Skip to content

release: 0.9.0-beta.8, the pin to bsv-rs 0.9.0-beta.8 - #43

Merged
Calgooon merged 3 commits into
betafrom
repin/bsv-rs-0.9.0-beta.8
Nov 17, 2026
Merged

Calgooon merged 3 commits into
betafrom
repin/bsv-rs-0.9.0-beta.8

Conversation

@Calgooon

@Calgooon Calgooon commented Nov 17, 2026 •

Copy link
Copy Markdown
Contributor

The beta pin to bsv-rs =0.9.0-beta.8 (on crates.io 2026-11-16T22:38:07Z, the tag at 31a6ccf0), the version 0.9.0-beta.8 and the changelog entry. One release commit, 7c3d5cb.

No code change. This crate names bsv_rs::transaction::MerklePath and MerklePathLeaf alone (src/compact.rs, src/evidence.rs); none of bsv-rs 0.9.0-beta.8's ten breaking changes reaches a call here. The one that names MerklePath, its writers returning Result (bsv-rs #817), meets no call: the crate writes a path through its own compact form. Cargo.lock moves the bsv-rs row and this crate's own row, nothing else.

Why now. bsv-rs 0.9.0-beta.8 carries the stable line's advisory fixes (0.8.5 through 0.8.11), and the toolbox and the CLI pin bsv-rs exactly: a build holds one bsv-rs, so their beta.8 pins need a tracker prerelease that pins it (bsv-wallet-cli#565).

Read at the pin, in a clone on the box, each as ci.yml runs it: cargo check --locked --all-features --all-targets finished; cargo test --locked: 64 passed, 0 failed; cargo doc --locked --no-deps generated; cargo fmt --all -- --check exit 0; cargo clippy --locked --all-targets -- -D warnings finished; the chain-event vectors 8 passed and the ARC status vector 1 passed; cargo build --locked --target wasm32-unknown-unknown --features wasm finished; cargo deny --locked check: advisories ok, bans ok, licenses ok, sources ok (rustc 1.98.1, cargo-deny 0.20.2).

Who lands. The beta line of this repository is held by the bsv-wallet-cli captain from this prerelease on (the owner's word of 2026-11-17, in the keeper's rulings and in the CLI captain's session). After one cold read and the six required checks: rebase-merge, the tag v0.9.0-beta.8 on beta's tip, the Release run, the index. main is not touched.

After the cold read (fix, 0.88), the fold 15c2a91. The read found that "no code change" was true of the source and incomplete as a record: bsv-rs #817 also changes the merkle walk behind MerklePath::new_unchecked and compute_root, which Proof::new and Proof::root call. A path whose walk pairs a hash holding a character that is not hex is now refused at the constructor (CheckError::InvalidProof); through 0.9.0-beta.7 it was built and failed at the header check (RootMismatch, Reask::ProofFailed). No word moves. The fold names it in CHANGELOG.md (Changed and Hosts; "types, words and compact bytes do not move" in place of "verdicts"), pins it in tests/transitions.rs (run on the box: Err(InvalidProof("merkle path error: Invalid hex string")); a sibling that is hex of another length still reduces, to another root), and says cargo deny runs outside ci.yml, which does not run it. The count stays 64 passed; no source, manifest or lock line moves in the fold.

After the re-read (fix, 0.82), the fold f168cad (CHANGELOG.md alone). The re-read found the sentence "a sibling that is hex of another length is still paired" false for an odd length: 0.9.0-beta.7 dropped that side, and 0.9.0-beta.8 pairs it with one leading 0, so Proof::root answers another root for it than it did (an even length is paired as before). That is the second answer that moves with the pin; either root is one the well-formed proof of the txid does not have, and no word moves. The changelog now says "Two answers", carries the re-read's own sentence, names Headers beside Unavailable, and cites the sixteen advisories bsv-rs's release note lists in place of a version range. The paragraph above this one says the other-length sibling "still reduces, to another root": true, and incomplete in the same way for the odd length. The test pins both lengths and did not move.

🤖 Generated with Claude Code

Calgooon and others added 3 commits November 17, 2026 08:22
No code change: the crate names bsv-rs's MerklePath and MerklePathLeaf
alone, and none of bsv-rs 0.9.0-beta.8's ten breaking changes reaches a
call here (its MerklePath writers returning Result meet no call). The lock
moves the bsv-rs row and this crate's own. bsv-rs 0.9.0-beta.8 carries the
stable line's advisory fixes; the toolbox and the CLI reach it only with a
tracker that pins it.

Read at the pin as ci.yml runs it: cargo test --locked (64 passed), the
documentation, Clippy -D warnings, the vector steps, the wasm32 build,
cargo deny.

Co-Authored-By: Claude <noreply@anthropic.com>
… Proof::new named and pinned

The cold read of PR 43 found one answer of this crate that moves with the
pin and that the changelog left unnamed. bsv-rs #817 changes the merkle
walk behind MerklePath::new_unchecked and compute_root, which Proof::new
and Proof::root call: a path whose walk pairs a hash holding a character
that is not hex is refused at the constructor (CheckError::InvalidProof),
where through 0.9.0-beta.7 it was built and failed at the header check with
RootMismatch and Reask::ProofFailed. No word moves.

- CHANGELOG.md: the moved answer under Changed and under Hosts; "types,
  words and compact bytes do not move" in place of "verdicts"; cargo deny
  named as run outside ci.yml, which does not run it.
- tests/transitions.rs: malformed_sdk_paths_are_rejected_before_root_reduction
  pins the refusal (run on the box: Err(InvalidProof("merkle path error:
  Invalid hex string"))) and that a sibling that is hex of another length
  still reduces, to a root the well-formed proof does not have. The count
  stays 64 passed.

No source line of the crate, no manifest line and no lock row moves in
this commit.

Co-Authored-By: Claude <noreply@anthropic.com>
… is the second moved answer

The re-read of PR 43 found one false sentence in the changelog: "a sibling
that is hex of another length is still paired". Of even length it is paired
as at 0.9.0-beta.7. Of odd length it was DROPPED at 0.9.0-beta.7 and is
paired with one leading 0 at 0.9.0-beta.8 (bsv-rs #817), so Proof::root
answers another root for it than it did. Either root is one the well-formed
proof of the txid does not have, and no word moves.

CHANGELOG.md alone: "Two answers" for "One answer"; the sentence replaced
by the re-read's own; `Headers` named beside `Unavailable` among what the
header check answered; the Security paragraph cites the sixteen advisories
bsv-rs's release note lists in place of a version range. The test already
pins both lengths and does not move.

Co-Authored-By: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant