fix: silence GCC 13 -Wstringop-overflow false positive in native tests

[cb1kenobi]

Summary

GCC 12/13 (e.g. 13.3.0 on Ubuntu 24.04) emit a -Wstringop-overflow false positive when compiling test/native/transaction_log_recovery_test.cc:

stl_algobase.h:437:30: warning: writing 8 bytes into a region of size 3 [-Wstringop-overflow=]
  inlined from 'void {anonymous}::LogImage::appendF64(double)' at ../test/native/transaction_log_recovery_test.cc:73:15,
  inlined from '{anonymous}::LogImage::LogImage()' at ../test/native/transaction_log_recovery_test.cc:25:12

This is the known GCC 12/13 false-positive family on std::vector<char>::insert(end, ptr, ptr+N) under inlining — the optimizer sizes the copy against the pre-grow capacity. The code is correct; only the diagnostic is wrong.

Fix

Rewrite the three LogImage append helpers (appendU8/appendU32/appendF64) to grow-then-write: resize() the vector, then write directly into bytes.data() + off via the existing rocksdb_js::writeUint8/writeUint32BE/writeDoubleBE helpers. Same bytes produced, but no range-insert for the optimizer to mis-model — more robust than a reserve() workaround, which would only mask this one inline chain.

transaction_log_validation_test.cc has a byte-for-byte identical helper. It happens not to warn today (its constructor takes runtime arguments, which blocks the const-propagation chain), but it gets the same rewrite so the sibling helpers stay consistent and out of this diagnostic family.

Verification

Fresh Ubuntu 24.04 container (podman), g++ 13.3.0-6ubuntu2~24.04.1, Node 24, full node scripts/native-test/run.mjs builds:

Build	stringop-overflow	Native tests
Pristine main (repro)	1 — exact warning above	104 ran, 103 passed, 1 skipped*
With this fix	0	104 ran, 103 passed, 1 skipped*

* FileLock.SharedHardFailsOnPermissionDenied skips when running as root.

Also verified on macOS (clang): both TUs compile clean, 104 ran, 101 passed + 3 platform-conditional madvise skips.

Pre-existing on main (confirmed at 3ea9a0f), unrelated to #780.

🤖 Generated with Claude Code

[gemini-code-assist]

Code Review

This pull request refactors the vector insertion logic in transaction_log_recovery_test.cc and transaction_log_validation_test.cc to use a grow-then-write pattern (resize followed by direct write) instead of insert. This change is accompanied by comments explaining that it avoids a GCC 12/13 -Wstringop-overflow false positive. There are no review comments to address, and I have no additional feedback to provide.

[github-actions]

📊 Benchmark Results

get-sync.bench.ts

getSync() > random keys - small key size (100 records)

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 lmdb	1	23.57K ops/sec	42.42	40.87	585.568	0.110	117,875
🥈 rocksdb	2	10.13K ops/sec	98.76	94.33	31,526.357	1.24	50,627

getSync() > sequential keys - small key size (100 records)

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 lmdb	1	28.43K ops/sec	35.17	33.68	550.505	0.108	142,149
🥈 rocksdb	2	10.66K ops/sec	93.79	88.97	2,470.215	0.107	53,313

ranges.bench.ts

getRange() > small range (100 records, 50 range)

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 lmdb	1	26.17K ops/sec	38.21	35.57	1,852.963	0.286	130,845
🥈 rocksdb	2	16.14K ops/sec	61.95	52.36	1,082.147	0.126	80,706

realistic-load.bench.ts

Realistic write load with workers > write variable records with transaction log

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 rocksdb	1	369.06 ops/sec	2,709.565	117.823	48,164.259	14.37	750
🥈 lmdb	2	26.42 ops/sec	37,856.778	390.856	1,178,721.385	135.244	64.00

transaction-log.bench.ts

Transaction log > read 100 iterators while write log with 100 byte records

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 rocksdb	1	38.80K ops/sec	25.77	12.06	13,751.32	0.579	194,011
🥈 lmdb	2	442.88 ops/sec	2,257.934	260.651	13,631.161	1.31	2,215

Transaction log > read one entry from random position from log with 1000 100 byte records

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 rocksdb	1	770.48K ops/sec	1.30	1.13	502.688	0.067	3,852,388
🥈 lmdb	2	432.15K ops/sec	2.31	1.10	2,971.676	0.327	2,160,760

worker-put-sync.bench.ts

putSync() > random keys - small key size (100 records, 10 workers)

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 rocksdb	1	827.98 ops/sec	1,207.762	1,053.281	2,104.34	0.347	1,656
🥈 lmdb	2	1.19 ops/sec	841,630.169	806,397.416	870,583.259	2.04	10.00

worker-transaction-log.bench.ts

Transaction log with workers > write log with 100 byte records

Implementation	Rank	Operations/sec	Mean (ms)	Min (ms)	Max (ms)	RME (%)	Samples
🥇 rocksdb	1	22.34K ops/sec	44.76	29.87	521.777	0.572	44,684
🥈 lmdb	2	823.88 ops/sec	1,213.763	308.48	15,371.6	5.29	1,648

---

Results from commit 40b2704