Skip to content

Latest commit

 

History

History
58 lines (43 loc) · 3.28 KB

1.1.1.md

File metadata and controls

58 lines (43 loc) · 3.28 KB

Lesscode 1.1.1

Prepared 2026-11-11. The tag v1.1.1 is pushed by the owner; the release workflow then builds, signs and publishes the archives (release signing).

Proposed tag message:

Lesscode 1.1.1: density measures syntax instead of lines; merge-annotations; corrected example

What is new

  • deconvolute density measures syntax. It counts tokens and named syntax nodes outside comments, with the grammar of the language, for Rust, Python, TypeScript and Java. Layout no longer changes the verdict. Receipt deconvolution-density-v2, with the thresholds in it: at least 20% fewer tokens, at least 400 tokens removed and at least 10% fewer nodes for pass. Lines are reported for information. A file that does not parse is refused with exit code 3.
  • deconvolute merge-annotations. Validates several annotation lists and writes the sites that at least N of them name (default 2), for use with run --annotations. A selection made by one model call varies from call to call; the sites several calls agree on vary less (R85).

Why the density change matters

Releases up to 1.1.0 counted lines and, under the name of statements, keywords and semicolons per line. A candidate written more densely passed as a reduction. The pair this README used as its example reported 40% fewer lines and pass; by syntax it has 7% fewer tokens and is marginal. All four specimens of this project were measured again: 6% to 19% fewer tokens, against 24% to 44% fewer lines reported before (R86, R87). If you kept a density receipt from an earlier release, measure the pair again.

What changed for existing users

  • density receipts have new fields (baseline_size, candidate_size, token_reduction_pct, tokens_removed, node_reduction_pct, thresholds); baseline_lines, candidate_lines, baseline_stmts, candidate_stmts and stmt_reduction_pct are gone. A pair that passed before may now be marginal.
  • density --lang python is a real measurement; before it counted lines only.
  • Library users: the line and keyword counters are removed (adapters::{rust,java,typescript}::metrics and the module density), with the tests that pinned the earlier compression figures.
  • Everything else is as in 1.1.0.

For the owner, before tagging

  1. Tag a commit for which CI is green on all five platforms.
  2. git tag -a v1.1.1 -m "<the tag message above>" and git push origin v1.1.1.
  3. Approve the release environment when the workflow asks.
  4. After publication: check signatures, checksums and the attestation, and that Zenodo archived it.

Verified after publication (2026-11-11)

Checked from the published assets, downloaded with gh release download:

  • All 12 entries of SHA256SUMS match the downloaded files.
  • SHA256SUMS.sig verifies with the project's public key (the key listed in scripts/release-public-keys.txt), and each of the five archives carries a valid signature.
  • gh attestation verify succeeds for the Windows archive.
  • The Windows binary reports the release's version.
  • Zenodo archived the release: version 1.1.1, DOI 10.5281/zenodo.24135840.- install.ps1 installs 1.1.1 into a directory of choice, and the installed binary has merge-annotations and the syntax-based density (the repository's Rust pair is reported as marginal).