test: stableswap parity and looping by mrq1911 · Pull Request #1408 · galacticcouncil/hydration-node

mrq1911 · 2026-04-03T16:35:50Z

No description provided.

github-actions · 2026-04-03T16:37:03Z

Crate versions that have been updated:

runtime-integration-tests: v1.75.1 -> v1.76.0

Runtime version has not been increased.

Copilot

Pull request overview

Adds a Curve-vs-Hydration stableswap math parity test suite by deploying a Vyper contract containing Curve’s StableSwap math onto the local EVM (Frontier) and comparing results across a broad set of scenarios, plus some looping “value extraction” checks.

Changes:

Add a pure Vyper “CurveStableSwapMath” contract exposing Curve StableSwap invariant/swap/liquidity math for testing.
Add Rust integration tests that deploy/call the Vyper contract and compare outputs against hydra_dx_math::stableswap.
Wire the new integration test module into the integration test crate.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 7 comments.

File	Description
scripts/test-contracts/vyper/CurveStableSwapMath.vy	New pure Vyper contract exposing Curve StableSwap math as callable functions for parity testing.
scripts/test-contracts/vyper/CurveStableSwapMath.bin	Compiled bytecode blob included for deployment from Rust tests.
integration-tests/src/stableswap_curve_comparison.rs	New integration tests deploying the Vyper contract and asserting parity/tolerances vs Hydration math + looping checks.
integration-tests/src/lib.rs	Registers the new test module.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-03T16:39:25Z

scripts/test-contracts/vyper/CurveStableSwapMath.vy

+    N_COINS: uint256 = len(xp)
+    S: uint256 = 0
+    for x: uint256 in xp:
+        S += x
+    if S == 0:
+        return 0
+
+    Dprev: uint256 = 0
+    D: uint256 = S
+    Ann: uint256 = amp * N_COINS
+    for _i: uint256 in range(255):
+        D_P: uint256 = D
+        for x: uint256 in xp:
+            D_P = D_P * D // (x * N_COINS)
+        Dprev = D
+        D = (Ann * S + D_P * N_COINS) * D // ((Ann - 1) * D + (N_COINS + 1) * D_P)


_get_D (and downstream _get_y / _get_y_D) can divide by zero when any element of xp is 0 (D_P * D // (x * N_COINS)). Since this contract exposes the math externally, it should explicitly validate inputs (e.g., assert N_COINS > 1 and assert x > 0 for all balances) or define a safe behavior for zero balances to avoid runtime reverts that are hard to diagnose.

Copilot · 2026-04-03T16:39:26Z

scripts/test-contracts/vyper/CurveStableSwapMath.vy

+    D2: uint256 = D1
+    if token_supply > 0 and fee > 0:
+        _fee: uint256 = fee * N_COINS // (4 * (N_COINS - 1))
+        adjusted: DynArray[uint256, 8] = []
+        for i: uint256 in range(8):
+            if i >= N_COINS:
+                break
+            ideal_balance: uint256 = D1 * old_balances[i] // D0
+            difference: uint256 = 0
+            if ideal_balance > new_balances[i]:
+                difference = ideal_balance - new_balances[i]
+            else:
+                difference = new_balances[i] - ideal_balance
+            adjusted.append(new_balances[i] - (_fee * difference // FEE_DENOMINATOR))
+        D2 = self._get_D(adjusted, amp)


Fee normalization uses fee * N_COINS // (4 * (N_COINS - 1)), which will revert for N_COINS == 1 (division by zero). Consider asserting N_COINS > 1 in calc_token_amount / calc_withdraw_one_coin (or guarding the fee path) so invalid pool sizes fail with a clear error.

Copilot · 2026-04-03T16:39:26Z

scripts/test-contracts/vyper/CurveStableSwapMath.vy

+    D0: uint256 = self._get_D(balances, amp)
+    D1: uint256 = D0 - token_amount * D0 // total_supply
+
+    new_y: uint256 = self._get_y_D(amp, i, balances, D1)
+    dy_0: uint256 = balances[i] - new_y
+
+    xp_reduced: DynArray[uint256, 8] = []
+    for j: uint256 in range(8):
+        if j >= N_COINS:
+            break
+        dx_expected: uint256 = 0
+        if j == i:
+            dx_expected = balances[j] * D1 // D0 - new_y
+        else:
+            dx_expected = balances[j] - balances[j] * D1 // D0
+        xp_reduced.append(balances[j] - (_fee * dx_expected // FEE_DENOMINATOR))
+
+    dy: uint256 = xp_reduced[i] - self._get_y_D(amp, i, xp_reduced, D1)
+    dy = dy - 1  # Withdraw less to account for rounding errors
+


calc_withdraw_one_coin can revert for valid-looking inputs due to missing guards:

total_supply can be 0 (division by zero)

token_amount can exceed total_supply (underflow in D1)

if token_amount * D0 // total_supply == 0, then D1 == D0 and later dy = dy - 1 will underflow.
Add explicit assertions (e.g., assert total_supply > 0, assert token_amount <= total_supply, assert D1 < D0 or early-return (0,0) when D1 == D0) to avoid these reverts.

Copilot · 2026-04-03T16:39:26Z

scripts/test-contracts/vyper/CurveStableSwapMath.vy

+    Curve's get_dy — calculate swap output amount.
+    Returns the amount of coin j received for dx of coin i, after fee.
+    fee is in units of FEE_DENOMINATOR (10^10). Pass 0 to disable fees.
+    """


get_dy lacks basic input validation (i != j, i/j < len(balances), dx > 0). Since it indexes balances[i]/balances[j] before any checks, invalid inputs will revert with an out-of-bounds error rather than a clear assertion. Consider adding the same index assertions used in _get_y plus assert dx > 0 to make failures explicit and consistent.

Suggested change

"""

"""

N_COINS: uint256 = len(balances)

assert i < N_COINS

assert j < N_COINS

assert i != j

assert dx > 0

Copilot · 2026-04-03T16:39:26Z

integration-tests/src/stableswap_curve_comparison.rs

+fn assert_parity(label: &str, hydra: u128, curve: u128, max_tolerance: u128, expect_hydra_gte: bool) {
+	let diff = hydra.abs_diff(curve);
+	assert!(
+		diff <= max_tolerance,
+		"{}: diff {} exceeds tolerance {} (hydra={}, curve={})",
+		label,
+		diff,
+		max_tolerance,
+		hydra,
+		curve
+	);
+	if expect_hydra_gte {
+		assert!(
+			hydra >= curve,
+			"{}: expected hydra ({}) >= curve ({})",
+			label,
+			hydra,
+			curve


assert_parity is defined as (hydra, curve, ...), but some callers pass (curve_out, hydra_out, ...) / (curve_shares, hydra_shares, ...). This makes assertion output misleading (it prints hydra= with the Curve value) and makes the expect_hydra_gte direction check easy to misuse. Consider standardizing call sites to always pass (hydra, curve) (or rename parameters / provide a dedicated helper for the “curve >= hydra” case).

Suggested change

fn assert_parity(label: &str, hydra: u128, curve: u128, max_tolerance: u128, expect_hydra_gte: bool) {

let diff = hydra.abs_diff(curve);

assert!(

diff <= max_tolerance,

"{}: diff {} exceeds tolerance {} (hydra={}, curve={})",

label,

diff,

max_tolerance,

hydra,

curve

);

if expect_hydra_gte {

assert!(

hydra >= curve,

"{}: expected hydra ({}) >= curve ({})",

label,

hydra,

curve

fn assert_parity(

label: &str,

first: u128,

second: u128,

max_tolerance: u128,

expect_first_gte_second: bool,

) {

let diff = first.abs_diff(second);

assert!(

diff <= max_tolerance,

"{}: diff {} exceeds tolerance {} (first={}, second={})",

label,

diff,

max_tolerance,

first,

second

);

if expect_first_gte_second {

assert!(

first >= second,

"{}: expected first ({}) >= second ({})",

label,

first,

second

Copilot · 2026-04-03T16:39:27Z

integration-tests/src/stableswap_curve_comparison.rs

+	// Hydration should give slightly less due to +2 bias and -1 rounding
+	assert_parity(
+		&format!("{} swap {}% no fee", label, 100 / dx_fraction),
+		curve_out,
+		hydra_out,
+		MAX_SWAP_TOLERANCE,
+		true, // curve_out >= hydra_out
+	);


The swap label uses 100 / dx_fraction as a percent. For dx_fraction = 10000 (0.01%) this formats as 0%, which makes failures harder to interpret. Consider formatting the swap size in basis points or as a fraction (e.g., dx_fraction itself) so the label matches the actual test case.

Copilot · 2026-04-03T16:39:27Z

integration-tests/src/stableswap_curve_comparison.rs

+use primitives::{AccountId, EvmAddress};
+use sp_core::U256;
+use sp_runtime::Permill;
+use xcm_emulator::{Network, TestExt};


Network is imported but never used (use xcm_emulator::{Network, TestExt};). This will produce an unused-import warning (and could fail CI if warnings are denied). Consider removing Network from the import list.

Suggested change

use xcm_emulator::{Network, TestExt};

use xcm_emulator::TestExt;

github-actions · 2026-04-03T17:52:45Z

Quick benchmark at commit f181c17 has been executed successfully.
View results

mrq1911 added 2 commits April 3, 2026 18:14

curve parity tests

a6257fa

add & remove cycle test

8c1fb82

Copilot AI review requested due to automatic review settings April 3, 2026 16:35

Copilot started reviewing on behalf of mrq1911 April 3, 2026 16:36 View session

Copilot AI reviewed Apr 3, 2026

View reviewed changes

mrq1911 added 2 commits April 3, 2026 18:51

fmt

1deaa0f

bump

f181c17

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test: stableswap parity and looping#1408

test: stableswap parity and looping#1408
mrq1911 wants to merge 4 commits intomasterfrom
curve-parity

mrq1911 commented Apr 3, 2026

Uh oh!

github-actions bot commented Apr 3, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

github-actions bot commented Apr 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-    """
+    """
+    N_COINS: uint256 = len(balances)
+    assert i < N_COINS
+    assert j < N_COINS
+    assert i != j
+    assert dx > 0

	use xcm_emulator::{Network, TestExt};
	use xcm_emulator::TestExt;

Conversation

mrq1911 commented Apr 3, 2026

Uh oh!

github-actions bot commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Apr 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

github-actions bot commented Apr 3, 2026 •

edited

Loading