04-Mar-2013 22:58, Namespace пишет:
Yes, I know. But I do not think I have a chance to get even closer to it. At least I don't know, how.
One thing that all of these lexers do (dmd's orignal and Dscanner) is get tokens as you need them - lazily. You would always be slower if taking time to append them in some array (no matter how). The idiomatic way as I said is to use ranges.
I'm on a AMD Athlon 64 X2 Dual Core Processor 5400+.
Mm that's not that far away from mine. But Brain has got something even more powerful :)
You can easily download and test Dscanner on the your CPU so that you get numbers that are actually comparable.
-- Dmitry Olshansky
