概要
SIPO DRepレビュー支援用の日本語作業訳です。正式な内容はHydra Voting上の英語原文を参照してください。
MLabsのCovenant 2提案は、Project Catalyst Fund13で開発されたCovenant compiler frameworkの次段階として、developer experience改善とcross-target / cross-language対応を行うものです。Covenantは高水準smart contract languagesとCardanoが実行するUPLCの間に位置し、analysis、optimization、code generationを担います。今回の提案では、error messages、interfaces、examples、tests、documentation、Van Rossum hard-fork builtin supportを改善し、Covenant coreをtarget-parameterizedにして、UPLCに加えてSolana sBPF parameterizationとc2sbpf code generatorを実装します。SIPOとしては、Cardano language ecosystemのcompiler pipeline重複を減らす公共インフラ性を評価します。一方で、Covenant 1のcloseout状況、Konma/HaskLedger以外の採用見込み、Aiken/Plutarch/Plinth等との連携、Solana sBPF対応をCardano Treasuryが負担する公共利益、本文中の約2.17M ADA表記と最新予算1,484,333 ADAの不整合、MLabs全体の同時提案負荷、NoWitnessLabs audit範囲、実際のTreasury返還条件を確認する必要があります。
英語原文を表示
Covenant is a Cardano-built compiler framework that sits between high-level smart contract languages (such as Aiken, Plutarch, and Plinth) and the scripts the Cardano blockchain actually executes. It does the heavy lifting — analysis, optimization, and code generation — so the maintainers of those languages can focus on what their users see, instead of the entire compilation pipeline. Covenant 1 was funded under Project Catalyst Fund 13 and is now feature-complete and in stabilization. Additionally, Covenant has been adopted by Konma, specifically for code generation of their DSL. As part of this work, Konma requested multiple improvements and features which could not be covered by the Fund13 budget. Notably, HaskLedger was intended to target Solana in addition to Cardano, but this wasn’t possible within that scope. This proposal is the natural next step. It does two things: **1. Improves the developer experience.** Better error messages, simpler interfaces, more examples, more tests, and clearer documentation. These are the unglamorous improvements that decide whether a tool is actually adopted. **2. Extends Covenant to support languages beyond Cardano** —. As HaskLedger is Covenant’s main adopter, the requirement to support compilation to Solana is essential for their use case. Furthermore, once Covenant can target multiple chains, it becomes a "universal glue" between high-level smart contract languages on one side and chains on the other. Thus, we can simultaneously support a current, concrete use case while also laying the groundwork for wider support of other tool chains and on-chain languages. The work will be fully open-source under Apache 2, builds on a delivered Catalyst project, and is delivered by the same MLabs team that built Covenant 1.