Implement sell=buy support for Sell orders #3894

m-sz · 2025-11-10T16:22:02Z

Sell=buy token initiative

The ability to swap the same token pairs unlocks the potential for CoW to generally become a transaction relay service. This feature will allow users to execute pre- and post- interaction hooks as gasless transactions. The overall implementation has certain caveats:

Each settlement has to go through the settlement contract anyway, so the sell=buy token transactions would require users to send X amount of sell token and receive X - cost of the token back. It is not realistic, with the current state of GPv2 contract for the user to just send the cost amount and save on gas this way.
The solvers can really compete only by batching to save on gas costs
There is little incentive for this feature to be supported by other solvers, but having this support in baseline solver could generally be enough.

Lift restrictions on sell=buy token

Based the changes on https://github.com/cowprotocol/services/pull/2513/files, adapted to current state of the codebase.

Adds a command-line argument to the orderbook API to allow for same sell and buy token swaps. The SameBuyAndSell token error is not raised in that case
Adds e2e test case that checks if sell=buy quoting and orders are possible
Liquidity fetching happens only if the sell and buy tokens are different
Relaxes the invariant on Tokens struct to allow for same token pairs
Minor tweak to Justfile to allow for custom filtering and extra args to nextest

Based on input from @fhenneke:

Adds a short-circuit behavior for solver to return empty route in case sell=buy
Returns sell and buy assets directly when solving for empty route

crates/driver/src/domain/quote.rs

crates/e2e/tests/e2e/place_order_with_quote.rs

squadgazzz · 2025-11-13T16:07:12Z

I think this PR needs to be converted to a draft, since more changes are expected. @m-sz , is that correct?

m-sz · 2025-11-13T17:10:14Z

Yep, had a bunch of 1:1s and will be adding feature flag to the whole change to control it in orderbook.

m-sz · 2025-11-14T16:55:12Z

I have updated the PR with CLI argument to orderbook to allow for same sell and buy tokens.
Changes to baseline solver, to support those transactions will come in a separate PR.

github-actions · 2025-11-22T00:09:15Z

This pull request has been marked as stale because it has been inactive a while. Please update this pull request or it will be automatically closed.

crates/driver/src/domain/quote.rs

crates/e2e/tests/e2e/place_order_with_quote.rs

crates/shared/src/order_validation.rs

crates/solvers/src/domain/solver.rs

crates/orderbook/src/run.rs

crates/shared/src/order_validation.rs

crates/solvers/src/domain/solver.rs

crates/e2e/tests/e2e/place_order_with_quote.rs

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

crates/orderbook/src/arguments.rs

crates/e2e/tests/e2e/submission.rs

MartinquaXD · 2025-12-12T10:05:03Z

crates/shared/src/price_estimation/trade_verifier/mod.rs

        .context("summary buy token is missing")?;

-    if *sell_token_lost >= sell_token_lost_limit || *buy_token_lost >= buy_token_lost_limit {
+    if (!sell_token_lost.is_zero() && *sell_token_lost >= sell_token_lost_limit)


So in other words "something that used to work now fails". I'd like a more thorough investigation than that. If adding this check is indeed the correct thing to do here there should be a convincing explanation.

crates/shared/src/price_estimation/sanitized.rs

MartinquaXD · 2025-12-12T10:14:12Z

crates/shared/src/order_validation.rs

+        // Check for orders selling wrapped native token for native token.
+        if &order.sell_token == native_token && order.buy_token == BUY_ETH_ADDRESS {
+            return Err(PartialValidationError::SameBuyAndSellToken);
+        }


Why is this always forbidden? With the new feature this should be allowed when the flag is enabled, no?

I will test it out

Is that a case we want to support? As I understand, right now if a user places such order we are simply unwrapping their WETH for ETH, same goes the other way around.

I tried if it works, and it seems that in order to support that, we would need to plug in native token everywhere where I check for sell=buy (so into the solver for example), and then compare the tokens to native or ETH buy address

MartinquaXD · 2025-12-12T10:18:30Z

crates/solvers/src/domain/solver.rs

+                        input,
+                        output,
+                        interactions: Vec::default(),
+                        gas: eth::Gas(U256::ZERO) + self.solution_gas_offset,


Did you check that there is logic somewhere that accounts for the order hooks? I'm a bit worried that the baseline solver will start bleeding money when it doesn't consider the gas costs correctly.

The routing (which is present in the non sell=buy case) takes a Request and max amount of hops.
Request contains wrappers (which I understand are the hooks), and they are present in the solution in both cases.

Routing itself cares about the sell and buy token, computing a path to solve for them, not taking into account wrappers. So I think we should be fine, but I will check it on staging with hooks.

I will introduce an E2E test case verifying the hooks gas usage in sell=buy orders in a follow-up PR.

crates/solvers/src/domain/solver.rs

Co-authored-by: Martin Magnus <martin.beckmann@protonmail.com>

m-sz · 2025-12-16T14:54:55Z

I have revisited how we calcualate the before and after balance. This now allows us to avoid the .is_zero checks while ensuring quoting accuracy and seems like a more correct approach.

Introduce special case to handle sell=buy when calculating the out amounts.

The out_amount had to be extended to I512, to account for negative outcomes of the previous calculation

m-sz · 2025-12-17T12:20:07Z

With @MartinquaXD's help, We have also came up with a correct way to account for sell and buy token's out_amount during trade verification steps. The previous could would have worked with current baseline solver (as the input and output amounts of sell=buy are the same), but if any external solver would (more realistically) provide a quote with output amount being smaller (not equal) than the input amount, the solution would break. The new approach uses signed arithmetic and fixes up the calculation if the tokens are the same.

crates/shared/src/price_estimation/trade_verifier/mod.rs

MartinquaXD · 2025-12-17T13:00:18Z

crates/shared/src/price_estimation/trade_verifier/mod.rs

            // It looks like the out_amount is 0 but only because the sell and buy tokens
            // are the same.


I think this comment can now be dropped in favor of the more specific one below.

crates/shared/src/price_estimation/trade_verifier/mod.rs

crates/number/src/conversions.rs

m-sz force-pushed the sell=buy branch 2 times, most recently from de09f52 to 858ac73 Compare November 12, 2025 15:54

m-sz marked this pull request as ready for review November 12, 2025 15:55

m-sz requested a review from a team as a code owner November 12, 2025 15:55

fafk reviewed Nov 13, 2025

View reviewed changes

crates/driver/src/domain/quote.rs Outdated Show resolved Hide resolved

fafk reviewed Nov 13, 2025

View reviewed changes

crates/e2e/tests/e2e/place_order_with_quote.rs Outdated Show resolved Hide resolved

m-sz marked this pull request as draft November 13, 2025 17:10

m-sz force-pushed the sell=buy branch from d921ea9 to 68f627d Compare November 14, 2025 16:49

m-sz marked this pull request as ready for review November 14, 2025 16:54

github-actions bot added the stale label Nov 22, 2025

github-actions bot closed this Nov 29, 2025

extrawurst reopened this Nov 29, 2025

extrawurst removed the stale label Nov 29, 2025

m-sz force-pushed the sell=buy branch from 68f627d to 2e7b53d Compare December 2, 2025 10:40

squadgazzz reviewed Dec 2, 2025

View reviewed changes

crates/driver/src/domain/quote.rs Outdated Show resolved Hide resolved

squadgazzz reviewed Dec 2, 2025

View reviewed changes

fafk reviewed Dec 3, 2025

View reviewed changes

crates/orderbook/src/run.rs Outdated Show resolved Hide resolved

fafk reviewed Dec 3, 2025

View reviewed changes

crates/shared/src/order_validation.rs Outdated Show resolved Hide resolved

fafk reviewed Dec 3, 2025

View reviewed changes

crates/shared/src/order_validation.rs Outdated Show resolved Hide resolved

fafk reviewed Dec 3, 2025

View reviewed changes

crates/solvers/src/domain/solver.rs Outdated Show resolved Hide resolved

m-sz marked this pull request as draft December 3, 2025 10:50

m-sz force-pushed the sell=buy branch 3 times, most recently from 405acc6 to 6fe22d1 Compare December 5, 2025 14:24

m-sz marked this pull request as ready for review December 5, 2025 14:25

squadgazzz reviewed Dec 5, 2025

View reviewed changes

crates/e2e/tests/e2e/place_order_with_quote.rs Outdated Show resolved Hide resolved

m-sz and others added 8 commits December 11, 2025 15:13

Update crates/shared/src/price_estimation/factory.rs

307e7fb

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

Update crates/shared/src/price_estimation/sanitized.rs

d2cbf6e

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

Update crates/shared/src/order_validation.rs

f690b29

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

Update crates/shared/src/order_validation.rs

75273a2

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

Update crates/solvers/src/domain/solver.rs

ec5a055

Co-authored-by: José Duarte <duarte.gmj@gmail.com>

fmt

348f584

Merge branch 'main' into sell=buy

5ab57d6

NonZeroU256 usage fix in tests after merging

2656b98

MartinquaXD reviewed Dec 12, 2025

View reviewed changes

m-sz and others added 5 commits December 12, 2025 12:33

Update crates/orderbook/src/arguments.rs

295e56b

Co-authored-by: Martin Magnus <martin.beckmann@protonmail.com>

Update crates/e2e/tests/e2e/submission.rs

1476ff9

Co-authored-by: Martin Magnus <martin.beckmann@protonmail.com>

Address comments

0ec4291

Merge branch 'main' into sell=buy

0644027

Merge branch 'main' into sell=buy

8efe448

m-sz requested a review from MartinquaXD December 15, 2025 11:55

m-sz added 4 commits December 16, 2025 15:57

Revert before/after balance calculation

f42b7ae

Introduce special case to handle sell=buy when calculating the out amounts.

Merge branch 'main' into sell=buy

b5c612c

clippy

f46c77d

Fix out_amount calculation for sell=buy

6a55fd9

The out_amount had to be extended to I512, to account for negative outcomes of the previous calculation

MartinquaXD approved these changes Dec 17, 2025

View reviewed changes

m-sz added 2 commits December 17, 2025 14:13

Minor changes: comments and log level

c99d351

Merge branch 'main' into sell=buy

56ad09a

m-sz enabled auto-merge December 17, 2025 13:17

m-sz added this pull request to the merge queue Dec 17, 2025

Merged via the queue into main with commit a1f6c2e Dec 17, 2025
18 checks passed

m-sz deleted the sell=buy branch December 17, 2025 13:37

github-actions bot locked and limited conversation to collaborators Dec 17, 2025

		// It looks like the out_amount is 0 but only because the sell and buy tokens
		// are the same.

Implement sell=buy support for Sell orders #3894

Implement sell=buy support for Sell orders #3894

Conversation

m-sz commented Nov 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Sell=buy token initiative

Lift restrictions on sell=buy token

Uh oh!

Uh oh!

Uh oh!

squadgazzz commented Nov 13, 2025

Uh oh!

m-sz commented Nov 13, 2025

Uh oh!

m-sz commented Nov 14, 2025

Uh oh!

github-actions bot commented Nov 22, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

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!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

m-sz commented Dec 16, 2025

Uh oh!

m-sz commented Dec 17, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

m-sz commented Nov 10, 2025 •

edited

Loading