feat: LiquidityParty updated adapter by timolson · Pull Request #1005 · propeller-heads/tycho

timolson · 2026-05-11T00:35:19Z

The swap() interface has changed, so a new Executor will need to be deployed.
The Substreams adapter handles Killed events by setting all pool token balances to zero. If we have to disable a pool, this workaround prevents solvers from attempting a swap that would revert, without us needing to support PC deletion.
swapToLimit() is no longer supported and was removed from the SwapAdapter.
swapAmounts() moved from PartyPool to PartyInfo
The new test pool has only 3 tokens instead of 10 for faster test runs.
Tightened bounds on LiquidityPartyExecutorTest with the exact amount outs expected.
Other trivial updates to contract addresses, swap test amounts, etc. PartyPoolSwapImpl is called PartyPoolExtraImpl now.
No major flags from slither just a loop warning
I absolutely could not get integration testing to run because Substreams keeps rejecting my JWT token for some reason. I created fresh ones. I used the CLI. I even tried hitting the API directly. Nothing. This happened to me before, and it's hella frustrating. Yes, I'm using JWT from The Graph and not API tokens from substreams.dev. All of the tests for each component pass, so I'm assuming integration testing will work, too.

claude

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

zach030 · 2026-05-13T02:24:54Z

+                .for_each(|component| match components_store.get_last(&component.id) {
+                    None => {
+                        substreams::log::info!(
+                            "killed component not found in store: {}",
+                            component.id
+                        );
+                    }
+                    Some(tokens_str) => match decode_addrs(&tokens_str) {
+                        Err(e) => {
+                            substreams::log::info!(
+                                "failed to decode token addresses for killed component {}: {}",
+                                component.id,
+                                e
+                            );
+                        }
+                        Ok(token_addrs) => {
+                            for token in token_addrs {
+                                builder.add_balance_change(&BalanceChange {
+                                    token,
+                                    balance: vec![0],
+                                    component_id: component.id.as_bytes().to_vec(),
+                                });
+                            }
+                        }
+                    },


Please remove the substreams::log.

I don’t think setting the TVL to zero is a good way to make this pool untradable in Tycho. If a pool is killed, you can update a mutable entity associated with it and filter it out in tycho-simulation.
You can use the Curve filter as a reference. FYI, static attributes are immutable, so they’re not suitable for this kind of kill case.

Ok, will do.

Any plans in the future to have a PC delete? Seems like the system was designed with CRUD in mind... I thought I saw a DELETE event somewhere although in the previous review I got the impression it wasn't supported. I'll match whatever Curve does.

Yes, it was designed around CRUD, and supporting deletion would take quite a bit of work. I wouldn’t expect it anytime soon.

zach030 · 2026-05-13T02:30:27Z

Iterating over all ERC20 Transfer events would add a lot of indexing latency. Does the Liquidity Party pool emit any Swap / Mint / Burn events we could use instead to make the filtering and balance detection more efficient?

Yes, all of those. I originally had a full suite of our events and switched it out for Transfers last time... You'll notice that this part of the code was not changed in this update. I'll rewrite this to use our custom events.

Yes, I didn’t notice that in my last review. Switching to the full suite of your events would be a big performance improvement.

Yah makes total sense. Done.

# Conflicts: # crates/tycho-execution/contracts/test/assets/calldata.txt

timolson · 2026-05-13T04:50:53Z

Removed all Substreams logging
Removed excessive sanity checks in the interest of lower latency
Balance tracking updated to use only Liquidity Party events. TVL excludes protocol fees sitting in the pool awaiting collection
Killed pools set a dynamic attribute and are filtered out in protocol_stream_processor.rs. The Curve filter doesn't seem to be wired, so I followed the balancer_v2_pool_filter example.
Killed pools no longer track balances or send any other Substreams messages. A killed pool's Substreams will report balances at killed time, even if they are subsequently withdrawn by LP's. (Kill still allows LP's to burn() but for Tycho purposes no more Substreams messages will be sent for the pool after the kill.)

I'm going to test the new balance tracking tomorrow. Our test pool already has an "exercise" series of transactions that invokes one of every operation, so I just need to run Substreams on that sequence and verify the final balances.

timolson · 2026-05-13T11:15:13Z

I found a few more nits in our DEX I'd like to clean up as long as we're re-deploying. The current Tycho adapter should be good to review, but I will probably make one more update to the singleton contract addresses before we're really ready to merge. But no changes to API's or anything like that, just updated deployment addresses.

timolson · 2026-05-13T21:54:33Z

I updated the contract addresses and start blocks. There was also an ABI touchup, but it doesn't affect anything in Tycho. That's all the updates I wanted to get in.

timolson · 2026-05-14T10:36:12Z

Our auditor found an issue and we're going to need to redeploy again. Minor issues, nothing close to value extraction. I'd understand if you want to pause review until the audit is complete; should be next week. But the ABI's are stable for swaps and there have been no issues found with swap. The adapter code should be stable as-is except for future contract addr changes.

zach030 · 2026-05-14T14:17:31Z

Our auditor found an issue and we're going to need to redeploy again. Minor issues, nothing close to value extraction. I'd understand if you want to pause review until the audit is complete; should be next week. But the ABI's are stable for swaps and there have been no issues found with swap. The adapter code should be stable as-is except for future contract addr changes.

Thanks for the heads up. I haven’t seen any other major issues so far, and I’ll take another look once the audit is done.

timolson · 2026-05-19T15:17:05Z

Turns out we do need to make some pretty significant changes. It's gonna be a few weeks, and I'll probably write a native simulator instead of the EVM simulator. We're moving the bulk of the LMSR solving off-chain and just using an on-chain validation approach instead. Basically we found regimes where our on-chain math wasn't mathing, and the fix is too much gas. So it will be a swap() ABI change to use an offchain solver + on-chain validator model. Don't expect a proper PR for a while.

zach030 · 2026-05-20T05:42:06Z

Turns out we do need to make some pretty significant changes. It's gonna be a few weeks, and I'll probably write a native simulator instead of the EVM simulator. We're moving the bulk of the LMSR solving off-chain and just using an on-chain validation approach instead. Basically we found regimes where our on-chain math wasn't mathing, and the fix is too much gas. So it will be a swap() ABI change to use an offchain solver + on-chain validator model. Don't expect a proper PR for a while.

Thanks for the heads up! Sounds like you’re planning a native integration?
Before you start integration, please read this integration guide first to make sure Tycho supports it.
Feel free to ping me if you have any questions.

timolson · 2026-06-01T16:25:08Z

Our audit will likely wrap up this week, so I should have an updated adapter ready for your review next week.

There will not be any significant changes to the adapter. Should just be contract addresses, start blocks, etc.

timolson · 2026-06-08T12:38:57Z

Not going to be this week; we have a little more work to do.

LiquidityParty updated adapter

d7b5c80

github-project-automation Bot added this to Tycho May 11, 2026

github-project-automation Bot moved this to Todo in Tycho May 11, 2026

claude Bot reviewed May 11, 2026

View reviewed changes

timolson added 2 commits May 10, 2026 20:55

tighten swap decoding test; nits

511b388

gas cost update

b3dc7a6

zach030 reviewed May 13, 2026

View reviewed changes

timolson added 4 commits May 12, 2026 23:47

Balance tracking uses LiqP-specific events

b9d2dad

Killed pools filter

99b712c

Merge remote-tracking branch 'tycho/main'

3855075

# Conflicts: # crates/tycho-execution/contracts/test/assets/calldata.txt

Update test_encode_liquidityparty calldata.txt

12d36a9

timolson and others added 3 commits May 13, 2026 17:48

Merge branch 'propeller-heads:main' into main

bf82a76

Update contract addresses and block numbers

50a5436

Update ABI's

2258b56

zach030 added the protocol integration label May 21, 2026

Conversation

timolson commented May 11, 2026

Uh oh!

claude Bot left a comment

Choose a reason for hiding this comment

Claude Code Review

Uh oh!

zach030 May 13, 2026

Choose a reason for hiding this comment

Uh oh!

timolson May 13, 2026

Choose a reason for hiding this comment

Uh oh!

zach030 May 13, 2026

Choose a reason for hiding this comment

Uh oh!

zach030 May 13, 2026

Choose a reason for hiding this comment

Uh oh!

timolson May 13, 2026

Choose a reason for hiding this comment

Uh oh!

zach030 May 13, 2026

Choose a reason for hiding this comment

Uh oh!

timolson May 13, 2026

Choose a reason for hiding this comment

Uh oh!

timolson commented May 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timolson commented May 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timolson commented May 13, 2026

Uh oh!

timolson commented May 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

zach030 commented May 14, 2026

Uh oh!

timolson commented May 19, 2026

Uh oh!

zach030 commented May 20, 2026

Uh oh!

timolson commented Jun 1, 2026

Uh oh!

timolson commented Jun 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

timolson commented May 13, 2026 •

edited

Loading

timolson commented May 13, 2026 •

edited

Loading

timolson commented May 14, 2026 •

edited

Loading