Speed up parser token scanning

[dduan]

Skip the unrolled space/tab scan unless the next byte is whitespace, cache value/bare-key tables during tokenization, and advance array string elements without redundant kind checks.

[github-actions]

Comparing results between 'main' and 'pull_request'

Host 'runnervmh13bl' with 4 'x86_64' processors with 15 GB memory, running:
#18~24.04.1-Ubuntu SMP Sat Jun 28 04:46:03 UTC 2025

TOMLDecoderBenchmarks

Decode toml.io example metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (μs) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	69	70	70	71	80	96	144	7125
pull_request	69	70	71	72	77	97	143	6993
Δ	0	0	1	1	-3	1	-1	-132
Improvement %	0	0	-1	-1	4	-1	1	-132

Retains: results within specified thresholds, fold down for details.

Retains *	p0	p25	p50	p75	p90	p99	p100	Samples
main	625	625	625	626	626	626	626	7125
pull_request	624	624	624	624	624	624	624	6993
Δ	-1	-1	-1	-2	-2	-2	-2	-132
Improvement %	0	0	0	0	0	0	0	-132

Parse toml.io example metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (ns) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	6772	6955	7023	7107	7227	12415	30186	10000
pull_request	6391	6523	6563	6615	6695	9631	30397	10000
Δ	-381	-432	-460	-492	-532	-2784	211	0
Improvement %	6	6	7	7	7	22	-1	0

Retains: results within specified thresholds, fold down for details.

Retains *	p0	p25	p50	p75	p90	p99	p100	Samples
main	8	9	9	9	9	9	9	10000
pull_request	8	9	9	9	9	9	9	10000
Δ	0	0	0	0	0	0	0	0
Improvement %	0	0	0	0	0	0	0	0

decode canada.toml metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (ms) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	271	271	273	273	275	275	275	4
pull_request	268	268	269	269	271	271	271	4
Δ	-3	-3	-4	-4	-4	-4	-4	0
Improvement %	1	1	1	1	1	1	1	0

Retains: results within specified thresholds, fold down for details.

Retains (K) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	1674	1674	1674	1674	1674	1674	1674	4
pull_request	1674	1674	1674	1674	1674	1674	1674	4
Δ	0	0	0	0	0	0	0	0
Improvement %	0	0	0	0	0	0	0	0

decode twitter.toml metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (μs) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	9652	9691	9716	9740	9789	10215	10268	102
pull_request	9334	9388	9404	9437	9478	9675	9955	106
Δ	-318	-303	-312	-303	-311	-540	-313	4
Improvement %	3	3	3	3	3	5	3	4

Retains: results within specified thresholds, fold down for details.

Retains (K) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	66	66	66	66	66	66	66	102
pull_request	66	66	66	66	66	66	66	106
Δ	0	0	0	0	0	0	0	4
Improvement %	0	0	0	0	0	0	0	4

parse canada.toml metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (ms) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	22	23	23	24	24	25	25	43
pull_request	17	18	18	19	19	21	21	55
Δ	-5	-5	-5	-5	-5	-4	-4	12
Improvement %	23	22	22	21	21	16	16	12

Retains: results within specified thresholds, fold down for details.

Retains *	p0	p25	p50	p75	p90	p99	p100	Samples
main	132	133	133	133	133	133	133	43
pull_request	132	133	133	133	133	133	133	55
Δ	0	0	0	0	0	0	0	12
Improvement %	0	0	0	0	0	0	0	12

parse twitter.toml metrics

Time (wall clock): results within specified thresholds, fold down for details.

Time (wall clock) (μs) *	p0	p25	p50	p75	p90	p99	p100	Samples
main	3416	3437	3443	3455	3469	3881	4038	286
pull_request	3070	3099	3111	3123	3146	3414	3573	315
Δ	-346	-338	-332	-332	-323	-467	-465	29
Improvement %	10	10	10	10	9	12	12	29

Retains: results within specified thresholds, fold down for details.

Retains *	p0	p25	p50	p75	p90	p99	p100	Samples
main	1292	1293	1293	1293	1293	1293	1293	286
pull_request	1292	1293	1293	1293	1293	1293	1293	315
Δ	0	0	0	0	0	0	0	29
Improvement %	0	0	0	0	0	0	0	29