Repository navigation
test(storage): the object store's refusals are exercised - #1069
Merged
Merged
Conversation
Four storage rows were Partial only because no test reached the branch that enforces them. Each now has a test on the assembly the node serves, on Postgres, and each test was observed red with its branch removed: - MR-STOR-0023 (spec §5.2): a put stating the address of other bytes is refused, and nothing is held at either address. - MR-STOR-0025 (§5.4): different bytes at an address are reported as corruption, and the held row is left as it was. - MR-STOR-0026 (§5.5): a held row damaged in place, in its payload or its namespace, is never served; an intact object beside it still is. - MR-STOR-0016 (§3.4): the same refusal on the node, with Core's re-hash (`wrong_bytes_never_count`) on the reader's side. The existing tests' helpers are untouched, so no pinned evidence moves; the new tests have their own (`member_and_store`, `put_stating`). CONFORMANCE §6.41 and §8, and the verification matrix, record the rows and their mutation controls; the totals are regenerated.
cryptskii
added a commit
that referenced
this pull request
Sep 30, 2026
A change that touches the storage node and not the SDK runs the `sdk-node-protocol` group, which named four SDK suites that no longer exist: - `b0x_integration` (deleted in #977) and `storage_full_lifecycle` (deleted in #932), integration targets. cargo refuses an unknown target, so the group failed on the first storage-only PR since then (#1069). - `sdk::storage_sync_sdk` and `sdk::storage_node_health`, lib filters. A filter that matches nothing runs nothing and passes, so these narrowed the group's coverage silently. The four names are removed. No SDK integration binary exercises the storage-node contract now, so the group runs one command: the four lib suites that do (`handlers::storage_routes`, `sdk::storage_node_sdk`, `sdk::b0x_sdk`, `sdk::storage_set`). The selector's self-test now refuses a lib filter that does not resolve to a module of dsm_sdk (`test_every_named_sdk_suite_is_a_module_that_exists`). Mutation control: `sdk::storage_sync_sdk` put back, the test is red.
Signed-off-by: Cryptskii <47649969+cryptskii@users.noreply.github.com>
cryptskii
added a commit
that referenced
this pull request
Sep 30, 2026
…numbered once (#1071) #1065, #1066 and #1069 merged together and left main (375506f) with three defects in its records. Evidence pins. #1066 moved code that 266 pins cover, and it merged before they were refreshed, so Code map is red on main and on every PR based on it. They are refreshed at 375506f: - against CI's code map of that commit, whose tree fingerprint equals this tree's; - the stale rows' evidence, 156 tests in 13 cargo runs, run here in release: 156 passed, 0 failed, 0 ignored; - 264 rows moved only in code, repinned as a code refresh; - MR-DSM-0114 and MR-DSM-0119 on `dsm::merkle::sparse_merkle_tree::DEFAULT_SMT_HEIGHT` (android) also moved in reading, REACHED_VIA_DISPATCH to REACHED: #1066 defines `const LEAF_LEVEL: usize = DEFAULT_SMT_HEIGHT as usize;` (sparse_merkle_tree.rs:154), and the map reaches the height through it. Accepted with --accept reading. Exactly those 266 rows changed. The comparator reads 0 failing rows and 581 of 581 pins PINNED. Totals. #1069's branch update kept both sides of the `All` row, and neither was the count. Regenerated by ci/conformance_evidence.py --write: 340 Met, 233 Partial, 117 Missing, 10 Violated, 64 Not code, 154 Deferred. §6 numbering. Three sections were numbered 6.41. They now run in merge order: - 6.41 is #1065; its six citations are unchanged; - 6.42 is #1066; nothing cites it; - 6.43 is #1069, moved after the other two, and its four row citations (MR-STOR-0016, 0023, 0025, 0026) now read §6.43. Two headings get back the blank line the merge dropped. Nothing else in the file changed. No code changes.
cryptskii
added a commit
that referenced
this pull request
Sep 30, 2026
…erified CONFORMANCE_GAPS.md conflicted in §6 and §7. Resolved: - §6 keeps main's order, 6.41 #1065, 6.42 #1066, 6.43 #1069, and this branch's section follows as 6.44. Its nine §8 references (MR-DSM-0017, 0019 twice, 0046, 0056, 0126, 0138, 0141, 0170) now read §6.44. The four MR-STOR rows keep §6.43, as on main. - §7 is regenerated by ci/conformance_evidence.py --write: 343 Met, 229 Partial, 118 Missing, 10 Violated, 64 Not code, 154 Deferred, the sum of the subtotal rows. Every section body is byte-identical on both sides; only numbers moved. From main this brings the refreshed pins (#1071). This branch still owes pins for its eight new manifest rows.
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.
What
Four storage-node requirement rows were Partial only because no test reached the branch that enforces them. This adds those tests. No production code changes.
dsm_storage_node::immutable_store_round_trip::…)a_put_stating_another_address_is_refused_and_nothing_is_helddifferent_bytes_at_an_address_are_reported_as_corruptiona_held_object_that_no_longer_hashes_to_its_address_is_not_serveddsm::sofi::storage::tests::wrong_bytes_never_counton the reader's sideThe damaged-row tests change a held row directly in Postgres, the way a failing disk would. That puts the store outside the fault model (§3), which is exactly the case these refusals exist for.
Mutation controls
Each mutant compiled, and its named test failed at the intended assertion. The source was restored to
main's after each.put_immutablenever refusesa_put_stating…: 201 where 422 is expectedget_immutableserves without recomputing the addressa_held_object…: 200 where 500 is expectedput_immutableanswers a conflict as an ackdifferent_bytes…: 200 where 500 is expectedEvidence pins
The existing tests' helpers (
member,put,get) are untouched, so no pinned evidence closure moves. The new tests have their own helpers,member_and_storeandput_stating.Records
ci/conformance_evidence.py --write).Runs
cargo test --locked -p dsm_storage_node --release -- --nocapture --test-threads=1: 64 passed, 0 failed, 0 ignored, on the rebased tree.cargo fmt --all -- --checkandcargo clippy --all-targets -- -D warnings(pinned 1.98.0): exit 0.make lintwas not run locally: this worktree has nonode_modules, and this change touches no frontend file.