Test migration #1115

josh-orca · 2025-10-17T09:41:49Z

Summary

This PR completes the test migration from Anchor-based integration tests to LiteSVM with Vitest across all categories. It removes the old Anchor integration test flow and introduces a single workspace-driven command that runs the LiteSVM test suite.

New test command: yarn workspace @orca-so/whirlpools-sdk-integration test

Completes in 171s!! 🥳🥳

What changed

Removed Anchor integration test usage and scripts; tests now run under Vitest + LiteSVM.
Added/updated scripts to run LiteSVM integration tests:
- legacy-sdk/whirlpool/package.json: "test": "vitest run tests/integration"
- legacy-sdk/integration/package.json: "test": "yarn workspace @orca-so/whirlpools-sdk test"
Migrated tests by suite:
- adaptive-fee
- migration
- multi-ix
- token-badge
- token-extensions
- v1-ix
- v2-ix
Replaced “anchor test” references in docs/config with the new workspace test command.

jshiohaha

oops - i was reviewing commit by commit, but it looks like some of the commits no longer exist 🙈 fortunately, it looks like GH still recognizes the comments and the general area in which they apply. i'm just posting comments for the first 2 commits before they become too outdated, and i will continue reviewing the rest of the commits 🌱

legacy-sdk/whirlpool/tests/utils/init-utils.ts

legacy-sdk/whirlpool/vitest.config.ts

legacy-sdk/whirlpool/tests/utils/init-utils.ts

jshiohaha · 2025-10-21T19:56:27Z

legacy-sdk/whirlpool/tests/utils/fixture.ts

-    ),
-  );
+  // Initialize tick arrays sequentially to avoid AccountBorrowFailed errors in LiteSVM
+  const results = [];


nit: you could probably add a utility to handle this sequential execution under the hood while maintaining a functional interface (if you want). under the hood, i imagine the utility could do just this or use smth like p-limit

haha! I was just thinking of this when I looking over the code earlier today! Will look into it

jshiohaha · 2025-10-22T03:31:03Z

legacy-sdk/whirlpool/tests/utils/litesvm.ts

+
+// Normalize logs so Anchor's EventParser can decode events reliably.
+// Solana RPC emits binary event data as: "Program data: <base64>".
+// Some LiteSVM builds may emit hex or numeric arrays instead; convert them to base64.


This is interesting! Could you explain more? Is behavior version dependent? Is it dependent on the underlying program?

and maybe dumb question, but is pretty_logs a utility that could help here? https://www.litesvm.com/docs/testing-your-program/executing-instructions#working-with-program-logs

This function is being defensive to handle all possible formats LiteSVM might emit, making tests robust across different environments. It's not dependent on the underlying program.

pretty_logs is for debugging/viewing logs, while convertPayloadToBase64 is for programmatically parsing events in your automated tests.

oh interesting.. i didn't realize LiteSVM would conditionally output different formats 🤔 i was thinking pretty_logs might be useful to convert the logs into the desired format w/o all the extra logic, e.g. pretty_logs -> split by new line -> parse events

legacy-sdk/whirlpool/tests/utils/litesvm.ts

jshiohaha

ok, i added comments for the rest of the commits! overall, it seems like most tests look good & many comments are related to styling or design. there are a few related to modified test logic that i wasn't sure about, so i tried to comment just to double check those cases

and apologies, i thought i added more comments last night, but it looks like they didn't actually get posted because they're still marked as pending 🤦🏻‍♂️

jshiohaha · 2025-10-23T05:39:52Z

legacy-sdk/whirlpool/tests/integration/multi-ix/splash_pool.test.ts

-      // SplashPool has only 2 TickArrays for negative and positive ticks
-      -1, +1,
-    ]))!.buildAndExecute();
+      const fullRange = TickUtil.getFullRangeTickIndex(


curious, is it only whitespace/formatting changes in this area? i'm having a little trouble seeing what changed

Ah yea git's diff does not make it clear. The only relevant changes are in the beforeAll and beforeEach callbacks.

jshiohaha · 2025-10-23T05:43:58Z

legacy-sdk/whirlpool/tests/integration/multi-ix/adaptive_fee.test.ts

+        resetLiteSVM();
+        await startLiteSVM();


how do you recommend we think about calling resetLiteSVM and startLiteSVM?

i see they're paired together here within beforeEach, but below it seems like sometimes it's just startLiteSVM in a beforeAll and resetLiteSVM in beforeEach (without startLiteSVM)

Some tests share the litesvm state, since startLiteSVM follows a singleton pattern. Each reset reloads multiple .so programs from disk and considerably increases test time, so I only reset state where necessary to prevent AccountAlreadyInUse errors.

ok, interesting! so your process was smth like don't reset b/c latency unless you run into issues?

legacy-sdk/whirlpool/tests/integration/multi-ix/adaptive_fee.test.ts

...y-sdk/whirlpool/tests/integration/migration/migrate_repurpose_reward_authority_space.test.ts

legacy-sdk/whirlpool/tests/integration/token-badge/delete_token_badge.test.ts

jshiohaha · 2025-10-24T04:39:37Z

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/initialize_pool_v2.test.ts

+              // Skip decimals assertions in event due to LiteSVM decoding quirks; verify via mints below
+              // LiteSVM event may misreport tickSpacing and initialSqrtPrice; verify on-chain below instead
+              // LiteSVM may also misdecode pubkeys in the event payload; verify on-chain below instead


wow, this is pretty interesting. is this due to the event payload being decoded incorrectly? iirc, there was some behavior related to logs requiring a bunch of custom logic

When a transaction is built:

It captures a blockhash from the blockchain

The transaction is serialized with that blockhash

That specific blockhash has a limited lifetime

After expireBlockhash():

The old blockhash is marked expired

LiteSVM strictly rejects transactions with expired blockhashs

The tx object still has the old, expired blockhash cached

solana-test-validator is more forgiving around using the same transaction object. In LiteSVM, once a transaction object has an expired blockhash, it required creating a completely new transaction object.

ok, thanks for explaining! although, i think this might be in response to this comment: #1115 (comment)

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/initialize_pool_v2.test.ts

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/initialize_reward_v2.test.ts

legacy-sdk/whirlpool/tests/integration/v2-ix/liquidity/increase_liquidity_v2.test.ts

legacy-sdk/whirlpool/tests/integration/v2-ix/two-hop-swap/two_hop_swap_v2.test.ts

jshiohaha

ok, i think i made it through all the files again. a couple last comments, but approving to merge after you address remaining comments at your discretion. maybe the most concerning (b/c i lack context) are the value changes here.

nice work! this is a beast 🐗

jshiohaha · 2025-10-31T05:42:19Z

.github/workflows/checks.yml

-      run: yarn install
-    - name: Build packages
-      run: yarn build --output-style static
+      - name: Checkout repository


revert ws changes?

ah yea I thought they had been altered by the yarn format command, but it was just my code editor - will do!

jshiohaha · 2025-10-31T05:45:48Z

legacy-sdk/whirlpool/tests/utils/litesvm.ts

+  });
+}
+
+export async function initializeLiteSVMEnvironment() {


thanks for adding, removes a lot of redundant lines from the code base + review 🙏🏻

yea I agreed that this was the bulk of the entire PR!

jshiohaha · 2025-10-31T05:54:04Z

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/set_reward_emissions_v2.test.ts

-            poolInitInfo.whirlpoolPda.publicKey,
-            IGNORE_CACHE,
-          )) as WhirlpoolData;
+          whirlpool = await (async () => {


curious - do we need the same for lines 111-119? i see it's doing something similar, i.e. setting emissions to a non-zero value first

Yes! After setting a non-zero emissions value, LiteSVM can also delay the pool update. We can use the same polling pattern there.

jshiohaha · 2025-10-31T05:59:07Z

legacy-sdk/whirlpool/tests/integration/v2-ix/two-hop-swap/two_hop_swap_v2.test.ts

+            startTickIndex: 25344,
            arrayCount: 12,
            aToB: false,
          });

+          // Add positions
+          const fundParams: FundedPositionV2Params[] = [
+            {
+              liquidityAmount: new anchor.BN(10_000_000),
+              tickLowerIndex: 29440,
+              tickUpperIndex: 33536,
+            },
+          ];
+          aqConfig.initPositionParams.push({ poolIndex: 0, fundParams });
+          aqConfig.initPositionParams.push({ poolIndex: 1, fundParams });
+


is there a reason to change some of these values?

we change them only to exercise specific scenarios; otherwise we keep them constant

jshiohaha · 2025-10-31T06:02:20Z

legacy-sdk/whirlpool/tests/sdk/whirlpools/utils/fetcher-util.test.ts

+      .getInfos()
+      .positions.slice(0, 5)
+      .map((p) => p.publicKey);
+    const posExt = fixture


maybe this update got lost amongst all the others? 😅

jshiohaha · 2025-10-31T06:03:04Z

legacy-sdk/whirlpool/tests/sdk/whirlpools/utils/fetcher-util.test.ts

+    const resultAllFalse = {
+      positions: new Map(),
+      positionsWithTokenExtensions: new Map(),
+    };


do you think it's worth adding a comment for context?

jshiohaha · 2025-10-31T06:04:37Z

legacy-sdk/whirlpool/tests/utils/init-utils.ts

+      const configKey =
+        configParams!.configInitInfo.whirlpoolsConfigKeypair.publicKey;


i vaguely remember that we might've updated configParams to have a default so we didn't need the shebang?

jshiohaha · 2025-10-31T06:09:49Z

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/initialize_pool_v2.test.ts

+          const actualBump = Array.isArray(whirlpool.whirlpoolBump)
+            ? whirlpool.whirlpoolBump[0]
+            : (whirlpool.whirlpoolBump as unknown as number);


what is this? 🤔

it verifies the program ignores a caller-supplied bump and uses the correct PDA-derived bump, preventing clients from influencing PDA derivation and ensuring deterministic, secure address generation.

jshiohaha · 2025-10-31T06:11:46Z

legacy-sdk/whirlpool/tests/integration/v2-ix/initialize/initialize_pool_v2.test.ts

+          let _eventVerified = false;
+          let _detectedSignature: string | null = null;
          const listener = ctx.program.addEventListener(
            "poolInitialized",
-            (event, _slot, signature) => {
-              detectedSignature = signature;
+            (_event, _slot, signature) => {
+              _detectedSignature = signature;
              // verify


it looks like this listener doesn't do anything either? worth removing?

yes you're right! removing!

josh-orca requested a review from jshiohaha October 17, 2025 09:41

josh-orca requested review from odcheung, wjthieme and yugure-orca as code owners October 17, 2025 09:41

calintje mentioned this pull request Oct 21, 2025

[TKN-703] Add test profile to docker-compose.yml (pending merge of PR 1115) #1123

Closed

josh-orca force-pushed the test-migration branch from a2c33c3 to 547cae9 Compare October 22, 2025 01:25

josh-orca requested a review from calintje as a code owner October 22, 2025 01:25

josh-orca force-pushed the test-migration branch 5 times, most recently from 3d75ad9 to 8ab1348 Compare October 22, 2025 03:42

jshiohaha reviewed Oct 22, 2025

View reviewed changes

josh-orca force-pushed the test-migration branch 5 times, most recently from b4707fb to 62f842b Compare October 22, 2025 21:48

jshiohaha reviewed Oct 24, 2025

View reviewed changes

josh-orca force-pushed the test-migration branch 4 times, most recently from f00da43 to 7c72880 Compare October 30, 2025 21:35

jshiohaha approved these changes Oct 31, 2025

View reviewed changes

josh-orca force-pushed the test-migration branch 6 times, most recently from 6e2ad75 to 14a4a1b Compare October 31, 2025 13:22

josh-orca force-pushed the test-migration branch 4 times, most recently from 0dbaff8 to 237eaad Compare October 31, 2025 14:35

wjthieme approved these changes Oct 31, 2025

View reviewed changes

josh-orca force-pushed the test-migration branch 4 times, most recently from 8627888 to ab304ec Compare October 31, 2025 15:52

josh-orca added 13 commits October 31, 2025 15:52

[Test Migration][1] Add litesvm configuration files

d71d7da

[Test Migration][2] Migrate adaptive-fee tests

e3af464

[Test Migration][3] Migrate migration tests

62b5984

[Test Migration][4] Migrate multi-ix tests

5f848da

[Test Migration][5] Migrate token-badge tests

6ff9cab

[Test Migration][6] Migrate token-extensions tests

bca3c8c

[Test Migration][7] Migrate v1-ix tests

51fcdee

[Test Migration][8] Migrate v2-ix tests

3ea7b49

[Test Migration][9] Migrate client-impl tests

a4b0062

[Test Migration][10] Migrate swap tests

dc256a1

[Test Migration][11] Migrate sdk utils

ab1ca74

[Test Migration][12] Remove anchor integration tests and update script

e3893a5

[Test Migration][13] Update changeset

ab304ec

josh-orca enabled auto-merge (squash) October 31, 2025 16:20

josh-orca disabled auto-merge October 31, 2025 16:21

[Test Migration][14] Temporarily disable test script

4af4615

josh-orca merged commit 94174ad into main Nov 3, 2025
7 checks passed

josh-orca deleted the test-migration branch November 3, 2025 18:04

		const configKey =
		configParams!.configInitInfo.whirlpoolsConfigKeypair.publicKey;

Test migration #1115

Test migration #1115

Uh oh!

Conversation

josh-orca commented Oct 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

What changed

Uh oh!

jshiohaha left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jshiohaha left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jshiohaha left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

josh-orca commented Oct 17, 2025 •

edited

Loading

jshiohaha left a comment •

edited

Loading