Skip to content

CHANGELOG.md

Latest commit

 

History

History
139 lines (114 loc) · 6.77 KB

CHANGELOG.md

File metadata and controls

139 lines (114 loc) · 6.77 KB
Raw

Changelog

All notable changes to defi-cli are documented here.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

Work landed on main after v1.0.13 (commit e3a3510). When this rolls into the next release, move these entries under a versioned heading.

Added

  • CLI: --asset alias on token and bridge commands (R1). Both flags resolve to the same handler argument; --asset makes the surface consistent with price / lending / yield. --token is kept as the long-standing primary so every existing script continues to work. Missing both flags now emits the explicit guard error "--token (or --asset) is required". (#R1)
  • CLI: --aggregator alias on swap --provider (R3). The flag value matches the prose in swap --help and project docs (KyberSwap / OpenOcean / LiquidSwap / LI.FI / Relay are "aggregators"). When both --provider and --aggregator are passed, --aggregator wins (last-flag-wins, the common commander convention). (#R3)
  • Test coverage across every package — see Tests below.
  • CLI command-based QA report (docs/qa-reports/2026-05-16-cli-command-qa.md) covering 40 commands across 5 chains (status / price / lending / yield / swap / bridge / token / lp / portfolio / wallet) — read-only RPC + external aggregator APIs + dry-run builders verified end-to-end.

Changed

  • CLI: swap / bridge help text now spells out dry-run semantics (R2). Both commands intentionally stay single-command (no quote / simulate subcommand) because dry-run is the default and already returns the quote + calldata; the .description() text now explains this: "Dry-run by default returns the quote + calldata; pass --broadcast to send the tx." No behaviour change — --help prose only.
  • CLI: yield --min-spread help text clarified that the flag is a decimal (e.g. 0.05 = 5%), not a percent. The numeric default ("1.0" ≈ 100%, which effectively disables plan_only) is left as-is to avoid breaking scripts; a follow-up PR is needed for a sane default change. (F2)
  • main.ts: factored the subcommand-detection logic into an exported pure function decideEntryPoint(rawArgs) plus a KNOWN_SUBCOMMANDS set so the routing branches are unit-testable. The bottom of main.ts now runs main() only when the file is the bin entrypoint (see Fixed for the symlink resolution).
  • mcp-server.ts: exported ok / err / getRegistry / resolveToken / makeExecutor / server so the JSON envelope helpers and registry factories are testable in isolation. The MCP stdio transport is now opened only when the file is invoked as the bin entrypoint, so importing mcp-server.ts from a test (or another module) no longer blocks on stdin.

Fixed

  • CI: bin entrypoint guards now resolve symlinks before comparing. Regression introduced when main.ts and mcp-server.ts got import guards in the previous two commits. The original guard compared fileURLToPath(import.meta.url) against process.argv[1] — which works for node dist/main.js but breaks for npx defi (and any npm bin invocation) because process.argv[1] is the symlink (node_modules/.bin/defi) and import.meta.url is the symlink target. The smoke test caught it with "Banner missing dynamic chain/protocol count". Both guards now run their paths through realpathSync before comparing. (commit 4ad8f7f)
  • price.ts: removed unreachable CDP oracle fan-out. The if (isWhype) branch asked the registry for CDP-category protocols, but no protocol in the current registry registers under the cdp category (felix is registered as lending via felix_morpho). The branch entered, iterated an empty array, and continued — dead code. Removed the import + the guard + the dead Promise.all loop. (F1)

Tests

Total monorepo tests went from 201 (session baseline) to 431 (+230, +114.4%):

Package Tests (baseline → current)
defi-core 32 → 49 (+17)
defi-protocols 63 → 147 (+84)
defi-cli 106 → 235 (+129)

Notable per-file coverage jumps:

  • commands/{price,setup,ows,portfolio,yield,price}.ts, portfolio-tracker.ts, cli.ts, landing.ts, main.ts, signer/index.ts, mcp-server.ts (helpers) — 0% → ~100% line coverage via mock-based unit tests with no real RPC.
  • defi-core/erc20.ts, multicall.ts, provider.ts — 0% → ~100%; covers ERC20 builders, multicall encode/decode + RPC mock, provider cache identity and SSOT 7.4 chain-anchor keying.
  • defi-protocols/factory.ts — branches 14.28% → 98.48%. Walks every dispatch branch across the 14 factory functions plus the inferRewardStrategy 5-branch fallback chain and the createKittenSwapFarming missing-contract guards.
  • defi-protocols/dex/algebra_v3.ts — branches 18.18% → 58.82%; V2/Integral dispatch + quote() decode paths via stubbed viem.
  • defi-protocols/dex/uniswap_v3.ts — branches 47.36% → 75%; Ramses + Slipstream CL-style mint dispatch.
  • defi-protocols/dex/thena_cl.ts — branches 42.85% → ~100%; constructor guard + remove_liquidity guards + reverse sort.
  • defi-protocols/lending/compound_v2.ts — branches 45.45% → 69.56%; constructor defaultVtoken fallback chain.

Refactor

  • commands/yield.ts: --min-spread flag carries an inline comment explaining the decimal semantics + default-as-disable behaviour. No behaviour change. (F2)
  • commands/price.ts: removed createOracleFromCdp import + WHYPE_ADDRESS constant after the CDP fan-out removal. (F1)

Internal

  • Per-handler test files now follow a consistent mock pattern: vi.mock("@hypurrquant/defi-protocols") for adapter factories, vi.mock("viem", importOriginal) for selective createPublicClient stubs, vi.spyOn(process, "exit") for error-path tests. Documented inline in each test header.

1.0.13 — 2026-05-08

Released by maintainer. See PR #32 for the full content.

  • Bridge feature complete (CCTP V2 --auto-receive, LI.FI, Relay, deBridge).
  • Production-grade verification across 4 chains × 35+ protocols.
  • Adapter fixes from the ULTRAQA sandbox cycle.
  • Per-package coverage baseline added via @vitest/coverage-v8.