Cex.C – Comprehensively EXtended C Language | Hacker News Reader

Cex.C – Comprehensively EXtended C Language

github.com

26 pointslifthrasiir9 comments

Noticed that it does have a website later: https://cex-c.org/

Name asides (meant to be pronounced sexy), I had the same approach in mind for a long time: single-file distribution, keep the surface syntax, use $ to extend the core language, transpiling for the win. Glad that I'm not alone.

> LEGAL NOTICE: Any intentional mispronunciation of Cex.C or cexy$ (build system), officially pronounced /ˈtsɛk.si/ ("tsek-see"), into an incorrect form may be considered intentional tseksual harassment of the project — which identifies itself with the code gender (it/its) — and may be subject to legal action under the MIT License. /LOL/

Nice :-)

I wonder how many sense-of-humour-impaired people are going to bristle at that

Can't wait for the Tsoding review video.

Yes, I really hope someone will share him the repo.indeed

When people try to create a better C, I always remind them, that such better C already exists and it's named C++.

> When people try to create a better C, I always remind them, that such better C already exists and it's named C++.

Somehow the idea of quadrupling the number of footguns is not as appealing an idea to some C programmers as you may think it is.

When people try to create a better spoon, I always remind them, that such a better spoon already exists and it's named spork.

I thought the idea of C++ as "a better C" died quite some time ago? It's quite clearly its own thing now.

In general I think the appeal of "a better C" is limited. The value of C is that it's simple limited thing and there's a compiler for every platform. Your "Better-C" won't have that. What seems like the most universally acknowledged problem with C (unexpected cases of UB) also largely stems from this. There's some effort in the newer standards to nail down some of this in exchange for dropping support for some of the oddball historical stuff like not assuming twos-complement or whatever. Probably the correct path to a better C, rather than adding things on top.

Separate are the efforts at creating new lean systems programming languages that enable you to do the same things as C. But there's not much point in trying to extend C itself at that point.

I think what could be a worthwhile approach is designing a source compatible language, that makes all the implicit rules defined in C explicit, for example owning/non-owning pointers or nullability. I like the concept of SPlint, although the implementation is garbage.

Reply on news.ycombinator.com