urieljareth/AgendaPro
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
4c19244df93d606409a620b3fa27e3c13ff876f5
AgendaPro/.superpowers/sdd/progress.md
T
AgendaPro DevandClaude Opus 5 4c19244df9 feat: public landing page with magic login, brand palette and demo build flag 
Adds a public funnel landing at `/` and moves the login to `/login`, rebuilt
around the panel's own colors instead of an invented palette.

Landing (`src/components/landing/`, one section per file, no props):
- Seven funnel sections composed by `LandingPage`. The hero's eight swatches are
  literally `PIE_COLORS` from `DashboardPage`, the same hex values the seed hands
  to avatars and services, so "your colors become your numbers" is literal.
- `BrandMark` becomes the single source for the logo, replicating
  `public/favicon.svg`. Blue is the action surface, orange only ever marks.
- `NotebookVisual` is the one deliberate exception to the palette: it is what the
  product replaces.
- Copy drops all system vocabulary; motion comes from `src/lib/motion.ts` with a
  single easing, and reduced-motion resolves `initial` to the final state so a
  never-firing `whileInView` cannot leave a section invisible forever.

Routing and bundle:
- `homePathFor` is the single definition of each role's destination.
- Landing, login, dashboard and calendar load lazily. Eager, the login dragged
  framer-motion (~40 KB gz) into every panel load and the landing downloaded
  recharts + FullCalendar (~187 KB gz) without charting anything. `clsx` is
  pinned to the `react` chunk because Rollup otherwise assigns it to `charts`,
  making the entry import 111 KB gz for a 200-byte utility.

Responsiveness (iPhone/iPad), verified with `npm run audit:responsive`:
- No touch form field below 16px, `dvh` height utilities, safe-area insets, and
  40px touch targets keyed off `pointer: coarse` rather than `sm:`.
- New `.ld-gutter`: `.safe-x` lives outside `@layer` and beats Tailwind's `px-*`,
  so it left the login's side padding at 0 on anything but an iPhone in
  landscape. No overflow check could see it — there was no overflow, just zero
  margin. The audit now guards it with a `gutter` check.
- `shell-height` no longer fires on pages that legitimately scroll; the static
  `raw-viewport-unit` scan covers those instead.

Production build:
- The Dockerfile now sets `VITE_DEMO_UI=1` as a build arg. `DEMO` is a build-time
  constant, so without it Vite eliminated the magic login and the "Ver como…"
  switcher: the deployed landing promised "no registration" and led to an empty
  form. Verified by building both ways and diffing the bundle.
- Service worker cache bumped to v2 so the orphaned pre-landing chunks get purged
  from returning visitors' caches.

Verified: typecheck clean; unit 39/39, e2e 33/33, admin 17/17, booking 12/12,
landing 13/13 (WebKit), PWA passed against the real production build.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
2026-07-28 14:25:58 -06:00

13 KiB
Raw Blame History

SDD Progress Ledger — Auto-assign specialist + booking redesign

  • Plan: docs/superpowers/plans/2026-07-26-auto-assign-specialist-booking-redesign.md
  • Spec: docs/superpowers/specs/2026-07-26-auto-assign-specialist-booking-redesign-design.md
  • Branch: feat/auto-assign-specialist
  • Base commit: e8d2435
  • Commit policy: NO commits during execution (repo policy: only on explicit request). Verify per task via typecheck + tests. Reviewer reads changed files + plan task.
  • Pre-existing WIP (do NOT clobber): src/components/AppointmentModal.tsx, src/pages/CalendarPage.tsx, FullCalendar section of src/index.css (~lines 136-160). Use targeted edit only.

Tasks

  • Task 1: complete — db.ts (migrateV3ToV4 + runMigrations call), shared/types.ts (Business/Employee/WorkingHoursMap). Review: APPROVED, no findings. typecheck 0 errors, schema_version=4, backfill verified.
  • Task 2: complete — server/lib/scheduling.ts (pure helpers, no imports/DB), server/lib/scheduling.test.ts (15 tests), package.json test:unit. Review: APPROVED. test:unit 15/15. Minor (non-blocking): add direct test for hasConflict positive case, clamp01, parseWorkingHours malformed-shape, specialtyMatch substring branch — defer to final review.
  • Task 3: complete — appended DB fns to scheduling.ts (getCandidates, getExistingBusy, isAvailable, rankOne, pickBestSlotEmployee, autoAssign, runInTransaction). Plan typo c.info.id→c.id fixed & verified. Review: APPROVED. typecheck 0, test:unit 15/15, tx smoke ok. Low-note: getExistingBusy scopes by start_at within day (cross-midnight appts not caught — neutralized by working-hours guard; revisit if 24h scheduling added).
  • Task 4: complete — booking.ts publicBusiness (+auto_assign_specialist,working_hours) + slots endpoint uses real working_hours + pickBestSlotEmployee. POST /book untouched. Review: APPROVED. typecheck 0, slots smoke ok. Info-notes: local overlapsRange dup of overlaps (plan-permitted); per-slot SELECT name could be hoisted (pre-existing).
  • Task 5: complete — booking.ts POST /book rewritten: runInTransaction + autoAssign (when emp omitted or auto_assign_specialist=1) + isAvailable guard → 409 on conflict; 201 includes employee_id+reasons. Review: APPROVED. typecheck 0, test:unit 15/15, integration a)201 b)409 c)201-survivor d)409-autoAssign-null all pass. Sound tx guard.
  • Task 6: complete — appointments.ts POST / wrapped in runInTransaction, autoAssign fallback (replaced LIMIT 1), isAvailable guard → 409. Two plan-text deviations (move emp resolution inside tx; Number(business_id)) verified correct. Review: APPROVED. typecheck 0, test:unit 15/15, conflict smoke 409/201 pass. FOLLOW-UP: e2e-test.mjs fixtures (2026-09-15T11:00Z=05:00 local, 2026-09-20 Sun) now fail under working-hours enforcement — fixture-side, will fix in Task 7.
  • Task 7: complete — settings.ts (+auto_assign_specialist,working_hours, JSON/0-1 coercion), employees.ts (+specialties/working_hours/efficiency_score), booking-e2e.mjs (new), e2e-test.mjs slot-aware fixture fix (no guards weakened). Review: APPROVED. typecheck 0, test:unit 15/15, test:e2e 25/25, test:booking 9/9.

== BACKEND COMPLETE (Tasks 1-7). Algorithm + guards + APIs all verified green. ==

  • Task 8: complete — index.css wrapped .input/.select/.textarea in @layer components; BookingPage DetailsForm 3× pl-9→pl-10. Verified directly by controller (build green, @layer confirmed at index.css:272, WIP untouched). Icon/placeholder overlap fixed app-wide.
  • Task 9: complete — publicApi.ts types (PublicBusiness.auto_assign_specialist; BookPayload.employee_id optional; BookResponse +employee_id/reasons). Verified directly (typecheck 0, no consumer breakage).
  • Task 10: complete — src/components/MonthCalendar.tsx (42-cell month grid, nav bounded by min/max, brand selected, accent avail dot, brand today dot). Verified directly (typecheck 0, build green). Will be integration-tested in Task 11.
  • Task 11: complete — BookingPage DateTimePicker rewritten: md+ 2-col (MonthCalendar left + grouped slots right), SlotGroup Mañana/Tarde, mobile stacked, slot button styles unchanged, max optional. Review: APPROVED. typecheck 0, build ok. AM/PM noon boundary verified.
  • Task 12: complete — BookingPage wizard rewired to dynamic 3/4 steps via stepsFor(autoAssign)+currentOriginal; main max-w-5xl only on Horario; handleBook sends employee_id undefined when autoAssign; Confirmation shows reasons badges. Review: APPROVED. typecheck 0, build ok, HTTP smoke both flows + 409. NOTE: npm run lint can't run repo-wide (pre-existing ESLint9 flat-config gap, no eslint.config.js) — pre-existing, will report at Task 15.
  • Task 13: complete — src/lib/workingHours.ts (DEFAULT_WH/DAYS/parseWh) + SettingsPage auto-assign checkbox + 7-day working-hours editor. Review: APPROVED. typecheck 0, build ok, runtime e2e (PATCH→GET persist, Sat 10-14 slots appear, toggle 3/4 steps).
  • Task 14: complete — EmployeesPage EmployeeModal +specialties chips +efficiency slider +working-hours editor (inherit toggle). Payload specialties[]/efficiency Number/working_hours null|map. Review: APPROVED. typecheck 0, build ok. Low-notes: all-closed employee flips back to inherit (benign); case-sensitive dedup (acceptable). Runtime persistence smoke deferred to Task 15.

== ALL 14 IMPLEMENTATION TASKS COMPLETE + REVIEWED. ==

== TASK 15: FINAL VERIFICATION ==

  • typecheck: 0 errors
  • test:unit: 15/15
  • build: OK (2462 modules)
  • test:e2e: 25/25
  • test:booking: 9/9 (auto-assign 201+reasons, double-book 409, auto_assign_specialist exposed)
  • employee persistence runtime check (Task 14 deferred): specialties/efficiency=87/working_hours custom→null(inherit) all persist ✓
  • lint: PRE-EXISTING repo-wide failure (no eslint.config.js, ESLint 9.39 requires flat config) — NOT introduced by this feature.
  • FINAL WHOLE-BRANCH REVIEW: READY TO MERGE. Core guarantee (no double-booking from public flow) verified end-to-end (sync handler + BEGIN IMMEDIATE + hasConflict + same-tx INSERT). No cross-task integration issues.

== POST-MERGE FOLLOW-UPS (non-blocking, deferred) ==

  1. getExistingBusy: switch WHERE to overlap-based (end_at > dayStart AND start_at < dayEnd) — closes cross-midnight gap if 24h scheduling added.
  2. Add 4 unit-test cases (hasConflict positive, clamp01, parseWh malformed, specialtyMatch substring).
  3. Replace local overlapsRange with imported overlaps.
  4. Hoist per-slot employee-name SELECT out of slots loop.
  5. MonthCalendar nav buttons → h-11 w-11 (44px touch target).

== DEMO EMAIL DOMAIN TASK ==

  • Task 1: complete — source constants, README credentials, active fixtures, and executable plan login example updated. Commits 6355d18..50dea33; review clean.
  • Task 2: complete — local data/agendapro.db migrated 2 users; second update changed 0; appointments 161 and clients 15 preserved; review clean.
  • Task 3: complete with limitation — typecheck/build/unit/admin passed; auth paths passed; e2e/booking slot assertions remain blocked by pre-existing null working-hours data, not the email change; review approved.
  • Task 4: complete — committed source/docs through 3e056a0, pushed to GitHub and Gitea, Coolify deployment wbnbq9b0kneba67sznu7p1qf finished, runtime container healthy.
  • Task 5: complete — production health/page passed, new admin/owner logins returned 200, old-domain logins returned 401; active DB was fresh and seeded with new emails, so targeted SQL was a no-op.
  • Final review: complete — final-fix review approved; regex acceptance fix review approved.

== AGENDAMAX PWA INSTALLABILITY TASK ==

  • Task 1: complete — manifest, iOS metadata, reproducible Playwright-generated 192/512 PNG icons, and package scripts. Commits 9d2ab39..de947f9; review clean.
  • Task 2: complete — versioned shell/static service worker and production-only registration. Commits de947f9..65233f2; review clean.
  • Task 3: complete — install prompt hook/component, iOS Safari guide, safe-area styling, and Login/AppShell/AdminShell integration; review fixes added prompt guard, appinstalled precedence, close-button label, and compact mobile layout. Commits 65233f2..9ded129; review clean.
  • Task 4: complete — production-server PWA endpoint/browser tests, online/offline API bypass checks, three-shell coverage, and install documentation; review clean. Commits 9ded129..43f5d13.
  • Final fix wave: complete — credential-aware cache policy, AgendaMax-scoped cleanup, session dismissal persistence, configurable PWA test credentials, dialog semantics, and dismissal regression coverage. Commit 0b466f3; final whole-feature review approved.

== LANDING PÚBLICA + ACCESO DE UN CLIC (2026-07-28) == Spec: docs/superpowers/specs/2026-07-28-landing-funnel-bi-design.md Plan: docs/superpowers/plans/2026-07-28-landing-funnel-bi.md

Ruteo: / sirve la landing (pública, React.lazy); el login pasa a /login. homePathFor(user) centraliza el destino por rol. /b/:slug intacto, fuera del AuthProvider.

Dos iteraciones de diseño a pedido del usuario:

  1. Primera versión oscura/sobria (criterio Apple). Entregada y verificada.
  2. Rediseño a claro, vívido y animado, con copy humano para dueñas de salón y público de 50+ (guía de dolores: Base de Conocimiento - IA Negocios y Dev/03-Preguntas-Guia.md, Q15/Q72/Q87: hablar de lo que gana el cliente, no de lo que hace el sistema). Se eliminó todo vocabulario de sistema del copy visible: ni «business intelligence», ni «datos», ni «métricas», ni «panel».
  3. Paleta rebasada sobre los colores REALES del panel a pedido del usuario: las 8 muestras del abanico del hero son PIE_COLORS de DashboardPage.tsx; la marca sale de public/favicon.svg vía src/components/BrandMark.tsx (única fuente); CTA = brand-500→brand-700 como .btn-primary.

Defectos encontrados y corregidos durante la verificación (no estaban en el plan):

  • El chunk de entrada arrastraba framer-motion (39 KB gz) a TODA carga del panel, porque LoginPage era eager. Se hizo lazy junto con la landing.
  • La landing descargaba recharts (111 KB gz) y FullCalendar (76 KB gz) sin usarlos, porque DashboardPage y CalendarPage eran eager. Ambas a lazy, con Suspense en el de AppShell.
  • clsx lo comparten lib/format.ts y recharts; Rollup lo asignaba al chunk charts, así que el entry importaba 111 KB para una utilidad de 200 bytes. Fijado al chunk react.
  • LOGIN CON EL TEXTO PEGADO AL BORDE en móvil: .safe-x está fuera de @layer y gana a px-5, dejando el padding en 0 con inset 0 (todo lo que no sea iPhone landscape). Nueva clase .ld-gutter. Ningún check existente lo detectaba (no hay desborde: hay cero margen) → añadido check gutter.
  • shell-height del audit daba falso positivo en toda página que scrollea: comparaba #root contra el viewport vía el fallback a body.firstElementChild. Acotado a [data-app-shell]; a cambio se añadió el check estático raw-viewport-unit sobre las fuentes públicas (verificado inyectando min-h-screen en ProofSection: dispara).
  • Rótulo «Como una de tus muchachas» en /login con Diego Castillo en la lista debajo. Reescrito sin asumir género («Como alguien de tu equipo»); igual en ObjectionsSection.
  • Anclas bajo el nav fijo (scroll-mt-24); degradado del cierre invisible tras -z-10; etiquetas del abanico ilegibles (eliminadas); avisos flotantes solapando el abanico (reubicados a las 3 zonas libres) y detalle truncado (acortado).

Audits ajustados por el cambio de ruteo (/ → /login): visual-audit.mjs, responsive-audit.mjs y los SIETE goto(baseUrl) de pwa-e2e.mjs que esperan InstallAppPrompt. La landing se añadió a las listas de páginas de los dos audits.

VERIFICACIÓN FINAL (todo en verde):

  • typecheck: 0 errores
  • test:landing (NUEVO, WebKit): 13/13 — ruteo, acceso de un clic, sin desborde a 390px y prefers-reduced-motion sin contenido invisible en las 7 secciones
  • audit:responsive: 0 hallazgos (9 dispositivos × 10 páginas)
  • audit:visual: 62 pantallas, 0 fallos, 0 desborde, 0 errores de consola
  • test:pwa sobre el build: passed
  • test:unit 39/39 · test:e2e 33/33 · test:admin 17/17 · test:booking 12/12
  • build: el chunk de entrada NO importa charts, framer ni calendar (verificado en dist/)
  • WebKit a 390/440/744/820/1024/1440 en / y /login: sin desborde, sin campos <16px, sin objetivos <40px, sin errores

Sin commits (política del repo). npm run lint sigue roto por falta de eslint.config.js (preexistente).

FOLLOW-UPS DIFERIDOS:

  1. /login en desktop deja un vacío bajo la tarjeta del formulario (columnas de altura muy distinta). No es un defecto medible; revisar si molesta en uso real.
  2. La landing no tiene contenido de precios ni captura de leads: en fase demo la conversión es entrar a la demo. Habrá que decidirlo antes de un lanzamiento real.
  3. Seguridad sin cambios: /login expone contraseñas de cuentas existentes bajo la bandera DEMO. Autorizado solo para esta fase; es lo único que separa esto de una fuga si se publica.
Reference in New Issue View Git Blame Copy Permalink