[shumkov]

Draft for discussion. Not meant to be merged. It holds the working plans for the end-to-end test system and the legacy JS retirement, so they can be reviewed with line comments. When the work starts, the plans turn into issues and implementation PRs, and this PR is closed.

What is in here

• docs/E2E_TEST_ARCHITECTURE_SPEC.md (revision 3) and its plain-language walkthrough: one e2e test system for Platform and every SDK.
• docs/LEGACY_JS_RETIREMENT_SPEC.md and its walkthrough: retiring the legacy JS stack (dash SDK, wallet-lib, js-dapi-client, wasm-dpp, dashcore-lib) and the old platform-test-suite. The e2e part of this plan is superseded by the e2e spec; the rest still stands as a track inside it.

The plan in short

• Platform behaviour is tested once, in Rust, against a real network. Every SDK we ship is a layer over the Rust SDK (JS through wasm, Swift and Kotlin through FFI), so each SDK only proves its own layer: offline by replaying recorded real responses, live with a small smoke test.
• Four levels: unit and in-process tests; protocol e2e in a new Rust crate (rs-platform-e2e) on the local dashmate network; recorded read fixtures replayed by every SDK through Rust; a small live smoke per SDK.
• One configuration for every SDK (DASH_E2E_* keys written by the existing setup) and one funding model.
• Built on test(platform-wallet): e2e framework + full test suite — triage pins, Found-*/PA-* guards, fail-closed persist, Stage-2 merge #3549's wallet-free core (about 2k lines, credited to its author), not from scratch: it already solves the shared runtime, stack size, contract registration, commit-wait and lagging-node problems. test(platform-wallet): e2e framework + full test suite — triage pins, Found-*/PA-* guards, fail-closed persist, Stage-2 merge #3549's wallet layer lands later on the same core.
• The legacy platform-test-suite is deleted after parity: 73 cases ported, 5 already covered elsewhere, 17 dropped with a reason each (§8).
• Phases: E1 network smoke on consensus PRs; E2 crate, framework, smoke set and CI; E3–E6 the families to parity, then delete the legacy suite; fixtures in parallel; SDK smoke alignment; E7–E9 tokens/groups/voting, remote runs, the wallet layer. Rough effort and CI cost are in §9 and §11.

Decisions so far

• Network smoke on every consensus PR (Q1): deferred to the last step.
• Browser reduction (Q2), local-only protocol e2e at deletion time (Q3), JS replay later and no Kotlin replay (Q4): accepted.
• test(platform-wallet): e2e framework + full test suite — triage pins, Found-*/PA-* guards, fail-closed persist, Stage-2 merge #3549's core as the base (Q5): the plan assumes it; it needs test(platform-wallet): e2e framework + full test suite — triage pins, Found-*/PA-* guards, fail-closed persist, Stage-2 merge #3549's author to agree before E2.
• Keeping the legacy suite alive until parity: dashcore-lib 0.26.1 (Core V24 support) is released and used by 5.0.2+ and v5.1-dev.

Open

• Agreement with test(platform-wallet): e2e framework + full test suite — triage pins, Found-*/PA-* guards, fail-closed persist, Stage-2 merge #3549's author on the copy, ownership and the rebase (§13 Q5).
• When to start E1/E2 relative to the rest of 5.1.

🤖 Generated with Claude Code

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[thepastaclaw]

🕓 Review not started yet because this PR is a draft.

• Request normal review — click when the PR is ready for review.
• Request priority review — click to move this review to the front of the queue.

Commit d571beb. Normal review starts when eligible; priority review starts as soon as a slot is available.