What's Changed

Mainnet E2E Validation

Mainnet sync with txindex=1 works. bitcoind syncs fully and serves tx queries through the FUSE layer. Also ran Mempool.space against it — the explorer indexes blocks and txs from the FUSE-served data without issue.

Release-Based Finalization

Moved block file finalization from the Write path to Release. A file is finalized only when the last write handle closes and a successor exists on disk. A per-file refcount tracks concurrent writers so state transitions don't fire early. Kills the races we hit under sustained IBD writes.

IBD Stability & Client Improvements

Several IBD fixes. Manifest loading hits the remote first, falling back to local cache on failure. Empty blk files are preserved — deleting them broke index alignment with bitcoind. Blockmap reads return ENODATA for out-of-range requests. The remote HTTP client got structured logging and configurable transport; a response-size cap guards against pathological upstream responses.

Note: Adaptive fetching may not correctly detect the IBD state on mainnet yet. If you are syncing mainnet from scratch, set general.remote-fetch-mode = file explicitly to enable file-level fetching and get the full speed benefit.

Full Changelog: v0.1.2...v0.1.5

Assets 7

18 Apr 16:50

github-actions

v0.1.4

795bded

v0.1.4 Pre-release

Pre-release

Full Changelog: v0.1.3...v0.1.4

Assets 7

18 Apr 16:48

github-actions

v0.1.3

5bf2742

v0.1.3 Pre-release

Pre-release

Full Changelog: v0.1.2...v0.1.3

Assets 7

09 Apr 00:37

github-actions

v0.1.2

6769797

v0.1.2 Pre-release

Pre-release

What's Changed

Testnet & Regtest Support

Supports testnet and regtest. Set general.chain in your config.

IBD Bug Fixes & E2E Validation

Fixed several bugs and panics encountered during Initial Block Download (IBD). An end-to-end test against testnet4 — including IBD, Electrs, and Mempool integration for transaction data queries — passes successfully.

Adaptive Fetching

Adaptive fetching to significantly reduce IBD time. Fetching blocks via individual HTTP requests would make a full IBD take months. Adaptive fetching dynamically switches between block requests and blk file downloads depending on the sync state, automatically switching to file-level fetching during IBD for a major speed improvement.

Note: Adaptive fetching may not correctly detect the IBD state on mainnet yet. If you are syncing mainnet from scratch, set general.remote-fetch-mode = file explicitly to enable file-level fetching and get the full speed benefit.

Full Changelog: v0.1.1...v0.1.2

Assets 7