ParIterableLike.scala

/*                     __                                               *\
**     ________ ___   / /  ___     Scala API                            **
**    / __/ __// _ | / /  / _ |    (c) 2003-2013, LAMP/EPFL             **
**  __\ \/ /__/ __ |/ /__/ __ |    http://scala-lang.org/               **
** /____/\___/_/ |_/____/_/ | |                                         **
**                          |/                                          **
\*                                                                      */

package scala
package collection.parallel

import scala.language.{ higherKinds, implicitConversions }

import scala.collection.mutable.Builder
import scala.collection.mutable.ArrayBuffer
import scala.collection.IterableLike
import scala.collection.Parallel
import scala.collection.Parallelizable
import scala.collection.CustomParallelizable
import scala.collection.generic._
import scala.collection.GenIterableLike
import scala.collection.GenIterable
import scala.collection.GenTraversableOnce
import scala.collection.GenTraversable
import immutable.HashMapCombiner
import scala.reflect.ClassTag

import scala.annotation.unchecked.uncheckedVariance

import scala.collection.parallel.ParallelCollectionImplicits._


/** A template trait for parallel collections of type `ParIterable[T]`.
 *
 *  $paralleliterableinfo
 *
 *  $sideeffects
 *
 *  @tparam T    the element type of the collection
 *  @tparam Repr the type of the actual collection containing the elements
 *
 *  @define paralleliterableinfo
 *  This is a base trait for Scala parallel collections. It defines behaviour
 *  common to all parallel collections. Concrete parallel collections should
 *  inherit this trait and `ParIterable` if they want to define specific combiner
 *  factories.
 *
 *  Parallel operations are implemented with divide and conquer style algorithms that
 *  parallelize well. The basic idea is to split the collection into smaller parts until
 *  they are small enough to be operated on sequentially.
 *
 *  All of the parallel operations are implemented as tasks within this trait. Tasks rely
 *  on the concept of splitters, which extend iterators. Every parallel collection defines:
 *
 *  {{{
 *     def splitter: IterableSplitter[T]
 *  }}}
 *
 *  which returns an instance of `IterableSplitter[T]`, which is a subtype of `Splitter[T]`.
 *  Splitters have a method `remaining` to check the remaining number of elements,
 *  and method `split` which is defined by splitters. Method `split` divides the splitters
 *  iterate over into disjunct subsets:
 *
 *  {{{
 *     def split: Seq[Splitter]
 *  }}}
 *
 *  which splits the splitter into a sequence of disjunct subsplitters. This is typically a
 *  very fast operation which simply creates wrappers around the receiver collection.
 *  This can be repeated recursively.
 *
 *  Tasks are scheduled for execution through a
 *  [[scala.collection.parallel.TaskSupport]] object, which can be changed
 *  through the `tasksupport` setter of the collection.
 *
 *  Method `newCombiner` produces a new combiner. Combiners are an extension of builders.
 *  They provide a method `combine` which combines two combiners and returns a combiner
 *  containing elements of both combiners.
 *  This method can be implemented by aggressively copying all the elements into the new combiner
 *  or by lazily binding their results. It is recommended to avoid copying all of
 *  the elements for performance reasons, although that cost might be negligible depending on
 *  the use case. Standard parallel collection combiners avoid copying when merging results,
 *  relying either on a two-step lazy construction or specific data-structure properties.
 *
 *  Methods:
 *
 *  {{{
 *     def seq: Sequential
 *     def par: Repr
 *  }}}
 *
 *  produce the sequential or parallel implementation of the collection, respectively.
 *  Method `par` just returns a reference to this parallel collection.
 *  Method `seq` is efficient - it will not copy the elements. Instead,
 *  it will create a sequential version of the collection using the same underlying data structure.
 *  Note that this is not the case for sequential collections in general - they may copy the elements
 *  and produce a different underlying data structure.
 *
 *  The combination of methods `toMap`, `toSeq` or `toSet` along with `par` and `seq` is a flexible
 *  way to change between different collection types.
 *
 *  Since this trait extends the `GenIterable` trait, methods like `size` must also
 *  be implemented in concrete collections, while `iterator` forwards to `splitter` by
 *  default.
 *
 *  Each parallel collection is bound to a specific fork/join pool, on which dormant worker
 *  threads are kept. The fork/join pool contains other information such as the parallelism
 *  level, that is, the number of processors used. When a collection is created, it is assigned the
 *  default fork/join pool found in the `scala.parallel` package object.
 *
 *  Parallel collections are not necessarily ordered in terms of the `foreach`
 *  operation (see `Traversable`). Parallel sequences have a well defined order for iterators - creating
 *  an iterator and traversing the elements linearly will always yield the same order.
 *  However, bulk operations such as `foreach`, `map` or `filter` always occur in undefined orders for all
 *  parallel collections.
 *
 *  Existing parallel collection implementations provide strict parallel iterators. Strict parallel iterators are aware
 *  of the number of elements they have yet to traverse. It's also possible to provide non-strict parallel iterators,
 *  which do not know the number of elements remaining. To do this, the new collection implementation must override
 *  `isStrictSplitterCollection` to `false`. This will make some operations unavailable.
 *
 *  To create a new parallel collection, extend the `ParIterable` trait, and implement `size`, `splitter`,
 *  `newCombiner` and `seq`. Having an implicit combiner factory requires extending this trait in addition, as
 *  well as providing a companion object, as with regular collections.
 *
 *  Method `size` is implemented as a constant time operation for parallel collections, and parallel collection
 *  operations rely on this assumption.
 *
 *  @author Aleksandar Prokopec
 *  @since 2.9
 *
 *  @define sideeffects
 *  The higher-order functions passed to certain operations may contain side-effects. Since implementations
 *  of bulk operations may not be sequential, this means that side-effects may not be predictable and may
 *  produce data-races, deadlocks or invalidation of state if care is not taken. It is up to the programmer
 *  to either avoid using side-effects or to use some form of synchronization when accessing mutable data.
 *
 *  @define pbfinfo
 *  An implicit value of class `CanCombineFrom` which determines the
 *  result class `That` from the current representation type `Repr` and
 *  and the new element type `B`. This builder factory can provide a parallel
 *  builder for the resulting collection.
 *
 *  @define abortsignalling
 *  This method will use `abort` signalling capabilities. This means
 *  that splitters may send and read `abort` signals.
 *
 *  @define indexsignalling
 *  This method will use `indexFlag` signalling capabilities. This means
 *  that splitters may set and read the `indexFlag` state.
 *  @define Coll `ParIterable`
 *  @define coll parallel iterable
 */
trait ParIterableLike[+T, +Repr <: ParIterable[T], +Sequential <: Iterable[T] with IterableLike[T, Sequential]]
extends GenIterableLike[T, Repr]
   with CustomParallelizable[T, Repr]
   with Parallel
   with HasNewCombiner[T, Repr]
{
self: ParIterableLike[T, Repr, Sequential] =>

  @transient
  @volatile
  private var _tasksupport = defaultTaskSupport

  protected def initTaskSupport() {
    _tasksupport = defaultTaskSupport
  }

  /** The task support object which is responsible for scheduling and
   *  load-balancing tasks to processors.
   *
   *  @see [[scala.collection.parallel.TaskSupport]]
   */
  def tasksupport = {
    val ts = _tasksupport
    if (ts eq null) {
      _tasksupport = defaultTaskSupport
      defaultTaskSupport
    } else ts
  }

  /** Changes the task support object which is responsible for scheduling and
   *  load-balancing tasks to processors.
   *
   *  A task support object can be changed in a parallel collection after it
   *  has been created, but only during a quiescent period, i.e. while there
   *  are no concurrent invocations to parallel collection methods.
   *
   *  Here is a way to change the task support of a parallel collection:
   *
   *  {{{
   *  import scala.collection.parallel._
   *  val pc = mutable.ParArray(1, 2, 3)
   *  pc.tasksupport = new ForkJoinTaskSupport(
   *    new java.util.concurrent.ForkJoinPool(2))
   *  }}}
   *
   *  @see [[scala.collection.parallel.TaskSupport]]
   */
  def tasksupport_=(ts: TaskSupport) = _tasksupport = ts

  def seq: Sequential

  def repr: Repr = this.asInstanceOf[Repr]

  final def isTraversableAgain = true

  def hasDefiniteSize = true

  def isEmpty = size == 0

  def nonEmpty = size != 0

  def head = iterator.next()

  def headOption = if (nonEmpty) Some(head) else None

  def tail = drop(1)

  def last = {
    var lst = head
    for (x <- this.seq) lst = x
    lst
  }

  def lastOption = if (nonEmpty) Some(last) else None

  def init = take(size - 1)

  /** Creates a new parallel iterator used to traverse the elements of this parallel collection.
   *  This iterator is more specific than the iterator of the returned by `iterator`, and augmented
   *  with additional accessor and transformer methods.
   *
   *  @return          a parallel iterator
   */
  protected[parallel] def splitter: IterableSplitter[T]

  /** Creates a new split iterator used to traverse the elements of this collection.
   *
   *  By default, this method is implemented in terms of the protected `splitter` method.
   *
   *  @return         a split iterator
   */
  def iterator: Splitter[T] = splitter

  override def par: Repr = repr

  /** Denotes whether this parallel collection has strict splitters.
   *
   *  This is true in general, and specific collection instances may choose to
   *  override this method. Such collections will fail to execute methods
   *  which rely on splitters being strict, i.e. returning a correct value
   *  in the `remaining` method.
   *
   *  This method helps ensure that such failures occur on method invocations,
   *  rather than later on and in unpredictable ways.
   */
  def isStrictSplitterCollection = true

  /** The `newBuilder` operation returns a parallel builder assigned to this collection's fork/join pool.
   *  This method forwards the call to `newCombiner`.
   */
  //protected[this] def newBuilder: scala.collection.mutable.Builder[T, Repr] = newCombiner

  /** Optionally reuses an existing combiner for better performance. By default it doesn't - subclasses may override this behaviour.
   *  The provided combiner `oldc` that can potentially be reused will be either some combiner from the previous computational task, or `None` if there
   *  was no previous phase (in which case this method must return `newc`).
   *
   *  @param oldc   The combiner that is the result of the previous task, or `None` if there was no previous task.
   *  @param newc   The new, empty combiner that can be used.
   *  @return       Either `newc` or `oldc`.
   */
  protected def reuse[S, That](oldc: Option[Combiner[S, That]], newc: Combiner[S, That]): Combiner[S, That] = newc

  type SSCTask[R, Tp] = StrictSplitterCheckTask[R, Tp]

  /* helper traits - to avoid structural invocations */

  trait TaskOps[R, Tp] {
    def mapResult[R1](mapping: R => R1): ResultMapping[R, Tp, R1]
    // public method with inaccessible types in parameters
    def compose[R3, R2, Tp2](t2: SSCTask[R2, Tp2])(resCombiner: (R, R2) => R3): SeqComposite[R, R2, R3, SSCTask[R, Tp], SSCTask[R2, Tp2]]
    def parallel[R3, R2, Tp2](t2: SSCTask[R2, Tp2])(resCombiner: (R, R2) => R3): ParComposite[R, R2, R3, SSCTask[R, Tp], SSCTask[R2, Tp2]]
  }

  trait BuilderOps[Elem, To] {
    trait Otherwise[Cmb] {
      def otherwise(notbody: => Unit)(implicit t: ClassTag[Cmb]): Unit
    }

    def ifIs[Cmb](isbody: Cmb => Unit): Otherwise[Cmb]
    def isCombiner: Boolean
    def asCombiner: Combiner[Elem, To]
  }

  trait SignallingOps[PI <: DelegatedSignalling] {
    def assign(cntx: Signalling): PI
  }

  /* convenience task operations wrapper */
  protected implicit def task2ops[R, Tp](tsk: SSCTask[R, Tp]) = new TaskOps[R, Tp] {
    def mapResult[R1](mapping: R => R1): ResultMapping[R, Tp, R1] = new ResultMapping[R, Tp, R1](tsk) {
      def map(r: R): R1 = mapping(r)
    }

    def compose[R3, R2, Tp2](t2: SSCTask[R2, Tp2])(resCombiner: (R, R2) => R3) = new SeqComposite[R, R2, R3, SSCTask[R, Tp], SSCTask[R2, Tp2]](tsk, t2) {
      def combineResults(fr: R, sr: R2): R3 = resCombiner(fr, sr)
    }

    def parallel[R3, R2, Tp2](t2: SSCTask[R2, Tp2])(resCombiner: (R, R2) => R3) = new ParComposite[R, R2, R3, SSCTask[R, Tp], SSCTask[R2, Tp2]](tsk, t2) {
      def combineResults(fr: R, sr: R2): R3 = resCombiner(fr, sr)
    }
  }

  protected def wrap[R](body: => R) = new NonDivisible[R] {
    def leaf(prevr: Option[R]) = result = body
    @volatile var result: R = null.asInstanceOf[R]
  }

  /* convenience signalling operations wrapper */
  protected implicit def delegatedSignalling2ops[PI <: DelegatedSignalling](it: PI) = new SignallingOps[PI] {
    def assign(cntx: Signalling): PI = {
      it.signalDelegate = cntx
      it
    }
  }

  protected implicit def builder2ops[Elem, To](cb: Builder[Elem, To]) = new BuilderOps[Elem, To] {
    def ifIs[Cmb](isbody: Cmb => Unit) = new Otherwise[Cmb] {
      def otherwise(notbody: => Unit)(implicit t: ClassTag[Cmb]) {
        if (cb.getClass == t.runtimeClass) isbody(cb.asInstanceOf[Cmb]) else notbody
      }
    }
    def isCombiner = cb.isInstanceOf[Combiner[_, _]]
    def asCombiner = cb.asInstanceOf[Combiner[Elem, To]]
  }

  protected[this] def bf2seq[S, That](bf: CanBuildFrom[Repr, S, That]) = new CanBuildFrom[Sequential, S, That] {
    def apply(from: Sequential) = bf.apply(from.par.asInstanceOf[Repr]) // !!! we only use this on `this.seq`, and know that `this.seq.par.getClass == this.getClass`
    def apply() = bf.apply()
  }

  protected[this] def sequentially[S, That <: Parallel](b: Sequential => Parallelizable[S, That]) = b(seq).par.asInstanceOf[Repr]

  def mkString(start: String, sep: String, end: String): String = seq.mkString(start, sep, end)

  def mkString(sep: String): String = seq.mkString("", sep, "")

  def mkString: String = seq.mkString("")

  override def toString = seq.mkString(stringPrefix + "(", ", ", ")")

  def canEqual(other: Any) = true

  /** Reduces the elements of this sequence using the specified associative binary operator.
   *
   *  $undefinedorder
   *
   *  Note this method has a different signature than the `reduceLeft`
   *  and `reduceRight` methods of the trait `Traversable`.
   *  The result of reducing may only be a supertype of this parallel collection's
   *  type parameter `T`.
   *
   *  @tparam U      A type parameter for the binary operator, a supertype of `T`.
   *  @param op       A binary operator that must be associative.
   *  @return         The result of applying reduce operator `op` between all the elements if the collection is nonempty.
   *  @throws UnsupportedOperationException
   *  if this $coll is empty.
   */
  def reduce[U >: T](op: (U, U) => U): U = {
    tasksupport.executeAndWaitResult(new Reduce(op, splitter) mapResult { _.get })
  }

  /** Optionally reduces the elements of this sequence using the specified associative binary operator.
   *
   *  $undefinedorder
   *
   *  Note this method has a different signature than the `reduceLeftOption`
   *  and `reduceRightOption` methods of the trait `Traversable`.
   *  The result of reducing may only be a supertype of this parallel collection's
   *  type parameter `T`.
   *
   *  @tparam U      A type parameter for the binary operator, a supertype of `T`.
   *  @param op      A binary operator that must be associative.
   *  @return        An option value containing result of applying reduce operator `op` between all
   *                 the elements if the collection is nonempty, and `None` otherwise.
   */
  def reduceOption[U >: T](op: (U, U) => U): Option[U] = if (isEmpty) None else Some(reduce(op))

  /** Folds the elements of this sequence using the specified associative binary operator.
   *  The order in which the elements are reduced is unspecified and may be nondeterministic.
   *
   *  Note this method has a different signature than the `foldLeft`
   *  and `foldRight` methods of the trait `Traversable`.
   *  The result of folding may only be a supertype of this parallel collection's
   *  type parameter `T`.
   *
   *  @tparam U      a type parameter for the binary operator, a supertype of `T`.
   *  @param z       a neutral element for the fold operation, it may be added to the result
   *                 an arbitrary number of times, not changing the result (e.g. `Nil` for list concatenation,
   *                 0 for addition, or 1 for multiplication)
   *  @param op      a binary operator that must be associative
   *  @return        the result of applying fold operator `op` between all the elements and `z`
   */
  def fold[U >: T](z: U)(op: (U, U) => U): U = {
    tasksupport.executeAndWaitResult(new Fold(z, op, splitter))
  }

  /** Aggregates the results of applying an operator to subsequent elements.
   *
   *  This is a more general form of `fold` and `reduce`. It has similar semantics, but does
   *  not require the result to be a supertype of the element type. It traverses the elements in
   *  different partitions sequentially, using `seqop` to update the result, and then
   *  applies `combop` to results from different partitions. The implementation of this
   *  operation may operate on an arbitrary number of collection partitions, so `combop`
   *  may be invoked arbitrary number of times.
   *
   *  For example, one might want to process some elements and then produce a `Set`. In this
   *  case, `seqop` would process an element and append it to the set, while `combop`
   *  would concatenate two sets from different partitions together. The initial value
   *  `z` would be an empty set.
   *
   *  {{{
   *    pc.aggregate(Set[Int]())(_ += process(_), _ ++ _)
   *  }}}
   *
   *  Another example is calculating geometric mean from a collection of doubles
   *  (one would typically require big doubles for this).
   *
   *  @tparam S        the type of accumulated results
   *  @param z         the initial value for the accumulated result of the partition - this
   *                   will typically be the neutral element for the `seqop` operator (e.g.
   *                   `Nil` for list concatenation or `0` for summation) and may be evaluated
   *                   more than once
   *  @param seqop     an operator used to accumulate results within a partition
   *  @param combop    an associative operator used to combine results from different partitions
   */
  def aggregate[S](z: =>S)(seqop: (S, T) => S, combop: (S, S) => S): S = {
    tasksupport.executeAndWaitResult(new Aggregate(() => z, seqop, combop, splitter))
  }

  def foldLeft[S](z: S)(op: (S, T) => S): S = seq.foldLeft(z)(op)

  def foldRight[S](z: S)(op: (T, S) => S): S = seq.foldRight(z)(op)

  def reduceLeft[U >: T](op: (U, T) => U): U = seq.reduceLeft(op)

  def reduceRight[U >: T](op: (T, U) => U): U = seq.reduceRight(op)

  def reduceLeftOption[U >: T](op: (U, T) => U): Option[U] = seq.reduceLeftOption(op)

  def reduceRightOption[U >: T](op: (T, U) => U): Option[U] = seq.reduceRightOption(op)

  /** Applies a function `f` to all the elements of $coll in an undefined order.
   *
   *  @tparam U    the result type of the function applied to each element, which is always discarded
   *  @param f     function applied to each element
   */
  def foreach[U](f: T => U) = {
    tasksupport.executeAndWaitResult(new Foreach(f, splitter))
  }

  def count(p: T => Boolean): Int = {
    tasksupport.executeAndWaitResult(new Count(p, splitter))
  }

  def sum[U >: T](implicit num: Numeric[U]): U = {
    tasksupport.executeAndWaitResult(new Sum[U](num, splitter))
  }

  def product[U >: T](implicit num: Numeric[U]): U = {
    tasksupport.executeAndWaitResult(new Product[U](num, splitter))
  }

  def min[U >: T](implicit ord: Ordering[U]): T = {
    tasksupport.executeAndWaitResult(new Min(ord, splitter) mapResult { _.get }).asInstanceOf[T]
  }

  def max[U >: T](implicit ord: Ordering[U]): T = {
    tasksupport.executeAndWaitResult(new Max(ord, splitter) mapResult { _.get }).asInstanceOf[T]
  }

  def maxBy[S](f: T => S)(implicit cmp: Ordering[S]): T = {
    if (isEmpty) throw new UnsupportedOperationException("empty.maxBy")

    reduce((x, y) => if (cmp.gteq(f(x), f(y))) x else y)
  }

  def minBy[S](f: T => S)(implicit cmp: Ordering[S]): T = {
    if (isEmpty) throw new UnsupportedOperationException("empty.minBy")

    reduce((x, y) => if (cmp.lteq(f(x), f(y))) x else y)
  }

  def map[S, That](f: T => S)(implicit bf: CanBuildFrom[Repr, S, That]): That = if (bf(repr).isCombiner) {
    tasksupport.executeAndWaitResult(new Map[S, That](f, combinerFactory(() => bf(repr).asCombiner), splitter) mapResult { _.resultWithTaskSupport })
  } else setTaskSupport(seq.map(f)(bf2seq(bf)), tasksupport)
  /*bf ifParallel { pbf =>
    tasksupport.executeAndWaitResult(new Map[S, That](f, pbf, splitter) mapResult { _.result })
  } otherwise seq.map(f)(bf2seq(bf))*/

  def collect[S, That](pf: PartialFunction[T, S])(implicit bf: CanBuildFrom[Repr, S, That]): That = if (bf(repr).isCombiner) {
    tasksupport.executeAndWaitResult(new Collect[S, That](pf, combinerFactory(() => bf(repr).asCombiner), splitter) mapResult { _.resultWithTaskSupport })
  } else setTaskSupport(seq.collect(pf)(bf2seq(bf)), tasksupport)
  /*bf ifParallel { pbf =>
    tasksupport.executeAndWaitResult(new Collect[S, That](pf, pbf, splitter) mapResult { _.result })
  } otherwise seq.collect(pf)(bf2seq(bf))*/

  def flatMap[S, That](f: T => GenTraversableOnce[S])(implicit bf: CanBuildFrom[Repr, S, That]): That = if (bf(repr).isCombiner) {
    tasksupport.executeAndWaitResult(new FlatMap[S, That](f, combinerFactory(() => bf(repr).asCombiner), splitter) mapResult { _.resultWithTaskSupport })
  } else setTaskSupport(seq.flatMap(f)(bf2seq(bf)), tasksupport)
  /*bf ifParallel { pbf =>
    tasksupport.executeAndWaitResult(new FlatMap[S, That](f, pbf, splitter) mapResult { _.result })
  } otherwise seq.flatMap(f)(bf2seq(bf))*/

  /** Tests whether a predicate holds for all elements of this $coll.
   *
   *  $abortsignalling
   *
   *  @param p       a predicate used to test elements
   *  @return        true if `p` holds for all elements, false otherwise
   */
  def forall(@deprecatedName('pred) p: T => Boolean): Boolean = {
    tasksupport.executeAndWaitResult(new Forall(p, splitter assign new DefaultSignalling with VolatileAbort))
  }

  /** Tests whether a predicate holds for some element of this $coll.
   *
   *  $abortsignalling
   *
   *  @param p       a predicate used to test elements
   *  @return        true if `p` holds for some element, false otherwise
   */
  def exists(@deprecatedName('pred) p: T => Boolean): Boolean = {
    tasksupport.executeAndWaitResult(new Exists(p, splitter assign new DefaultSignalling with VolatileAbort))
  }

  /** Finds some element in the collection for which the predicate holds, if such
   *  an element exists. The element may not necessarily be the first such element
   *  in the iteration order.
   *
   *  If there are multiple elements obeying the predicate, the choice is nondeterministic.
   *
   *  $abortsignalling
   *
   *  @param p        predicate used to test the elements
   *  @return         an option value with the element if such an element exists, or `None` otherwise
   */
  def find(@deprecatedName('pred) p: T => Boolean): Option[T] = {
    tasksupport.executeAndWaitResult(new Find(p, splitter assign new DefaultSignalling with VolatileAbort))
  }

  /** Creates a combiner factory. Each combiner factory instance is used
   *  once per invocation of a parallel transformer method for a single
   *  collection.
   *
   *  The default combiner factory creates a new combiner every time it
   *  is requested, unless the combiner is thread-safe as indicated by its
   *  `canBeShared` method. In this case, the method returns a factory which
   *  returns the same combiner each time. This is typically done for
   *  concurrent parallel collections, the combiners of which allow
   *  thread safe access.
   */
  protected[this] def combinerFactory = {
    val combiner = newCombiner
    combiner.combinerTaskSupport = tasksupport
    if (combiner.canBeShared) new CombinerFactory[T, Repr] {
      val shared = combiner
      def apply() = shared
      def doesShareCombiners = true
    } else new CombinerFactory[T, Repr] {
      def apply() = newCombiner
      def doesShareCombiners = false
    }
  }

  protected[this] def combinerFactory[S, That](cbf: () => Combiner[S, That]) = {
    val combiner = cbf()
    combiner.combinerTaskSupport = tasksupport
    if (combiner.canBeShared) new CombinerFactory[S, That] {
      val shared = combiner
      def apply() = shared
      def doesShareCombiners = true
    } else new CombinerFactory[S, That] {
      def apply() = cbf()
      def doesShareCombiners = false
    }
  }

  def withFilter(pred: T => Boolean): Repr = filter(pred)

  def filter(pred: T => Boolean): Repr = {
    tasksupport.executeAndWaitResult(new Filter(pred, combinerFactory, splitter) mapResult { _.resultWithTaskSupport })
  }

  def filterNot(pred: T => Boolean): Repr = {
    tasksupport.executeAndWaitResult(new FilterNot(pred, combinerFactory, splitter) mapResult { _.resultWithTaskSupport })
  }

  def ++[U >: T, That](that: GenTraversableOnce[U])(implicit bf: CanBuildFrom[Repr, U, That]): That = {
    if (that.isParallel && bf.isParallel) {
      // println("case both are parallel")
      val other = that.asParIterable
      val pbf = bf.asParallel
      val cfactory = combinerFactory(() => pbf(repr))
      val copythis = new Copy(cfactory, splitter)
      val copythat = wrap {
        val othtask = new other.Copy(cfactory, other.splitter)
        tasksupport.executeAndWaitResult(othtask)
      }
      val task = (copythis parallel copythat) { _ combine _ } mapResult {
        _.resultWithTaskSupport
      }
      tasksupport.executeAndWaitResult(task)
    } else if (bf(repr).isCombiner) {
      // println("case parallel builder, `that` not parallel")
      val copythis = new Copy(combinerFactory(() => bf(repr).asCombiner), splitter)
      val copythat = wrap {
        val cb = bf(repr).asCombiner
        for (elem <- that.seq) cb += elem
        cb
      }
      tasksupport.executeAndWaitResult((copythis parallel copythat) { _ combine _ } mapResult { _.resultWithTaskSupport })
    } else {
      // println("case not a parallel builder")
      val b = bf(repr)
      this.splitter.copy2builder[U, That, Builder[U, That]](b)
      for (elem <- that.seq) b += elem
      setTaskSupport(b.result(), tasksupport)
    }
  }

  def partition(pred: T => Boolean): (Repr, Repr) = {
    tasksupport.executeAndWaitResult(
      new Partition(pred, combinerFactory, combinerFactory, splitter) mapResult {
        p => (p._1.resultWithTaskSupport, p._2.resultWithTaskSupport)
      }
    )
  }

  def groupBy[K](f: T => K): immutable.ParMap[K, Repr] = {
    val r = tasksupport.executeAndWaitResult(new GroupBy(f, () => HashMapCombiner[K, T], splitter) mapResult {
      rcb => rcb.groupByKey(() => combinerFactory())
    })
    setTaskSupport(r, tasksupport)
  }

  def take(n: Int): Repr = {
    val actualn = if (size > n) n else size
    if (actualn < MIN_FOR_COPY) take_sequential(actualn)
    else tasksupport.executeAndWaitResult(new Take(actualn, combinerFactory, splitter) mapResult {
      _.resultWithTaskSupport
    })
  }

  private def take_sequential(n: Int) = {
    val cb = newCombiner
    cb.sizeHint(n)
    val it = splitter
    var left = n
    while (left > 0) {
      cb += it.next
      left -= 1
    }
    cb.resultWithTaskSupport
  }

  def drop(n: Int): Repr = {
    val actualn = if (size > n) n else size
    if ((size - actualn) < MIN_FOR_COPY) drop_sequential(actualn)
    else tasksupport.executeAndWaitResult(new Drop(actualn, combinerFactory, splitter) mapResult { _.resultWithTaskSupport })
  }

  private def drop_sequential(n: Int) = {
    val it = splitter drop n
    val cb = newCombiner
    cb.sizeHint(size - n)
    while (it.hasNext) cb += it.next
    cb.resultWithTaskSupport
  }

  override def slice(unc_from: Int, unc_until: Int): Repr = {
    val from = unc_from min size max 0
    val until = unc_until min size max from
    if ((until - from) <= MIN_FOR_COPY) slice_sequential(from, until)
    else tasksupport.executeAndWaitResult(new Slice(from, until, combinerFactory, splitter) mapResult { _.resultWithTaskSupport })
  }

  private def slice_sequential(from: Int, until: Int): Repr = {
    val cb = newCombiner
    var left = until - from
    val it = splitter drop from
    while (left > 0) {
      cb += it.next
      left -= 1
    }
    cb.resultWithTaskSupport
  }

  def splitAt(n: Int): (Repr, Repr) = {
    tasksupport.executeAndWaitResult(
      new SplitAt(n, combinerFactory, combinerFactory, splitter) mapResult {
        p => (p._1.resultWithTaskSupport, p._2.resultWithTaskSupport)
      }
    )
  }

  /** Computes a prefix scan of the elements of the collection.
   *
   *  Note: The neutral element `z` may be applied more than once.
   *
   *  @tparam U         element type of the resulting collection
   *  @tparam That      type of the resulting collection
   *  @param z          neutral element for the operator `op`
   *  @param op         the associative operator for the scan
   *  @param bf         $bfinfo
   *  @return           a collection containing the prefix scan of the elements in the original collection
   *
   *  @usecase def scan(z: T)(op: (T, T) => T): $Coll[T]
   *    @inheritdoc
   *
   *    @return           a new $coll containing the prefix scan of the elements in this $coll
   */
  def scan[U >: T, That](z: U)(op: (U, U) => U)(implicit bf: CanBuildFrom[Repr, U, That]): That = if (bf(repr).isCombiner) {
    if (tasksupport.parallelismLevel > 1) {
      if (size > 0) tasksupport.executeAndWaitResult(new CreateScanTree(0, size, z, op, splitter) mapResult {
        tree => tasksupport.executeAndWaitResult(new FromScanTree(tree, z, op, combinerFactory(() => bf(repr).asCombiner)) mapResult {
          cb => cb.resultWithTaskSupport
        })
      }) else setTaskSupport((bf(repr) += z).result(), tasksupport)
    } else setTaskSupport(seq.scan(z)(op)(bf2seq(bf)), tasksupport)
  } else setTaskSupport(seq.scan(z)(op)(bf2seq(bf)), tasksupport)

  def scanLeft[S, That](z: S)(op: (S, T) => S)(implicit bf: CanBuildFrom[Repr, S, That]) = setTaskSupport(seq.scanLeft(z)(op)(bf2seq(bf)), tasksupport)

  def scanRight[S, That](z: S)(op: (T, S) => S)(implicit bf: CanBuildFrom[Repr, S, That]) = setTaskSupport(seq.scanRight(z)(op)(bf2seq(bf)), tasksupport)

  /** Takes the longest prefix of elements that satisfy the predicate.
   *
   *  $indexsignalling
   *  The index flag is initially set to maximum integer value.
   *
   *  @param pred   the predicate used to test the elements
   *  @return       the longest prefix of this $coll of elements that satisfy the predicate `pred`
   */
  def takeWhile(pred: T => Boolean): Repr = {
    val cbf = combinerFactory
    if (cbf.doesShareCombiners) {
      val parseqspan = toSeq.takeWhile(pred)
      tasksupport.executeAndWaitResult(new Copy(combinerFactory, parseqspan.splitter) mapResult {
        _.resultWithTaskSupport
      })
    } else {
      val cntx = new DefaultSignalling with AtomicIndexFlag
      cntx.setIndexFlag(Int.MaxValue)
      tasksupport.executeAndWaitResult(new TakeWhile(0, pred, combinerFactory, splitter assign cntx) mapResult {
        _._1.resultWithTaskSupport
      })
    }
  }

  /** Splits this $coll into a prefix/suffix pair according to a predicate.
   *
   *  $indexsignalling
   *  The index flag is initially set to maximum integer value.
   *
   *  @param pred   the predicate used to test the elements
   *  @return       a pair consisting of the longest prefix of the collection for which all
   *                the elements satisfy `pred`, and the rest of the collection
   */
  def span(pred: T => Boolean): (Repr, Repr) = {
    val cbf = combinerFactory
    if (cbf.doesShareCombiners) {
      val (xs, ys) = toSeq.span(pred)
      val copyxs = new Copy(combinerFactory, xs.splitter) mapResult { _.resultWithTaskSupport }
      val copyys = new Copy(combinerFactory, ys.splitter) mapResult { _.resultWithTaskSupport }
      val copyall = (copyxs parallel copyys) {
        (xr, yr) => (xr, yr)
      }
      tasksupport.executeAndWaitResult(copyall)
    } else {
      val cntx = new DefaultSignalling with AtomicIndexFlag
      cntx.setIndexFlag(Int.MaxValue)
      tasksupport.executeAndWaitResult(new Span(0, pred, combinerFactory, combinerFactory, splitter assign cntx) mapResult {
        p => (p._1.resultWithTaskSupport, p._2.resultWithTaskSupport)
      })
    }
  }

  /** Drops all elements in the longest prefix of elements that satisfy the predicate,
   *  and returns a collection composed of the remaining elements.
   *
   *  $indexsignalling
   *  The index flag is initially set to maximum integer value.
   *
   *  @param pred   the predicate used to test the elements
   *  @return       a collection composed of all the elements after the longest prefix of elements
   *                in this $coll that satisfy the predicate `pred`
   */
  def dropWhile(pred: T => Boolean): Repr = {
    val cntx = new DefaultSignalling with AtomicIndexFlag
    cntx.setIndexFlag(Int.MaxValue)
    tasksupport.executeAndWaitResult(
      new Span(0, pred, combinerFactory, combinerFactory, splitter assign cntx) mapResult {
        _._2.resultWithTaskSupport
      }
    )
  }

  def copyToArray[U >: T](xs: Array[U]) = copyToArray(xs, 0)

  def copyToArray[U >: T](xs: Array[U], start: Int) = copyToArray(xs, start, xs.length - start)

  def copyToArray[U >: T](xs: Array[U], start: Int, len: Int) = if (len > 0) {
    tasksupport.executeAndWaitResult(new CopyToArray(start, len, xs, splitter))
  }

  def sameElements[U >: T](that: GenIterable[U]) = seq.sameElements(that)

  def zip[U >: T, S, That](that: GenIterable[S])(implicit bf: CanBuildFrom[Repr, (U, S), That]): That = if (bf(repr).isCombiner && that.isParSeq) {
    val thatseq = that.asParSeq
    tasksupport.executeAndWaitResult(new Zip(combinerFactory(() => bf(repr).asCombiner), splitter, thatseq.splitter) mapResult { _.resultWithTaskSupport })
  } else setTaskSupport(seq.zip(that)(bf2seq(bf)), tasksupport)

  def zipWithIndex[U >: T, That](implicit bf: CanBuildFrom[Repr, (U, Int), That]): That = this zip immutable.ParRange(0, size, 1, inclusive = false)

  def zipAll[S, U >: T, That](that: GenIterable[S], thisElem: U, thatElem: S)(implicit bf: CanBuildFrom[Repr, (U, S), That]): That = if (bf(repr).isCombiner && that.isParSeq) {
    val thatseq = that.asParSeq
    tasksupport.executeAndWaitResult(
      new ZipAll(size max thatseq.length, thisElem, thatElem, combinerFactory(() => bf(repr).asCombiner), splitter, thatseq.splitter) mapResult {
        _.resultWithTaskSupport
      }
    )
  } else setTaskSupport(seq.zipAll(that, thisElem, thatElem)(bf2seq(bf)), tasksupport)

  protected def toParCollection[U >: T, That](cbf: () => Combiner[U, That]): That = {
    tasksupport.executeAndWaitResult(new ToParCollection(combinerFactory(cbf), splitter) mapResult { _.resultWithTaskSupport })
  }

  protected def toParMap[K, V, That](cbf: () => Combiner[(K, V), That])(implicit ev: T <:< (K, V)): That = {
    tasksupport.executeAndWaitResult(new ToParMap(combinerFactory(cbf), splitter)(ev) mapResult { _.resultWithTaskSupport })
  }

  @deprecated("use .seq.view instead", "2.11.0")
  def view = seq.view

  override def toArray[U >: T: ClassTag]: Array[U] = {
    val arr = new Array[U](size)
    copyToArray(arr)
    arr
  }

  override def toList: List[T] = seq.toList

  override def toIndexedSeq: scala.collection.immutable.IndexedSeq[T] = seq.toIndexedSeq

  override def toStream: Stream[T] = seq.toStream

  override def toIterator: Iterator[T] = splitter

  // the methods below are overridden

  override def toBuffer[U >: T]: scala.collection.mutable.Buffer[U] = seq.toBuffer // have additional, parallel buffers?

  override def toTraversable: GenTraversable[T] = this.asInstanceOf[GenTraversable[T]]

  override def toIterable: ParIterable[T] = this.asInstanceOf[ParIterable[T]]

  override def toSeq: ParSeq[T] = toParCollection[T, ParSeq[T]](() => ParSeq.newCombiner[T])

  override def toSet[U >: T]: immutable.ParSet[U] = toParCollection[U, immutable.ParSet[U]](() => immutable.ParSet.newCombiner[U])

  override def toMap[K, V](implicit ev: T <:< (K, V)): immutable.ParMap[K, V] = toParMap[K, V, immutable.ParMap[K, V]](() => immutable.ParMap.newCombiner[K, V])

  override def toVector: Vector[T] = to[Vector]

  override def to[Col[_]](implicit cbf: CanBuildFrom[Nothing, T, Col[T @uncheckedVariance]]): Col[T @uncheckedVariance] = if (cbf().isCombiner) {
    toParCollection[T, Col[T]](() => cbf().asCombiner)
  } else seq.to(cbf)

  /* tasks */

  protected trait StrictSplitterCheckTask[R, Tp] extends Task[R, Tp] {
    def requiresStrictSplitters = false
    if (requiresStrictSplitters && !isStrictSplitterCollection)
      throw new UnsupportedOperationException("This collection does not provide strict splitters.")
  }

  /** Standard accessor task that iterates over the elements of the collection.
   *
   *  @tparam R    type of the result of this method (`R` for result).
   *  @tparam Tp   the representation type of the task at hand.
   */
  protected trait Accessor[R, Tp]
  extends StrictSplitterCheckTask[R, Tp] {
    protected[this] val pit: IterableSplitter[T]
    protected[this] def newSubtask(p: IterableSplitter[T]): Accessor[R, Tp]
    def shouldSplitFurther = pit.shouldSplitFurther(self.repr, tasksupport.parallelismLevel)
    def split = pit.splitWithSignalling.map(newSubtask(_)) // default split procedure
    private[parallel] override def signalAbort = pit.abort()
    override def toString = this.getClass.getSimpleName + "(" + pit.toString + ")(" + result + ")(supername: " + super.toString + ")"
  }

  protected[this] trait NonDivisibleTask[R, Tp] extends StrictSplitterCheckTask[R, Tp] {
    def shouldSplitFurther = false
    def split = throw new UnsupportedOperationException("Does not split.")
  }

  protected[this] trait NonDivisible[R] extends NonDivisibleTask[R, NonDivisible[R]]

  protected[this] abstract class Composite[FR, SR, R, First <: StrictSplitterCheckTask[FR, _], Second <: StrictSplitterCheckTask[SR, _]]
    (val ft: First, val st: Second)
  extends NonDivisibleTask[R, Composite[FR, SR, R, First, Second]] {
    def combineResults(fr: FR, sr: SR): R
    @volatile var result: R = null.asInstanceOf[R]
    private[parallel] override def signalAbort() {
      ft.signalAbort()
      st.signalAbort()
    }
    protected def mergeSubtasks() {
      ft mergeThrowables st
      if (throwable eq null) result = combineResults(ft.result, st.result)
    }
    override def requiresStrictSplitters = ft.requiresStrictSplitters || st.requiresStrictSplitters
  }

  /** Sequentially performs one task after another. */
  protected[this] abstract class SeqComposite[FR, SR, R, First <: StrictSplitterCheckTask[FR, _], Second <: StrictSplitterCheckTask[SR, _]]
  (f: First, s: Second)
  extends Composite[FR, SR, R, First, Second](f, s) {
    def leaf(prevr: Option[R]) = {
      tasksupport.executeAndWaitResult(ft) : Any
      tasksupport.executeAndWaitResult(st) : Any
      mergeSubtasks()
    }
  }

  /** Performs two tasks in parallel, and waits for both to finish. */
  protected[this] abstract class ParComposite[FR, SR, R, First <: StrictSplitterCheckTask[FR, _], Second <: StrictSplitterCheckTask[SR, _]]
  (f: First, s: Second)
  extends Composite[FR, SR, R, First, Second](f, s) {
    def leaf(prevr: Option[R]) = {
      val ftfuture: () => Any = tasksupport.execute(ft)
      tasksupport.executeAndWaitResult(st) : Any
      ftfuture()
      mergeSubtasks()
    }
  }

  protected[this] abstract class ResultMapping[R, Tp, R1](val inner: StrictSplitterCheckTask[R, Tp])
  extends NonDivisibleTask[R1, ResultMapping[R, Tp, R1]] {
    @volatile var result: R1 = null.asInstanceOf[R1]
    def map(r: R): R1
    def leaf(prevr: Option[R1]) = {
      val initialResult = tasksupport.executeAndWaitResult(inner)
      result = map(initialResult)
    }
    private[parallel] override def signalAbort() {
      inner.signalAbort()
    }
    override def requiresStrictSplitters = inner.requiresStrictSplitters
  }

  protected trait Transformer[R, Tp] extends Accessor[R, Tp]

  protected[this] class Foreach[S](op: T => S, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Unit, Foreach[S]] {
    @volatile var result: Unit = ()
    def leaf(prevr: Option[Unit]) = pit.foreach(op)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Foreach[S](op, p)
  }

  protected[this] class Count(pred: T => Boolean, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Int, Count] {
    // val pittxt = pit.toString
    @volatile var result: Int = 0
    def leaf(prevr: Option[Int]) = result = pit.count(pred)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Count(pred, p)
    override def merge(that: Count) = result = result + that.result
    // override def toString = "CountTask(" + pittxt + ")"
  }

  protected[this] class Reduce[U >: T](op: (U, U) => U, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Option[U], Reduce[U]] {
    @volatile var result: Option[U] = None
    def leaf(prevr: Option[Option[U]]) = if (pit.remaining > 0) result = Some(pit.reduce(op))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Reduce(op, p)
    override def merge(that: Reduce[U]) =
      if (this.result == None) result = that.result
      else if (that.result != None) result = Some(op(result.get, that.result.get))
    override def requiresStrictSplitters = true
  }

  protected[this] class Fold[U >: T](z: U, op: (U, U) => U, protected[this] val pit: IterableSplitter[T])
  extends Accessor[U, Fold[U]] {
    @volatile var result: U = null.asInstanceOf[U]
    def leaf(prevr: Option[U]) = result = pit.fold(z)(op)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Fold(z, op, p)
    override def merge(that: Fold[U]) = result = op(result, that.result)
  }

  protected[this] class Aggregate[S](z: () => S, seqop: (S, T) => S, combop: (S, S) => S, protected[this] val pit: IterableSplitter[T])
  extends Accessor[S, Aggregate[S]] {
    @volatile var result: S = null.asInstanceOf[S]
    def leaf(prevr: Option[S]) = result = pit.foldLeft(z())(seqop)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Aggregate(z, seqop, combop, p)
    override def merge(that: Aggregate[S]) = result = combop(result, that.result)
  }

  protected[this] class Sum[U >: T](num: Numeric[U], protected[this] val pit: IterableSplitter[T])
  extends Accessor[U, Sum[U]] {
    @volatile var result: U = null.asInstanceOf[U]
    def leaf(prevr: Option[U]) = result = pit.sum(num)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Sum(num, p)
    override def merge(that: Sum[U]) = result = num.plus(result, that.result)
  }

  protected[this] class Product[U >: T](num: Numeric[U], protected[this] val pit: IterableSplitter[T])
  extends Accessor[U, Product[U]] {
    @volatile var result: U = null.asInstanceOf[U]
    def leaf(prevr: Option[U]) = result = pit.product(num)
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Product(num, p)
    override def merge(that: Product[U]) = result = num.times(result, that.result)
  }

  protected[this] class Min[U >: T](ord: Ordering[U], protected[this] val pit: IterableSplitter[T])
  extends Accessor[Option[U], Min[U]] {
    @volatile var result: Option[U] = None
    def leaf(prevr: Option[Option[U]]) = if (pit.remaining > 0) result = Some(pit.min(ord))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Min(ord, p)
    override def merge(that: Min[U]) =
      if (this.result == None) result = that.result
      else if (that.result != None) result = if (ord.lteq(result.get, that.result.get)) result else that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class Max[U >: T](ord: Ordering[U], protected[this] val pit: IterableSplitter[T])
  extends Accessor[Option[U], Max[U]] {
    @volatile var result: Option[U] = None
    def leaf(prevr: Option[Option[U]]) = if (pit.remaining > 0) result = Some(pit.max(ord))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Max(ord, p)
    override def merge(that: Max[U]) =
      if (this.result == None) result = that.result
      else if (that.result != None) result = if (ord.gteq(result.get, that.result.get)) result else that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class Map[S, That](f: T => S, cbf: CombinerFactory[S, That], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[S, That], Map[S, That]] {
    @volatile var result: Combiner[S, That] = null
    def leaf(prev: Option[Combiner[S, That]]) = result = pit.map2combiner(f, reuse(prev, cbf()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Map(f, cbf, p)
    override def merge(that: Map[S, That]) = result = result combine that.result
  }

  protected[this] class Collect[S, That]
  (pf: PartialFunction[T, S], pbf: CombinerFactory[S, That], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[S, That], Collect[S, That]] {
    @volatile var result: Combiner[S, That] = null
    def leaf(prev: Option[Combiner[S, That]]) = result = pit.collect2combiner[S, That](pf, pbf())
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Collect(pf, pbf, p)
    override def merge(that: Collect[S, That]) = result = result combine that.result
  }

  protected[this] class FlatMap[S, That]
  (f: T => GenTraversableOnce[S], pbf: CombinerFactory[S, That], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[S, That], FlatMap[S, That]] {
    @volatile var result: Combiner[S, That] = null
    def leaf(prev: Option[Combiner[S, That]]) = result = pit.flatmap2combiner(f, pbf())
    protected[this] def newSubtask(p: IterableSplitter[T]) = new FlatMap(f, pbf, p)
    override def merge(that: FlatMap[S, That]) = {
      //debuglog("merging " + result + " and " + that.result)
      result = result combine that.result
      //debuglog("merged into " + result)
    }
  }

  protected[this] class Forall(pred: T => Boolean, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Boolean, Forall] {
    @volatile var result: Boolean = true
    def leaf(prev: Option[Boolean]) = { if (!pit.isAborted) result = pit.forall(pred); if (result == false) pit.abort() }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Forall(pred, p)
    override def merge(that: Forall) = result = result && that.result
  }

  protected[this] class Exists(pred: T => Boolean, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Boolean, Exists] {
    @volatile var result: Boolean = false
    def leaf(prev: Option[Boolean]) = { if (!pit.isAborted) result = pit.exists(pred); if (result == true) pit.abort() }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Exists(pred, p)
    override def merge(that: Exists) = result = result || that.result
  }

  protected[this] class Find[U >: T](pred: T => Boolean, protected[this] val pit: IterableSplitter[T])
  extends Accessor[Option[U], Find[U]] {
    @volatile var result: Option[U] = None
    def leaf(prev: Option[Option[U]]) = { if (!pit.isAborted) result = pit.find(pred); if (result != None) pit.abort() }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Find(pred, p)
    override def merge(that: Find[U]) = if (this.result == None) result = that.result
  }

  protected[this] class Filter[U >: T, This >: Repr](pred: T => Boolean, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, This], Filter[U, This]] {
    @volatile var result: Combiner[U, This] = null
    def leaf(prev: Option[Combiner[U, This]]) = {
      result = pit.filter2combiner(pred, reuse(prev, cbf()))
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Filter(pred, cbf, p)
    override def merge(that: Filter[U, This]) = result = result combine that.result
  }

  protected[this] class FilterNot[U >: T, This >: Repr](pred: T => Boolean, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, This], FilterNot[U, This]] {
    @volatile var result: Combiner[U, This] = null
    def leaf(prev: Option[Combiner[U, This]]) = {
      result = pit.filterNot2combiner(pred, reuse(prev, cbf()))
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new FilterNot(pred, cbf, p)
    override def merge(that: FilterNot[U, This]) = result = result combine that.result
  }

  protected class Copy[U >: T, That](cfactory: CombinerFactory[U, That], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, That], Copy[U, That]] {
    @volatile var result: Combiner[U, That] = null
    def leaf(prev: Option[Combiner[U, That]]) = result = pit.copy2builder[U, That, Combiner[U, That]](reuse(prev, cfactory()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Copy[U, That](cfactory, p)
    override def merge(that: Copy[U, That]) = result = result combine that.result
  }

  protected[this] class Partition[U >: T, This >: Repr]
  (pred: T => Boolean, cbfTrue: CombinerFactory[U, This], cbfFalse: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[(Combiner[U, This], Combiner[U, This]), Partition[U, This]] {
    @volatile var result: (Combiner[U, This], Combiner[U, This]) = null
    def leaf(prev: Option[(Combiner[U, This], Combiner[U, This])]) = result = pit.partition2combiners(pred, reuse(prev.map(_._1), cbfTrue()), reuse(prev.map(_._2), cbfFalse()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = new Partition(pred, cbfTrue, cbfFalse, p)
    override def merge(that: Partition[U, This]) = result = (result._1 combine that.result._1, result._2 combine that.result._2)
  }

  protected[this] class GroupBy[K, U >: T](
    f: U => K,
    mcf: () => HashMapCombiner[K, U],
    protected[this] val pit: IterableSplitter[T]
  ) extends Transformer[HashMapCombiner[K, U], GroupBy[K, U]] {
    @volatile var result: Result = null
    final def leaf(prev: Option[Result]) = {
      // note: HashMapCombiner doesn't merge same keys until evaluation
      val cb = mcf()
      while (pit.hasNext) {
        val elem = pit.next()
        cb += f(elem) -> elem
      }
      result = cb
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new GroupBy(f, mcf, p)
    override def merge(that: GroupBy[K, U]) = {
      // note: this works because we know that a HashMapCombiner doesn't merge same keys until evaluation
      // --> we know we're not dropping any mappings
      result = (result combine that.result).asInstanceOf[HashMapCombiner[K, U]]
    }
  }

  protected[this] class Take[U >: T, This >: Repr]
  (n: Int, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, This], Take[U, This]] {
    @volatile var result: Combiner[U, This] = null
    def leaf(prev: Option[Combiner[U, This]]) = {
      result = pit.take2combiner(n, reuse(prev, cbf()))
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      val sizes = pits.scanLeft(0)(_ + _.remaining)
      for ((p, untilp) <- pits zip sizes; if untilp <= n) yield {
        if (untilp + p.remaining < n) new Take(p.remaining, cbf, p)
        else new Take(n - untilp, cbf, p)
      }
    }
    override def merge(that: Take[U, This]) = result = result combine that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class Drop[U >: T, This >: Repr]
  (n: Int, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, This], Drop[U, This]] {
    @volatile var result: Combiner[U, This] = null
    def leaf(prev: Option[Combiner[U, This]]) = result = pit.drop2combiner(n, reuse(prev, cbf()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      val sizes = pits.scanLeft(0)(_ + _.remaining)
      for ((p, withp) <- pits zip sizes.tail; if withp >= n) yield {
        if (withp - p.remaining > n) new Drop(0, cbf, p)
        else new Drop(n - withp + p.remaining, cbf, p)
      }
    }
    override def merge(that: Drop[U, This]) = result = result combine that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class Slice[U >: T, This >: Repr]
  (from: Int, until: Int, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, This], Slice[U, This]] {
    @volatile var result: Combiner[U, This] = null
    def leaf(prev: Option[Combiner[U, This]]) = result = pit.slice2combiner(from, until, reuse(prev, cbf()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      val sizes = pits.scanLeft(0)(_ + _.remaining)
      for ((p, untilp) <- pits zip sizes; if untilp + p.remaining >= from || untilp <= until) yield {
        val f = (from max untilp) - untilp
        val u = (until min (untilp + p.remaining)) - untilp
        new Slice(f, u, cbf, p)
      }
    }
    override def merge(that: Slice[U, This]) = result = result combine that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class SplitAt[U >: T, This >: Repr]
  (at: Int, cbfBefore: CombinerFactory[U, This], cbfAfter: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[(Combiner[U, This], Combiner[U, This]), SplitAt[U, This]] {
    @volatile var result: (Combiner[U, This], Combiner[U, This]) = null
    def leaf(prev: Option[(Combiner[U, This], Combiner[U, This])]) = result = pit.splitAt2combiners(at, reuse(prev.map(_._1), cbfBefore()), reuse(prev.map(_._2), cbfAfter()))
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      val sizes = pits.scanLeft(0)(_ + _.remaining)
      for ((p, untilp) <- pits zip sizes) yield new SplitAt((at max untilp min (untilp + p.remaining)) - untilp, cbfBefore, cbfAfter, p)
    }
    override def merge(that: SplitAt[U, This]) = result = (result._1 combine that.result._1, result._2 combine that.result._2)
    override def requiresStrictSplitters = true
  }

  protected[this] class TakeWhile[U >: T, This >: Repr]
  (pos: Int, pred: T => Boolean, cbf: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[(Combiner[U, This], Boolean), TakeWhile[U, This]] {
    @volatile var result: (Combiner[U, This], Boolean) = null
    def leaf(prev: Option[(Combiner[U, This], Boolean)]) = if (pos < pit.indexFlag) {
      result = pit.takeWhile2combiner(pred, reuse(prev.map(_._1), cbf()))
      if (!result._2) pit.setIndexFlagIfLesser(pos)
    } else result = (reuse(prev.map(_._1), cbf()), false)
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      for ((p, untilp) <- pits zip pits.scanLeft(0)(_ + _.remaining)) yield new TakeWhile(pos + untilp, pred, cbf, p)
    }
    override def merge(that: TakeWhile[U, This]) = if (result._2) {
      result = (result._1 combine that.result._1, that.result._2)
    }
    override def requiresStrictSplitters = true
  }

  protected[this] class Span[U >: T, This >: Repr]
  (pos: Int, pred: T => Boolean, cbfBefore: CombinerFactory[U, This], cbfAfter: CombinerFactory[U, This], protected[this] val pit: IterableSplitter[T])
  extends Transformer[(Combiner[U, This], Combiner[U, This]), Span[U, This]] {
    @volatile var result: (Combiner[U, This], Combiner[U, This]) = null
    def leaf(prev: Option[(Combiner[U, This], Combiner[U, This])]) = if (pos < pit.indexFlag) {
      // val lst = pit.toList
      // val pa = mutable.ParArray(lst: _*)
      // val str = "At leaf we will iterate: " + pa.splitter.toList
      result = pit.span2combiners(pred, cbfBefore(), cbfAfter()) // do NOT reuse old combiners here, lest ye be surprised
      // println("\nAt leaf result is: " + result)
      if (result._2.size > 0) pit.setIndexFlagIfLesser(pos)
    } else {
      result = (reuse(prev.map(_._2), cbfBefore()), pit.copy2builder[U, This, Combiner[U, This]](reuse(prev.map(_._2), cbfAfter())))
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      for ((p, untilp) <- pits zip pits.scanLeft(0)(_ + _.remaining)) yield new Span(pos + untilp, pred, cbfBefore, cbfAfter, p)
    }
    override def merge(that: Span[U, This]) = result = if (result._2.size == 0) {
      (result._1 combine that.result._1, that.result._2)
    } else {
      (result._1, result._2 combine that.result._1 combine that.result._2)
    }
    override def requiresStrictSplitters = true
  }

  protected[this] class Zip[U >: T, S, That](pbf: CombinerFactory[(U, S), That], protected[this] val pit: IterableSplitter[T], val othpit: SeqSplitter[S])
  extends Transformer[Combiner[(U, S), That], Zip[U, S, That]] {
    @volatile var result: Result = null
    def leaf(prev: Option[Result]) = result = pit.zip2combiner[U, S, That](othpit, pbf())
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      val sizes = pits.map(_.remaining)
      val opits = othpit.psplitWithSignalling(sizes: _*)
      (pits zip opits) map { p => new Zip(pbf, p._1, p._2) }
    }
    override def merge(that: Zip[U, S, That]) = result = result combine that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class ZipAll[U >: T, S, That]
  (len: Int, thiselem: U, thatelem: S, pbf: CombinerFactory[(U, S), That], protected[this] val pit: IterableSplitter[T], val othpit: SeqSplitter[S])
  extends Transformer[Combiner[(U, S), That], ZipAll[U, S, That]] {
    @volatile var result: Result = null
    def leaf(prev: Option[Result]) = result = pit.zipAll2combiner[U, S, That](othpit, thiselem, thatelem, pbf())
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = if (pit.remaining <= len) {
      val pits = pit.splitWithSignalling
      val sizes = pits.map(_.remaining)
      val opits = othpit.psplitWithSignalling(sizes: _*)
      ((pits zip opits) zip sizes) map { t => new ZipAll(t._2, thiselem, thatelem, pbf, t._1._1, t._1._2) }
    } else {
      val opits = othpit.psplitWithSignalling(pit.remaining)
      val diff = len - pit.remaining
      Seq(
        new ZipAll(pit.remaining, thiselem, thatelem, pbf, pit, opits(0)), // nothing wrong will happen with the cast below - elem T is never accessed
        new ZipAll(diff, thiselem, thatelem, pbf, immutable.repetition(thiselem, diff).splitter.asInstanceOf[IterableSplitter[T]], opits(1))
      )
    }
    override def merge(that: ZipAll[U, S, That]) = result = result combine that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class CopyToArray[U >: T, This >: Repr](from: Int, len: Int, array: Array[U], protected[this] val pit: IterableSplitter[T])
  extends Accessor[Unit, CopyToArray[U, This]] {
    @volatile var result: Unit = ()
    def leaf(prev: Option[Unit]) = pit.copyToArray(array, from, len)
    protected[this] def newSubtask(p: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      for ((p, untilp) <- pits zip pits.scanLeft(0)(_ + _.remaining); if untilp < len) yield {
        val plen = p.remaining min (len - untilp)
        new CopyToArray[U, This](from + untilp, plen, array, p)
      }
    }
    override def requiresStrictSplitters = true
  }

  protected[this] class ToParCollection[U >: T, That](cbf: CombinerFactory[U, That], protected[this] val pit: IterableSplitter[T])
  extends Transformer[Combiner[U, That], ToParCollection[U, That]] {
    @volatile var result: Result = null
    def leaf(prev: Option[Combiner[U, That]]) {
      result = cbf()
      while (pit.hasNext) result += pit.next
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new ToParCollection[U, That](cbf, p)
    override def merge(that: ToParCollection[U, That]) = result = result combine that.result
  }

  protected[this] class ToParMap[K, V, That](cbf: CombinerFactory[(K, V), That], protected[this] val pit: IterableSplitter[T])(implicit ev: T <:< (K, V))
  extends Transformer[Combiner[(K, V), That], ToParMap[K, V, That]] {
    @volatile var result: Result = null
    def leaf(prev: Option[Combiner[(K, V), That]]) {
      result = cbf()
      while (pit.hasNext) result += pit.next
    }
    protected[this] def newSubtask(p: IterableSplitter[T]) = new ToParMap[K, V, That](cbf, p)(ev)
    override def merge(that: ToParMap[K, V, That]) = result = result combine that.result
  }

  protected[this] class CreateScanTree[U >: T](from: Int, len: Int, z: U, op: (U, U) => U, protected[this] val pit: IterableSplitter[T])
  extends Transformer[ScanTree[U], CreateScanTree[U]] {
    @volatile var result: ScanTree[U] = null
    def leaf(prev: Option[ScanTree[U]]) = if (pit.remaining > 0) {
      val trees = ArrayBuffer[ScanTree[U]]()
      var i = from
      val until = from + len
      val blocksize = scanBlockSize
      while (i < until) {
        trees += scanBlock(i, scala.math.min(blocksize, pit.remaining))
        i += blocksize
      }

      // merge trees
      result = mergeTrees(trees, 0, trees.length)
    } else result = null // no elements to scan (merge will take care of `null`s)
    private def scanBlock(from: Int, len: Int): ScanTree[U] = {
      val pitdup = pit.dup
      new ScanLeaf(pitdup, op, from, len, None, pit.reduceLeft(len, op))
    }
    private def mergeTrees(trees: ArrayBuffer[ScanTree[U]], from: Int, howmany: Int): ScanTree[U] = if (howmany > 1) {
      val half = howmany / 2
      ScanNode(mergeTrees(trees, from, half), mergeTrees(trees, from + half, howmany - half))
    } else trees(from)
    protected[this] def newSubtask(pit: IterableSplitter[T]) = throw new UnsupportedOperationException
    override def split = {
      val pits = pit.splitWithSignalling
      for ((p, untilp) <- pits zip pits.scanLeft(from)(_ + _.remaining)) yield {
        new CreateScanTree(untilp, p.remaining, z, op, p)
      }
    }
    override def merge(that: CreateScanTree[U]) = if (this.result != null) {
      if (that.result != null) result = ScanNode(result, that.result)
    } else result = that.result
    override def requiresStrictSplitters = true
  }

  protected[this] class FromScanTree[U >: T, That]
  (tree: ScanTree[U], z: U, op: (U, U) => U, cbf: CombinerFactory[U, That])
  extends StrictSplitterCheckTask[Combiner[U, That], FromScanTree[U, That]] {
    @volatile var result: Combiner[U, That] = null
    def leaf(prev: Option[Combiner[U, That]]) {
      val cb = reuse(prev, cbf())
      iterate(tree, cb)
      result = cb
    }
    private def iterate(tree: ScanTree[U], cb: Combiner[U, That]): Unit = tree match {
      case ScanNode(left, right) =>
        iterate(left, cb)
        iterate(right, cb)
      case ScanLeaf(p, _, _, len, Some(prev), _) =>
        p.scanToCombiner(len, prev.acc, op, cb)
      case ScanLeaf(p, _, _, len, None, _) =>
        cb += z
        p.scanToCombiner(len, z, op, cb)
    }
    def split = tree match {
      case ScanNode(left, right) => Seq(
        new FromScanTree(left, z, op, cbf),
        new FromScanTree(right, z, op, cbf)
      )
      case _ => throw new UnsupportedOperationException("Cannot be split further")
    }
    def shouldSplitFurther = tree match {
      case ScanNode(_, _) => true
      case ScanLeaf(_, _, _, _, _, _) => false
    }
    override def merge(that: FromScanTree[U, That]) = result = result combine that.result
  }

  /* scan tree */

  protected[this] def scanBlockSize = (thresholdFromSize(size, tasksupport.parallelismLevel) / 2) max 1

  protected[this] trait ScanTree[U >: T] {
    def beginsAt: Int
    def pushdown(v: U): Unit
    def leftmost: ScanLeaf[U]
    def rightmost: ScanLeaf[U]
    def print(depth: Int = 0): Unit
  }

  protected[this] case class ScanNode[U >: T](left: ScanTree[U], right: ScanTree[U]) extends ScanTree[U] {
    right.pushdown(left.rightmost.acc)
    right.leftmost.prev = Some(left.rightmost)

    val leftmost = left.leftmost
    val rightmost = right.rightmost

    def beginsAt = left.beginsAt
    def pushdown(v: U) {
      left.pushdown(v)
      right.pushdown(v)
    }
    def print(depth: Int) {
      println((" " * depth) + "ScanNode, begins at " + beginsAt)
      left.print(depth + 1)
      right.print(depth + 1)
    }
  }

  protected[this] case class ScanLeaf[U >: T]
  (pit: IterableSplitter[U], op: (U, U) => U, from: Int, len: Int, var prev: Option[ScanLeaf[U]], var acc: U)
  extends ScanTree[U] {
    def beginsAt = from
    def pushdown(v: U) = {
      acc = op(v, acc)
    }
    def leftmost = this
    def rightmost = this
    def print(depth: Int) = println((" " * depth) + this)
  }

  /* alias methods */

  def /:[S](z: S)(op: (S, T) => S): S = foldLeft(z)(op)

  def :\[S](z: S)(op: (T, S) => S): S = foldRight(z)(op)

  /* debug information */

  private[parallel] def debugInformation = "Parallel collection: " + this.getClass

  private[parallel] def brokenInvariants = Seq[String]()

  // private val dbbuff = ArrayBuffer[String]()
  // def debugBuffer: ArrayBuffer[String] = dbbuff
  def debugBuffer: ArrayBuffer[String] = null

  private[parallel] def debugclear() = synchronized {
    debugBuffer.clear()
  }

  private[parallel] def debuglog(s: String) = synchronized {
    debugBuffer += s
  }

  import scala.collection.DebugUtils._
  private[parallel] def printDebugBuffer() = println(buildString {
    append =>
    for (s <- debugBuffer) {
      append(s)
    }
  })
}