Wismill
Wismill
Ok for me, but I have no time for this.
When we do the split, I propose the following new version scheme: `U.B.M`, where: - `U` is the supported Unicode _major_ version; i.e. 15 for Unicode 15.0.0. - `B` is...
We can bump the version on master to ensure we are not using a released one, but following you example I would bump only to `0.2.5` and not to `0.3.0`,...
@adithyaov We can try. I propose #96 to facilitate the process.
There is a [PR](https://github.com/fourmolu/fourmolu/pull/95), but not yet merged.
Latest benchmark All unicode-transforms-text NFD/AllChars: OK (2.76s) 5.41 ms ± 126 μs, 7% faster than baseline NFD/Deutsch: OK (0.59s) 2.27 ms ± 66 μs, 27% faster than baseline NFD/Devanagari: OK...
@harendra-kumar I managed to obtain _nice improvements:_ Benchmark results All unicode-transforms-text NFD/AllChars: OK (1.31s) 5.13 ms ± 66 μs, 9.2 MB allocated, 815 B copied, 31 MB peak memory, 0.72x,...
I updated the benchmark results to include allocated memory & diff with `text-icu.`.
> I see the following benchmarks getting slower: > > ``` > All > unicode-transforms-text > NFD/Korean: OK (0.20s) > 13.5 ms ± 983 μs, 32% slower than baseline >...
Sure thing. I tested on a laptop equipped with AMD Ryzen 5 2500U @ 2.0GHz, on Linux.