Re: do { quit; } else { }

Em 4/4/2025 5:18 PM, bart escreveu:

On 04/04/2025 20:23, Thiago Adams wrote:

   What do you think of this control block?
>
   do
   {
      FILE f = fopen("file.txt", "r");
>
      if (f == NULL) quit; /*goes to else part*/
>
      /*success here*/
      for (int i =0; i < 10; i++){
          ...
          if (error) quit;
      }
>
   }
   else
   {
      /*some error*/
   }
>

 As a new language feature?
 I think 'do {}' would be confused 'do {} while' (a parser - or a human - won't know this isn't a loop until the 'else' is seen. And then there's a possibility that that 'else' belongs to an 'if' but an omission or inclusion of a braces has screwed things up.
 

Thanks for the comments.
I agree with this part..
"a possibility that that 'else' belongs to an 'if'"

Perhaps a new keyword ('try' maybe) is better.

  try { quit; } else { } ?
  try { quit; } except { } ?

You'd also need to specify how nested do-else blocks will work. Or will 'quit' (so you have a new keyword anyway) only work at this level?
>

do {
   do {
       quit; //goes to 1
   }
   else{
        //1
        quit; //goes to 2
   }
}
else{
    //2
}

What about chained versions: do {} else do {} else ...; I guess 'quit' will just go to the next else block rather than quit the whole chain?
 Note that this can be emulated in C now using:
     if
    {
      goto quit001;
    }
    else
    {quit001:
    }
 Perhaps what C needs is a way to do that jump without an explicit goto:
     if
    {
      quit001;
    }
    else
    {quit001:
    }
 You can require that such jumps can only go forwards.

I'm already doing this with macros. The problem is defining the right keywords. In the past, I used keywords like try, catch, and throw, but people didn’t like the feature. throw is indeed a little confusing since it’s not actually throwing anything.
So, I was trying to disassociate it from throw and see if that changes opinions about it.