Hacker News Reader

Comment by owlstuffing | Hacker News Reader

Overall, this is one of the better C killers I’ve seen. But keeping macros while ditching conditional compilation? That seems completely backward to me. Oh well.

View on HN

What kind of conditional compilation are you missing?

For adapting code to different versions of libraries for one thing:

    #if defined(__SunOS)
        presult = getprotobyname_r(proto,&result,tmp,sizeof(tmp));
        if (presult == NULL)
          return luaL_error(L,"protocol: %s",strerror(errno));
    #elif defined(__linux__)
        if (getprotobyname_r(proto,&result,tmp,sizeof(tmp),&presult) != 0)
          return luaL_error(L,"protocol: %s",strerror(errno));
    #else
        presult = getprotobyname(proto);
        if (presult == NULL)
          return luaL_error(L,"protocol: %s",strerror(errno));
        result = *presult;
    #endif

The sometimes annoyingly small differences between platforms.

Oh, I think you missed something then:

There is both `$if` and `$switch` compile time statements for this: https://c3-lang.org/generic-programming/compiletime/#if-and-...

At the top level and `@if` attribute is used to achieve the same thing: https://c3-lang.org/language-common/attributes/#if

Ah. The top-level lang description claims “No preprocessor”, but my definition of that word doesn’t appear to be the same as yours :/

The difference here is that a preprocessor runs before parsing and semantic analysis. In C3 compile time if runs in the analysis step, so after parsing.

So the macros and compile time execution occurs after parsing in C3, but in C everything happens at lexing, before the code is parsed.

That strategy will backfire when the grammar changes in a later release. A pre-parse step is necessary for code that targets different compiler releases.

Now I'm confused, what do you mean? What grammar changes?

Imagine v2.0 introduces a new feature that requires a parser change—one that v1.0 wouldn't be able to parse.

  $if $defined(C3_V2_PLUS):  
      // new feature
      fn f = \ -> foo(); // this won't parse in v1.0
  $else
      Callback f = ...;
  $endif

This is also an issue if a future version of C3 introduces language level support, allowing newer compilers to compile code as if it were written for an earlier version. While this approach works well when teams standardize on a specific version’s feature set, they may still want to take advantage of performance improvements in the latest compiler.

That said, this is a niche case and not something I’d consider a dealbreaker.

Ah, that is indeed true. However, the plan is to have the language fixed at 1.0, only updating stdlib and tooling after that.

Well, if your language gains traction that plan may not succeed. For your sake, I hope that's the case :)

I know it’s hard.

That one was missing a comma: "I know, it's hard". I'll do my best.

You don't need an ifdef.

You just need a CLI option like java's --source, where you specify the source compatibility with different language versions.

You do, that's the point here. You need ifdefs in this example to selectively compile the same code with different compilers, or different language levels with the same compiler. In either case C3 will fail while parsing a newer feature with an older compiler or lang level.

This exact situation is there in Java, using compiler args, after lambda functions were introduced. Older compilers would not be able to handle it, and the newer compilers would not break backward compatibility.

And. . . Java does not have conditional compilation.

Reply on news.ycombinator.com