[nigrosimone]

@ValoChet's idea from #1353: only a string reads the thread-local pool, binary data never does, and pool being set replaces the allocated flag.

A JS handler that passes Buffers (cork, writeStatus, two writeHeader, end), user-space instructions per request on arm64: 9435 to 8990 (-4.7%), the same build twice within 0.1%. With strings it does not move (10580 to 10571).

[github-actions]

Benchmark: d9d5f78 against 9f90f38

Scenario	Base req/s	Head req/s	Head / base	Rounds	Noise	Busy	CPU per request
http/hello-world	116650	115151	0.989x (-1.1%)	0.92 to 0.99	±13.5%	100% / 100%	8.56 / 8.68 us (1.011x)
http/headers	103793	101238	0.995x (-0.5%)	0.97 to 1.02	±5.8%	100% / 100%	9.63 / 9.87 us (1.005x)
http/json-post	84161	83296	0.998x (-0.2%)	0.98 to 1.00	±3.5%	100% / 100%	11.87 / 11.99 us (1.001x)
http/cached	127992	127721	0.998x (-0.2%)	0.99 to 1.00	±0.8%	100% / 100%	7.81 / 7.83 us (1.002x)
ws/echo-20b	124938	120858	0.992x (-0.8%)	0.94 to 0.99	±5.7%	100% / 100%	8.00 / 8.27 us (1.008x)
ws/echo-4kb	106096	105841	0.996x (-0.4%)	0.99 to 1.00	±7.8%	100% / 100%	9.42 / 9.44 us (1.004x)

4 rounds of 3s per scenario, each round loading base, head and a second process of each one, from a copy of its build, after the other, in an order that changes every round. "Head / base" is the median of the per-round ratios of req/s and "Rounds" their range. "Noise" is how far base/base and head/head, the same code on both sides, got from 1 in this same run: that is what the machine did, so a row is marked only when the median moved further than that, at least 2%, and every round moved the same way: 👀 slower, 🏆 faster. "Busy" is the cpu time of the server process over the round and "CPU per request" that time divided by the requests it answered, with the head/base ratio judged against its own same-code band and bold when it cleared it. When busy is well under 100% the load generator set the pace, the req/s are its and the cpu per request is the column to read. Only the ratios are comparable across runs, the absolute figures depend on the runner.

Node v26.11.1, AMD EPYC 7763 64-Core Processor, 4 cores (server on cpu 0, load on 2,3), http_load_test and load_test with 100 connections.