alkblobs/tasks/suite-column-mem.md

id, name, status, depends_on, scope, risk, impact, level, tags

id	name	status	depends_on	scope	risk	impact	level	tags
suite-column-mem	Conformance suite — factory binding and the mem column (the reference column)	completed	store-read-paths store-put-paths gc-protocol	moderate	medium	phase	implementation	wave-2 suite mem

Description

Bind the conformance harness to the real facade and grow the mem column — the contract-reference column that defines the required set every other engine's rows mirror (store-api.md invariant 7; ADR-011 §2).

• The factory binding (the skeleton's anticipated wave-2 moment): bind StoreFactory::Store to the real facade type — most plausibly the associated type collapsing to the concrete Store across all factories. If binding reveals a shape correction, make it and record it in tests/conformance/factory.rs + README.md per the skeleton's note. The mem factory vends fresh, isolated in-memory stores (no shared state between rows) with idempotent teardown; mem never skips (README §SKIP posture — SKIP machinery stays unneeded until the SQL waves).
• The mem column's rows, each ADR-stamped per the convention (Contract stamp: citing the pinning spec §/ADR §), tracing to the backlog entries:
  • INV-1 round-trips: whole-blob put/get byte-identical under the canonical derivation, over a content matrix looping both put paths (known-length, unknown-length) in-body; the git-interop test remains the derivation's source of truth.
  • INV-2 dedup: identical content twice (same and different put path) is one pool entry; the re-put's Pin protects exactly as a fresh put's.
  • INV-3 stage hygiene: exact counts over both put paths and both failure modes (explicit discard, failed commit) — zero residue on failure, exactly one committed entry on success (the degenerate kv stage form).
  • INV-4 sweep safety: single-node exact-count family — happy shape (exact SweepReport counts), abort shapes (ProtectFailed, NoLivenessSources — nothing deleted, Ok(SweepReport) always), and the sweep-vs-pinned-put race.
  • INV-5 list correctness: asserted through sweep outcomes (exact counts), never by comparing enumerated lists directly.
  • INV-6 stat discipline: stat returns length + algorithm without fetching content.
  • Constructor-mode rows (facade contract, ADR-011 §4 / ADR-012 §6.2/§6.3): mem-only mode; kv-only(mem) over-threshold put-time rejection (known-length immediate, unknown-length mid-stream).
• Rows take factories, never stores (factory.rs contract); a failing assertion panics the row; teardown is idempotent against that path.

Acceptance Criteria

• StoreFactory::Store bound to the real facade; any shape correction recorded in factory.rs + README
• The mem column green: INV-1/2/3/4/5/6 rows plus constructor-mode rows, every row ADR-stamped and traceable to the backlog
• The mem column's row set is the reference set (INV-7's basis) — recorded in README as what other columns mirror
• Factory isolation + idempotent teardown hold (fresh store per call; teardown safe against a panicking row)
• cargo test --features mem green (the column runs under its feature); clippy/fmt gates green

References

• tests/conformance/README.md (the conventions this grows against)
• tests/conformance/backlog.rs (INV-1…7 entries and their first-column plan)
• docs/architecture/store-api.md (§Invariants — the test gate)
• docs/architecture/decisions/011-vocabulary-tiers-engines-and-fleet-state.md §2 (mem's reference role)
• docs/plans/implementation.md (Decided points — suite home; §Wave 2 — suite columns)

Notes

Filled by the implementation agent (2026-10-11).

• The factory binding was already made and recorded by the store-construction task: open() vends the concrete alkblobs::Store facade (the skeleton's Store associated type collapsed), recorded in factory.rs + README §The factory contract before this task ran. Verified; no further shape correction was needed. The wave-1 binding demonstrator (binding.rs) remains as the proof-of-implementability tests.
• The reference row set lives in tests/conformance/rows.rs (engine-uniform, factory-generic, every row ADR-stamped and traced to the backlog): INV-1 (content matrix × both put paths, in-body), INV-2 (same-path + cross-path legs), INV-3 (both paths × both injectable failure modes — mid-stream Io failure and declared-len Verification mismatch; the kv stage form is degenerate, so residue is the pool itself, observed through exact sweep outcomes), INV-4 (happy shape over the full protection set; all three GcAbortCause abort shapes — NoLivenessSources, ProtectFailed, SweeperLock, the last completing the single-node family per ADR-013 §2/ADR-012 §5; the sweep-vs-pinned- put race over both put paths), INV-5 (exact sweep outcomes, never list comparison — honored by every row), INV-6, and the two constructor-mode rows.
• INV-6's slice through the facade: stat's shape (len + algorithm), purity, and None-on-miss; the engine-level non-content-load half rides the tier size probes' own bodies (ADR-008 §1) — noted in the row doc; the mismatch-based proof needs a plant seam the factory contract deliberately lacks.
• The kv-only row is threshold-agnostic by discovery: a known-length rejection reads nothing, so Unreadable probes distinguish rejection from acceptance and binary search finds the factory's rejection point (~64 free probes); the row then asserts immediate/mid-stream rejections and both at-threshold-lands legs. Configuration contract recorded in the README: threshold < u64::MAX and materializable (≤ 8 MiB, MAX_BOUNDARY_MATERIALIZATION). Any column's kv-only factory mirrors the row unchanged.
• Mem factories (mem_column.rs): MemOnlyFactory (the reference host for the INV rows) and KvOnlyMemFactory at a column-declared threshold (1024) — fresh isolated store per open by construction, teardown a no-op (mem holds no external state), trivially idempotent and panicking-row-safe. Structural wiring tests (not backlog rows) demonstrate factory isolation and teardown-against-panic: factory_open_vends_fresh_isolated_stores, factory_teardown_is_safe_against_a_panicking_row.
• Row lifecycle convention recorded: the row body owns open/assert/teardown; a failing assertion panics before teardown (README §The factory contract grew the bullet). The column runs under feature mem; rows.rs compiles ungated so the reference rows stay type-checked in default builds. No SKIP machinery added.

Summary

Filled on completion (2026-10-11).

Landed the mem column — the conformance suite's contract-reference column (ADR-011 §2). tests/conformance/rows.rs carries the reference row set: INV-1…6 families (12 rows total, incl. all three sweep-abort shapes and the sweep-vs-put race) plus the two constructor-mode rows (mem-only posture; kv-only put-time rejection), each ADR-stamped and traceable to the backlog. tests/conformance/mem_column.rs binds the factories (mem-only, kv-only(mem)) and runs every row; the factory contract's isolation and idempotent-teardown guarantees are demonstrated structurally, including the panicking-row path. README records the reference row set (INV-7's basis — what other columns mirror), the row-lifecycle convention, and the new current state. The wave-2 factory binding (concrete Store facade) was already made and recorded upstream of this task; verified, nothing further to correct.

Verification: cargo test --features mem green (97 lib + 18 conformance + 4 git-interop), cargo test (default) and cargo test --all-features green, cargo clippy --all-features --all-targets -- -D warnings green, cargo fmt --check green.