Uh oh!
There was an error while loading. Please reload this page.
style(docs): deno fmt CLAUDE.md and README.md - #6
Conversation
`rs-static` fails on the `denofmt` pre-commit hook, which now checks markdown: both files carry hand-wrapped prose that `deno fmt` reflows. One hunk is not cosmetic. CLAUDE.md's Rust/npm paragraph wrapped onto a line beginning "+ `scripts/` + `test_js/` ...", and a `+ ` at the start of a line is a CommonMark bullet, which interrupts the paragraph. GitHub already renders a stray list item there, and `deno fmt` was going to bless that reading by rewriting the tail of the paragraph as list continuation. The sentence is reworded to "`crates/float-wasm`, `scripts/` and `test_js/` build ..." so it stays one paragraph. Everything else is reflow only; no claim changes.
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Warning Review limit reached
Next review available in:41 minutes Limit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. How can I continue?Wait for the limit to reset, then comment An organization admin can change what happens after included review limits in Billing. How do review limits work?CodeRabbit enforces per-developer PR review limits within each organization. For paid Pro and Pro+ reviews, CodeRabbit uses a developer's included PR review attempts over the past 7 days to set the current hourly allowance. At typical activity levels, the full plan allowance applies. Higher sustained activity can lower the allowance until earlier attempts leave the 7-day window. Please refer docs for additional details. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
The Releases section claimed "The addresses already exist on-chain (deterministic Zoltu), so the first release attests existing pins rather than deploying fresh". That is false for the DecimalFloat half: eth_getCode at ZOLTU_DEPLOYED_DECIMAL_FLOAT_ADDRESS (0x799632d282178e770C7465cad54aDA1021A913D6) returns 0x on arbitrum, base and flare. Only the log tables at 0xc51a...f5F5 are actually deployed. It is also the same fact the five red testProdDeployment* tests report. Beyond being wrong today, it was a statement about deployment status living in prose that nothing re-checks, so it could only ever go stale. Replaced with the invariant that does hold — a Zoltu pin is derived from bytecode, so it is fixed before any broadcast — plus a pointer to LibDecimalFloatDeployProdTest, which is the thing that actually asserts a chain carries the code. deno fmt --check clean (deno 2.9.4, CI's version). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The failure
rs-staticfails, and after #3 removes thevm.skipthe only thing leftfailing in it is the
denofmtpre-commit hook:Neither file has changed since the initial commit. What changed is the hook:
on the 2026-08-20 runs
denofmtreported(no files to check) Skipped, and itnow checks markdown. Both files carry hand-wrapped prose that
deno fmtreflows to its own 80-column wrap, so the hook rewrites them and the job goes
red.
The fix
deno fmt CLAUDE.md README.md, plus one wording change that is notcosmetic.
CLAUDE.md's Rust/npm paragraph was hand-wrapped so that a line began:
A
+at the start of a line is a CommonMark bullet, and a bullet interruptsan open paragraph. GitHub already renders a stray list item there today. Left
alone,
deno fmtaccepts that reading and rewrites the whole tail of theparagraph as list continuation — a six-line indented bullet where prose was
intended. The sentence is instead reworded to
which is the meaning the
+was standing in for, has no line-initial+, andformats as one paragraph.
Every other hunk in both files is pure reflow. No claim, path, name or number
changes.
Verification
Ran
deno fmtlocally with deno 2.9.4 and diffed against what CI printed.The README.md result is byte-identical to CI's — same git blob hash on both
sides (
index 3e40130..abddac0), which is what establishes that the localformatter matches the one in the job. The CLAUDE.md result differs from CI's
only in the reworded sentence and its consequence, exactly as intended.
deno fmt --check CLAUDE.md README.mdthen exits 0, so the hook is idempotenton the committed text rather than merely quiet once.
QA
discriminating check is the CI hook itself:
deno fmt --checkon these twofiles exits 1 on
main(it is what turnsrs-staticred) and exits0 on this branch. Run both ways locally.
available was applied and is the reason the reword exists: leaving the
line-initial
+in place and lettingdeno fmtdecide produces asix-line bullet in place of the paragraph — the formatter's own output is the
evidence that the source was ambiguous, and the reword is what kills it.
Re-running
deno fmton the reworded text is a fixed point.deno fmtitself, cross-checked against CI's printed diff. TheREADME blob hash matching CI's on both sides of the diff is an oracle
independent of my local toolchain choice. For the bullet question the oracle
is CommonMark:
+opens a list item and a list item interrupts a paragraph.denofmtis now the only red hook inrs-staticand this is the whole ofit.
rainix-sol / testis red for an unrelated reason —DecimalFloatis notdeployed on chain — and is untouched here.
Scope and stacking
Deliberately separate from #3 and #5. #3 is the
vm.skipremoval, #5 is the[rpc_endpoints]gap; this is a docs reflow that happens to share a job with#3. Keeping them apart means each PR's red checks stay attributable.
Based on #3's branch, not
main.rs-staticruns the orgno-ignored-testsgate before pre-commit, so with #3 unmerged the job exitsat the
vm.skipin 31 seconds and never reaches thedenofmthook — this PR'sfix would have been unverifiable from CI, and the red check would have read as
this PR's fault. Stacked on #3 the job runs through to the hook, so the green
here is real. GitHub will retarget this to
mainwhen #3 merges. Merge #3first.
rainix-sol / teststays red on this branch and on every branch:DecimalFloathas no code at its Zoltu address on any chain, which is a deploy that has not
happened, not something a PR can fix.