Alter perf benchmark fixture

[necolas]

Alters the large perf benchmark fixture so it can be used to test #20

[github-actions]

workflow: benchmarks/size

Comparison of minified (terser) and compressed (brotli) size results, measured in bytes. Smaller is better.

[email protected] benchmark:compare
node benchmark/compare.js /tmp/tmp.sX8c1Aa0tb /tmp/tmp.JqGiZ8NWZt

Results	Base	Patch	Ratio
styleq.js
· compressed	600	600	1.00
· minified	1,280	1,280	1.00

[github-actions]

workflow: benchmarks/perf

Comparison of performance test results, measured in operations per second. Larger is better.

[email protected] benchmark:compare
node benchmark/compare.js /tmp/tmp.VCLLbxkKf4 /tmp/tmp.QrbGrww93D

Results	Base	Patch	Ratio
styleq
· small object	15,388,959	16,200,202	1.05	!!
· small object (cache miss)	1,376,222	1,321,802	0.96	-
· small object (cache disabled)	10,400,554	10,454,178	1.01	+
· large object	9,447,488	9,245,801	0.98	-
· large object (cache miss)	392,364	340,580	0.87	!!
· large object (cache disabled)	1,451,173	1,187,587	0.82	!!
· small merge	8,971,548	8,984,454	1.00	+
· small merge (cache miss)	964,092	961,919	1.00	-
· small merge (cache disabled)	7,276,018	7,045,190	0.97	-
· large merge	1,664,042	1,539,610	0.93	!!
· large merge (cache disabled)	514,071	422,232	0.82	!!
· large merge (transform)	1,366,120	1,284,269	0.94	!!
· small inline style	14,272,808	14,121,186	0.99	-
· large inline style	1,243,196	1,261,320	1.01	+
· merged inline style	1,170,919	1,175,563	1.00	+
· merged inline style (mix disabled)	1,323,759	1,323,535	1.00	-

[necolas]

The "regressions" are expected because the fixture was changed