Closed kkebo closed 4 months ago
In the case of html5ever,
https://github.com/servo/html5ever/blob/master/html5ever/benches/html5ever.rs
In the case of html5gum,
https://github.com/untitaker/html5gum/blob/main/benches/patterns.rs
In the case of html-parse,
In the case of html-tokenizer,
https://github.com/greim/html-tokenizer/blob/master/benchmark/benchmark.js
I think that I don't need a large HTML for the benchmark test of Tokenizer
because other libraries also use small pieces of HTML for that purpose. Moreover, I can do the benchmark test for the whole browser engine not only for Tokenizer
when I implement the main module.
Anyway, I should finish #39 first.
I'm currently measuring performance by using the following small HTML, but I should replace it with a bigger HTML for more accurate benchmarks.
https://github.com/kkk669/zyphy/blob/b95c9d11fe4641dba3257c9e75ed412fb492049e/Benchmarks/Benchmarks/MyBenchmark/MyBenchmark.swift#L15-L26
And I would like to compare Zyphy's tokenizer with other high-performance HTML tokenizers if possible.