConfig.findFirst() check if botConfigRow is falsy and call a
logger (e.g., this.logger.warn or this.logger.error) with a descriptive message
including context like "reloadConfig: botConfig row missing after insert" and
any available diagnostics (e.g., this.db transaction id, schema version, or the
inserted result) before returning so that the unexpected state where configCache
remains null is observable; update references in reloadConfig, botConfigRow, and
configCache accordingly.
In `@packages/api/src/routers/config.ts`:
- Around line 378-380: The update calls like
ctx.db.update(schema.timerConfig).set(input).returning() must be scoped to the
singleton row; add an explicit .where(eq(schema.timerConfig.id, "singleton")) to
that call and to the other five similar mutation calls (e.g., where you call
ctx.db.update(schema.<Table>).set(...).returning()). Also ensure the eq symbol
is imported/available in the file so you can reference eq(schema.<Table>.id,
"singleton") for each update. This encodes the singleton invariant and prevents
accidental multi-row updates.
In `@packages/api/src/routers/overlay.ts`:
- Around line 53-115: Duplicate DB-read + mapping logic in onTimerState,
onTaskList (and the separate getTimerState path) should be extracted into small
helpers to avoid drift; add async functions like readTimerSnapshot(ctx) and
readTaskSnapshot(ctx) that perform the Promise.all queries
(timerState/timerConfig/timerStyle and task/findMany/taskStyle respectively) and
return the mapped shape (using buildTimerConfig/buildTimerStylesConfig and
buildTaskStylesConfig). Replace the initial-yield blocks and the for-await loop
bodies in onTimerState and onTaskList (and call sites such as getTimerState) to
invoke these helpers and yield their result, ensuring all yields share the same
mapping logic.
In `@packages/api/src/routers/timer.ts`:
- Around line 89-91: The update calls like
ctx.db.update(schema.timerState).set(data).returning() currently lack a
predicate and will update all rows; add an explicit WHERE predicate such as
.where(eq(schema.timerState.id, "singleton")) to each update call to document
intent and guard the singleton invariant. Locate the update expressions (e.g.,
ctx.db.update(schema.timerState).set(data).returning()) and add
.where(eq(schema.timerState.id, "singleton")) (import eq if necessary), and
apply the same change to the other update sites in this file (the similar update
calls referenced in the review).
In `@packages/auth/src/__tests__/allowlist.test.ts`:
- Around line 1-71: Test file name no longer matches its contents: rename the
test file from allowlist.test.ts to has-owner.test.ts (or owner.test.ts) so it
reflects the single exported function under test (hasOwner) and improves
discoverability; ensure any CI/test patterns or docs referencing the old
filename are updated and keep the existing describe/it blocks as-is since they
already target hasOwner.
In `@packages/db/drizzle/meta/0000_snapshot.json`:
- Around line 1-1695: This change replaces migration history with a clean-slate
snapshot (0000_snapshot.json + replaced 0000_init.sql) so confirm there are no
production/staging databases or any contributor local DBs that must be
preserved, and update documentation (README or CLAUDE.md) with explicit
instructions to drop/recreate local dev databases before running drizzle-kit
migrate; also verify the snapshot schema (instance_config table, user.is_owner
column, task index changes) exactly matches packages/db/src/schema/app.ts and
mention the required manual upgrade/drop steps in the PR description or docs so
users are not surprised.
In `@packages/db/src/schema/auth.ts`:
- Line 14: isOwner column is correctly not-null with default false and the
single-owner invariant is enforced in the Better Auth hook user.create.before in
packages/auth/src/index.ts; if you want DB-level defense-in-depth add a
migration that creates a partial unique index (e.g. name it
user_single_owner_idx) on the "user" table for rows WHERE is_owner = true so the
database will prevent more than one owner in addition to the application-level
check.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 217cd40d-1fdb-49d7-a3ab-30974787bdc8
⛔ Files ignored due to path filters (1)
bun.lockis excluded by!**/*.lock
📒 Files selected for processing (34)
.env.exampleCLAUDE.mdREADME.mdapps/web/instrumentation.tsapps/web/src/app/(app)/api/bot/callback/twitch/route.tsapps/web/src/app/(app)/home-content.tsxapps/web/src/app/(app)/page.tsxapps/web/src/app/(app)/setup/page.tsxapps/web/src/app/(app)/setup/setup-content.tsxpackages/api/src/__tests__/events.test.tspackages/api/src/bot/commands.tspackages/api/src/bot/index.tspackages/api/src/events.tspackages/api/src/routers/__tests__/config-roundtrip.test.tspackages/api/src/routers/__tests__/config.test.tspackages/api/src/routers/bot.tspackages/api/src/routers/config.tspackages/api/src/routers/overlay.tspackages/api/src/routers/task.tspackages/api/src/routers/timer.tspackages/api/src/routers/user.tspackages/auth/package.jsonpackages/auth/src/__tests__/allowlist.test.tspackages/auth/src/index.tspackages/db/drizzle/0000_init.sqlpackages/db/drizzle/0001_pretty_jocasta.sqlpackages/db/drizzle/meta/0000_snapshot.jsonpackages/db/drizzle/meta/0001_snapshot.jsonpackages/db/drizzle/meta/_journal.jsonpackages/db/src/index.tspackages/db/src/schema/app.tspackages/db/src/schema/auth.tspackages/db/src/schema/index.tspackages/env/src/server.ts
💤 Files with no reviewable changes (5)
- CLAUDE.md
- .env.example
- packages/env/src/server.ts
- packages/db/drizzle/0001_pretty_jocasta.sql
- packages/db/drizzle/meta/0001_snapshot.json
- OAuth error codes: replace hardcoded error=instance_claimed with error=signin_failed on both setup and home sign-in buttons; add toast handler for signin_failed on home page. The instance_claimed code is reserved for future server-side detection; generic failures now get a neutral message instead of a misleading claim toast. - regenerateOverlayToken: replace insert...onConflictDoUpdate with a bare update(...).where(eq(id, 'singleton')) to avoid silently minting the other token via schema $defaultFn when instanceConfig hasn't been provisioned yet. - Single-owner TOCTOU: add partial unique index user_single_owner_idx ON "user"(is_owner) WHERE is_owner = true to enforce the single-owner invariant at the DB level. The app-level count check still runs for a friendly FORBIDDEN in the common path; the DB constraint handles concurrent races. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
userIdFK from all config/state tables (timerState,timerConfig,timerStyle,taskStyle,botConfig,botAccount); each usesid='singleton'as the primary keyinstanceConfigtable: holdsoverlayTimerTokenandoverlayTasksToken(moved offusertable), provisioned automatically on first logindatabaseHooks.user.create.beforecounts existing users — first signup setsisOwner=true, any subsequent attempt throwsFORBIDDEN;ALLOWED_TWITCH_IDSenv var removed entirely/setuproute: server component callsnotFound()if an owner already exists; client shows "Claim this instance" UI with Twitch sign-in; home page shows a toast on?error=instance_claimed:userIdsuffix from all event constants (timerStateChange,taskListChange,botConfigChange)userId/ownerIdfilters); overlay token lookup viainstanceConfig;ensureUserConfigrenamedensureSingletonsstart(db)signature (no userId); broadcaster check via unscopeduser.findFirst(); task queries dropownerId0000_init.sql(app not yet released — clean slate)drizzle-ormadded to@dirework/authdependencies (new import forhasOwner()count query)buncommands and/setupclaim step; CLAUDE.md env var section cleaned upTest plan
bun run db:start && bun run db:push && bun run dev:web→ visithttp://localhost:3001→ redirects to/setup/dashboard; DB showsuser.isOwner=true, all singleton rows exist,instanceConfighas tokens/setup→ 404/?error=instance_claimed→ toast showninstanceConfig; SSE pushes work on timer/task mutationsbun run check-types→ 0 errorsbun run test→ 175 tests passingSKIP_ENV_VALIDATION=true bun run next build→ all 15 routes compiled🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Bug Fixes
Documentation
buninstead ofpnpmChores