Repeatedly walk the array comparing adjacent pairs and swapping any that are out of order. After pass k, the k largest elements are in their final positions at the end. Stop early when a full pass makes zero swaps.

Algorithm

Canonical input {5, 1, 4, 2, 8} finishes after three passes: two with swaps, then a clean pass that triggers the early exit. Final array {1, 2, 4, 5, 8}.

Basic Implementation

basic.lua
local arr = {5, 1, 4, 2, 8}
local n = #arr
local i = 1
local done = false
while i <= n - 1 and not done do
	local swapped = false
	local j = 1
	while j <= n - i do
		if arr[j] > arr[j + 1] then
			local tmp = arr[j]
			arr[j] = arr[j + 1]
			arr[j + 1] = tmp
			swapped = true
		end
		j = j + 1
	end
	if not swapped then
		done = true
	end
	i = i + 1
end
io.write("[")
for k = 1, #arr do
	if k > 1 then io.write(", ") end
	io.write(tostring(arr[k]))
end
io.write("]\n")

Complexity

  • Time: O(n^2) worst and average; O(n) best (already sorted with early exit)
  • Space: O(1)
  • Stable: yes

Implementation notes

  • Lua: explicit while loops with i, j, done, and swapped so the early-exit flow stays visible. The stdlib table.sort(arr) would hide the comparison-and-swap the lesson is teaching, and a goto continue inside a for loop would collapse the done flag into a control-flow shortcut.
  • The explicit local tmp = arr[j]; arr[j] = arr[j+1]; arr[j+1] = tmp three-line swap keeps the move visible without leaning on Lua's parallel assignment arr[j], arr[j+1] = arr[j+1], arr[j].
  • The replay distinguishes compare frames from swap frames so the moving pivot value is visible. The pass number and swapped flag appear in the trace.
adjacent-pair compare and swap Inner loop walks `j` from `1` to `n - i` comparing `arr[j]` and `arr[j + 1]`.
early exit A `swapped` flag set `false` at the start of each pass. If no swap happened, flip a `done` flag and break out of the outer loop.