// also sweep pages (e.g., for a large allocation), it can pass a
// non-zero callerSweepPages to leave that many pages unswept.
//
// deductSweepCredit makes a worst-case assumption that all spanBytes
// bytes of the ultimately allocated span will be available for object
// allocation.
//
// deductSweepCredit is the core of the "proportional sweep" system.

						//   never gets executed.
						// - in worst case, make both ends (of the intersection) a big number,i.e.,
						//   M intersected with N instead of M with 1 (or 1 with N)
						// - in random case, each pod failed by a random plugin, and also the moveEvent
						//   is randomized.
						if tt.name == "baseline" {
							podInfo = q.newQueuedPodInfo(p)
						} else if tt.name == "worst" {
							// Each pod failed by all plugins.

That's the entirety of long division: scale the inputs, and then loop over
each output position, guessing, checking, and correcting the next output digit.

For a 2n-digit number divided by an n-digit number (the worst size-n case for
division complexity), this algorithm uses n+1 iterations, each of which must do
at least the 1-by-n-digit multiplication q̂·v. That's O(n) iterations of
O(n) time each, so O(n²) time overall.

   * negation is cheaper to precompute than the matcher itself; it tries to build small hash tables
   * for matchers that only match a few characters, and so on. In the worst-case scenario, it
   * constructs an eight-kilobyte bit array and queries that. In many situations this produces a
   * matcher which is faster to query than the original.
   */
  @GwtIncompatible // SmallCharMatcher

// Complexity depends on the type of operations that were performed
// since the last checkpoint; each Set* operation creates an undo
// pass which Undo has to revert with a worst-case complexity of O(n).
func (po *poset) Undo() {
	if len(po.undo) == 0 {
		panic("empty undo stack")
	}
	if debugPoset {
		defer po.CheckIntegrity()
	}

	for len(po.undo) > 0 {

	userArenaChunkPages = userArenaChunkBytes / pageSize

	// userArenaChunkMaxAllocBytes is the maximum size of an object that can
	// be allocated from an arena. This number is chosen to cap worst-case
	// fragmentation of user arenas to 25%. Larger allocations are redirected
	// to the heap.
	userArenaChunkMaxAllocBytes = userArenaChunkBytes / 4
)

func init() {

   * negation is cheaper to precompute than the matcher itself; it tries to build small hash tables
   * for matchers that only match a few characters, and so on. In the worst-case scenario, it
   * constructs an eight-kilobyte bit array and queries that. In many situations this produces a
   * matcher which is faster to query than the original.
   */
  @GwtIncompatible // SmallCharMatcher

	// for File Header, Auxiliary Header, and Section Headers.
	// May waste some.
	XCOFFHDRRESERVE = FILHSZ_64 + AOUTHSZ_EXEC64 + SCNHSZ_64*23

	// base on dump -o, then rounded from 32B to 64B to
	// match worst case elf text section alignment on ppc64.
	XCOFFSECTALIGN int64 = 64

	// XCOFF binaries should normally have all its sections position-independent.

	i scavengeIndex
}

func NewScavengeIndex(min, max ChunkIdx) *ScavengeIndex {
	s := new(ScavengeIndex)
	// This is a bit lazy but we easily guarantee we'll be able
	// to reference all the relevant chunks. The worst-case
	// memory usage here is 512 MiB, but tests generally use
	// small offsets from BaseChunkIdx, which results in ~100s
	// of KiB in memory use.
	//
	// This may still be worth making better, at least by sharing

		// to determine whether a given feature is supported,
		// we may receive a positive response even if the feature is not supported.
		//
		// In this very rare scenario, the worst case will be that this
		// request will wait for 3 seconds before it fails.
		if etcdfeature.DefaultFeatureSupportChecker.Supports(storage.RequestWatchProgress) && c.watchCache.notFresh(requestedWatchRV) {