Fold duplicated campanha access, loaders, list/filter/form shells, and scripts into shared modules; reorganize utilities by domain; stabilize e2e; and document prevention-first audit workflow so the same smells cannot creep back in.
Bookmarks the canonical list recorte beside Limpar, syncs via useSyncExternalStore, and matches active URLs with a pure href comparator so the app layout never imports the heavy URL serializer.
WebAuthn platform passkeys: opt-in enrollment after a password login, a
password-less sign-in that mints the same 14-day campaign-token, and
revocation from the profile. The session is minted through Payload's public
pieces (checkLoginPermission, addSessionToUser, getFieldsToSign, jwtSign)
inside a transaction with an advisory lock, so a passkey login neither
bypasses a lockout nor evicts the account's other sessions. The rpID is
derived from NEXT_PUBLIC_SITE_URL and resolves to null when the request host
disagrees, so previews never offer a ceremony the browser would refuse.
The picker shipped as a capability of CampaignTable rather than a municípios
feature, so every list that already uses the shared table got it: pass
`columnVisibility` and the table mounts the menu and filters the columns.
Persistence is a cookie, not localStorage, and the reason is a type rather
than a preference: CampaignTableColumn.cell is a function, so the column array
never crosses the server/client boundary and the client cannot rebuild the
table without it. The server filters, so a hidden column does not reach the RSC
payload either. `campaign_columns` carries a listId -> hidden ids map in a
compact form with no JSON and no percent-encoding, because document.cookie and
Next's cookies() disagree about decoding.
Closing the menu is the commit. The first cut debounced at 400 ms, which
batches a burst rather than a reading pace, so unchecking four columns while
reading their labels was four full-route refreshes — on /campanha/liderancas
each one re-pays the ~229 ms / 371-statement render already ledgered P1. The
timer survives at 3 s as a safety net for a menu left open; the price is that
the table behind the menu no longer updates per checkbox.
`defaultVisible` is gone: it was true on 10 columns and false on none, and
storing only the hidden set is what makes a newly shipped column visible to
everyone who never touched the picker. In exchange `label` is required, which
made `head` optional — 24 columns had a head that only restated the new label.
Each surface declares one mandatory column, re-filtered on the server so a
stale cookie cannot leave a row without its only linked cell.
The bug review caught is this codebase's recurring one: the RSC payload of
toggle A arrives inside the window of toggle B, and adopting the server there
undid B and then committed the set without it. Reconciliation now runs only
when no write is pending, pinned red against the previous code.
Municipality column ids are typed against MunicipalityListColumnId now, so the
id, the B22 description and the picker label are one key; the union member
`nivel` was renamed `level` to match the id and the ?level= filter. The
picker's own content comparison is sameIdSet, widened to string ids.
The "Municípios" column was dead text; it is now chips that navigate to the
município, with a territory chip when a leadership covers a whole TI, and it is
editable in place. What picks inline vs. Drawer is the `pointer:` media query,
not viewport width: on a fine pointer the editor is always mounted (the "×" sits
at the chip's corner, revealed on hover/focus) and on a coarse pointer a tap
anywhere in the cell opens a Drawer with 44 px targets.
`AdvisorMunicipalityCell` is promoted to `shared/MunicipalityPortfolioCell` and
/campanha/assessores migrates onto it in the same change, so the second call site
is born with the first. One batch action only; the single-id advisor twin is
deleted with its int tests migrated. The anti-goal on always-mounted list inputs
is amended for chip-based relations on fine pointers, not reopened.
Also fixes a latent bug: `CampaignTableHead`'s `description` seam broke any RSC
caller, because its two helpers lived in a 'use client' module.
`municipality.engagementLevel` (staff-only, no default) records what the
campaign decided to invest in a município, next to the E8/E10 metrics that
describe what it knows. Moving it is a reallocation decision, so only
unrestricted staff may, every move carries a rationale plus the signals that
would reverse it, and the write files an `allocationDecision` in the same
transaction under an advisory lock taken before the current level is read —
otherwise the snapshot's `from` is a guess. The enum gains `movimento`: a level
move is neither `aceita` nor `descarta`.
No backfill. Null is "Sem nível", and filtering by it is the triage queue.
That is also why the list sorts `nivel` in memory rather than in the query:
Postgres puts NULLs first on DESC, so a native sort would open "N4 primeiro"
on a page of municípios nobody has decided about yet. The filter stays native.
The hysteresis rules from research §6.8 are pure and client-safe, so the
Popover explains a block before submitting and the server re-runs them anyway;
a blocked move answers 409 with typed violations and the override is recorded
with exactly the rules it contradicted.
Tendência, Assessores and Votos estimados now share one container
(CampaignCellEditOverlay: Popover on md+, bottom Drawer below), the card drops
to the table's density, and the whole card navigates through a stretched link
instead of a second "Abrir município" button.
- Popover + NativeSelect sempre montado na coluna Status de
/campanha/liderancas, sem toggle de modo de edição
- Auto-save (~150ms) via POST /campanha/liderancas/support-status,
reusando isSameOriginRequest/campaignJsonMutationRoute (B24/B27),
sem revalidatePath
- Extrai leadershipStaffEditSafeMessages (adotado também em
liderancas/formActions.ts e nova/formActions.ts) e
utilities/leadershipLabels.ts (3º call site de supportStatusLabels,
agora fonte única também para SupportStatusBadge)
- Novo e2e tests/e2e/campaignLeaderships.e2e.spec.ts; setup.e2e.spec.ts
prewarma as rotas novas
- Fecha o Adiado com gatilho do B29; docs/plans, notebook e roadmap
atualizados
Gate: tsc/lint/format/knip(P3 pré-existente)/check:cycles/unit+int/e2e
verdes; two /simplify passes (labels/estado duplicado/isSupportStatus;
flush-on-close ao fechar o popover, comentário órfão de summary/
campos summary mortos removidos de SupportStatusBadge).
Salvador era um polígono só carregando os votos da cidade inteira, então o
coroplético não conseguia responder "qual pedaço de Salvador está fraco / é
meu". As 19 ZE ganham polígono próprio, dissolvido da malha de bairros do IBGE
(Censo 2022) pela lista de circunscrição da RA 02/2017 do TRE-BA, e o mapa passa
a pintar por map key (slug para as zonas, ibgeCode para o resto) em vez de
codarea. O polígono municipal de Salvador continua desenhado por baixo, sem dado
e sem interação, para não abrir buraco no contorno do estado.
Aproximação declarada na copy e no cabeçalho do script: o TRE publica a
circunscrição como lista de bairros, então o bairro é o átomo geométrico —
Periperi, que a resolução corta na Rua das Pedrinhas, vai inteiro para a ZE 17.
competitiveRank segue por cidade (o artefato TSE só tem federalRankByIbgeCode).
Sem migration, sem collection, sem action, sem Consent.
Terceira rodada de /simplify sobre o B14, com dois enunciados errados corrigidos:
- A propriedade da fronteira compartilhada NAO vem da ordem do `.find()`, e sim
da convencao meio-aberta do ray casting (`point.lng < crossingLng`): a aresta
pertence ao poligono cujo interior fica a leste (ao norte, se horizontal). Tres
comentarios diziam o contrario, e o do unit test dizia "oeste" enquanto
afirmava `east`.
- O plano dizia que o `Promise.all` esconde o decode da malha atras do fix de
GPS. Nao esconde: o decode roda no topo do modulo, quando o chunk executa. O
paralelo continua certo por outro motivo (decode ~2 ms; sequencial jogaria os
42 KB no caminho critico quando o fix vem do cache).
Simplificacoes: `unsupported` virou estado proprio, entao `failureCopy` deixa de
ser anulavel e caem duas leituras defensivas inalcancaveis; "Voce esta em " era
escrito duas vezes (frase sr-only + JSX), o que anulava justamente a invariante
que o comentario alegava; `zoneCityHrefs` virou `Partial` porque a busca falha
para os 416 municipios sem zona e o guard existia sem o tipo exigir;
`MunicipalityForIbgeIndex` era duplicata campo-a-campo de `AccessibleMunicipality`;
`kmFormatter` duplicava `oneDecimalFormatter`; `RecentlyVisitedCard` carregava
tres props que nenhum caller passa (dez condicionais mortas); e o bbox de feature
estava escrito em dois specs (agora `tests/helpers/featureBounds.ts`, compartilhado
por vitest e Playwright).
Copy: fora da Bahia sem nada por perto dizia a mesma coisa em duas frases e nunca
dizia que voce estava fora do estado. Agora o headline afirma o fato e o detalhe
reusa "Nenhum municipio da sua carteira por perto.".
Tambem limpa um achado de knip herdado do main (`MunicipalityPoliticalTrendViewModel`
exportado sem uso) e registra no plano o caminho afiado do debito de payload RSC
mais dois adiados com gatilho.
Card "Onde estou" no dashboard staff: point-in-polygon contra a malha
municipal que o painel do mapa ja carrega, sem artefato de centroides,
sem migration, sem collection, sem server action e sem Consent -- a
posicao e lida por navigator.geolocation e casada no browser, nunca sai
do aparelho.
- src/lib/municipalityProximity.ts: haversine, ray casting, centroide
ponderado por area e uniao discriminada
(inScope | zoneCity | outOfScope | outsideBahia)
- src/utilities/campaignGeolocation.ts: flag de sessao, leitura guardada
de permissao e requestCurrentPosition com falha tipada
- Salvador nao ganha ZE chutada: link para a lista filtrada, com href
serializado no servidor (importar o serializador canonico no cliente
custava 21 kB de First Load JS)
- fallback fora da carteira ordenado por centroide, teto de 150 km
- fix grosseiro (>10 km) mostra o resultado com ressalva de precisao
/campanha First Load JS: 265 -> 268 kB.
Replaces the checkbox-list + "Salvar assessores" button in the
Assessors popover (/campanha/municipios) with removable chips + a
Command combobox, delta-based auto-save (one advisor add/remove per
request, never the whole array), and optimistic UI with request
sequencing to resolve out-of-order responses correctly.
- New POST /campanha/municipios/advisors JSON route (mirrors
expected-votes/), backed by setMunicipalityAdvisorMembership(Record)
in actions/municipality.ts (text advisory lock + transaction).
- nextAdvisorIdsAfterMembership moved to
src/lib/municipalityAdvisorMembership.ts with unit tests.
- Shared isSameOriginRequest (src/utilities/sameOriginRequest.ts) and
uniqueRelationshipIds (src/utilities/relationship.ts) helpers,
de-duplicating the same-origin CSRF check and the
Set+map(relationshipId)+filter pattern across advisor call sites.
- MunicipalityListAdvisorsControl.tsx rewritten: chips, Command
combobox, per-item pending affordance, rollback on error, and a
request-sequence guard so only the most recently sent delta's
response is adopted when concurrent optimistic updates settle.
- e2e coverage for the combobox auto-save flow; setup prewarms the new
POST-only route to avoid Playwright dev-server HMR aborting the
first request.
- Plan and roadmap marked as delivered (B27).
Collection actionPlan -> activity (labels "Atividade"/"Atividades"), campo
campaignDemand.actionPlan -> activity, rota /campanha/planos ->
/campanha/atividades (criacao em atividades/nova), sidebar "Atividades" e
rotulos alinhados a entidade ("Tipo de atividade", "Origem da atividade",
"Resultado da atividade"). Valores de enum kind/status/origin inalterados.
Schema movido por migration data-preserving escrita a mao
(20260725_213000_rename_action_plan_to_activity): 4 tabelas, 3 enums, 2
colunas, 39 indices (4 pkeys inclusos), 14 FKs e 2 sequences, cada statement
guardado na existencia do objeto de origem, com down() simetrico. O gerador do
Payload emitiria DROP+CREATE e nao ve o indice parcial escrito a mao
action_plan_upcoming_start_at_idx, entao a lista veio de
pg_indexes/pg_constraint/pg_type/pg_sequences e o snapshot .json foi derivado
do anterior por transformacao textual -- validado por migrate:create __probe
sem mudanca de schema. A mesma migration alinha os 8 indices fosseis %plaza%
que a remodelagem de municipios deixou (renomeou as colunas plaza_id, nao os
indices).
Guard: o describe "banned campaign terminology" de
codebaseConventions.unit.spec.ts virou tabela de termos varrendo src+tests+
scripts (migrations fora -- historia congelada), banindo os compostos
actionPlan, a copy "plano de acao" e a rota antiga, mas nao plan/plano isolado
(colidiria com os planos de assinatura da Fase 2). Allowlist: o proprio spec,
scripts/generate-remodel-municipalities-migration.mjs e o comentario historico
em campaignMigrationReconciliation.int.spec.ts.
Plano: docs/plans/renomear-plano-acao-para-atividade.md
- ~70 strings visiveis + labels de admin (headers de coluna, APIErrors, empty states, dashboard, mapa, formularios, filtros) com concordancia de genero corrigida
- apoiadores desambiguado: cidade do contato vira 'Cidade'; a unidade operacional assume 'Municipio' (filtro, coluna, ficha)
- comentarios de codigo atualizados; payload-types regenerado
- fora por decisao D6: textos provisorios de consent do Onda 0 (lote juridico) e migrations congeladas
- e2e do plano de acao acompanha o label ('Municipio *')
A lista de municípios vira a fila de decisão, sem rota nova, sem coluna nova
e sem migration.
Fase 0 corrige a meta sintética do E8, que repartia a meta estadual
proporcionalmente só ao teto do campo e ignorava a votação própria (meta 2.911
para Vitória da Conquista, de 5.005 votos em 2022; 813 para Campo Formoso, de
47). `deriveSuggestedGoalsByScenario` substitui `decomposeStateGoal` e ancora a
meta na votação de 2022 por cenário — pessimista = base × (1 − margin), central
= base, otimista = base × (stateGoal ÷ Σ base), com clamp. `goalCoverage` passa
a receber `SuggestedGoalByScenario`, deixando `expectedVotes[S] ??
suggestedGoal[S]` simétrico. Só descrições do global mudaram.
Fase 1 adiciona frescor e ordenações: `declaredAt`/`estimatedAt` no select do
agregado de pledges (`lastPledgeAt`), `lastSignalAt` no view model via
`resolveMunicipalityLastSignalAt` (mesmo helper para o comparador do servidor e
a célula), e as sort keys `deficit` (novo default do staff) e `frescor`,
derivadas em memória no cenário central.
Fase 2 dobra os sinais nas células existentes: "há N dias" na coluna de última
atualização (frio ≥ 21 dias, em âmbar — vermelho já é da prioridade), badge
"sem responsável" em prioridade alta sem assessor, coluna da vergonha no
detalhe da métrica de assessoria, tooltip de cobertura nomeando o cenário e
copy Praças→Municípios.
Cortes: votos em jogo → B13, LQ/captura → E10, coluna dedicada de déficit
desnecessária.
Nova rota staff-only /campanha/conceitos (glossário curado em
src/lib/campaignIntelligenceConcepts.ts, revisado junto das fórmulas
que descreve em municipalityPotential.ts/goalCoverage.ts) com
descoberta por "Saiba mais" nos tooltips do card "Conta da cadeira" e
entrada fixa no rodapé da sidebar (staffSecondaryNav). Corrige defeito
E8 de formatElectionNumber (meta fracionária lida como cem mil em
pt-BR). Ajustes pós-/simplify: MetricExplanation.conceptID obrigatório,
href da sidebar desacoplado do módulo de conteúdo (client bundle).
Meta Bom/Regular/Mínimo duplicated the estimate scenarios — keep a single municipality series (pessimista/média/otimista), backfill then drop the old columns, and map the spreadsheet onto expectedVotes.
- knip (files/deps at error level; unused exports ledgered as warn) with
config for Payload/Next entries; deleted the dead modules it confirmed:
RecentlyVisited + UpcomingActionPlansCard + actionPlanUpcomingPreview,
lazy-dialog helpers, tseZone, federalCandidateTotalsAggregate (test-only),
unused ui kit files, one-shot rename script; pruned unused radix deps and
typescript-eslint; added pg/@types/pg (pinned to drizzle's 8.10.2)
- GitHub Actions CI on the mirror: lint, typecheck, knip, unit, migrate, int
against a Postgres 17 service
- eslint: unused-vars with ^_ ignore patterns (migrations args exempt),
repo-wide ban on `as never` (no-restricted-syntax), lint gate at
--max-warnings=0; fixed all 54 pre-existing warnings (dead imports,
hooks deps, tracking-pixel img disable with impeccable config exception)
- tsconfig: include consolidated from 219 lines to 5 (wildcard covers all
.next dist variants), noUnusedLocals on
- typed replacements for every as-never: hookFilledCreateData (actions
create data completed by collection hooks), drizzle scopeColumn narrowing,
tests' stub<T> helper (property-level checking kept, completeness waived)
- test infra: leased-consent generalization used by onda0 spec (shared
leases over the stable consent keys; window race ledgered), purge fix
- docs: IMPROVE-CODE-QUALITY-PLAN.md tracker, TECH-DEBT.md ledger,
TESTING.md safety-net map (package.json also gains the Phase 2
build:election-aggregates script entry; the script lands next commit)
Replace the `Function`-typed payload param (no-unsafe-function-type lint error
blocking pre-push) with `Pick<Payload, 'find' | 'create'>`, which also surfaced
missing Lexical `version` fields in the consent fixture — now added.
Rename the operational unit "Praça"/"Núcleo" to "Município" (code identifier
`municipality`) and land the 2026-07-23 course changes:
- New `candidate` role with unrestricted visibility (`isCampaignStaff` now
includes it) alongside coordinator/advisor; leader is locked down to the
supporter contact tool only.
- Demandas become staff-only; dobradinhas modeled via a `stateDeputy` entity.
- Camaçari consolidated to a single município (ZE 170/171 aggregated);
Salvador keeps its 19 zone municípios.
- Map moves to the staff dashboard; `/campanha/municipios` is list/overview.
- Vote pledges keep the declared vs estimated (scenario) asymmetry; leaders
never see estimates.
Destructive migration for the campaign vertical (drops nucleus/plaza-era
tables, wipes campaign rows, seeds 435 municípios). Public-site data untouched.
Includes the post-delivery /simplify cleanups: shared reloadStaffActor /
reloadCoordinatorActor helpers, req.context cache for getFreshCampaignUser,
parallelized dashboard loads, candidate dashboard-copy fix, and dead
leader-dashboard / roleKind code removal.
436 seeded Praças (municipality-based; per-TSE-zone in Salvador/Camaçari) become
the operational unit; roles renamed to coordinator/advisor/leader with advisors
scoped to administered Praças. Leadership is one record per person across many
Praças/organizations; vote pledges split leader-declared vs staff-estimated
(estimates never serialized to leaders). Adds organizations, demand workflow
with staff-only cost/receipts, action-plan results with deputy presence, manual
political trend, multi-candidate comparison and the diverging red/white/blue
map mode. Migration 20260721_020109_remodel_plazas deliberately resets the
campaign vertical in production (public site untouched); test suite ported to
the new model (unit 187 / int 306 / e2e 9 green).
Align the campaign desk with quieter KPI strips, queues-first dashboard, shared shells, and advanced filters; capture Impeccable design context and the remaining simplify/critique debts on the roadmap.
Phase 1 UX: StrictCombobox municipality in SupporterForm, AlertDialog in
RemoveSupporterDataButton. Phase 2 DRY: campaignListUrl + CampaignListPagination
+ campaignFormFields + mapCampaignFormActionError shared across nuclei,
supporters and action plans. Phase 3: single SQL aggregate for supporter list
KPIs via drizzle in the Payload txn. Phase 4: bulk drizzle insert for CSV import
inside the existing Payload txn with ON CONFLICT DO NOTHING, plus a fail-closed
skipContactPhoneInvariant context flag. Phase 5: HMAC-signed single-use import
token backed by a new supporterImportBatch staging collection so the full ok-row
set never crosses the Server Action boundary twice.
Replace scalar region/city/neighborhood with hasMany arrays so nuclei can span multiple Bahia municipalities while keeping neighborhoods scoped to a single city.
Adds electoral nuclei, leaderships, updates, and WhatsApp invite flows behind campaign auth, plus the consolidated schema migration that Vercel applies on deploy.
