refactor(runtime): use lamports per signature constant

[leviathanbeak]

reopening #1586

I missed the notification for #1586, here's the updated version (didn't have the privileges to reopen the closed one directly)

thanks, for the review

[dnut]

I'm reopening the other one to keep the discussion history unified

[dnut]

never mind, github won't allow it because the branch was force pushed or recreated

[Copilot]

Pull request overview

This PR refactors fee handling by introducing a shared LAMPORTS_PER_SIGNATURE constant in the runtime fee-check logic and replacing several hard-coded 5000 usages with that constant (or with a derived lamports-per-signature value where already available), while also simplifying SvmGateway initialization by removing a now-unneeded lamports_per_signature init parameter.

Changes:

• Add sig.runtime.check_transactions.LAMPORTS_PER_SIGNATURE and use it in runtime execution/tests and mock transfer logic.
• Remove lamports_per_signature from SvmGateway.Params and update all init call sites accordingly.
• Update fee computation in RPC getFeeForMessage to use the blockhash-queue-derived lamports-per-signature instead of a hard-coded value.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
src/transaction_sender/MockTransferService.zig	Replace hard-coded transfer fee with runtime LAMPORTS_PER_SIGNATURE.
src/rpc/hook_contexts/SendTransaction.zig	Update SvmGateway.init call after removing lamports_per_signature param.
src/rpc/hook_contexts/Account.zig	Use bq_lps (blockhash queue lamports-per-signature) when computing FeeDetails.
src/replay/svm_gateway.zig	Remove lamports_per_signature from gateway params; set env fee rate from runtime constant.
src/replay/execution.zig	Update replay/test scaffolding to match new SvmGateway.Params.
shared/runtime/transaction_execution.zig	Thread lamports-per-signature into fee payer checking; update tests to use constant.
shared/runtime/check_transactions.zig	Add LAMPORTS_PER_SIGNATURE constant; accept lamports_per_signature as an input to fee payer checks; update tests.
conformance/src/txn_execute.zig	Use runtime constant for transaction execution environment fee configuration.

---

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

[leviathanbeak]

0947ac8

[leviathanbeak]

0947ac8

[leviathanbeak]

c7d526e

[dnut]

Hey @leviathanbeak, any thoughts on this feedback from the original PR? #1586 (review)

The removal of lamports_per_signature as an field/argument may not be something we want to do given the current solana spec.

Keeping the fields/arguments wouldn't be just because agave is currently doing it as well as using a constant variable, but more so because their FeeRateGovernor struct definition the doc comment on lamports_per_signature mentions that this amount can change over time.

A change to how this is calculated, could ultimately lead to needing to re-thread the parameter through each caller and function.

Replacing the hardcoded values with the new LAMPORTS_PER_SIGNATURE is definitely something we should keep here, since it eliminates the use of a magic number.

Also it looks like you have some ci failures.

[leviathanbeak]

Hey @leviathanbeak, any thoughts on this feedback from the original PR? #1586 (review)

The removal of lamports_per_signature as an field/argument may not be something we want to do given the current solana spec.
Keeping the fields/arguments wouldn't be just because agave is currently doing it as well as using a constant variable, but more so because their FeeRateGovernor struct definition the doc comment on lamports_per_signature mentions that this amount can change over time.
A change to how this is calculated, could ultimately lead to needing to re-thread the parameter through each caller and function.
Replacing the hardcoded values with the new LAMPORTS_PER_SIGNATURE is definitely something we should keep here, since it eliminates the use of a magic number.

Also it looks like you have some ci failures.

thanks @dnut , I think everything from original comment is resolved now, will check if there are further CI issues related to this PR

[leviathanbeak]

So given the feedback on the closed PR, all in all this PR

• addresses part of Simplify fee handling #1133 by replacing hardcoded 5000 fee values with LAMPORTS_PER_SIGNATURE.
• keeps lamports_per_signature plumbing where it already exists per review feedback
• checkFeePayer still uses the default fee-structure value rather than blockhash/nonce fee rates.

[dnut]

the files were moved so you'll need to resolve the conflicts before we can merge this

[leviathanbeak]

the files were moved so you'll need to resolve the conflicts before we can merge this

@dnut done