The x-Done Loop

This is a construct that is very useful to avoid either a redundant check or a goto. This...
for(a;b;c) {
 expr1;
 if(d) {
  expr4;
  goto done;
 }
 expr2;
}
 expr3;
done:
...becomes...
for(a;b;c) {
 expr1;
 if(d) {
  expr4;
  break;
 }
 expr2;
} done {
 expr3;
}

where the intent is for statements in a done clause immediately following a loop to be executed if and only if the loop was exited via its termination condition being reached. This also generalises for while() and do-while, here using some common examples.

while(*j) {
 if(*j++ == k) {
  u = v + k;
  break;
 }
} done
 u = v - k;

do {
 if(f(a[i]))
  break;
} while(a[i++]) done {
 g();
}

Discuss.

>>54,55
How so? I'd say it's far cleaner to write a separate function like 'searchuranus' then branch depending on its return value, rather than using what you've proposed. Performing an extra comparison on a null pointer really isn't a huge issue, especially with high level languages where clarity and the use of idiomatic forms is far more important than unnoticeable performance gains. If you run a large program through a profiler and find that it performs poorly, chances are it's because an inefficient data structure was used, rather than using two comparisons where one would have been sufficient.