feat(e2e): Token Lane Refactor

[rodrigombsoares]

Summary

Second stacked PR in the token-lane refactor (pr2-token-lane on top of pr1-load-metrics). Centralizes token pool config in TokenLane, adds env-scoped TOML, supports Canton fixed-point amounts and optional transfer instruments, and updates SetupSend / MintTokens to use *big.Rat.
Devenv-only paths are exercised; [prod-testnet.*] TOML blocks exist but stay inert until PR4.

Reviewer story: Token tests resolve lanes from TOML instead of ad-hoc setup; Canton token sends support multiple instruments via fixed-point amounts. SetupSend + SetSequentialSends correctly pick initial holdings for multi-send batches so rotation can advance send 2+.

Features

Constants & env

• ccip/devenv/constants.go — Shared constants moved out of tests/constants.go:
  • CantonToEVMFeeAmount = 50 (message-only)
  • CantonToEVMTokenTransferFeeAmount = 130 (token path)
  • CantonFixedPointScale (10¹⁰) and CantonFixedPointToEVMScale (10⁸), replacing EVMDecimalsScale
• ccip/devenv/tests/env.go — Minimal CCIPEnv API: EnvDevenv, EnvProdTestnet, ParseCCIPEnv, IsRemote()

Token lane

• token_transfer_config.toml — Flat [evm_to_canton] / [canton_to_evm] → nested [devenv.*] and [prod-testnet.*] skeleton
• token_lane.go — ResolveTokenLane(t, env, …) with env-scoped TOML, TransferInstrument / TransferInstrumentID, Canton instrument resolution from datastore labels
• token_lane_test.go — Unit tests for env selection, TOML parsing, scale conversion, instrument helpers

Core impl

• impl.go — Token hunks only:
  • SetupSend(ctx, feePerMessage, transferPerMessage *big.Rat, optional instrument) — per-send amounts; callers pass 1× values
  • SetSequentialSends + batch-aware initial holding selection — call SetSequentialSends(N) before SetupSend when running N sends in a batch. Setup multiplies fee/transfer holding mins by sendsLeft (perMessage × N) so the first UTXO can split across sends; stored feePerMessage and rotation mins stay per-send (1×). Fixes Canton→EVM 2-send e2e failing on send 2 when setup picked a dust UTXO (exact 1× transfer) with no change Created for rotation
  • selectHolding logging at setup/rotate (Source, Candidates, MinRequired) for holding-debug
  • MintTokens(ctx, *big.Rat) with nil/zero guard
  • Fixed-point SendMessage via CantonFixedPointScale
  • FTF BlockDepth: 1 on token pool configs
  • setNextHoldings nil/exhaustion semantics with *big.Rat; skip rotation after last planned send in batch
  • LINKInstrument constant
• manual_execution.go — Idempotent DeployCCIPSender (CANTON_SENDER_INSTANCE_ID, default e2e-ccipsender)
• deployment/sequences/token_pools.go — Copy Labels from deploy input onto token ref in datastore (required for ResolveTokenLane instrument resolution)

Tests

• E2E and load tests use ResolveTokenLane(t, EnvDevenv, …) and cantondevenv.* constants
• Token e2e/load: SetSequentialSends(N) before SetupSend; per-send transferPerMessage only (no batch totals in callers)
• Inline devenv bootstrap kept (no BootstrapE2E yet — PR3)
• setupCantonTokenLoadHoldings updated for token transfer fee, fixed-point minting, and optional TransferInstrument

Removed

• ccip/devenv/tests/constants.go — Superseded by ccip/devenv/constants.go

Testing

On CI:

1. Check e2e tests already run by CI here
2. Use github action to run load tests pointing to this branch https://github.com/smartcontractkit/chainlink-canton/actions/workflows/ccip-load-tests.yml

example run: https://github.com/smartcontractkit/chainlink-canton/actions/runs/28914300723/job/85777914380

With local devenv running:

1. make start-devenv
2. Token e2e — Run token subtests in both directions (canton2evm, evm2canton); Canton→EVM token subtest sends 2 messages sequentially
3. Token load — make run-canton2evm-token-load and make run-evm2canton-token-load
4. Compile / unit — e.g. go test ./ccip/devenv/... and token_lane unit tests

[github-actions]

👋 rodrigombsoares, thanks for creating this pull request!

To help reviewers, please consider creating future PRs as drafts first. This allows you to self-review and make any final changes before notifying the team.

Once you're ready, you can mark it as "Ready for review" to request feedback. Thanks!