
Ethereum Glamsterdam Test Flaw Lets Transactions Skip Payments in Simulations
Developer-published simulations of Ethereum’s proposed Glamsterdam upgrade have exposed an Ethereum Glamsterdam test flaw, even though it appeared only on a test network and no funds were lost. According to KuCoin News, tests carried out on November 14 examined contracts that move payments between projects, and they found four Juicebox V6 payment paths in which a transaction could succeed even though an internal payment inside it had failed. That is the kind of result upgrade testing is designed to catch: a quiet mismatch between what a transaction appears to do and what actually happens to the money inside it. The finding does not mean Ethereum’s main network is broken, or that anyone has exploited the flaw to steal funds.
What the simulations found
The simulations were published by a developer known as “mejango”, and they focused on Juicebox, a protocol that projects use to raise funds and manage payments on Ethereum. According to KuCoin News, the November 14 tests examined contracts that transfer payments between projects on the Sepolia test network, one of Ethereum’s public testing environments. The results included a case in which a buyer retained an NFT despite the failure of the associated payment to another project, a striking outcome because the purchase and the payment are supposed to move together. In a separate cash-out simulation, a withdrawal was allowed without collecting a 2.5% fee at sampled gas limits between 679,622 and 817,622. Gas limits are the maximum amount of computational work a transaction is allowed to use, so the fee being skipped within that range matters for understanding when the behaviour appears. The developer said the simulations invoke deployed contract code through eth_simulateV1, a method that lets developers run a transaction in a simulated environment and inspect the result without actually sending it on the network.
Juicebox V6 payment paths explained
To understand why the finding matters, it helps to understand what Juicebox V6 payment paths actually do. Juicebox is a funding and payments system that lets projects receive money, split it between different recipients, and pay it out again under rules written into smart contracts. A payment path is the route money follows through those contracts: a person pays into one project, and part of that payment may then be forwarded to another project, a contributor, or a shared treasury. The four paths identified in the simulations are cases where that forwarding step failed but the overall transaction still reported success. In a blockchain system, the problem is sharper because transactions are supposed to be atomic, meaning either everything in them happens or nothing does. When a payment inside a transaction fails silently while the rest completes, the ledger records a state nobody intended. The Juicebox V6 payment paths at the centre of this story are therefore not a minor edge case: the cost of each step determines whether the whole transaction has enough gas to finish.
What EIP-8037 and EIP-8038 would change
The proposed changes at the heart of the simulations are known as EIP-8037 and EIP-8038, together often written as EIP-8037 EIP-8038. An EIP, or Ethereum Improvement Proposal, is a formal document that describes a proposed change to the network, and both proposals deal with how Ethereum charges for using its storage. EIP-8037 proposes increasing the cost of creating persistent blockchain data and accounting separately for state creation and execution, which means writing new information that must be stored permanently would be priced differently from simply running code. EIP-8038 changes costs for state access and state writes, the operations that read existing stored data and update it. Together, the two proposals would reshape the economics of interacting with Ethereum’s state, the shared record of every account balance, contract and piece of stored information. The intention is usually to make pricing fairer and to protect the network from operations that are cheap for a user but expensive for every node that must store the result forever. The risk, as the simulations suggest, is that changing those prices can also change which steps inside a complex transaction run out of gas first.
How eth_simulateV1 and Sepolia testing work
The testing itself is as important as the result, because it shows how Ethereum catches problems before they reach real users. Sepolia is a test network, or testnet, that runs the same software as Ethereum’s main network but uses tokens with no real-world value, so developers can experiment freely without putting anyone’s money at risk. New upgrades are deployed there first, applications are run against them, and anything strange is reported before the upgrade is considered for the main network. The method used here, eth_simulateV1, lets a developer ask a node to simulate a transaction and return a detailed trace of what would happen, including which internal calls succeeded and which failed. A developer can run it in a sandbox and inspect every step, which is how “mejango” showed that a payment to another project failed while the surrounding transaction still succeeded. The tests also used deployed contract code, meaning they ran against the real Juicebox contracts rather than a simplified copy.
Why a testnet finding is not a mainnet exploit
It is essential to be clear about what this story is not. This is a test-network simulation of a proposed change, not a mainnet exploit, and no funds were lost. Glamsterdam has not been deployed to Ethereum’s main network, and the behaviour was observed on Sepolia, where tokens are valueless and the whole point is to find problems safely. There is no report of anyone losing money, of a live Juicebox project being drained, or of an attacker using the flaw in the wild. That distinction matters for readers who hold ether or use Ethereum applications, because a headline about a flaw can easily be misread as a warning that the network itself is unsafe today. The honest reading is the opposite: the testing process worked as intended, and a potential problem was surfaced while it was still cheap to fix.
What happens next: November 16, November 19 and Hoodi
The calendar around this finding shows how Ethereum’s upgrade process moves from a test result to a decision. According to KuCoin News, Ethereum’s November 16 testing-call agenda includes Juicebox’s compatibility concerns and proofs of concept under investigations into EIP-8037 and EIP-8038, and the findings are meant to inform Thursday’s execution-layer discussion. The execution layer is the part of Ethereum that runs smart contracts and processes transactions, so a discussion there is exactly where a question about gas costs and transaction outcomes belongs. The Ethereum Foundation’s November 9 update recommended examining public audit findings before decisions about specification changes or the Hoodi testnet schedule, and it schedules discussion of findings and next steps for November 19 without committing to a decision. Hoodi is another testnet, and the schedule for deploying an upgrade there is a milestone developers watch closely, because a testnet deployment is a step towards a mainnet release but not a promise of one.
Why it matters for Ethereum users and developers
For ordinary users, the immediate practical impact is limited, because nothing on the main network has changed and no action is required. The longer-term significance is larger, because Glamsterdam is Ethereum’s next major upgrade after Fusaka, and it will shape how much everyday transactions cost and how reliably complex applications behave. If a pricing change can cause a payment inside a transaction to fail silently, then wallets, decentralised exchanges, funding platforms and other applications that chain several steps together all need to be checked against the new rules. For developers, the lesson is concrete: assumptions about gas costs that were safe under the old pricing may not hold under the new one, and testing with tools such as eth_simulateV1 is the way to find out before users do. For the wider ecosystem, the episode is a reminder that Ethereum’s upgrade process is deliberately slow and public, and that openness lets independent developers such as “mejango” publish simulations anyone can inspect and reproduce.
Risks and open questions
Several questions remain open, and honest reporting should name them rather than pretend they are settled. The first is scope: the simulations identified four Juicebox V6 payment paths, but it is not yet clear how many other applications with similar multi-step payment flows could be affected in the same way. The second is the exact mechanism: whether the fee-skipping behaviour is a direct consequence of the proposed pricing in EIP-8037 and EIP-8038, or an interaction with how Juicebox itself handles gas, will matter for deciding whether the fix belongs in the upgrade specification or in the application. The third is the remedy: options could include adjusting the proposed costs, clarifying how failed internal payments should be handled, or updating Juicebox’s contracts. The fourth is timing: with discussion scheduled for November 19 and no decision committed, readers should not expect an instant verdict.
Frequently asked questions
What is the Ethereum Glamsterdam test flaw? The Ethereum Glamsterdam test flaw is a behaviour found in simulations of the proposed Glamsterdam upgrade, in which transactions succeeded on the Sepolia test network despite an internal payment failing, including a case where a buyer kept an NFT while the associated payment did not go through.
Did anyone lose money because of this flaw? No. The finding came from simulations on the Sepolia test network using valueless test tokens, and the source material reports no loss of funds, no mainnet exploit and no live project being affected.
What are EIP-8037 and EIP-8038? EIP-8037 and EIP-8038 are Ethereum Improvement Proposals that would change how the network charges for creating, accessing and writing persistent blockchain data, with EIP-8037 separating the cost of state creation from execution and EIP-8038 changing the costs of state access and state writes.
What happens next for Glamsterdam? The findings are on the agenda for Ethereum’s November 16 testing call and are meant to inform an execution-layer discussion, while the Ethereum Foundation has scheduled discussion of findings and next steps for November 19 without committing to a decision.
Source: KuCoin News.
Related reading on MM Crypto Analysis Web
- Ethereum Sinks 26% in a Month as Whales Buy Near $1,700
- MegaETH Parent Cuts 20% of Staff and Pivots to AI Inference Market
- Ethereum Foundation Sold 7,500 ETH in Four Days as Price Breaks Below $1,750
More in Ethereum on MM Crypto Analysis Web.
Pingback: BitMine Buys 7653 ETH as Holdings Hit 6.08M