You might guess that something changed from "revisiting" in the headline, but there's nothing particularly new in the article. It describes the same Zig feature

You might guess that something changed from "revisiting" in the headline, but there's nothing particularly new in the article. It describes the same Zig features we've been hearing about all along and they're probably why you've heard about Zig in the first place. To talk about one of them, it makes sense that Zig ignores compile-time branches that aren't chosen because otherwise they wouldn't be a replacement for ifdefs. However, I wonder if some kind of abstract interpretation mode that checks both sides of a branch would make sense?