Using goto to jump out of nested loops
suggest changeJumping out of nested loops would usually require use of a boolean variable with a check for this variable in the loops. Supposing we are iterating over i
and j
, it could look like this
size_t i,j;
for (i = 0; i < myValue && !breakout_condition; ++i) {
for (j = 0; j < mySecondValue && !breakout_condition; ++j) {
... /* Do something, maybe modifying breakout_condition */
/* When breakout_condition == true the loops end */
}
}
But the C language offers the goto
clause, which can be useful in this case. By using it with a label declared after the loops, we can easily break out of the loops.
size_t i,j;
for (i = 0; i < myValue; ++i) {
for (j = 0; j < mySecondValue; ++j) {
...
if(breakout_condition)
goto final;
}
}
final:
However, often when this need comes up a return
could be better used instead. This construct is also considered “unstructured” in structural programming theory.
Another situation where goto
might be useful is for jumping to an error-handler:
ptr = malloc(N * x);
if(!ptr)
goto out_of_memory;
/* normal processing */
free(ptr);
return SUCCESS;
out_of_memory:
free(ptr); /* harmless, and necessary if we have further errors */
return FAILURE;
Use of goto
keeps error flow separate from normal program control flow. It is however also considered “unstructured” in the technical sense.
Found a mistake? Have a question or improvement idea?
Let me know.
Table Of Contents