Repository navigation
docs(readme): recapture the daily entry, reports and sales screenshots - #865
Merged
Merged
Conversation
|
Important Review skippedReview was skipped due to path filters ⛔ Files ignored due to path filters (3)
CodeRabbit blocks several paths by default. You can override this behavior by explicitly including those paths in the path filters. For example, including ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Advanced Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
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 |
Owner
Author
|
@coderabbitai review |
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Owner
Author
|
@coderabbitai review |
|
mforce
added a commit
that referenced
this pull request
Sep 14, 2026
## Why `docs/images/dashboard.png` is the one README image #865 could not refresh. Its capture fails the spec's own guard (`screenshots.spec.ts`, bar heights must vary) because on the simulation fixture every day in the window is a **partial** day: the fixture seeds ~100 catalog flocks for the picker (#627) which are placed, active and never file, so every day owes a count nobody filed, `DayStripData.max` is null, and all fourteen bars render as the 2% floor stub. #865 states that gap and leaves the image at its 2026-09-02 state, which predates #781's bar strip and #791. The product rule is right and the fixture's counts are pinned by the picker-paging specs, k6 and the e2e suite. So the fixture stays as it is and the **capture** moves: the sim stack now carries a second, small farm seeded with the demo profile, and the dashboard image is taken from that. Closes nothing — no issue exists. It follows from #865's stated gap. ## Scope **`seed --profile demo --farm-code <slug>`** (`SeedCliCommand`, `DemoDataSeeder`). The code is resolved by slug exactly as `rename-account` and the lifecycle verbs resolve theirs (`AccountSlugLookup`), after the migrate and before the seed; an unknown code exits 1 naming `list-accounts`. `DemoDataSeeder.SeedAsync` takes the target account as an explicit `Guid?` parameter, never ambient state — `TenantContext` is single-assignment, so a seeder reading the tenant instead of setting it could only run where somebody else had already resolved one. Default behaviour is unchanged: every existing caller passes nothing and gets `SeedDefaults.AccountId`. `--profile simulation` **refuses** the flag: its manifest, its cast emails and the counts k6 and the e2e suite pin are all default-farm facts, so honouring it would need a second decision, not a parameter. Two things came along with that file. Every stderr path in the verb now routes through one sanitizing sink (#560) instead of two of five, because the shape where only the messages that quote argv get fixed is the shape `rename-account` was corrected out of. And the demo seeder's prerequisite messages no longer say "the default account", which stopped being true. **The sim harness** (#370 — all three files considered, and it says so below). `reset.sh` provisions `readme-farm` with `provision-account`, rotates its Owner off the printed one-time password onto a stable one, demo-seeds it, and preflights that the farm is signable and holds the demo fixture's three flocks. The timezone passed is `Simulation__TimeZoneId`, not a literal, so the two farms on one stack cannot end up on different clocks. - The stable password is generated in `bootstrap.sh` beside `SIM_ADMIN_PASSWORD` and read back by `reset.sh`, **mirroring the existing pattern** rather than minting one in `reset.sh`. Generating it in `reset.sh` would produce a new credential on every reset and leave `.sim-cast.json` describing the previous one. - It lands in `.sim-cast.json` under a top-level `readmeFarm` key, outside the `cast` array, because every entry there signs into `default-farm` and a driver iterating the cast must not have to ask which farm each member belongs to. - The rotation block is now **one shell function with two callers** rather than two copies of a credential-rotation block. - Re-running converges. `reset.sh`'s own flow never reaches the already-exists branch (`down -v` ran at the top), but `provision-account`'s duplicate behaviour is *not* a no-op like `bootstrap-admin`'s — it exits 1 with `Provision.SlugTaken*` and prints no password — so the branch checks the stable credential still signs in and carries on, and fails loudly on any other failure. - `verify-harness.sh` fails closed on a missing or blank `README_*` value and on a cast file that predates the `readmeFarm` key. - **`docker-compose.sim.yml` needs no change**, and that is a considered answer rather than an omission: the `README_*` vars carry no `__`, exactly like `SIM_ADMIN_*`, so they are script-level values `reset.sh` greps out of `.env.sim` and never app configuration. Nothing new reaches the container's `environment:` block. For the same reason there is no `src/Cluckwork.AppHost/Program.cs` change under #565 — no new required config key exists. **The e2e suite.** `cast.ts` exposes `readmeFarmOwner()`, whose return type widens `farmCode` from optional to required; `signIn` and the API sign-in helper take the code from the member, falling back to `default-farm`, so every persona written before this one is untouched. Only the dashboard capture uses it. **Docs.** `AGENTS.md`, `tools/simulation/README.md` ("Two farms on this stack" + the `.env.sim` parameter row + the `reset.sh` chain), `tools/simulation/ui/README.md`, and the dev-database runbook. **No GLOSSARY or Help change**, deliberately: no user-visible concept changed — the flag is an operator CLI argument and the second farm exists only inside the sim harness. ## Blast Radius `seed --profile demo` with no flag behaves exactly as before, which is what every existing caller does. `--profile simulation` gains one refusal on an argument nothing passes today. Per #394 the write contract is unchanged, so no caller under `tools/simulation/k6/` or `specs/` needed a change; the one Playwright caller that did (`signIn`) is in this diff, and `session-races.spec.ts`'s own hardcoded `default-farm` is correct as written because it drives a sim-cast persona. The `readme-farm` account exists only in a throwaway `cluckwork-sim` database. A regenerated `.env.sim`/`.sim-cast.json` is required — run `bootstrap.sh --force`, then `reset.sh`; `verify-harness.sh` says so by name if you forget. Two capture fixes ride along. The #780 readout assertion moved **below** the screenshot, because focusing a day leaves a focus ring and a readout balloon that `capture()`'s blur does not dismiss, and the first capture published both. And the `1280x1180` frame is now held open by the **Owner's sidebar** (its content ends at 1164px, measured on the rendered page) rather than by the main column, which on this farm ends at 700px — anything shorter clips the navigation mid-list. The comment in `playwright.screenshots.config.ts` says so, because the visible empty space below the panels otherwise invites a shrink that breaks the sidebar. ## Verification Everything below ran in the worktree, against the real stack. - `dotnet build Cluckwork.sln` — 0 warnings, 0 errors. - `bash tools/simulation/bootstrap.sh --force` then `bash tools/simulation/reset.sh` — up, migrated, both farms seeded, all four preflights green, and the temporary password redacted on both provisioning paths (checked in the log). - `cd tools/simulation/ui && npm ci && npm run screenshots` — **4 passed**, including the dashboard capture whose guard fails on the simulation fixture. `npm run typecheck` clean. - `SeedCommandTests` (8, up from 5) plus `SimulationSeedCommandTests` — 11 passed. Registry readers found by grepping `CliDispatcher.Commands|ProcessRoles.OneShotVerbs` under `tests/` rather than from memory: `CliDispatcherTests`, `OneShotVerbMinimalConfigTests`, `ProcessRoleRegistryTests`, run with `DemoSeedTests` and `DemoSeedActorTests` — 29 passed. - Both `PostgresImagePin_IsOneIdenticalString*` guards run before the markdown was committed — 2 passed. - **Mutation-checked, not asserted.** Reverting the CLI's account routing to `SeedAsync()` turns `SeedCommand_Demo_WithFarmCode_SeedsThatFarmAndLeavesTheDefaultEmpty` red. Deleting `readmeFarm` from the cast file, and blanking its password, each fail `verify-harness.sh` with exit 1; so does a blank `README_OWNER_EMAIL` in `.env.sim`. - The new seed test runs against **its own Postgres**, not the class fixture's, and that is the assertion rather than tidiness: "nothing landed under the default farm" is only meaningful on a database no sibling `[Fact]` has demo-seeded, and xUnit guarantees no order within a class. - `dotnet test Cluckwork.sln` — result in a comment below. The `RealSourceTree_AllBypassesAreAllowListed` guard fired on the `SeedAsync` signature change, which is #632's registry working: the entry is keyed by the enclosing symbol including its parameters, so adding one demanded a re-read. The justification is re-written rather than re-pinned — the `AccountId` predicate on those pre-`tenant.Resolve` queries is now the caller's account rather than always `SeedDefaults.AccountId`, and the bypass is still what lets the preflight see the target farm at all. ## Two judgment calls worth a reviewer's eye 1. **The function is `readmeFarmOwner()`, not `readmeFarm()`.** It returns a persona, like `owner()` and `restrictedWorker()` beside it, and `readmeFarm()` reads as though it returns the farm. 2. **The farm name was truncated and is now fixed.** "Meadowlark Farm" ellipsised to "Meadowlark F…" at the sidebar's 244px; commit 71f95ad names the farm "Meadowlark" (`README_FARM_NAME` in `bootstrap.sh`) and recaptures the image after a full reset. The comment below carries the new capture. ## Screenshot Before and after below. Same screen, same 1280x1180 frame; the before is the committed image this PR replaces.   <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit - **New Features** - The demo seed command can now target a specific farm with `--farm-code <slug>`. - Unknown farm codes return a clear error and guidance. - Simulation profiles explicitly reject the farm-selection option. - **Bug Fixes** - Sign-in and simulation tooling now correctly support members of non-default farms. - **Documentation** - Updated runbooks and simulation guidance describe multi-farm seeding and screenshot workflows. - **Tests** - Added coverage for targeted seeding, invalid farm codes, and unsupported simulation options. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Co-authored-by: mforce <cleyva@clvc.net>
mforce
pushed a commit
that referenced
this pull request
Sep 16, 2026
🤖 I have created a release *beep* *boop* --- ## [0.1.2](v0.1.1...v0.1.2) (2026-09-16) ### Features * **data:** standardize business record chronology ([#820](#820)) ([6231b31](6231b31)) * **infra:** optional leader-lease endpoint for pooled deploys ([#869](#869)) ([e9bc6a7](e9bc6a7)) * **sim:** seed a second farm for the README dashboard capture ([#867](#867)) ([de407c6](de407c6)) * **web:** adopt MUI, themed from the farm palette tokens ([#674](#674)) ([#860](#860)) ([6c83c5c](6c83c5c)) * **web:** convert Daily entry to MUI, field-first on the phone ([#888](#888)) ([b66f8b8](b66f8b8)) * **web:** convert the Dashboard and app shell to MUI ([#829](#829)) ([#883](#883)) ([2e94277](2e94277)) * **web:** retire the Slack-blue link colour for ink + a rule underline ([#884](#884)) ([c08f9d8](c08f9d8)) * **web:** serve a per-request CSP nonce so Emotion's styles apply under style-src 'self' ([#874](#874)) ([ba4e6f3](ba4e6f3)) * **web:** visual language theme overrides for the MUI revamp ([#864](#864)) ([#882](#882)) ([0bb6b73](0bb6b73)) * **web:** whole-app MUI baseline, theme policy guard and the [#740](#740) phone action rule ([#823](#823)) ([#871](#871)) ([af565e4](af565e4)) ### Bug fixes * **auth:** fail closed on unresolved flock-scope actors ([#787](#787)) ([#868](#868)) ([16d0350](16d0350)) * **auth:** make farm configuration owner-only ([#870](#870)) ([42f9036](42f9036)) * **e2e:** repoint the canary at the markup two PRs replaced ([#844](#844)) ([18b45dc](18b45dc)) * **i18n:** tl glossary uses the standard passive of ilagay ([#813](#813)) ([20dec10](20dec10)), closes [#738](#738) * **sim:** stop the k6-baseline EXIT trap masking a clean run as failed ([#838](#838)) ([f5ec96f](f5ec96f)) * **web:** declare the rule tokens the Dashboard reads, and guard undeclared custom properties ([#885](#885)) ([5bead1f](5bead1f)) ### Performance * **ci:** start the serialized integration collection first ([#861](#861)) ([1dcc7f6](1dcc7f6)), closes [#839](#839) ### Documentation * **auth:** record the OAuth 2.1 decision for MCP authentication ([#801](#801)) ([0510854](0510854)) * **designs:** MUI revamp design doc, component map, layout system, IA ([#862](#862)) ([da49481](da49481)) * **readme:** recapture the daily entry, reports and sales screenshots ([#865](#865)) ([f18e336](f18e336)) * **specs:** correct the sales_order_items column list in §10.5 ([#812](#812)) ([afe4a02](afe4a02)), closes [#737](#737) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: cluckwork-lockfix[bot] <309265648+cluckwork-lockfix[bot]@users.noreply.github.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.
Why
The README's four screenshots were captured on 2026-09-02, before #781 rebuilt both dashboard charts and #791 changed the capture tiles, and nobody recaptured. Three of the four are refreshed here from a sim stack rebuilt at
main(6c83c5c), reset and reseeded per the capture procedure intools/simulation/ui/README.md.Scope
docs/images/daily-entry.png,docs/images/reports.png,docs/images/sales.png: recaptured bynpm run screenshots.docs/images/dashboard.png. Its capture fails the spec's own guard (screenshots.spec.ts:108, bar heights must vary) because on the simulation fixture every one of the last 14 days is a partial day: the fixture carries 102 flocks and only two file entries, so no day is complete, the complete-day peak is null, and every bar is a 2% floor stub with no average. That is the honest picture of the fixture, and it is not a picture for the README. The dashboard image therefore stays at its 2026-09-02 state until the fixture or the partial-day rule changes; see the discussion on the PR.Blast Radius
Documentation only. The image job skips (#782); the tracked-file pin guard and GitGuardian still run.
Verification
bash tools/simulation/reset.shatmain: up, migrated, seeded (fingerprint8987971d), verified.npm run screenshots: 3 passed, 1 failed (dashboard, as above), and the three passing captures are the files in this diff.