Eliminate CI Disk Pressure by Using Ripr Binary Directly

Описание и применение

CI теперь использует готовый бинарник Ripr, устраняя необходимость компиляции xtask и уменьшая нагрузку на диск.

Задача

CI runs compile the repo’s xtask orchestration layer on every run, creating a 5‑8 GB build tree that causes disk‑full failures on self‑hosted runners.

Результат

CI runs now use the prebuilt Ripr binary, removing the 5‑8 GB build tree, eliminating disk‑full failures, and reducing latency.

Рабочий процесс

• Audit each cargo xtask ripr-* step in .github/workflows/ripr.yml and map it to an equivalent ripr <subcommand> call.
• Replace cargo xtask ripr-* steps with direct ripr <subcommand> commands in the workflow.
• For orchestration logic not covered by Ripr, either add a new Ripr subcommand or build a slim standalone gate binary once and cache it.
• Configure the workflow to use a pinned Ripr release asset or the setup-ripr action instead of cargo install ripr.
• Verify that the job produces the same evidence artifacts (target/ripr/** and target/receipts/**).
• Confirm that the CI no longer compiles the workspace and that disk pressure is eliminated.

Доказательства

The recurring CX43 self-hosted-runner disk pressure is a downstream integration-shape problem in this repo, not a ripr deficiency.

Источник

The ripr binary is fully capable and already does the static analysis — our xtask/src/tasks/ripr_evidence.rs shells out to it.

Источник