access.PolicyStore: PutPolicy/ListPolicies/DeletePolicy, plus version and seeding methods
Each repeats the insert/get/list/update/delete-by-id shape with per-entity validation and sentinel errors. Consumer tests hand-write a per-entity fake.
Wanted: one generic CRUD[E] interface in storecontract — Insert(ctx, E) (string, error), Get(ctx, id) (E, error), List(ctx), Update(ctx, E), Delete(ctx, id) — with per-entity extras (Approve/Pause/Resume, Set*) staying on the specific interfaces, and per-entity sentinels carried by the entity's adapter (e.g. an errors hook) rather than by method names. Plus one generic in-memory fake for tests. Postgres implementations stay per-entity (sqlc generates per-table SQL; the generic interface removes the Go-side duplication, not the SQL).
Done when:
CRUD[E] exists and at least MappingStore and EventTypeStore are expressed as CRUD[...] + extras
one shared fake serves the consumer tests that today hand-roll fakes
task check green
From the 2026-08 architecture review follow-up (complements #1781, #1783).
**What is wrong:** several store interfaces are the same CRUD operation set repeated under entity-prefixed method names:
- `MappingStore`: InsertMapping/GetMapping/ListMappings/UpdateMapping/DeleteMapping
- `BindingStore`: the same five + Approve/Pause/Resume
- `SourceStore`: Insert/Get/List/Delete + Set* mutators
- `EventTypeStore`: Insert/Update/Delete/List
- `access.PolicyStore`: PutPolicy/ListPolicies/DeletePolicy, plus version and seeding methods
Each repeats the insert/get/list/update/delete-by-id shape with per-entity validation and sentinel errors. Consumer tests hand-write a per-entity fake.
**Wanted:** one generic `CRUD[E]` interface in `storecontract` — `Insert(ctx, E) (string, error)`, `Get(ctx, id) (E, error)`, `List(ctx)`, `Update(ctx, E)`, `Delete(ctx, id)` — with per-entity extras (Approve/Pause/Resume, Set*) staying on the specific interfaces, and per-entity sentinels carried by the entity's adapter (e.g. an errors hook) rather than by method names. Plus one generic in-memory fake for tests. Postgres implementations stay per-entity (sqlc generates per-table SQL; the generic interface removes the Go-side duplication, not the SQL).
**Done when:**
- `CRUD[E]` exists and at least MappingStore and EventTypeStore are expressed as `CRUD[...]` + extras
- one shared fake serves the consumer tests that today hand-roll fakes
- `task check` green
sam
added the discovery label 2026-11-02 15:26:38 +00:00
sam
added the medium label 2026-11-14 14:45:42 +00:00
sam
added the deferred label 2026-11-15 17:26:12 +00:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
From the 2026-08 architecture review follow-up (complements #1781, #1783).
What is wrong: several store interfaces are the same CRUD operation set repeated under entity-prefixed method names:
MappingStore: InsertMapping/GetMapping/ListMappings/UpdateMapping/DeleteMappingBindingStore: the same five + Approve/Pause/ResumeSourceStore: Insert/Get/List/Delete + Set* mutatorsEventTypeStore: Insert/Update/Delete/Listaccess.PolicyStore: PutPolicy/ListPolicies/DeletePolicy, plus version and seeding methodsEach repeats the insert/get/list/update/delete-by-id shape with per-entity validation and sentinel errors. Consumer tests hand-write a per-entity fake.
Wanted: one generic
CRUD[E]interface instorecontract—Insert(ctx, E) (string, error),Get(ctx, id) (E, error),List(ctx),Update(ctx, E),Delete(ctx, id)— with per-entity extras (Approve/Pause/Resume, Set*) staying on the specific interfaces, and per-entity sentinels carried by the entity's adapter (e.g. an errors hook) rather than by method names. Plus one generic in-memory fake for tests. Postgres implementations stay per-entity (sqlc generates per-table SQL; the generic interface removes the Go-side duplication, not the SQL).Done when:
CRUD[E]exists and at least MappingStore and EventTypeStore are expressed asCRUD[...]+ extrastask checkgreen