@prologic That is strange.. i wonder if there is another change that is causing it. Benchmarks are thinking the other way :|\n\n\nBenchmarkAll/retwt-16 1 4940172200 ns/op 587319376 B/op 2587159 allocs/op\nBenchmarkAll/lextwt-16 1 775764020 ns/op 9223088 B/op 197557 allocs/op\nBenchmarkParse/retwt-16 1 591158277 ns/op 67539096 B/op 230841 allocs/op\nBenchmarkParse/lextwt-16 1 716961837 ns/op 5450448 B/op 130290 allocs/op\nBenchmarkOutput/retwt-html-16 1 8358103017 ns/op 918709168 B/op 4692292 allocs/op\nBenchmarkOutput/lextwt-html-16 1 822033267 ns/op 14280112 B/op 261795 allocs/op\nBenchmarkOutput/retwt-markdown-16 1 8114225415 ns/op 929928384 B/op 4693004 allocs/op\nBenchmarkOutput/lextwt-markdown-16 1 806554306 ns/op 14332608 B/op 270905 allocs/op\nBenchmarkOutput/retwt-text-16 1 8098215897 ns/op 923862192 B/op 4684739 allocs/op\nBenchmarkOutput/lextwt-text-16 1 745064673 ns/op 12689784 B/op 252476 allocs/op\nBenchmarkOutput/retwt-literal-16 1 4068799822 ns/op 409517880 B/op 2372471 allocs/op\nBenchmarkOutput/lextwt-literal-16 1 754793627 ns/op 9834520 B/op 214931 allocs/op\n