Skip to content

landing: name v0.2.0, and state the CI claim the way ci.yml does - #6

Merged
tannevaled merged 1 commit into
mainfrom
docs/v0.2.0-refresh
Sep 23, 2026
Merged

tannevaled merged 1 commit into
mainfrom
docs/v0.2.0-refresh

Conversation

@tannevaled

Copy link
Copy Markdown
Contributor

The parser shipped v0.2.0 on 2026-09-21. This page named no parser version and no release date anywhere — its prose was last touched on 2026-08-06, and the version-bearing copy had never carried one at all.

Changes

A version pill, and the lead now names v0.2.0 with the thing worth saying about it: every one of the 68 files in the pinned ruby/spec language/ corpus parses under it, thirteen of which had not parsed at all before. Those files had been contributing nothing to either column of any conformance measurement — not failing, invisible.

The CI claim. The page said:

100% coverage enforced in CI across the six 64-bit targets.

That conflates two different gates. Per the parser's own ci.yml, which I read:

  • the coverage gate runs on the three OS lanes — ubuntu, macOS, windows (awk … exit !(t+0 >= 100.0));
  • the six 64-bit targets run go test ./... without it — amd64 and arm64 natively, riscv64/loong64/ppc64le/s390x under qemu-user-static;
  • and there are js/wasm and wasip1/wasm lanes the page did not mention.

Both facts are now stated, each attributed to the lane that actually does it. The same sentence in hugo.toml's repo blurb is corrected the same way.

A note on the link

I linked /tree/v0.2.0, not /releases/tag/v0.2.0. The repo publishes no GitHub Releases at all — gh release list is empty, gh api …/releases returns [] — so v0.2.0 exists only as a git tag and the releases URL renders an empty page. Publishing the releases would be worth doing, but that is a repo-process change rather than a doc correction, so it is not in this PR.

Left alone deliberately

  • Every other claim on the page — Ruby 4.0, MRI 4.0.5 differential-tested, CGO=0, Go 1.26.4+, pattern matching, heredocs, stateful lexer, recursive-descent + Pratt, status = "ready". I checked each against the parser repo; they still hold.
  • The upstream Performance numbers (1.5×/2.1× vs RubyVM::AbstractSyntaxTree.parse, ~6× vs Ripper.sexp, 24–30× on the lexer) are absent from this site. That is a gap, not a wrong claim, and I could not reproduce those benchmarks in this pass — so I did not import figures I have not measured.
  • HUGO_VERSION: 0.135.0 in the deploy workflow is the oldest pin in the repo. Dependency hygiene, not doc drift.

No merge. No screenshots.

🤖 Generated with Claude Code

The page named no parser version and no release date anywhere. The parser
shipped v0.2.0 on 2026-09-21 with fourteen grammar fixes, three months after
this page's prose was last touched.

* a version pill, and the lead now names v0.2.0 with what it closed -- every
  one of the 68 files in the pinned ruby/spec language corpus parses, thirteen
  of which had not parsed at all before
* "100% coverage enforced in CI across the six 64-bit targets" conflated two
  different gates. Per the parser's ci.yml the coverage gate is on the three OS
  lanes; the six 64-bit targets (amd64/arm64 native, riscv64/loong64/ppc64le/
  s390x under QEMU) and the js/wasm and wasip1/wasm lanes run the suite without
  it. Both are now stated, each attributed to the lane that does it.
* the same claim in hugo.toml's repo blurb

Linked to the tag tree rather than /releases/tag/v0.2.0: the repo publishes no
GitHub Releases, so that page renders empty.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@tannevaled
tannevaled merged commit 1ddb7d2 into main Sep 23, 2026
2 checks passed
@tannevaled
tannevaled deleted the docs/v0.2.0-refresh branch September 23, 2026 17:51
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