Skip to content

dintoken-tests.md

Latest commit

 

History

History
130 lines (100 loc) · 6.25 KB

dintoken-tests.md

File metadata and controls

130 lines (100 loc) · 6.25 KB
Raw

DIN Token CLI Unit Tests (tests/test_dintoken.py)

This document describes the unit test suite for the DIN token commands in dincli/cli/dintoken.py: buy_dintokens (ETH → DIN via DinCoordinator.depositAndMint), stake_dintokens (ERC20 approve + DinValidatorStake.stake), and read_dintoken_stake, plus the Typer wiring that exposes them as dincli dintoken buy|stake|read-stake and as the legacy role-scoped copies under dincli aggregator dintoken ... / dincli auditor dintoken ....

Unlike the integration harness (tests/dincli/, see dincli-testing-guide.md), these are pure unit tests: no chain, no RPC. The entire web3/contract surface is replaced by hand-rolled dummy objects, and build_and_send_tx is monkeypatched so no transaction is ever built or signed.

Running

cd /path/to/devnet
pytest tests/test_dintoken.py -v

# single test
pytest tests/test_dintoken.py -k test_buy_sends_eth_value

No services or environment variables are required.

What the suite verifies

buy — ETH deposit and mint

Test Behavior pinned down
test_buy_sends_eth_value buy_dintokens(ctx, 1.5) calls DinCoordinator.functions.depositAndMint() exactly once and passes it to build_and_send_tx with tx_params={"value": 1.5 ETH in wei} — i.e. the ETH amount rides along as transaction value, not as a function argument

stake — approve-then-stake flow

MIN_STAKE is 10 DIN (10 * 10**18 wei, defined in dincli/cli/utils.py). Validation order in the command is: minimum-stake check first, then token-balance check, then the two transactions.

Test Behavior pinned down
test_stake_uses_requested_amount stake_dintokens(ctx, 12) converts 12 DIN to wei once and uses the same amount for both legs: first DinToken.approve(stake_contract.address, amount), then DinValidatorStake.stake(amount) — two build_and_send_tx calls in that order
test_stake_exits_when_balance_is_insufficient With a balance of 11 DIN, staking 12 exits (typer.Exit) before any approve/stake call is made
test_stake_exits_when_amount_is_below_minimum Staking 9 (< MIN_STAKE = 10) exits with no contract interaction, regardless of balance

read-stake

Test Behavior pinned down
test_read_stake_reads_active_account_stake read_dintoken_stake queries DinValidatorStake.getStake(...) with the active wallet's address

Typer wiring

Test Behavior pinned down
test_top_level_dintoken_help_exposes_commands dincli dintoken --help lists buy, stake, read-stake
test_legacy_role_dintoken_help_exposes_commands[...] The role-scoped duplicates aggregator dintoken --help and auditor dintoken --help (separate Typer sub-apps in aggregator.py/auditor.py that wrap the same functions with a role label) expose the same three commands

Isolation patterns

Conventions to reuse when adding tests to this suite:

  • Dummy contract graph — DummyContextObj fakes the DinContext surface the commands touch (get_en_w3_account_console, get_deployed_din_token_contract / _stake_ / _coordinator_). Each dummy contract (DummyToken, DummyStake, DummyCoordinator) records invocations (approvals, stakes, deposit_calls, balance_of_calls, get_stake_calls) and returns sentinel values (e.g. ("approve", address, amount)) from its functions.* methods. Tests then assert both sides: the contract mock saw the call, and the sentinel arrived as build_and_send_tx's second positional argument — proving the command sent exactly the contract call it constructed. View calls return a DummyCall whose .call() yields a canned value (token balance, current stake).

  • build_and_send_tx capture — patched via monkeypatch.setattr(dintoken, "build_and_send_tx", ...) (on the consuming module, since it's imported by name) with a closure appending (args, kwargs) to a list and returning a fake receipt (SimpleNamespace(transactionHash=...)). Nothing web3-shaped is needed.

  • No real sleeping — stake_dintokens sleeps 5 s between approve and stake; every stake test patches dintoken.time.sleep to a no-op.

  • Constructor-injected scenario state — make_ctx(token_balance=..., current_stake=...) parameterizes the dummy state per test instead of mutating shared fixtures.

  • CliRunner only for wiring — command bodies are called directly with the fake ctx; CliRunner().invoke(main_app, ...) is used only for the --help wiring tests, which don't need a wallet or network.

Coverage gaps

Behavior of the dintoken commands this suite does not yet exercise:

  • Failure paths swallow errors — both buy_dintokens and stake_dintokens wrap their transaction calls in try/except Exception and only print the error (no typer.Exit, exit code stays 0). There is no test pinning this down — nor one deciding whether it's even the desired behavior (a failed approve leaves the command "successful").
  • MIN_STAKE boundary — only 9 (below) and 12 (above) are tested; staking exactly 10 (== MIN_STAKE, should pass) is unasserted.
  • Check ordering — minimum-stake is checked before balance; no test distinguishes the order (e.g. amount both below minimum and above balance).
  • Fractional amounts — the stake CLI command accepts a float, but stake_dintokens is typed amount: int and Web3.to_wei is trusted to convert; fractional stakes (e.g. 10.5) are untested.
  • buy output/balance display — the before/after balanceOf reads are recorded by the dummy but never asserted; console output (balances, tx hash) is unchecked throughout.
  • Command bodies via the CLI layer — buy/stake/read-stake are only invoked as plain functions; no CliRunner test drives them end-to-end through the Typer layer (argument parsing/conversion included).
  • Role-scoped behavior beyond help — the aggregator/auditor wrappers pass name="Aggregator"/"Auditor" labels; nothing asserts the label actually appears in output. (The client role has no dintoken sub-app, so aggregator/auditor is the complete set.)