Pre-size FormData serialization buffer

[Brooooooklyn]

Summary

• calculate the exact multipart FormData serialization size with overflow checks
• pre-size the output vector to avoid repeated reallocations and the final release copy
• add serializer coverage for escaped names and explicit/default file content types
• add a focused FormData serialization benchmark

Benchmark

Warm median results from 5 repetitions:

Payload	Before	After	Improvement
1 KiB	0.224 us	0.0865 us	2.59x
1 MiB	0.0387 ms	0.0127 ms	3.06x
16 MiB	0.704 ms	0.214 ms	3.28x

The 16 MiB case improves from 22.2 GiB/s to 72.9 GiB/s.

Tests

• bazel test //src/workerd/api/tests:form-data-test@
• python3 tools/cross/format.py --check git

[codspeed-hq]

Merging this PR will improve performance by 9.89%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
✅ 71 untouched benchmarks
🆕 3 new benchmarks
⏩ 129 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	bodyWithHeaders[Response]	34.3 µs	31.2 µs	+9.89%
🆕	singleFile[FormDataSerialization][1024]	N/A	5.9 µs	N/A
🆕	singleFile[FormDataSerialization][1048576]	N/A	1 ms	N/A
🆕	singleFile[FormDataSerialization][16777216]	N/A	15.9 ms	N/A

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing Brooooooklyn:optimize-formdata-serialization (c1d2a0b) with main (4654087)}

Footnotes

1. 129 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩