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 @coderabbitai help to get the list of available commands and usage tips.}

@coderabbitai resolving your two comments:

1. Cargo.toml:68 — zccache-artifact = "1.8" (Critical: "appears not to exist on crates.io")

This is a false positive. zccache-artifact 1.8.2 was published to crates.io today (2026-05-22):

$ curl -s https://crates.io/api/v1/crates/zccache-artifact | jq '.crate | {max_version, max_stable_version, updated_at}'
{
  "max_version": "1.8.2",
  "max_stable_version": "1.8.2",
  "updated_at": "2026-05-22T14:28:54.540420Z"
}

Cargo.lock resolves it cleanly, and the workspace check on all three OSes (Linux/macOS/Windows) passes on this PR. Your web search likely hit stale cache. No change needed.

2. .github/workflows/bench-205.yml:38-43 — replace target-cache-profile: thin-v1 with target-cache-mode: hot (Major)

This comment is stale — it was filed against commit 0d29481. I reverted that change two commits ago (commit 1bd3b2b) because the modernized target-cache-profile input triggered a different upstream regression (the new default-on prebuild-deps: cargo-chef introduced in zackees/setup-soldr#140). The current PR head uses the deprecated-but-functional target-cache: true everywhere, plus an explicit prebuild-deps: none to work around the cargo-chef issue. See commit 2f7f1a3.

Both items addressed — proceeding to merge.

#!/bin/bash
# List review comments made by coderabbitai on PR `#253`
gh api repos/FastLED/fbuild/pulls/253/comments --paginate \
  | jq '[.[] | select(.user.login == "coderabbitai") | {id, path, body: .body[:120]}]'

#!/bin/bash
# Check review threads
gh api repos/FastLED/fbuild/pulls/253/reviews --paginate \
  | jq '[.[] | select(.user.login == "coderabbitai") | {id, state, body: .body[:120]}]'

echo "---ISSUE COMMENTS---"
gh api repos/FastLED/fbuild/issues/253/comments --paginate \
  | jq '[.[] | select(.user.login == "coderabbitai") | {id, body: .body[:200]}]'

@zackees: Thanks for the detailed clarification on both points — my previous comments appear to have already been resolved (likely auto-dismissed when the commits were pushed). To summarize:

1. zccache-artifact = "1.8" (false positive) — Confirmed, zccache-artifact 1.8.2 is live on crates.io as of today. My earlier flag was based on stale lookup data. No action needed.

2. bench-205.yml stale comment — Understood. The target-cache-profile: thin-v1 approach was reverted in commit 1bd3b2b due to the upstream prebuild-deps: cargo-chef regression introduced in zackees/setup-soldr#140, and the current head correctly uses target-cache: true + prebuild-deps: none as a workaround. The comment was filed against a superseded commit and should not have persisted.

🐇 ₍ᐢ..ᐢ₎ ♡

Learnt from: zackees
URL: https://github.com/FastLED/fbuild/pull/253

Timestamp: 2026-05-22T21:14:12.156Z
Learning: In the FastLED/fbuild repository, when checking crates.io availability for Rust workspace dependencies (e.g., in Cargo.toml), do not flag a crate version as missing based on cached/stale web search results. Always verify against the live crates.io API before raising a critical comment, especially for crates that may have been published very recently.

Learnt from: zackees
URL: https://github.com/FastLED/fbuild/pull/253

Timestamp: 2026-05-22T21:14:12.156Z
Learning: In the FastLED/fbuild repository, the `zackees/setup-soldr` action's `target-cache-profile: thin-v1` input can trigger a regression when the default-on `prebuild-deps: cargo-chef` is active (introduced in zackees/setup-soldr#140). The workaround is to use the deprecated-but-functional `target-cache: true` together with `prebuild-deps: none`. Do not suggest switching to `target-cache-profile` or `target-cache-mode` inputs without checking upstream setup-soldr changelog for regressions.