Benchmarking

[JoviDeCroock]

Currently the benchmarks seem geared towards testing the JS-execution and less of the CSSOM impact on the DOM, what would you think about moving to a more similar benchmark to styled-components. The SC-benchmark is considered to be the most solid one at drawing out the actual perf impact.

The currently used one will only draw out the JS-impact of the styling while that concern is less prevalent in a DOM-world.

[lukejacksonn]

Ahh yes, I think @cristianbote mentioned this to me once before and it makes sense. It would be nice to know the end-to-end impact so to speak. I will try pulling in their test suite and then adding some Twind specific tests. I presume that is the approach you are suggesting we take?

[JoviDeCroock]

Yes, it's widely seen as one of the accurate ones since it's doing it in an actual browser and the bench itself is performing operations you'll see return in applications.

[sastan]

Sounds good.

[lukejacksonn]

Managed to get this setup today, here are some preliminary results: