Walter Bright wrote:
Andrei Alexandrescu wrote:
It's no secret that string et al. are not a magic recipe for writing correct Unicode code.

I'm concerned it would be slow. Most operations on strings do not need to decode the unicode characters, for example, find, startsWith, etc., do not. Decoding then doing find, startsWith, etc., will be considerably slower.

I thought you're going to say that, but fortunately it's easy to special-case certain algorithms for strings during compilation. In fact I already did - for example, Boyer-Moore searching would be very difficult to rewrite for variable-length characters, but there's no need for it. I special-cased that algorithm.

I believe this is a good strategy.

Andrei

Reply via email to