perf(rpc): dedupe preconfirmed subscriptions before adaptation

[danielntmd]

User description

Performs preconfirmed subscription deduplication before RPC response adaptation, avoiding unnecessary transaction and receipt object creation for entries already emitted in the same preconfirmed round.

---

PR Type

Enhancement

---

Description

• Dedupe preconfirmed entries before adaptation

• Avoid unnecessary transaction object creation

• Avoid unnecessary receipt object creation

• Apply pattern across RPC v9 and v10

---

File Walkthrough

Relevant files

Enhancement

subscription_receipts.go Dedupe preconfirmed receipts before adaptation                      rpc/v10/subscription_receipts.go Adds shouldSend callback to receiptsOf Marks receipts as sent before adaptation Skips already-sent preconfirmed receipts Passes nil callback for canonical blocks +12/-7    subscription_transactions.go Dedupe preconfirmed transactions before adaptation              rpc/v10/subscription_transactions.go Adds shouldSend callback to transactionsOf Marks transactions as sent before adaptation Uses txn.Hash() for deduplication Passes nil callback for canonical blocks +11/-7    subscription_receipts.go Dedupe preconfirmed receipts before adaptation                      rpc/v9/subscription_receipts.go Adds shouldSend callback to receiptsOf Marks receipts as sent before adaptation Skips already-sent preconfirmed receipts Uses status variable for canonical blocks +12/-7    subscription_transactions.go Dedupe preconfirmed transactions before adaptation              rpc/v9/subscription_transactions.go Adds shouldSend callback to transactionsOf Marks transactions as sent before adaptation Uses txn.Hash() for deduplication Passes nil callback for canonical blocks +11/-7

---

[github-actions]

Claude finished @danielntmd's task in 3m 33s —— View job

---

Review: perf(rpc): dedupe preconfirmed subscriptions before adaptation

Verdict: clean. The refactor moves the deduper.MarkSent check ahead of AdaptReceiptWithBlockInfo/AdaptTransaction by threading a shouldSend func(core.Transaction) bool predicate into receiptsOf/transactionsOf, so already-emitted preconfirmed entries are skipped before the (relatively expensive) adaptation work rather than after. This is applied consistently across rpc/v9 and rpc/v10 for both receipts and transactions.

Checked:

• Hash equivalence: switching from the adapted response's Hash field to core.Transaction.Hash() (called on the raw block transaction before adaptation) yields the same *felt.Felt value, so dedup keys are unchanged.
• Filter ordering: sender-address filtering still happens before the shouldSend check inside the iterator, matching the original semantics — no change in which transactions get marked/consumed by the deduper.
• onNewHead call sites: correctly pass nil for shouldSend, preserving the existing comment that canonical blocks bypass the deduper entirely.
• Deduper itself (rpc/rpccore/preconfirmed_deduper.go) is untouched and remains single-goroutine-only, consistent with receiptsSubscriberState/transactionsSubscriberState usage.
• No test files call receiptsOf/transactionsOf directly, so the new parameter doesn't break existing coverage; go build and go vet pass on rpc/v9 and rpc/v10.

No inline comments — didn't find correctness, concurrency, or resource issues to flag.

[github-actions]

PR Reviewer Guide 🔍

(Review updated until commit 00fc30a)

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🏅 Score: 90

🧪 No relevant tests

🔒 No security concerns identified

⚡ No major issues detected

[github-actions]

Claude finished @danielntmd's task in 3m 44s —— View job

---

Review: perf(rpc): dedupe preconfirmed subscriptions before adaptation

Verdict: clean. No blocker/important/nit findings — no inline comments posted.

Summary of what changed: receiptsOf/transactionsOf (v9 + v10) now take an optional shouldSend func(core.Transaction) bool and run it before AdaptReceiptWithBlockInfo/AdaptTransaction, so already-emitted preconfirmed entries are skipped before the adaptation work instead of after. onNewHead passes nil (canonical blocks still bypass the deduper); onPreConfirmed passes a closure wrapping deduper.MarkSent.

Checked and confirmed correct:

• Dedup key equivalence: the closure now keys on core.Transaction.Hash() (the raw transaction, pre-adaptation) instead of the previously-adapted response's Hash field. Every concrete core.Transaction implementation (InvokeTransaction, DeclareTransaction, DeployTransaction, DeployAccountTransaction, L1HandlerTransaction) exposes Hash() *felt.Felt, and this is the same hash the adapters derive their Hash field from — no change in dedup semantics.
• Ordering preserved: filterTxBySender still runs before shouldSend, and shouldSend still runs before adaptation, in both receiptsOf and transactionsOf, for both v9 and v10 — consistent across all 4 files.
• onNewHead call sites correctly pass nil, preserving the "canonical blocks bypass the deduper" comment/behavior.
• Deduper (rpc/rpccore/preconfirmed_deduper.go) is untouched; receiptsSubscriberState/transactionsSubscriberState remain single-goroutine-owned per their existing doc comments, so no new concurrency concerns from calling MarkSent earlier in the pipeline.
• go build ./rpc/... and go vet ./rpc/v9/... ./rpc/v10/... both pass. Was not able to execute go test in this sandbox (command required approval I don't have in this context); no test files call receiptsOf/transactionsOf directly, so existing coverage is exercised only through the subscriber handlers, which are unchanged in interface.

[github-actions]

Persistent review updated to latest commit 00fc30a

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.19%. Comparing base (121146c) to head (00fc30a).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4036      +/-   ##
==========================================
- Coverage   79.31%   79.19%   -0.12%     
==========================================
  Files         464      464              
  Lines       35793    35793              
==========================================
- Hits        28389    28348      -41     
- Misses       7395     7436      +41     
  Partials        9        9              

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[brbrr]

LGTM!