Loops, labels and do-while

break leaves the nearest loop, continue skips to the next round, and a label (outer:) lets either one target an outer loop. do { ... } while (cond) always runs its body at least once.

A mental model

Walking through the rooms of a building floor by floor: continue is skipping a room, break is leaving the floor, and break outer is leaving the whole building.

How it works

for when the number of rounds is known, while when it depends on a condition, do-while when the body must run first (menus, retry until valid input), and for-each for collections and arrays.

break ends the innermost loop (or a switch case). continue jumps to the next iteration (running the for update step first).

outer: for (...) { for (...) { if (found) break outer; } } stops both loops. Labels are rare, and often a sign that the code should be a separate method with return.

Infinite loops written on purpose: while (true) { ... if (done) break; } or for (;;); a thread loop should also check the interrupt flag.

Going deeper

Off-by-one errors live here: i < n versus i <= n, starting at 0 or 1, and updating the loop variable inside the body. Reading a loop out loud with a tiny example ("n = 3: i = 0, 1, 2") catches most of them.

A for loop variable declared in the header is scoped to the loop; a variable used after the loop must be declared before it.

Extracting the loop body into a method with early return is usually more readable than labelled breaks.

Under the hood

A for loop compiles to a conditional jump back to the start; a labelled break is a plain goto to the end of the labelled statement in bytecode. The JIT unrolls and optimises simple counting loops heavily.

Common mistakes

• Off-by-one in the loop bound.
• Modifying the collection while looping.
• An infinite loop from a missing update.
• Overusing labels.