Closed PonteIneptique closed 4 years ago
Gain is very low (fraction of a second over 194k words...) but does it have a big cost to implement, I don't think so. What do you think @emanjavacas ?
Hey. Thanks for looking into this. However, I'd be very reluctant to change stuff that works and is not problematic. As you see the time penalty is negligible, and we risk breaking things, making debugging and mantainence harder, etc...
Yeah. Unlike #70, this might not be necessary :)
For the same reason as #70, I had a look at the weight of encoding words even if you do not use wemb.
Gain is minimal, but code weight as well so... What do you think ?
Set-up
Current situation
Proposed removal of wemb when not used