about summary refs log tree commit diff
path: root/third_party/bazel/rules_haskell/examples/vector/Data
diff options
context:
space:
mode:
Diffstat (limited to 'third_party/bazel/rules_haskell/examples/vector/Data')
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector.hs1719
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle.hs655
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Monadic.hs1106
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Size.hs121
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Stream/Monadic.hs1639
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Util.hs60
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic.hs2206
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Base.hs140
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable.hs1034
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable/Base.hs145
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/New.hs178
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Internal/Check.hs152
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Mutable.hs416
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive.hs1393
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive/Mutable.hs366
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable.hs1489
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Internal.hs33
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Mutable.hs543
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed.hs1488
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Base.hs408
-rw-r--r--third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Mutable.hs307
21 files changed, 0 insertions, 15598 deletions
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector.hs
deleted file mode 100644
index 21b61960ca40..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector.hs
+++ /dev/null
@@ -1,1719 +0,0 @@
-{-# LANGUAGE CPP
-           , DeriveDataTypeable
-           , FlexibleInstances
-           , MultiParamTypeClasses
-           , TypeFamilies
-           , Rank2Types
-           , BangPatterns
-  #-}
-
--- |
--- Module      : Data.Vector
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- A library for boxed vectors (that is, polymorphic arrays capable of
--- holding any Haskell value). The vectors come in two flavours:
---
---  * mutable
---
---  * immutable
---
--- and support a rich interface of both list-like operations, and bulk
--- array operations.
---
--- For unboxed arrays, use "Data.Vector.Unboxed"
---
-
-module Data.Vector (
-  -- * Boxed vectors
-  Vector, MVector,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Indexing
-  (!), (!?), head, last,
-  unsafeIndex, unsafeHead, unsafeLast,
-
-  -- ** Monadic indexing
-  indexM, headM, lastM,
-  unsafeIndexM, unsafeHeadM, unsafeLastM,
-
-  -- ** Extracting subvectors (slicing)
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- * Construction
-
-  -- ** Initialisation
-  empty, singleton, replicate, generate, iterateN,
-
-  -- ** Monadic initialisation
-  replicateM, generateM, iterateNM, create, createT,
-
-  -- ** Unfolding
-  unfoldr, unfoldrN,
-  unfoldrM, unfoldrNM,
-  constructN, constructrN,
-
-  -- ** Enumeration
-  enumFromN, enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- ** Concatenation
-  cons, snoc, (++), concat,
-
-  -- ** Restricting memory usage
-  force,
-
-  -- * Modifying vectors
-
-  -- ** Bulk updates
-  (//), update, update_,
-  unsafeUpd, unsafeUpdate, unsafeUpdate_,
-
-  -- ** Accumulations
-  accum, accumulate, accumulate_,
-  unsafeAccum, unsafeAccumulate, unsafeAccumulate_,
-
-  -- ** Permutations
-  reverse, backpermute, unsafeBackpermute,
-
-  -- ** Safe destructive updates
-  modify,
-
-  -- * Elementwise operations
-
-  -- ** Indexing
-  indexed,
-
-  -- ** Mapping
-  map, imap, concatMap,
-
-  -- ** Monadic mapping
-  mapM, imapM, mapM_, imapM_, forM, forM_,
-
-  -- ** Zipping
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  izipWith, izipWith3, izipWith4, izipWith5, izipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- ** Monadic zipping
-  zipWithM, izipWithM, zipWithM_, izipWithM_,
-
-  -- ** Unzipping
-  unzip, unzip3, unzip4, unzip5, unzip6,
-
-  -- * Working with predicates
-
-  -- ** Filtering
-  filter, ifilter, uniq,
-  mapMaybe, imapMaybe,
-  filterM,
-  takeWhile, dropWhile,
-
-  -- ** Partitioning
-  partition, unstablePartition, span, break,
-
-  -- ** Searching
-  elem, notElem, find, findIndex, findIndices, elemIndex, elemIndices,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1, foldr', foldr1',
-  ifoldl, ifoldl', ifoldr, ifoldr',
-
-  -- ** Specialised folds
-  all, any, and, or,
-  sum, product,
-  maximum, maximumBy, minimum, minimumBy,
-  minIndex, minIndexBy, maxIndex, maxIndexBy,
-
-  -- ** Monadic folds
-  foldM, ifoldM, foldM', ifoldM',
-  fold1M, fold1M',foldM_, ifoldM_,
-  foldM'_, ifoldM'_, fold1M_, fold1M'_,
-
-  -- ** Monadic sequencing
-  sequence, sequence_,
-
-  -- * Prefix sums (scans)
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl', scanl1, scanl1',
-  iscanl, iscanl',
-  prescanr, prescanr',
-  postscanr, postscanr',
-  scanr, scanr', scanr1, scanr1',
-  iscanr, iscanr',
-
-  -- * Conversions
-
-  -- ** Lists
-  toList, Data.Vector.fromList, Data.Vector.fromListN,
-
-  -- ** Other vector types
-  G.convert,
-
-  -- ** Mutable vectors
-  freeze, thaw, copy, unsafeFreeze, unsafeThaw, unsafeCopy
-) where
-
-import qualified Data.Vector.Generic as G
-import           Data.Vector.Mutable  ( MVector(..) )
-import           Data.Primitive.Array
-import qualified Data.Vector.Fusion.Bundle as Bundle
-
-import Control.DeepSeq ( NFData, rnf )
-import Control.Monad ( MonadPlus(..), liftM, ap )
-import Control.Monad.ST ( ST )
-import Control.Monad.Primitive
-
-
-import Control.Monad.Zip
-
-import Prelude hiding ( length, null,
-                        replicate, (++), concat,
-                        head, last,
-                        init, tail, take, drop, splitAt, reverse,
-                        map, concatMap,
-                        zipWith, zipWith3, zip, zip3, unzip, unzip3,
-                        filter, takeWhile, dropWhile, span, break,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        all, any, and, or, sum, product, minimum, maximum,
-                        scanl, scanl1, scanr, scanr1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_, sequence, sequence_ )
-
-#if MIN_VERSION_base(4,9,0)
-import Data.Functor.Classes (Eq1 (..), Ord1 (..), Read1 (..), Show1 (..))
-#endif
-
-import Data.Typeable  ( Typeable )
-import Data.Data      ( Data(..) )
-import Text.Read      ( Read(..), readListPrecDefault )
-import Data.Semigroup ( Semigroup(..) )
-
-import qualified Control.Applicative as Applicative
-import qualified Data.Foldable as Foldable
-import qualified Data.Traversable as Traversable
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Monoid   ( Monoid(..) )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-import qualified GHC.Exts as Exts (IsList(..))
-#endif
-
-
--- | Boxed vectors, supporting efficient slicing.
-data Vector a = Vector {-# UNPACK #-} !Int
-                       {-# UNPACK #-} !Int
-                       {-# UNPACK #-} !(Array a)
-        deriving ( Typeable )
-
-instance NFData a => NFData (Vector a) where
-    rnf (Vector i n arr) = rnfAll i
-        where
-          rnfAll ix | ix < n    = rnf (indexArray arr ix) `seq` rnfAll (ix+1)
-                    | otherwise = ()
-
-instance Show a => Show (Vector a) where
-  showsPrec = G.showsPrec
-
-instance Read a => Read (Vector a) where
-  readPrec = G.readPrec
-  readListPrec = readListPrecDefault
-
-#if MIN_VERSION_base(4,9,0)
-instance Show1 Vector where
-    liftShowsPrec = G.liftShowsPrec
-
-instance Read1 Vector where
-    liftReadsPrec = G.liftReadsPrec
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-
-instance Exts.IsList (Vector a) where
-  type Item (Vector a) = a
-  fromList = Data.Vector.fromList
-  fromListN = Data.Vector.fromListN
-  toList = toList
-#endif
-
-instance Data a => Data (Vector a) where
-  gfoldl       = G.gfoldl
-  toConstr _   = error "toConstr"
-  gunfold _ _  = error "gunfold"
-  dataTypeOf _ = G.mkType "Data.Vector.Vector"
-  dataCast1    = G.dataCast
-
-type instance G.Mutable Vector = MVector
-
-instance G.Vector Vector a where
-  {-# INLINE basicUnsafeFreeze #-}
-  basicUnsafeFreeze (MVector i n marr)
-    = Vector i n `liftM` unsafeFreezeArray marr
-
-  {-# INLINE basicUnsafeThaw #-}
-  basicUnsafeThaw (Vector i n arr)
-    = MVector i n `liftM` unsafeThawArray arr
-
-  {-# INLINE basicLength #-}
-  basicLength (Vector _ n _) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice j n (Vector i _ arr) = Vector (i+j) n arr
-
-  {-# INLINE basicUnsafeIndexM #-}
-  basicUnsafeIndexM (Vector i _ arr) j = indexArrayM arr (i+j)
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector i n dst) (Vector j _ src)
-    = copyArray dst i src j n
-
--- See http://trac.haskell.org/vector/ticket/12
-instance Eq a => Eq (Vector a) where
-  {-# INLINE (==) #-}
-  xs == ys = Bundle.eq (G.stream xs) (G.stream ys)
-
-  {-# INLINE (/=) #-}
-  xs /= ys = not (Bundle.eq (G.stream xs) (G.stream ys))
-
--- See http://trac.haskell.org/vector/ticket/12
-instance Ord a => Ord (Vector a) where
-  {-# INLINE compare #-}
-  compare xs ys = Bundle.cmp (G.stream xs) (G.stream ys)
-
-  {-# INLINE (<) #-}
-  xs < ys = Bundle.cmp (G.stream xs) (G.stream ys) == LT
-
-  {-# INLINE (<=) #-}
-  xs <= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= GT
-
-  {-# INLINE (>) #-}
-  xs > ys = Bundle.cmp (G.stream xs) (G.stream ys) == GT
-
-  {-# INLINE (>=) #-}
-  xs >= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= LT
-
-#if MIN_VERSION_base(4,9,0)
-instance Eq1 Vector where
-  liftEq eq xs ys = Bundle.eqBy eq (G.stream xs) (G.stream ys)
-
-instance Ord1 Vector where
-  liftCompare cmp xs ys = Bundle.cmpBy cmp (G.stream xs) (G.stream ys)
-#endif
-
-instance Semigroup (Vector a) where
-  {-# INLINE (<>) #-}
-  (<>) = (++)
-
-  {-# INLINE sconcat #-}
-  sconcat = G.concatNE
-
-instance Monoid (Vector a) where
-  {-# INLINE mempty #-}
-  mempty = empty
-
-  {-# INLINE mappend #-}
-  mappend = (++)
-
-  {-# INLINE mconcat #-}
-  mconcat = concat
-
-instance Functor Vector where
-  {-# INLINE fmap #-}
-  fmap = map
-
-instance Monad Vector where
-  {-# INLINE return #-}
-  return = Applicative.pure
-
-  {-# INLINE (>>=) #-}
-  (>>=) = flip concatMap
-
-  {-# INLINE fail #-}
-  fail _ = empty
-
-instance MonadPlus Vector where
-  {-# INLINE mzero #-}
-  mzero = empty
-
-  {-# INLINE mplus #-}
-  mplus = (++)
-
-instance MonadZip Vector where
-  {-# INLINE mzip #-}
-  mzip = zip
-
-  {-# INLINE mzipWith #-}
-  mzipWith = zipWith
-
-  {-# INLINE munzip #-}
-  munzip = unzip
-
-
-instance Applicative.Applicative Vector where
-  {-# INLINE pure #-}
-  pure = singleton
-
-  {-# INLINE (<*>) #-}
-  (<*>) = ap
-
-instance Applicative.Alternative Vector where
-  {-# INLINE empty #-}
-  empty = empty
-
-  {-# INLINE (<|>) #-}
-  (<|>) = (++)
-
-instance Foldable.Foldable Vector where
-  {-# INLINE foldr #-}
-  foldr = foldr
-
-  {-# INLINE foldl #-}
-  foldl = foldl
-
-  {-# INLINE foldr1 #-}
-  foldr1 = foldr1
-
-  {-# INLINE foldl1 #-}
-  foldl1 = foldl1
-
-#if MIN_VERSION_base(4,6,0)
-  {-# INLINE foldr' #-}
-  foldr' = foldr'
-
-  {-# INLINE foldl' #-}
-  foldl' = foldl'
-#endif
-
-#if MIN_VERSION_base(4,8,0)
-  {-# INLINE toList #-}
-  toList = toList
-
-  {-# INLINE length #-}
-  length = length
-
-  {-# INLINE null #-}
-  null = null
-
-  {-# INLINE elem #-}
-  elem = elem
-
-  {-# INLINE maximum #-}
-  maximum = maximum
-
-  {-# INLINE minimum #-}
-  minimum = minimum
-
-  {-# INLINE sum #-}
-  sum = sum
-
-  {-# INLINE product #-}
-  product = product
-#endif
-
-instance Traversable.Traversable Vector where
-  {-# INLINE traverse #-}
-  traverse f xs = Data.Vector.fromList Applicative.<$> Traversable.traverse f (toList xs)
-
-  {-# INLINE mapM #-}
-  mapM = mapM
-
-  {-# INLINE sequence #-}
-  sequence = sequence
-
--- Length information
--- ------------------
-
--- | /O(1)/ Yield the length of the vector
-length :: Vector a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | /O(1)/ Test whether a vector is empty
-null :: Vector a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Indexing
--- --------
-
--- | O(1) Indexing
-(!) :: Vector a -> Int -> a
-{-# INLINE (!) #-}
-(!) = (G.!)
-
--- | O(1) Safe indexing
-(!?) :: Vector a -> Int -> Maybe a
-{-# INLINE (!?) #-}
-(!?) = (G.!?)
-
--- | /O(1)/ First element
-head :: Vector a -> a
-{-# INLINE head #-}
-head = G.head
-
--- | /O(1)/ Last element
-last :: Vector a -> a
-{-# INLINE last #-}
-last = G.last
-
--- | /O(1)/ Unsafe indexing without bounds checking
-unsafeIndex :: Vector a -> Int -> a
-{-# INLINE unsafeIndex #-}
-unsafeIndex = G.unsafeIndex
-
--- | /O(1)/ First element without checking if the vector is empty
-unsafeHead :: Vector a -> a
-{-# INLINE unsafeHead #-}
-unsafeHead = G.unsafeHead
-
--- | /O(1)/ Last element without checking if the vector is empty
-unsafeLast :: Vector a -> a
-{-# INLINE unsafeLast #-}
-unsafeLast = G.unsafeLast
-
--- Monadic indexing
--- ----------------
-
--- | /O(1)/ Indexing in a monad.
---
--- The monad allows operations to be strict in the vector when necessary.
--- Suppose vector copying is implemented like this:
---
--- > copy mv v = ... write mv i (v ! i) ...
---
--- For lazy vectors, @v ! i@ would not be evaluated which means that @mv@
--- would unnecessarily retain a reference to @v@ in each element written.
---
--- With 'indexM', copying can be implemented like this instead:
---
--- > copy mv v = ... do
--- >                   x <- indexM v i
--- >                   write mv i x
---
--- Here, no references to @v@ are retained because indexing (but /not/ the
--- elements) is evaluated eagerly.
---
-indexM :: Monad m => Vector a -> Int -> m a
-{-# INLINE indexM #-}
-indexM = G.indexM
-
--- | /O(1)/ First element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-headM :: Monad m => Vector a -> m a
-{-# INLINE headM #-}
-headM = G.headM
-
--- | /O(1)/ Last element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-lastM :: Monad m => Vector a -> m a
-{-# INLINE lastM #-}
-lastM = G.lastM
-
--- | /O(1)/ Indexing in a monad without bounds checks. See 'indexM' for an
--- explanation of why this is useful.
-unsafeIndexM :: Monad m => Vector a -> Int -> m a
-{-# INLINE unsafeIndexM #-}
-unsafeIndexM = G.unsafeIndexM
-
--- | /O(1)/ First element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeHeadM :: Monad m => Vector a -> m a
-{-# INLINE unsafeHeadM #-}
-unsafeHeadM = G.unsafeHeadM
-
--- | /O(1)/ Last element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeLastM :: Monad m => Vector a -> m a
-{-# INLINE unsafeLastM #-}
-unsafeLastM = G.unsafeLastM
-
--- Extracting subvectors (slicing)
--- -------------------------------
-
--- | /O(1)/ Yield a slice of the vector without copying it. The vector must
--- contain at least @i+n@ elements.
-slice :: Int   -- ^ @i@ starting index
-                 -> Int   -- ^ @n@ length
-                 -> Vector a
-                 -> Vector a
-{-# INLINE slice #-}
-slice = G.slice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty.
-init :: Vector a -> Vector a
-{-# INLINE init #-}
-init = G.init
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty.
-tail :: Vector a -> Vector a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | /O(1)/ Yield at the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case it is returned unchanged.
-take :: Int -> Vector a -> Vector a
-{-# INLINE take #-}
-take = G.take
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case an empty vector is returned.
-drop :: Int -> Vector a -> Vector a
-{-# INLINE drop #-}
-drop = G.drop
-
--- | /O(1)/ Yield the first @n@ elements paired with the remainder without copying.
---
--- Note that @'splitAt' n v@ is equivalent to @('take' n v, 'drop' n v)@
--- but slightly more efficient.
-{-# INLINE splitAt #-}
-splitAt :: Int -> Vector a -> (Vector a, Vector a)
-splitAt = G.splitAt
-
--- | /O(1)/ Yield a slice of the vector without copying. The vector must
--- contain at least @i+n@ elements but this is not checked.
-unsafeSlice :: Int   -- ^ @i@ starting index
-                       -> Int   -- ^ @n@ length
-                       -> Vector a
-                       -> Vector a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty but this is not checked.
-unsafeInit :: Vector a -> Vector a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty but this is not checked.
-unsafeTail :: Vector a -> Vector a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector must
--- contain at least @n@ elements but this is not checked.
-unsafeTake :: Int -> Vector a -> Vector a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector
--- must contain at least @n@ elements but this is not checked.
-unsafeDrop :: Int -> Vector a -> Vector a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
--- Initialisation
--- --------------
-
--- | /O(1)/ Empty vector
-empty :: Vector a
-{-# INLINE empty #-}
-empty = G.empty
-
--- | /O(1)/ Vector with exactly one element
-singleton :: a -> Vector a
-{-# INLINE singleton #-}
-singleton = G.singleton
-
--- | /O(n)/ Vector of the given length with the same value in each position
-replicate :: Int -> a -> Vector a
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | /O(n)/ Construct a vector of the given length by applying the function to
--- each index
-generate :: Int -> (Int -> a) -> Vector a
-{-# INLINE generate #-}
-generate = G.generate
-
--- | /O(n)/ Apply function n times to value. Zeroth element is original value.
-iterateN :: Int -> (a -> a) -> a -> Vector a
-{-# INLINE iterateN #-}
-iterateN = G.iterateN
-
--- Unfolding
--- ---------
-
--- | /O(n)/ Construct a vector by repeatedly applying the generator function
--- to a seed. The generator function yields 'Just' the next element and the
--- new seed or 'Nothing' if there are no more elements.
---
--- > unfoldr (\n -> if n == 0 then Nothing else Just (n,n-1)) 10
--- >  = <10,9,8,7,6,5,4,3,2,1>
-unfoldr :: (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldr #-}
-unfoldr = G.unfoldr
-
--- | /O(n)/ Construct a vector with at most @n@ elements by repeatedly applying
--- the generator function to a seed. The generator function yields 'Just' the
--- next element and the new seed or 'Nothing' if there are no more elements.
---
--- > unfoldrN 3 (\n -> Just (n,n-1)) 10 = <10,9,8>
-unfoldrN :: Int -> (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldrN #-}
-unfoldrN = G.unfoldrN
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrM :: (Monad m) => (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrM #-}
-unfoldrM = G.unfoldrM
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrNM :: (Monad m) => Int -> (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrNM #-}
-unfoldrNM = G.unfoldrNM
-
--- | /O(n)/ Construct a vector with @n@ elements by repeatedly applying the
--- generator function to the already constructed part of the vector.
---
--- > constructN 3 f = let a = f <> ; b = f <a> ; c = f <a,b> in f <a,b,c>
---
-constructN :: Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructN #-}
-constructN = G.constructN
-
--- | /O(n)/ Construct a vector with @n@ elements from right to left by
--- repeatedly applying the generator function to the already constructed part
--- of the vector.
---
--- > constructrN 3 f = let a = f <> ; b = f<a> ; c = f <b,a> in f <c,b,a>
---
-constructrN :: Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructrN #-}
-constructrN = G.constructrN
-
--- Enumeration
--- -----------
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+1@
--- etc. This operation is usually more efficient than 'enumFromTo'.
---
--- > enumFromN 5 3 = <5,6,7>
-enumFromN :: Num a => a -> Int -> Vector a
-{-# INLINE enumFromN #-}
-enumFromN = G.enumFromN
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc. This operations is usually more efficient than 'enumFromThenTo'.
---
--- > enumFromStepN 1 0.1 5 = <1,1.1,1.2,1.3,1.4>
-enumFromStepN :: Num a => a -> a -> Int -> Vector a
-{-# INLINE enumFromStepN #-}
-enumFromStepN = G.enumFromStepN
-
--- | /O(n)/ Enumerate values from @x@ to @y@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromN' instead.
-enumFromTo :: Enum a => a -> a -> Vector a
-{-# INLINE enumFromTo #-}
-enumFromTo = G.enumFromTo
-
--- | /O(n)/ Enumerate values from @x@ to @y@ with a specific step @z@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: Enum a => a -> a -> a -> Vector a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo = G.enumFromThenTo
-
--- Concatenation
--- -------------
-
--- | /O(n)/ Prepend an element
-cons :: a -> Vector a -> Vector a
-{-# INLINE cons #-}
-cons = G.cons
-
--- | /O(n)/ Append an element
-snoc :: Vector a -> a -> Vector a
-{-# INLINE snoc #-}
-snoc = G.snoc
-
-infixr 5 ++
--- | /O(m+n)/ Concatenate two vectors
-(++) :: Vector a -> Vector a -> Vector a
-{-# INLINE (++) #-}
-(++) = (G.++)
-
--- | /O(n)/ Concatenate all vectors in the list
-concat :: [Vector a] -> Vector a
-{-# INLINE concat #-}
-concat = G.concat
-
--- Monadic initialisation
--- ----------------------
-
--- | /O(n)/ Execute the monadic action the given number of times and store the
--- results in a vector.
-replicateM :: Monad m => Int -> m a -> m (Vector a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | /O(n)/ Construct a vector of the given length by applying the monadic
--- action to each index
-generateM :: Monad m => Int -> (Int -> m a) -> m (Vector a)
-{-# INLINE generateM #-}
-generateM = G.generateM
-
--- | /O(n)/ Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: Monad m => Int -> (a -> m a) -> a -> m (Vector a)
-{-# INLINE iterateNM #-}
-iterateNM = G.iterateNM
-
--- | Execute the monadic action and freeze the resulting vector.
---
--- @
--- create (do { v \<- new 2; write v 0 \'a\'; write v 1 \'b\'; return v }) = \<'a','b'\>
--- @
-create :: (forall s. ST s (MVector s a)) -> Vector a
-{-# INLINE create #-}
--- NOTE: eta-expanded due to http://hackage.haskell.org/trac/ghc/ticket/4120
-create p = G.create p
-
--- | Execute the monadic action and freeze the resulting vectors.
-createT :: Traversable.Traversable f => (forall s. ST s (f (MVector s a))) -> f (Vector a)
-{-# INLINE createT #-}
-createT p = G.createT p
-
-
-
--- Restricting memory usage
--- ------------------------
-
--- | /O(n)/ Yield the argument but force it not to retain any extra memory,
--- possibly by copying it.
---
--- This is especially useful when dealing with slices. For example:
---
--- > force (slice 0 2 <huge vector>)
---
--- Here, the slice retains a reference to the huge vector. Forcing it creates
--- a copy of just the elements that belong to the slice and allows the huge
--- vector to be garbage collected.
-force :: Vector a -> Vector a
-{-# INLINE force #-}
-force = G.force
-
--- Bulk updates
--- ------------
-
--- | /O(m+n)/ For each pair @(i,a)@ from the list, replace the vector
--- element at position @i@ by @a@.
---
--- > <5,9,2,7> // [(2,1),(0,3),(2,8)] = <3,9,8,7>
---
-(//) :: Vector a   -- ^ initial vector (of length @m@)
-                -> [(Int, a)] -- ^ list of index/value pairs (of length @n@)
-                -> Vector a
-{-# INLINE (//) #-}
-(//) = (G.//)
-
--- | /O(m+n)/ For each pair @(i,a)@ from the vector of index/value pairs,
--- replace the vector element at position @i@ by @a@.
---
--- > update <5,9,2,7> <(2,1),(0,3),(2,8)> = <3,9,8,7>
---
-update :: Vector a        -- ^ initial vector (of length @m@)
-       -> Vector (Int, a) -- ^ vector of index/value pairs (of length @n@)
-       -> Vector a
-{-# INLINE update #-}
-update = G.update
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @a@ from the value vector, replace the element of the
--- initial vector at position @i@ by @a@.
---
--- > update_ <5,9,2,7>  <2,0,2> <1,3,8> = <3,9,8,7>
---
--- The function 'update' provides the same functionality and is usually more
--- convenient.
---
--- @
--- update_ xs is ys = 'update' xs ('zip' is ys)
--- @
-update_ :: Vector a   -- ^ initial vector (of length @m@)
-        -> Vector Int -- ^ index vector (of length @n1@)
-        -> Vector a   -- ^ value vector (of length @n2@)
-        -> Vector a
-{-# INLINE update_ #-}
-update_ = G.update_
-
--- | Same as ('//') but without bounds checking.
-unsafeUpd :: Vector a -> [(Int, a)] -> Vector a
-{-# INLINE unsafeUpd #-}
-unsafeUpd = G.unsafeUpd
-
--- | Same as 'update' but without bounds checking.
-unsafeUpdate :: Vector a -> Vector (Int, a) -> Vector a
-{-# INLINE unsafeUpdate #-}
-unsafeUpdate = G.unsafeUpdate
-
--- | Same as 'update_' but without bounds checking.
-unsafeUpdate_ :: Vector a -> Vector Int -> Vector a -> Vector a
-{-# INLINE unsafeUpdate_ #-}
-unsafeUpdate_ = G.unsafeUpdate_
-
--- Accumulations
--- -------------
-
--- | /O(m+n)/ For each pair @(i,b)@ from the list, replace the vector element
--- @a@ at position @i@ by @f a b@.
---
--- > accum (+) <5,9,2> [(2,4),(1,6),(0,3),(1,7)] = <5+3, 9+6+7, 2+4>
-accum :: (a -> b -> a) -- ^ accumulating function @f@
-      -> Vector a      -- ^ initial vector (of length @m@)
-      -> [(Int,b)]     -- ^ list of index/value pairs (of length @n@)
-      -> Vector a
-{-# INLINE accum #-}
-accum = G.accum
-
--- | /O(m+n)/ For each pair @(i,b)@ from the vector of pairs, replace the vector
--- element @a@ at position @i@ by @f a b@.
---
--- > accumulate (+) <5,9,2> <(2,4),(1,6),(0,3),(1,7)> = <5+3, 9+6+7, 2+4>
-accumulate :: (a -> b -> a)  -- ^ accumulating function @f@
-            -> Vector a       -- ^ initial vector (of length @m@)
-            -> Vector (Int,b) -- ^ vector of index/value pairs (of length @n@)
-            -> Vector a
-{-# INLINE accumulate #-}
-accumulate = G.accumulate
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @b@ from the the value vector,
--- replace the element of the initial vector at
--- position @i@ by @f a b@.
---
--- > accumulate_ (+) <5,9,2> <2,1,0,1> <4,6,3,7> = <5+3, 9+6+7, 2+4>
---
--- The function 'accumulate' provides the same functionality and is usually more
--- convenient.
---
--- @
--- accumulate_ f as is bs = 'accumulate' f as ('zip' is bs)
--- @
-accumulate_ :: (a -> b -> a) -- ^ accumulating function @f@
-            -> Vector a      -- ^ initial vector (of length @m@)
-            -> Vector Int    -- ^ index vector (of length @n1@)
-            -> Vector b      -- ^ value vector (of length @n2@)
-            -> Vector a
-{-# INLINE accumulate_ #-}
-accumulate_ = G.accumulate_
-
--- | Same as 'accum' but without bounds checking.
-unsafeAccum :: (a -> b -> a) -> Vector a -> [(Int,b)] -> Vector a
-{-# INLINE unsafeAccum #-}
-unsafeAccum = G.unsafeAccum
-
--- | Same as 'accumulate' but without bounds checking.
-unsafeAccumulate :: (a -> b -> a) -> Vector a -> Vector (Int,b) -> Vector a
-{-# INLINE unsafeAccumulate #-}
-unsafeAccumulate = G.unsafeAccumulate
-
--- | Same as 'accumulate_' but without bounds checking.
-unsafeAccumulate_
-  :: (a -> b -> a) -> Vector a -> Vector Int -> Vector b -> Vector a
-{-# INLINE unsafeAccumulate_ #-}
-unsafeAccumulate_ = G.unsafeAccumulate_
-
--- Permutations
--- ------------
-
--- | /O(n)/ Reverse a vector
-reverse :: Vector a -> Vector a
-{-# INLINE reverse #-}
-reverse = G.reverse
-
--- | /O(n)/ Yield the vector obtained by replacing each element @i@ of the
--- index vector by @xs'!'i@. This is equivalent to @'map' (xs'!') is@ but is
--- often much more efficient.
---
--- > backpermute <a,b,c,d> <0,3,2,3,1,0> = <a,d,c,d,b,a>
-backpermute :: Vector a -> Vector Int -> Vector a
-{-# INLINE backpermute #-}
-backpermute = G.backpermute
-
--- | Same as 'backpermute' but without bounds checking.
-unsafeBackpermute :: Vector a -> Vector Int -> Vector a
-{-# INLINE unsafeBackpermute #-}
-unsafeBackpermute = G.unsafeBackpermute
-
--- Safe destructive updates
--- ------------------------
-
--- | Apply a destructive operation to a vector. The operation will be
--- performed in place if it is safe to do so and will modify a copy of the
--- vector otherwise.
---
--- @
--- modify (\\v -> write v 0 \'x\') ('replicate' 3 \'a\') = \<\'x\',\'a\',\'a\'\>
--- @
-modify :: (forall s. MVector s a -> ST s ()) -> Vector a -> Vector a
-{-# INLINE modify #-}
-modify p = G.modify p
-
--- Indexing
--- --------
-
--- | /O(n)/ Pair each element in a vector with its index
-indexed :: Vector a -> Vector (Int,a)
-{-# INLINE indexed #-}
-indexed = G.indexed
-
--- Mapping
--- -------
-
--- | /O(n)/ Map a function over a vector
-map :: (a -> b) -> Vector a -> Vector b
-{-# INLINE map #-}
-map = G.map
-
--- | /O(n)/ Apply a function to every element of a vector and its index
-imap :: (Int -> a -> b) -> Vector a -> Vector b
-{-# INLINE imap #-}
-imap = G.imap
-
--- | Map a function over a vector and concatenate the results.
-concatMap :: (a -> Vector b) -> Vector a -> Vector b
-{-# INLINE concatMap #-}
-concatMap = G.concatMap
-
--- Monadic mapping
--- ---------------
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results
-mapM :: Monad m => (a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE mapM #-}
-mapM = G.mapM
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, yielding a vector of results
-imapM :: Monad m => (Int -> a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE imapM #-}
-imapM = G.imapM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results
-mapM_ :: Monad m => (a -> m b) -> Vector a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ = G.mapM_
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, ignoring the results
-imapM_ :: Monad m => (Int -> a -> m b) -> Vector a -> m ()
-{-# INLINE imapM_ #-}
-imapM_ = G.imapM_
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results. Equivalent to @flip 'mapM'@.
-forM :: Monad m => Vector a -> (a -> m b) -> m (Vector b)
-{-# INLINE forM #-}
-forM = G.forM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results. Equivalent to @flip 'mapM_'@.
-forM_ :: Monad m => Vector a -> (a -> m b) -> m ()
-{-# INLINE forM_ #-}
-forM_ = G.forM_
-
--- Zipping
--- -------
-
--- | /O(min(m,n))/ Zip two vectors with the given function.
-zipWith :: (a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE zipWith #-}
-zipWith = G.zipWith
-
--- | Zip three vectors with the given function.
-zipWith3 :: (a -> b -> c -> d) -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE zipWith3 #-}
-zipWith3 = G.zipWith3
-
-zipWith4 :: (a -> b -> c -> d -> e)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE zipWith4 #-}
-zipWith4 = G.zipWith4
-
-zipWith5 :: (a -> b -> c -> d -> e -> f)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f
-{-# INLINE zipWith5 #-}
-zipWith5 = G.zipWith5
-
-zipWith6 :: (a -> b -> c -> d -> e -> f -> g)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f -> Vector g
-{-# INLINE zipWith6 #-}
-zipWith6 = G.zipWith6
-
--- | /O(min(m,n))/ Zip two vectors with a function that also takes the
--- elements' indices.
-izipWith :: (Int -> a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE izipWith #-}
-izipWith = G.izipWith
-
--- | Zip three vectors and their indices with the given function.
-izipWith3 :: (Int -> a -> b -> c -> d)
-          -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE izipWith3 #-}
-izipWith3 = G.izipWith3
-
-izipWith4 :: (Int -> a -> b -> c -> d -> e)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE izipWith4 #-}
-izipWith4 = G.izipWith4
-
-izipWith5 :: (Int -> a -> b -> c -> d -> e -> f)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f
-{-# INLINE izipWith5 #-}
-izipWith5 = G.izipWith5
-
-izipWith6 :: (Int -> a -> b -> c -> d -> e -> f -> g)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f -> Vector g
-{-# INLINE izipWith6 #-}
-izipWith6 = G.izipWith6
-
--- | Elementwise pairing of array elements.
-zip :: Vector a -> Vector b -> Vector (a, b)
-{-# INLINE zip #-}
-zip = G.zip
-
--- | zip together three vectors into a vector of triples
-zip3 :: Vector a -> Vector b -> Vector c -> Vector (a, b, c)
-{-# INLINE zip3 #-}
-zip3 = G.zip3
-
-zip4 :: Vector a -> Vector b -> Vector c -> Vector d
-     -> Vector (a, b, c, d)
-{-# INLINE zip4 #-}
-zip4 = G.zip4
-
-zip5 :: Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-     -> Vector (a, b, c, d, e)
-{-# INLINE zip5 #-}
-zip5 = G.zip5
-
-zip6 :: Vector a -> Vector b -> Vector c -> Vector d -> Vector e -> Vector f
-     -> Vector (a, b, c, d, e, f)
-{-# INLINE zip6 #-}
-zip6 = G.zip6
-
--- Unzipping
--- ---------
-
--- | /O(min(m,n))/ Unzip a vector of pairs.
-unzip :: Vector (a, b) -> (Vector a, Vector b)
-{-# INLINE unzip #-}
-unzip = G.unzip
-
-unzip3 :: Vector (a, b, c) -> (Vector a, Vector b, Vector c)
-{-# INLINE unzip3 #-}
-unzip3 = G.unzip3
-
-unzip4 :: Vector (a, b, c, d) -> (Vector a, Vector b, Vector c, Vector d)
-{-# INLINE unzip4 #-}
-unzip4 = G.unzip4
-
-unzip5 :: Vector (a, b, c, d, e)
-       -> (Vector a, Vector b, Vector c, Vector d, Vector e)
-{-# INLINE unzip5 #-}
-unzip5 = G.unzip5
-
-unzip6 :: Vector (a, b, c, d, e, f)
-       -> (Vector a, Vector b, Vector c, Vector d, Vector e, Vector f)
-{-# INLINE unzip6 #-}
-unzip6 = G.unzip6
-
--- Monadic zipping
--- ---------------
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and yield a
--- vector of results
-zipWithM :: Monad m => (a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE zipWithM #-}
-zipWithM = G.zipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and yield a vector of results
-izipWithM :: Monad m => (Int -> a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE izipWithM #-}
-izipWithM = G.izipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and ignore the
--- results
-zipWithM_ :: Monad m => (a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ = G.zipWithM_
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and ignore the results
-izipWithM_ :: Monad m => (Int -> a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE izipWithM_ #-}
-izipWithM_ = G.izipWithM_
-
--- Filtering
--- ---------
-
--- | /O(n)/ Drop elements that do not satisfy the predicate
-filter :: (a -> Bool) -> Vector a -> Vector a
-{-# INLINE filter #-}
-filter = G.filter
-
--- | /O(n)/ Drop elements that do not satisfy the predicate which is applied to
--- values and their indices
-ifilter :: (Int -> a -> Bool) -> Vector a -> Vector a
-{-# INLINE ifilter #-}
-ifilter = G.ifilter
-
--- | /O(n)/ Drop repeated adjacent elements.
-uniq :: (Eq a) => Vector a -> Vector a
-{-# INLINE uniq #-}
-uniq = G.uniq
-
--- | /O(n)/ Drop elements when predicate returns Nothing
-mapMaybe :: (a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE mapMaybe #-}
-mapMaybe = G.mapMaybe
-
--- | /O(n)/ Drop elements when predicate, applied to index and value, returns Nothing
-imapMaybe :: (Int -> a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE imapMaybe #-}
-imapMaybe = G.imapMaybe
-
--- | /O(n)/ Drop elements that do not satisfy the monadic predicate
-filterM :: Monad m => (a -> m Bool) -> Vector a -> m (Vector a)
-{-# INLINE filterM #-}
-filterM = G.filterM
-
--- | /O(n)/ Yield the longest prefix of elements satisfying the predicate
--- without copying.
-takeWhile :: (a -> Bool) -> Vector a -> Vector a
-{-# INLINE takeWhile #-}
-takeWhile = G.takeWhile
-
--- | /O(n)/ Drop the longest prefix of elements that satisfy the predicate
--- without copying.
-dropWhile :: (a -> Bool) -> Vector a -> Vector a
-{-# INLINE dropWhile #-}
-dropWhile = G.dropWhile
-
--- Parititioning
--- -------------
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't. The
--- relative order of the elements is preserved at the cost of a sometimes
--- reduced performance compared to 'unstablePartition'.
-partition :: (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE partition #-}
-partition = G.partition
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't.
--- The order of the elements is not preserved but the operation is often
--- faster than 'partition'.
-unstablePartition :: (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE unstablePartition #-}
-unstablePartition = G.unstablePartition
-
--- | /O(n)/ Split the vector into the longest prefix of elements that satisfy
--- the predicate and the rest without copying.
-span :: (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE span #-}
-span = G.span
-
--- | /O(n)/ Split the vector into the longest prefix of elements that do not
--- satisfy the predicate and the rest without copying.
-break :: (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE break #-}
-break = G.break
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | /O(n)/ Check if the vector contains an element
-elem :: Eq a => a -> Vector a -> Bool
-{-# INLINE elem #-}
-elem = G.elem
-
-infix 4 `notElem`
--- | /O(n)/ Check if the vector does not contain an element (inverse of 'elem')
-notElem :: Eq a => a -> Vector a -> Bool
-{-# INLINE notElem #-}
-notElem = G.notElem
-
--- | /O(n)/ Yield 'Just' the first element matching the predicate or 'Nothing'
--- if no such element exists.
-find :: (a -> Bool) -> Vector a -> Maybe a
-{-# INLINE find #-}
-find = G.find
-
--- | /O(n)/ Yield 'Just' the index of the first element matching the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: (a -> Bool) -> Vector a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex = G.findIndex
-
--- | /O(n)/ Yield the indices of elements satisfying the predicate in ascending
--- order.
-findIndices :: (a -> Bool) -> Vector a -> Vector Int
-{-# INLINE findIndices #-}
-findIndices = G.findIndices
-
--- | /O(n)/ Yield 'Just' the index of the first occurence of the given element or
--- 'Nothing' if the vector does not contain the element. This is a specialised
--- version of 'findIndex'.
-elemIndex :: Eq a => a -> Vector a -> Maybe Int
-{-# INLINE elemIndex #-}
-elemIndex = G.elemIndex
-
--- | /O(n)/ Yield the indices of all occurences of the given element in
--- ascending order. This is a specialised version of 'findIndices'.
-elemIndices :: Eq a => a -> Vector a -> Vector Int
-{-# INLINE elemIndices #-}
-elemIndices = G.elemIndices
-
--- Folding
--- -------
-
--- | /O(n)/ Left fold
-foldl :: (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl #-}
-foldl = G.foldl
-
--- | /O(n)/ Left fold on non-empty vectors
-foldl1 :: (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1 #-}
-foldl1 = G.foldl1
-
--- | /O(n)/ Left fold with strict accumulator
-foldl' :: (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl' #-}
-foldl' = G.foldl'
-
--- | /O(n)/ Left fold on non-empty vectors with strict accumulator
-foldl1' :: (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1' #-}
-foldl1' = G.foldl1'
-
--- | /O(n)/ Right fold
-foldr :: (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr #-}
-foldr = G.foldr
-
--- | /O(n)/ Right fold on non-empty vectors
-foldr1 :: (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1 #-}
-foldr1 = G.foldr1
-
--- | /O(n)/ Right fold with a strict accumulator
-foldr' :: (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr' #-}
-foldr' = G.foldr'
-
--- | /O(n)/ Right fold on non-empty vectors with strict accumulator
-foldr1' :: (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1' #-}
-foldr1' = G.foldr1'
-
--- | /O(n)/ Left fold (function applied to each element and its index)
-ifoldl :: (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl #-}
-ifoldl = G.ifoldl
-
--- | /O(n)/ Left fold with strict accumulator (function applied to each element
--- and its index)
-ifoldl' :: (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl' #-}
-ifoldl' = G.ifoldl'
-
--- | /O(n)/ Right fold (function applied to each element and its index)
-ifoldr :: (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr #-}
-ifoldr = G.ifoldr
-
--- | /O(n)/ Right fold with strict accumulator (function applied to each
--- element and its index)
-ifoldr' :: (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr' #-}
-ifoldr' = G.ifoldr'
-
--- Specialised folds
--- -----------------
-
--- | /O(n)/ Check if all elements satisfy the predicate.
-all :: (a -> Bool) -> Vector a -> Bool
-{-# INLINE all #-}
-all = G.all
-
--- | /O(n)/ Check if any element satisfies the predicate.
-any :: (a -> Bool) -> Vector a -> Bool
-{-# INLINE any #-}
-any = G.any
-
--- | /O(n)/ Check if all elements are 'True'
-and :: Vector Bool -> Bool
-{-# INLINE and #-}
-and = G.and
-
--- | /O(n)/ Check if any element is 'True'
-or :: Vector Bool -> Bool
-{-# INLINE or #-}
-or = G.or
-
--- | /O(n)/ Compute the sum of the elements
-sum :: Num a => Vector a -> a
-{-# INLINE sum #-}
-sum = G.sum
-
--- | /O(n)/ Compute the produce of the elements
-product :: Num a => Vector a -> a
-{-# INLINE product #-}
-product = G.product
-
--- | /O(n)/ Yield the maximum element of the vector. The vector may not be
--- empty.
-maximum :: Ord a => Vector a -> a
-{-# INLINE maximum #-}
-maximum = G.maximum
-
--- | /O(n)/ Yield the maximum element of the vector according to the given
--- comparison function. The vector may not be empty.
-maximumBy :: (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE maximumBy #-}
-maximumBy = G.maximumBy
-
--- | /O(n)/ Yield the minimum element of the vector. The vector may not be
--- empty.
-minimum :: Ord a => Vector a -> a
-{-# INLINE minimum #-}
-minimum = G.minimum
-
--- | /O(n)/ Yield the minimum element of the vector according to the given
--- comparison function. The vector may not be empty.
-minimumBy :: (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE minimumBy #-}
-minimumBy = G.minimumBy
-
--- | /O(n)/ Yield the index of the maximum element of the vector. The vector
--- may not be empty.
-maxIndex :: Ord a => Vector a -> Int
-{-# INLINE maxIndex #-}
-maxIndex = G.maxIndex
-
--- | /O(n)/ Yield the index of the maximum element of the vector according to
--- the given comparison function. The vector may not be empty.
-maxIndexBy :: (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE maxIndexBy #-}
-maxIndexBy = G.maxIndexBy
-
--- | /O(n)/ Yield the index of the minimum element of the vector. The vector
--- may not be empty.
-minIndex :: Ord a => Vector a -> Int
-{-# INLINE minIndex #-}
-minIndex = G.minIndex
-
--- | /O(n)/ Yield the index of the minimum element of the vector according to
--- the given comparison function. The vector may not be empty.
-minIndexBy :: (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE minIndexBy #-}
-minIndexBy = G.minIndexBy
-
--- Monadic folds
--- -------------
-
--- | /O(n)/ Monadic fold
-foldM :: Monad m => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM #-}
-foldM = G.foldM
-
--- | /O(n)/ Monadic fold (action applied to each element and its index)
-ifoldM :: Monad m => (a -> Int -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE ifoldM #-}
-ifoldM = G.ifoldM
-
--- | /O(n)/ Monadic fold over non-empty vectors
-fold1M :: Monad m => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M #-}
-fold1M = G.fold1M
-
--- | /O(n)/ Monadic fold with strict accumulator
-foldM' :: Monad m => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM' #-}
-foldM' = G.foldM'
-
--- | /O(n)/ Monadic fold with strict accumulator (action applied to each
--- element and its index)
-ifoldM' :: Monad m => (a -> Int -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE ifoldM' #-}
-ifoldM' = G.ifoldM'
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
-fold1M' :: Monad m => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = G.fold1M'
-
--- | /O(n)/ Monadic fold that discards the result
-foldM_ :: Monad m => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM_ #-}
-foldM_ = G.foldM_
-
--- | /O(n)/ Monadic fold that discards the result (action applied to each
--- element and its index)
-ifoldM_ :: Monad m => (a -> Int -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE ifoldM_ #-}
-ifoldM_ = G.ifoldM_
-
--- | /O(n)/ Monadic fold over non-empty vectors that discards the result
-fold1M_ :: Monad m => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M_ #-}
-fold1M_ = G.fold1M_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
-foldM'_ :: Monad m => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM'_ #-}
-foldM'_ = G.foldM'_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
--- (action applied to each element and its index)
-ifoldM'_ :: Monad m => (a -> Int -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE ifoldM'_ #-}
-ifoldM'_ = G.ifoldM'_
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
--- that discards the result
-fold1M'_ :: Monad m => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M'_ #-}
-fold1M'_ = G.fold1M'_
-
--- Monadic sequencing
--- ------------------
-
--- | Evaluate each action and collect the results
-sequence :: Monad m => Vector (m a) -> m (Vector a)
-{-# INLINE sequence #-}
-sequence = G.sequence
-
--- | Evaluate each action and discard the results
-sequence_ :: Monad m => Vector (m a) -> m ()
-{-# INLINE sequence_ #-}
-sequence_ = G.sequence_
-
--- Prefix sums (scans)
--- -------------------
-
--- | /O(n)/ Prescan
---
--- @
--- prescanl f z = 'init' . 'scanl' f z
--- @
---
--- Example: @prescanl (+) 0 \<1,2,3,4\> = \<0,1,3,6\>@
---
-prescanl :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl #-}
-prescanl = G.prescanl
-
--- | /O(n)/ Prescan with strict accumulator
-prescanl' :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl' #-}
-prescanl' = G.prescanl'
-
--- | /O(n)/ Scan
---
--- @
--- postscanl f z = 'tail' . 'scanl' f z
--- @
---
--- Example: @postscanl (+) 0 \<1,2,3,4\> = \<1,3,6,10\>@
---
-postscanl :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl #-}
-postscanl = G.postscanl
-
--- | /O(n)/ Scan with strict accumulator
-postscanl' :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl' #-}
-postscanl' = G.postscanl'
-
--- | /O(n)/ Haskell-style scan
---
--- > scanl f z <x1,...,xn> = <y1,...,y(n+1)>
--- >   where y1 = z
--- >         yi = f y(i-1) x(i-1)
---
--- Example: @scanl (+) 0 \<1,2,3,4\> = \<0,1,3,6,10\>@
---
-scanl :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl #-}
-scanl = G.scanl
-
--- | /O(n)/ Haskell-style scan with strict accumulator
-scanl' :: (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl' #-}
-scanl' = G.scanl'
-
--- | /O(n)/ Scan over a vector with its index
-iscanl :: (Int -> a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE iscanl #-}
-iscanl = G.iscanl
-
--- | /O(n)/ Scan over a vector (strictly) with its index
-iscanl' :: (Int -> a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE iscanl' #-}
-iscanl' = G.iscanl'
-
--- | /O(n)/ Scan over a non-empty vector
---
--- > scanl f <x1,...,xn> = <y1,...,yn>
--- >   where y1 = x1
--- >         yi = f y(i-1) xi
---
-scanl1 :: (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1 #-}
-scanl1 = G.scanl1
-
--- | /O(n)/ Scan over a non-empty vector with a strict accumulator
-scanl1' :: (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1' #-}
-scanl1' = G.scanl1'
-
--- | /O(n)/ Right-to-left prescan
---
--- @
--- prescanr f z = 'reverse' . 'prescanl' (flip f) z . 'reverse'
--- @
---
-prescanr :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr #-}
-prescanr = G.prescanr
-
--- | /O(n)/ Right-to-left prescan with strict accumulator
-prescanr' :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr' #-}
-prescanr' = G.prescanr'
-
--- | /O(n)/ Right-to-left scan
-postscanr :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr #-}
-postscanr = G.postscanr
-
--- | /O(n)/ Right-to-left scan with strict accumulator
-postscanr' :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr' #-}
-postscanr' = G.postscanr'
-
--- | /O(n)/ Right-to-left Haskell-style scan
-scanr :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr #-}
-scanr = G.scanr
-
--- | /O(n)/ Right-to-left Haskell-style scan with strict accumulator
-scanr' :: (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr' #-}
-scanr' = G.scanr'
-
--- | /O(n)/ Right-to-left scan over a vector with its index
-iscanr :: (Int -> a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE iscanr #-}
-iscanr = G.iscanr
-
--- | /O(n)/ Right-to-left scan over a vector (strictly) with its index
-iscanr' :: (Int -> a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE iscanr' #-}
-iscanr' = G.iscanr'
-
--- | /O(n)/ Right-to-left scan over a non-empty vector
-scanr1 :: (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1 #-}
-scanr1 = G.scanr1
-
--- | /O(n)/ Right-to-left scan over a non-empty vector with a strict
--- accumulator
-scanr1' :: (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1' #-}
-scanr1' = G.scanr1'
-
--- Conversions - Lists
--- ------------------------
-
--- | /O(n)/ Convert a vector to a list
-toList :: Vector a -> [a]
-{-# INLINE toList #-}
-toList = G.toList
-
--- | /O(n)/ Convert a list to a vector
-fromList :: [a] -> Vector a
-{-# INLINE fromList #-}
-fromList = G.fromList
-
--- | /O(n)/ Convert the first @n@ elements of a list to a vector
---
--- @
--- fromListN n xs = 'fromList' ('take' n xs)
--- @
-fromListN :: Int -> [a] -> Vector a
-{-# INLINE fromListN #-}
-fromListN = G.fromListN
-
--- Conversions - Mutable vectors
--- -----------------------------
-
--- | /O(1)/ Unsafe convert a mutable vector to an immutable one without
--- copying. The mutable vector may not be used after this operation.
-unsafeFreeze :: PrimMonad m => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE unsafeFreeze #-}
-unsafeFreeze = G.unsafeFreeze
-
--- | /O(1)/ Unsafely convert an immutable vector to a mutable one without
--- copying. The immutable vector may not be used after this operation.
-unsafeThaw :: PrimMonad m => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE unsafeThaw #-}
-unsafeThaw = G.unsafeThaw
-
--- | /O(n)/ Yield a mutable copy of the immutable vector.
-thaw :: PrimMonad m => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE thaw #-}
-thaw = G.thaw
-
--- | /O(n)/ Yield an immutable copy of the mutable vector.
-freeze :: PrimMonad m => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE freeze #-}
-freeze = G.freeze
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length. This is not checked.
-unsafeCopy :: PrimMonad m => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length.
-copy :: PrimMonad m => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE copy #-}
-copy = G.copy
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle.hs
deleted file mode 100644
index 6b6b6236d7cb..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle.hs
+++ /dev/null
@@ -1,655 +0,0 @@
-{-# LANGUAGE CPP, FlexibleInstances, Rank2Types, BangPatterns #-}
-
--- |
--- Module      : Data.Vector.Fusion.Bundle
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Bundles for stream fusion
---
-
-module Data.Vector.Fusion.Bundle (
-  -- * Types
-  Step(..), Chunk(..), Bundle, MBundle,
-
-  -- * In-place markers
-  inplace,
-
-  -- * Size hints
-  size, sized,
-
-  -- * Length information
-  length, null,
-
-  -- * Construction
-  empty, singleton, cons, snoc, replicate, generate, (++),
-
-  -- * Accessing individual elements
-  head, last, (!!), (!?),
-
-  -- * Substreams
-  slice, init, tail, take, drop,
-
-  -- * Mapping
-  map, concatMap, flatten, unbox,
-
-  -- * Zipping
-  indexed, indexedR,
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- * Filtering
-  filter, takeWhile, dropWhile,
-
-  -- * Searching
-  elem, notElem, find, findIndex,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1,
-
-  -- * Specialised folds
-  and, or,
-
-  -- * Unfolding
-  unfoldr, unfoldrN, iterateN,
-
-  -- * Scans
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl',
-  scanl1, scanl1',
-
-  -- * Enumerations
-  enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- * Conversions
-  toList, fromList, fromListN, unsafeFromList, lift,
-  fromVector, reVector, fromVectors, concatVectors,
-
-  -- * Monadic combinators
-  mapM, mapM_, zipWithM, zipWithM_, filterM, foldM, fold1M, foldM', fold1M',
-
-  eq, cmp, eqBy, cmpBy
-) where
-
-import Data.Vector.Generic.Base ( Vector )
-import Data.Vector.Fusion.Bundle.Size
-import Data.Vector.Fusion.Util
-import Data.Vector.Fusion.Stream.Monadic ( Stream(..), Step(..) )
-import Data.Vector.Fusion.Bundle.Monadic ( Chunk(..) )
-import qualified Data.Vector.Fusion.Bundle.Monadic as M
-import qualified Data.Vector.Fusion.Stream.Monadic as S
-
-import Prelude hiding ( length, null,
-                        replicate, (++),
-                        head, last, (!!),
-                        init, tail, take, drop,
-                        map, concatMap,
-                        zipWith, zipWith3, zip, zip3,
-                        filter, takeWhile, dropWhile,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        and, or,
-                        scanl, scanl1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_ )
-
-#if MIN_VERSION_base(4,9,0)
-import Data.Functor.Classes (Eq1 (..), Ord1 (..))
-#endif
-
-import GHC.Base ( build )
-
--- Data.Vector.Internal.Check is unused
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- | The type of pure streams
-type Bundle = M.Bundle Id
-
--- | Alternative name for monadic streams
-type MBundle = M.Bundle
-
-inplace :: (forall m. Monad m => S.Stream m a -> S.Stream m b)
-        -> (Size -> Size) -> Bundle v a -> Bundle v b
-{-# INLINE_FUSED inplace #-}
-inplace f g b = b `seq` M.fromStream (f (M.elements b)) (g (M.size b))
-
-{-# RULES
-
-"inplace/inplace [Vector]"
-  forall (f1 :: forall m. Monad m => S.Stream m a -> S.Stream m a)
-         (f2 :: forall m. Monad m => S.Stream m a -> S.Stream m a)
-         g1 g2 s.
-  inplace f1 g1 (inplace f2 g2 s) = inplace (f1 . f2) (g1 . g2) s   #-}
-
-
-
--- | Convert a pure stream to a monadic stream
-lift :: Monad m => Bundle v a -> M.Bundle m v a
-{-# INLINE_FUSED lift #-}
-lift (M.Bundle (Stream step s) (Stream vstep t) v sz)
-    = M.Bundle (Stream (return . unId . step) s)
-               (Stream (return . unId . vstep) t) v sz
-
--- | 'Size' hint of a 'Bundle'
-size :: Bundle v a -> Size
-{-# INLINE size #-}
-size = M.size
-
--- | Attach a 'Size' hint to a 'Bundle'
-sized :: Bundle v a -> Size -> Bundle v a
-{-# INLINE sized #-}
-sized = M.sized
-
--- Length
--- ------
-
--- | Length of a 'Bundle'
-length :: Bundle v a -> Int
-{-# INLINE length #-}
-length = unId . M.length
-
--- | Check if a 'Bundle' is empty
-null :: Bundle v a -> Bool
-{-# INLINE null #-}
-null = unId . M.null
-
--- Construction
--- ------------
-
--- | Empty 'Bundle'
-empty :: Bundle v a
-{-# INLINE empty #-}
-empty = M.empty
-
--- | Singleton 'Bundle'
-singleton :: a -> Bundle v a
-{-# INLINE singleton #-}
-singleton = M.singleton
-
--- | Replicate a value to a given length
-replicate :: Int -> a -> Bundle v a
-{-# INLINE replicate #-}
-replicate = M.replicate
-
--- | Generate a stream from its indices
-generate :: Int -> (Int -> a) -> Bundle v a
-{-# INLINE generate #-}
-generate = M.generate
-
--- | Prepend an element
-cons :: a -> Bundle v a -> Bundle v a
-{-# INLINE cons #-}
-cons = M.cons
-
--- | Append an element
-snoc :: Bundle v a -> a -> Bundle v a
-{-# INLINE snoc #-}
-snoc = M.snoc
-
-infixr 5 ++
--- | Concatenate two 'Bundle's
-(++) :: Bundle v a -> Bundle v a -> Bundle v a
-{-# INLINE (++) #-}
-(++) = (M.++)
-
--- Accessing elements
--- ------------------
-
--- | First element of the 'Bundle' or error if empty
-head :: Bundle v a -> a
-{-# INLINE head #-}
-head = unId . M.head
-
--- | Last element of the 'Bundle' or error if empty
-last :: Bundle v a -> a
-{-# INLINE last #-}
-last = unId . M.last
-
-infixl 9 !!
--- | Element at the given position
-(!!) :: Bundle v a -> Int -> a
-{-# INLINE (!!) #-}
-s !! i = unId (s M.!! i)
-
-infixl 9 !?
--- | Element at the given position or 'Nothing' if out of bounds
-(!?) :: Bundle v a -> Int -> Maybe a
-{-# INLINE (!?) #-}
-s !? i = unId (s M.!? i)
-
--- Substreams
--- ----------
-
--- | Extract a substream of the given length starting at the given position.
-slice :: Int   -- ^ starting index
-      -> Int   -- ^ length
-      -> Bundle v a
-      -> Bundle v a
-{-# INLINE slice #-}
-slice = M.slice
-
--- | All but the last element
-init :: Bundle v a -> Bundle v a
-{-# INLINE init #-}
-init = M.init
-
--- | All but the first element
-tail :: Bundle v a -> Bundle v a
-{-# INLINE tail #-}
-tail = M.tail
-
--- | The first @n@ elements
-take :: Int -> Bundle v a -> Bundle v a
-{-# INLINE take #-}
-take = M.take
-
--- | All but the first @n@ elements
-drop :: Int -> Bundle v a -> Bundle v a
-{-# INLINE drop #-}
-drop = M.drop
-
--- Mapping
--- ---------------
-
--- | Map a function over a 'Bundle'
-map :: (a -> b) -> Bundle v a -> Bundle v b
-{-# INLINE map #-}
-map = M.map
-
-unbox :: Bundle v (Box a) -> Bundle v a
-{-# INLINE unbox #-}
-unbox = M.unbox
-
-concatMap :: (a -> Bundle v b) -> Bundle v a -> Bundle v b
-{-# INLINE concatMap #-}
-concatMap = M.concatMap
-
--- Zipping
--- -------
-
--- | Pair each element in a 'Bundle' with its index
-indexed :: Bundle v a -> Bundle v (Int,a)
-{-# INLINE indexed #-}
-indexed = M.indexed
-
--- | Pair each element in a 'Bundle' with its index, starting from the right
--- and counting down
-indexedR :: Int -> Bundle v a -> Bundle v (Int,a)
-{-# INLINE_FUSED indexedR #-}
-indexedR = M.indexedR
-
--- | Zip two 'Bundle's with the given function
-zipWith :: (a -> b -> c) -> Bundle v a -> Bundle v b -> Bundle v c
-{-# INLINE zipWith #-}
-zipWith = M.zipWith
-
--- | Zip three 'Bundle's with the given function
-zipWith3 :: (a -> b -> c -> d) -> Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-{-# INLINE zipWith3 #-}
-zipWith3 = M.zipWith3
-
-zipWith4 :: (a -> b -> c -> d -> e)
-                    -> Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                    -> Bundle v e
-{-# INLINE zipWith4 #-}
-zipWith4 = M.zipWith4
-
-zipWith5 :: (a -> b -> c -> d -> e -> f)
-                    -> Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                    -> Bundle v e -> Bundle v f
-{-# INLINE zipWith5 #-}
-zipWith5 = M.zipWith5
-
-zipWith6 :: (a -> b -> c -> d -> e -> f -> g)
-                    -> Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                    -> Bundle v e -> Bundle v f -> Bundle v g
-{-# INLINE zipWith6 #-}
-zipWith6 = M.zipWith6
-
-zip :: Bundle v a -> Bundle v b -> Bundle v (a,b)
-{-# INLINE zip #-}
-zip = M.zip
-
-zip3 :: Bundle v a -> Bundle v b -> Bundle v c -> Bundle v (a,b,c)
-{-# INLINE zip3 #-}
-zip3 = M.zip3
-
-zip4 :: Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                -> Bundle v (a,b,c,d)
-{-# INLINE zip4 #-}
-zip4 = M.zip4
-
-zip5 :: Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                -> Bundle v e -> Bundle v (a,b,c,d,e)
-{-# INLINE zip5 #-}
-zip5 = M.zip5
-
-zip6 :: Bundle v a -> Bundle v b -> Bundle v c -> Bundle v d
-                -> Bundle v e -> Bundle v f -> Bundle v (a,b,c,d,e,f)
-{-# INLINE zip6 #-}
-zip6 = M.zip6
-
--- Filtering
--- ---------
-
--- | Drop elements which do not satisfy the predicate
-filter :: (a -> Bool) -> Bundle v a -> Bundle v a
-{-# INLINE filter #-}
-filter = M.filter
-
--- | Longest prefix of elements that satisfy the predicate
-takeWhile :: (a -> Bool) -> Bundle v a -> Bundle v a
-{-# INLINE takeWhile #-}
-takeWhile = M.takeWhile
-
--- | Drop the longest prefix of elements that satisfy the predicate
-dropWhile :: (a -> Bool) -> Bundle v a -> Bundle v a
-{-# INLINE dropWhile #-}
-dropWhile = M.dropWhile
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | Check whether the 'Bundle' contains an element
-elem :: Eq a => a -> Bundle v a -> Bool
-{-# INLINE elem #-}
-elem x = unId . M.elem x
-
-infix 4 `notElem`
--- | Inverse of `elem`
-notElem :: Eq a => a -> Bundle v a -> Bool
-{-# INLINE notElem #-}
-notElem x = unId . M.notElem x
-
--- | Yield 'Just' the first element matching the predicate or 'Nothing' if no
--- such element exists.
-find :: (a -> Bool) -> Bundle v a -> Maybe a
-{-# INLINE find #-}
-find f = unId . M.find f
-
--- | Yield 'Just' the index of the first element matching the predicate or
--- 'Nothing' if no such element exists.
-findIndex :: (a -> Bool) -> Bundle v a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex f = unId . M.findIndex f
-
--- Folding
--- -------
-
--- | Left fold
-foldl :: (a -> b -> a) -> a -> Bundle v b -> a
-{-# INLINE foldl #-}
-foldl f z = unId . M.foldl f z
-
--- | Left fold on non-empty 'Bundle's
-foldl1 :: (a -> a -> a) -> Bundle v a -> a
-{-# INLINE foldl1 #-}
-foldl1 f = unId . M.foldl1 f
-
--- | Left fold with strict accumulator
-foldl' :: (a -> b -> a) -> a -> Bundle v b -> a
-{-# INLINE foldl' #-}
-foldl' f z = unId . M.foldl' f z
-
--- | Left fold on non-empty 'Bundle's with strict accumulator
-foldl1' :: (a -> a -> a) -> Bundle v a -> a
-{-# INLINE foldl1' #-}
-foldl1' f = unId . M.foldl1' f
-
--- | Right fold
-foldr :: (a -> b -> b) -> b -> Bundle v a -> b
-{-# INLINE foldr #-}
-foldr f z = unId . M.foldr f z
-
--- | Right fold on non-empty 'Bundle's
-foldr1 :: (a -> a -> a) -> Bundle v a -> a
-{-# INLINE foldr1 #-}
-foldr1 f = unId . M.foldr1 f
-
--- Specialised folds
--- -----------------
-
-and :: Bundle v Bool -> Bool
-{-# INLINE and #-}
-and = unId . M.and
-
-or :: Bundle v Bool -> Bool
-{-# INLINE or #-}
-or = unId . M.or
-
--- Unfolding
--- ---------
-
--- | Unfold
-unfoldr :: (s -> Maybe (a, s)) -> s -> Bundle v a
-{-# INLINE unfoldr #-}
-unfoldr = M.unfoldr
-
--- | Unfold at most @n@ elements
-unfoldrN :: Int -> (s -> Maybe (a, s)) -> s -> Bundle v a
-{-# INLINE unfoldrN #-}
-unfoldrN = M.unfoldrN
-
--- | Apply function n-1 times to value. Zeroth element is original value.
-iterateN :: Int -> (a -> a) -> a -> Bundle v a
-{-# INLINE iterateN #-}
-iterateN = M.iterateN
-
--- Scans
--- -----
-
--- | Prefix scan
-prescanl :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE prescanl #-}
-prescanl = M.prescanl
-
--- | Prefix scan with strict accumulator
-prescanl' :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE prescanl' #-}
-prescanl' = M.prescanl'
-
--- | Suffix scan
-postscanl :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE postscanl #-}
-postscanl = M.postscanl
-
--- | Suffix scan with strict accumulator
-postscanl' :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE postscanl' #-}
-postscanl' = M.postscanl'
-
--- | Haskell-style scan
-scanl :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE scanl #-}
-scanl = M.scanl
-
--- | Haskell-style scan with strict accumulator
-scanl' :: (a -> b -> a) -> a -> Bundle v b -> Bundle v a
-{-# INLINE scanl' #-}
-scanl' = M.scanl'
-
--- | Scan over a non-empty 'Bundle'
-scanl1 :: (a -> a -> a) -> Bundle v a -> Bundle v a
-{-# INLINE scanl1 #-}
-scanl1 = M.scanl1
-
--- | Scan over a non-empty 'Bundle' with a strict accumulator
-scanl1' :: (a -> a -> a) -> Bundle v a -> Bundle v a
-{-# INLINE scanl1' #-}
-scanl1' = M.scanl1'
-
-
--- Comparisons
--- -----------
-
--- | Check if two 'Bundle's are equal
-eq :: (Eq a) => Bundle v a -> Bundle v a -> Bool
-{-# INLINE eq #-}
-eq = eqBy (==)
-
-eqBy :: (a -> b -> Bool) -> Bundle v a -> Bundle v b -> Bool
-{-# INLINE eqBy #-}
-eqBy e x y = unId (M.eqBy e x y)
-
--- | Lexicographically compare two 'Bundle's
-cmp :: (Ord a) => Bundle v a -> Bundle v a -> Ordering
-{-# INLINE cmp #-}
-cmp = cmpBy compare
-
-cmpBy :: (a ->  b -> Ordering) -> Bundle v a -> Bundle v b -> Ordering
-{-# INLINE cmpBy #-}
-cmpBy c x y = unId (M.cmpBy c x y)
-
-instance Eq a => Eq (M.Bundle Id v a) where
-  {-# INLINE (==) #-}
-  (==) = eq
-
-instance Ord a => Ord (M.Bundle Id v a) where
-  {-# INLINE compare #-}
-  compare = cmp
-
-#if MIN_VERSION_base(4,9,0)
-instance Eq1 (M.Bundle Id v) where
-  {-# INLINE liftEq #-}
-  liftEq = eqBy
-
-instance Ord1 (M.Bundle Id v) where
-  {-# INLINE liftCompare #-}
-  liftCompare = cmpBy
-#endif
-
--- Monadic combinators
--- -------------------
-
--- | Apply a monadic action to each element of the stream, producing a monadic
--- stream of results
-mapM :: Monad m => (a -> m b) -> Bundle v a -> M.Bundle m v b
-{-# INLINE mapM #-}
-mapM f = M.mapM f . lift
-
--- | Apply a monadic action to each element of the stream
-mapM_ :: Monad m => (a -> m b) -> Bundle v a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ f = M.mapM_ f . lift
-
-zipWithM :: Monad m => (a -> b -> m c) -> Bundle v a -> Bundle v b -> M.Bundle m v c
-{-# INLINE zipWithM #-}
-zipWithM f as bs = M.zipWithM f (lift as) (lift bs)
-
-zipWithM_ :: Monad m => (a -> b -> m c) -> Bundle v a -> Bundle v b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ f as bs = M.zipWithM_ f (lift as) (lift bs)
-
--- | Yield a monadic stream of elements that satisfy the monadic predicate
-filterM :: Monad m => (a -> m Bool) -> Bundle v a -> M.Bundle m v a
-{-# INLINE filterM #-}
-filterM f = M.filterM f . lift
-
--- | Monadic fold
-foldM :: Monad m => (a -> b -> m a) -> a -> Bundle v b -> m a
-{-# INLINE foldM #-}
-foldM m z = M.foldM m z . lift
-
--- | Monadic fold over non-empty stream
-fold1M :: Monad m => (a -> a -> m a) -> Bundle v a -> m a
-{-# INLINE fold1M #-}
-fold1M m = M.fold1M m . lift
-
--- | Monadic fold with strict accumulator
-foldM' :: Monad m => (a -> b -> m a) -> a -> Bundle v b -> m a
-{-# INLINE foldM' #-}
-foldM' m z = M.foldM' m z . lift
-
--- | Monad fold over non-empty stream with strict accumulator
-fold1M' :: Monad m => (a -> a -> m a) -> Bundle v a -> m a
-{-# INLINE fold1M' #-}
-fold1M' m = M.fold1M' m . lift
-
--- Enumerations
--- ------------
-
--- | Yield a 'Bundle' of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc.
-enumFromStepN :: Num a => a -> a -> Int -> Bundle v a
-{-# INLINE enumFromStepN #-}
-enumFromStepN = M.enumFromStepN
-
--- | Enumerate values
---
--- /WARNING:/ This operations can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromTo :: Enum a => a -> a -> Bundle v a
-{-# INLINE enumFromTo #-}
-enumFromTo = M.enumFromTo
-
--- | Enumerate values with a given step.
---
--- /WARNING:/ This operations is very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: Enum a => a -> a -> a -> Bundle v a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo = M.enumFromThenTo
-
--- Conversions
--- -----------
-
--- | Convert a 'Bundle' to a list
-toList :: Bundle v a -> [a]
-{-# INLINE toList #-}
--- toList s = unId (M.toList s)
-toList s = build (\c n -> toListFB c n s)
-
--- This supports foldr/build list fusion that GHC implements
-toListFB :: (a -> b -> b) -> b -> Bundle v a -> b
-{-# INLINE [0] toListFB #-}
-toListFB c n M.Bundle{M.sElems = Stream step t} = go t
-  where
-    go s = case unId (step s) of
-             Yield x s' -> x `c` go s'
-             Skip    s' -> go s'
-             Done       -> n
-
--- | Create a 'Bundle' from a list
-fromList :: [a] -> Bundle v a
-{-# INLINE fromList #-}
-fromList = M.fromList
-
--- | Create a 'Bundle' from the first @n@ elements of a list
---
--- > fromListN n xs = fromList (take n xs)
-fromListN :: Int -> [a] -> Bundle v a
-{-# INLINE fromListN #-}
-fromListN = M.fromListN
-
-unsafeFromList :: Size -> [a] -> Bundle v a
-{-# INLINE unsafeFromList #-}
-unsafeFromList = M.unsafeFromList
-
-fromVector :: Vector v a => v a -> Bundle v a
-{-# INLINE fromVector #-}
-fromVector = M.fromVector
-
-reVector :: Bundle u a -> Bundle v a
-{-# INLINE reVector #-}
-reVector = M.reVector
-
-fromVectors :: Vector v a => [v a] -> Bundle v a
-{-# INLINE fromVectors #-}
-fromVectors = M.fromVectors
-
-concatVectors :: Vector v a => Bundle u (v a) -> Bundle v a
-{-# INLINE concatVectors #-}
-concatVectors = M.concatVectors
-
--- | Create a 'Bundle' of values from a 'Bundle' of streamable things
-flatten :: (a -> s) -> (s -> Step s b) -> Size -> Bundle v a -> Bundle v b
-{-# INLINE_FUSED flatten #-}
-flatten mk istep sz = M.flatten (return . mk) (return . istep) sz . lift
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Monadic.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Monadic.hs
deleted file mode 100644
index 46f4a165f88d..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Monadic.hs
+++ /dev/null
@@ -1,1106 +0,0 @@
-{-# LANGUAGE CPP, ExistentialQuantification, MultiParamTypeClasses, FlexibleInstances, Rank2Types, BangPatterns, KindSignatures, GADTs, ScopedTypeVariables #-}
-
--- |
--- Module      : Data.Vector.Fusion.Bundle.Monadic
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Monadic bundles.
---
-
-module Data.Vector.Fusion.Bundle.Monadic (
-  Bundle(..), Chunk(..),
-
-  -- * Size hints
-  size, sized,
-
-  -- * Length
-  length, null,
-
-  -- * Construction
-  empty, singleton, cons, snoc, replicate, replicateM, generate, generateM, (++),
-
-  -- * Accessing elements
-  head, last, (!!), (!?),
-
-  -- * Substreams
-  slice, init, tail, take, drop,
-
-  -- * Mapping
-  map, mapM, mapM_, trans, unbox, concatMap, flatten,
-
-  -- * Zipping
-  indexed, indexedR, zipWithM_,
-  zipWithM, zipWith3M, zipWith4M, zipWith5M, zipWith6M,
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- * Comparisons
-  eqBy, cmpBy,
-
-  -- * Filtering
-  filter, filterM, takeWhile, takeWhileM, dropWhile, dropWhileM,
-
-  -- * Searching
-  elem, notElem, find, findM, findIndex, findIndexM,
-
-  -- * Folding
-  foldl, foldlM, foldl1, foldl1M, foldM, fold1M,
-  foldl', foldlM', foldl1', foldl1M', foldM', fold1M',
-  foldr, foldrM, foldr1, foldr1M,
-
-  -- * Specialised folds
-  and, or, concatMapM,
-
-  -- * Unfolding
-  unfoldr, unfoldrM,
-  unfoldrN, unfoldrNM,
-  iterateN, iterateNM,
-
-  -- * Scans
-  prescanl, prescanlM, prescanl', prescanlM',
-  postscanl, postscanlM, postscanl', postscanlM',
-  scanl, scanlM, scanl', scanlM',
-  scanl1, scanl1M, scanl1', scanl1M',
-
-  -- * Enumerations
-  enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- * Conversions
-  toList, fromList, fromListN, unsafeFromList,
-  fromVector, reVector, fromVectors, concatVectors,
-  fromStream, chunks, elements
-) where
-
-import Data.Vector.Generic.Base
-import qualified Data.Vector.Generic.Mutable.Base as M
-import Data.Vector.Fusion.Bundle.Size
-import Data.Vector.Fusion.Util ( Box(..), delay_inline )
-import Data.Vector.Fusion.Stream.Monadic ( Stream(..), Step(..) )
-import qualified Data.Vector.Fusion.Stream.Monadic as S
-import Control.Monad.Primitive
-
-import qualified Data.List as List
-import Data.Char      ( ord )
-import GHC.Base       ( unsafeChr )
-import Control.Monad  ( liftM )
-import Prelude hiding ( length, null,
-                        replicate, (++),
-                        head, last, (!!),
-                        init, tail, take, drop,
-                        map, mapM, mapM_, concatMap,
-                        zipWith, zipWith3, zip, zip3,
-                        filter, takeWhile, dropWhile,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        and, or,
-                        scanl, scanl1,
-                        enumFromTo, enumFromThenTo )
-
-import Data.Int  ( Int8, Int16, Int32 )
-import Data.Word ( Word8, Word16, Word32, Word64 )
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Word ( Word )
-#endif
-
-#include "vector.h"
-#include "MachDeps.h"
-
-#if WORD_SIZE_IN_BITS > 32
-import Data.Int  ( Int64 )
-#endif
-
-data Chunk v a = Chunk Int (forall m. (PrimMonad m, Vector v a) => Mutable v (PrimState m) a -> m ())
-
--- | Monadic streams
-data Bundle m v a = Bundle { sElems  :: Stream m a
-                           , sChunks :: Stream m (Chunk v a)
-                           , sVector :: Maybe (v a)
-                           , sSize   :: Size
-                           }
-
-fromStream :: Monad m => Stream m a -> Size -> Bundle m v a
-{-# INLINE fromStream #-}
-fromStream (Stream step t) sz = Bundle (Stream step t) (Stream step' t) Nothing sz
-  where
-    step' s = do r <- step s
-                 return $ fmap (\x -> Chunk 1 (\v -> M.basicUnsafeWrite v 0 x)) r
-
-chunks :: Bundle m v a -> Stream m (Chunk v a)
-{-# INLINE chunks #-}
-chunks = sChunks
-
-elements :: Bundle m v a -> Stream m a
-{-# INLINE elements #-}
-elements = sElems
-
--- | 'Size' hint of a 'Bundle'
-size :: Bundle m v a -> Size
-{-# INLINE size #-}
-size = sSize
-
--- | Attach a 'Size' hint to a 'Bundle'
-sized :: Bundle m v a -> Size -> Bundle m v a
-{-# INLINE_FUSED sized #-}
-sized s sz = s { sSize = sz }
-
--- Length
--- ------
-
--- | Length of a 'Bundle'
-length :: Monad m => Bundle m v a -> m Int
-{-# INLINE_FUSED length #-}
-length Bundle{sSize = Exact n}  = return n
-length Bundle{sChunks = s} = S.foldl' (\n (Chunk k _) -> n+k) 0 s
-
--- | Check if a 'Bundle' is empty
-null :: Monad m => Bundle m v a -> m Bool
-{-# INLINE_FUSED null #-}
-null Bundle{sSize = Exact n} = return (n == 0)
-null Bundle{sChunks = s} = S.foldr (\(Chunk n _) z -> n == 0 && z) True s
-
--- Construction
--- ------------
-
--- | Empty 'Bundle'
-empty :: Monad m => Bundle m v a
-{-# INLINE_FUSED empty #-}
-empty = fromStream S.empty (Exact 0)
-
--- | Singleton 'Bundle'
-singleton :: Monad m => a -> Bundle m v a
-{-# INLINE_FUSED singleton #-}
-singleton x = fromStream (S.singleton x) (Exact 1)
-
--- | Replicate a value to a given length
-replicate :: Monad m => Int -> a -> Bundle m v a
-{-# INLINE_FUSED replicate #-}
-replicate n x = Bundle (S.replicate n x)
-                       (S.singleton $ Chunk len (\v -> M.basicSet v x))
-                       Nothing
-                       (Exact len)
-  where
-    len = delay_inline max n 0
-
--- | Yield a 'Bundle' of values obtained by performing the monadic action the
--- given number of times
-replicateM :: Monad m => Int -> m a -> Bundle m v a
-{-# INLINE_FUSED replicateM #-}
--- NOTE: We delay inlining max here because GHC will create a join point for
--- the call to newArray# otherwise which is not really nice.
-replicateM n p = fromStream (S.replicateM n p) (Exact (delay_inline max n 0))
-
-generate :: Monad m => Int -> (Int -> a) -> Bundle m v a
-{-# INLINE generate #-}
-generate n f = generateM n (return . f)
-
--- | Generate a stream from its indices
-generateM :: Monad m => Int -> (Int -> m a) -> Bundle m v a
-{-# INLINE_FUSED generateM #-}
-generateM n f = fromStream (S.generateM n f) (Exact (delay_inline max n 0))
-
--- | Prepend an element
-cons :: Monad m => a -> Bundle m v a -> Bundle m v a
-{-# INLINE cons #-}
-cons x s = singleton x ++ s
-
--- | Append an element
-snoc :: Monad m => Bundle m v a -> a -> Bundle m v a
-{-# INLINE snoc #-}
-snoc s x = s ++ singleton x
-
-infixr 5 ++
--- | Concatenate two 'Bundle's
-(++) :: Monad m => Bundle m v a -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED (++) #-}
-Bundle sa ta _ na ++ Bundle sb tb _ nb = Bundle (sa S.++ sb) (ta S.++ tb) Nothing (na + nb)
-
--- Accessing elements
--- ------------------
-
--- | First element of the 'Bundle' or error if empty
-head :: Monad m => Bundle m v a -> m a
-{-# INLINE_FUSED head #-}
-head = S.head . sElems
-
--- | Last element of the 'Bundle' or error if empty
-last :: Monad m => Bundle m v a -> m a
-{-# INLINE_FUSED last #-}
-last = S.last . sElems
-
-infixl 9 !!
--- | Element at the given position
-(!!) :: Monad m => Bundle m v a -> Int -> m a
-{-# INLINE (!!) #-}
-b !! i = sElems b S.!! i
-
-infixl 9 !?
--- | Element at the given position or 'Nothing' if out of bounds
-(!?) :: Monad m => Bundle m v a -> Int -> m (Maybe a)
-{-# INLINE (!?) #-}
-b !? i = sElems b S.!? i
-
--- Substreams
--- ----------
-
--- | Extract a substream of the given length starting at the given position.
-slice :: Monad m => Int   -- ^ starting index
-                 -> Int   -- ^ length
-                 -> Bundle m v a
-                 -> Bundle m v a
-{-# INLINE slice #-}
-slice i n s = take n (drop i s)
-
--- | All but the last element
-init :: Monad m => Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED init #-}
-init Bundle{sElems = s, sSize = sz} = fromStream (S.init s) (sz-1)
-
--- | All but the first element
-tail :: Monad m => Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED tail #-}
-tail Bundle{sElems = s, sSize = sz} = fromStream (S.tail s) (sz-1)
-
--- | The first @n@ elements
-take :: Monad m => Int -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED take #-}
-take n Bundle{sElems = s, sSize = sz} = fromStream (S.take n s) (smaller (Exact n) sz)
-
--- | All but the first @n@ elements
-drop :: Monad m => Int -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED drop #-}
-drop n Bundle{sElems = s, sSize = sz} =
-  fromStream (S.drop n s) (clampedSubtract sz (Exact n))
-
--- Mapping
--- -------
-
-instance Monad m => Functor (Bundle m v) where
-  {-# INLINE fmap #-}
-  fmap = map
-
--- | Map a function over a 'Bundle'
-map :: Monad m => (a -> b) -> Bundle m v a -> Bundle m v b
-{-# INLINE map #-}
-map f = mapM (return . f)
-
--- | Map a monadic function over a 'Bundle'
-mapM :: Monad m => (a -> m b) -> Bundle m v a -> Bundle m v b
-{-# INLINE_FUSED mapM #-}
-mapM f Bundle{sElems = s, sSize = n} = fromStream (S.mapM f s) n
-
--- | Execute a monadic action for each element of the 'Bundle'
-mapM_ :: Monad m => (a -> m b) -> Bundle m v a -> m ()
-{-# INLINE_FUSED mapM_ #-}
-mapM_ m = S.mapM_ m . sElems
-
--- | Transform a 'Bundle' to use a different monad
-trans :: (Monad m, Monad m') => (forall z. m z -> m' z)
-                             -> Bundle m v a -> Bundle m' v a
-{-# INLINE_FUSED trans #-}
-trans f Bundle{sElems = s, sChunks = cs, sVector = v, sSize = n}
-  = Bundle { sElems = S.trans f s, sChunks = S.trans f cs, sVector = v, sSize = n }
-
-unbox :: Monad m => Bundle m v (Box a) -> Bundle m v a
-{-# INLINE_FUSED unbox #-}
-unbox Bundle{sElems = s, sSize = n} = fromStream (S.unbox s) n
-
--- Zipping
--- -------
-
--- | Pair each element in a 'Bundle' with its index
-indexed :: Monad m => Bundle m v a -> Bundle m v (Int,a)
-{-# INLINE_FUSED indexed #-}
-indexed Bundle{sElems = s, sSize = n} = fromStream (S.indexed s) n
-
--- | Pair each element in a 'Bundle' with its index, starting from the right
--- and counting down
-indexedR :: Monad m => Int -> Bundle m v a -> Bundle m v (Int,a)
-{-# INLINE_FUSED indexedR #-}
-indexedR m Bundle{sElems = s, sSize = n} = fromStream (S.indexedR m s) n
-
--- | Zip two 'Bundle's with the given monadic function
-zipWithM :: Monad m => (a -> b -> m c) -> Bundle m v a -> Bundle m v b -> Bundle m v c
-{-# INLINE_FUSED zipWithM #-}
-zipWithM f Bundle{sElems = sa, sSize = na}
-           Bundle{sElems = sb, sSize = nb} = fromStream (S.zipWithM f sa sb) (smaller na nb)
-
--- FIXME: This might expose an opportunity for inplace execution.
-{-# RULES
-
-"zipWithM xs xs [Vector.Bundle]" forall f xs.
-  zipWithM f xs xs = mapM (\x -> f x x) xs   #-}
-
-
-zipWithM_ :: Monad m => (a -> b -> m c) -> Bundle m v a -> Bundle m v b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ f sa sb = S.zipWithM_ f (sElems sa) (sElems sb)
-
-zipWith3M :: Monad m => (a -> b -> c -> m d) -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-{-# INLINE_FUSED zipWith3M #-}
-zipWith3M f Bundle{sElems = sa, sSize = na}
-            Bundle{sElems = sb, sSize = nb}
-            Bundle{sElems = sc, sSize = nc}
-  = fromStream (S.zipWith3M f sa sb sc) (smaller na (smaller nb nc))
-
-zipWith4M :: Monad m => (a -> b -> c -> d -> m e)
-                     -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                     -> Bundle m v e
-{-# INLINE zipWith4M #-}
-zipWith4M f sa sb sc sd
-  = zipWithM (\(a,b) (c,d) -> f a b c d) (zip sa sb) (zip sc sd)
-
-zipWith5M :: Monad m => (a -> b -> c -> d -> e -> m f)
-                     -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                     -> Bundle m v e -> Bundle m v f
-{-# INLINE zipWith5M #-}
-zipWith5M f sa sb sc sd se
-  = zipWithM (\(a,b,c) (d,e) -> f a b c d e) (zip3 sa sb sc) (zip sd se)
-
-zipWith6M :: Monad m => (a -> b -> c -> d -> e -> f -> m g)
-                     -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                     -> Bundle m v e -> Bundle m v f -> Bundle m v g
-{-# INLINE zipWith6M #-}
-zipWith6M fn sa sb sc sd se sf
-  = zipWithM (\(a,b,c) (d,e,f) -> fn a b c d e f) (zip3 sa sb sc)
-                                                  (zip3 sd se sf)
-
-zipWith :: Monad m => (a -> b -> c) -> Bundle m v a -> Bundle m v b -> Bundle m v c
-{-# INLINE zipWith #-}
-zipWith f = zipWithM (\a b -> return (f a b))
-
-zipWith3 :: Monad m => (a -> b -> c -> d)
-                    -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-{-# INLINE zipWith3 #-}
-zipWith3 f = zipWith3M (\a b c -> return (f a b c))
-
-zipWith4 :: Monad m => (a -> b -> c -> d -> e)
-                    -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                    -> Bundle m v e
-{-# INLINE zipWith4 #-}
-zipWith4 f = zipWith4M (\a b c d -> return (f a b c d))
-
-zipWith5 :: Monad m => (a -> b -> c -> d -> e -> f)
-                    -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                    -> Bundle m v e -> Bundle m v f
-{-# INLINE zipWith5 #-}
-zipWith5 f = zipWith5M (\a b c d e -> return (f a b c d e))
-
-zipWith6 :: Monad m => (a -> b -> c -> d -> e -> f -> g)
-                    -> Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                    -> Bundle m v e -> Bundle m v f -> Bundle m v g
-{-# INLINE zipWith6 #-}
-zipWith6 fn = zipWith6M (\a b c d e f -> return (fn a b c d e f))
-
-zip :: Monad m => Bundle m v a -> Bundle m v b -> Bundle m v (a,b)
-{-# INLINE zip #-}
-zip = zipWith (,)
-
-zip3 :: Monad m => Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v (a,b,c)
-{-# INLINE zip3 #-}
-zip3 = zipWith3 (,,)
-
-zip4 :: Monad m => Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                -> Bundle m v (a,b,c,d)
-{-# INLINE zip4 #-}
-zip4 = zipWith4 (,,,)
-
-zip5 :: Monad m => Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                -> Bundle m v e -> Bundle m v (a,b,c,d,e)
-{-# INLINE zip5 #-}
-zip5 = zipWith5 (,,,,)
-
-zip6 :: Monad m => Bundle m v a -> Bundle m v b -> Bundle m v c -> Bundle m v d
-                -> Bundle m v e -> Bundle m v f -> Bundle m v (a,b,c,d,e,f)
-{-# INLINE zip6 #-}
-zip6 = zipWith6 (,,,,,)
-
--- Comparisons
--- -----------
-
--- | Check if two 'Bundle's are equal
-eqBy :: (Monad m) => (a -> b -> Bool) -> Bundle m v a -> Bundle m v b -> m Bool
-{-# INLINE_FUSED eqBy #-}
-eqBy eq x y = S.eqBy eq (sElems x) (sElems y)
-
--- | Lexicographically compare two 'Bundle's
-cmpBy :: (Monad m) => (a -> b -> Ordering) -> Bundle m v a -> Bundle m v b -> m Ordering
-{-# INLINE_FUSED cmpBy #-}
-cmpBy cmp x y = S.cmpBy cmp (sElems x) (sElems y)
-
--- Filtering
--- ---------
-
--- | Drop elements which do not satisfy the predicate
-filter :: Monad m => (a -> Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE filter #-}
-filter f = filterM (return . f)
-
--- | Drop elements which do not satisfy the monadic predicate
-filterM :: Monad m => (a -> m Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED filterM #-}
-filterM f Bundle{sElems = s, sSize = n} = fromStream (S.filterM f s) (toMax n)
-
--- | Longest prefix of elements that satisfy the predicate
-takeWhile :: Monad m => (a -> Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE takeWhile #-}
-takeWhile f = takeWhileM (return . f)
-
--- | Longest prefix of elements that satisfy the monadic predicate
-takeWhileM :: Monad m => (a -> m Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED takeWhileM #-}
-takeWhileM f Bundle{sElems = s, sSize = n} = fromStream (S.takeWhileM f s) (toMax n)
-
--- | Drop the longest prefix of elements that satisfy the predicate
-dropWhile :: Monad m => (a -> Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE dropWhile #-}
-dropWhile f = dropWhileM (return . f)
-
--- | Drop the longest prefix of elements that satisfy the monadic predicate
-dropWhileM :: Monad m => (a -> m Bool) -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED dropWhileM #-}
-dropWhileM f Bundle{sElems = s, sSize = n} = fromStream (S.dropWhileM f s) (toMax n)
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | Check whether the 'Bundle' contains an element
-elem :: (Monad m, Eq a) => a -> Bundle m v a -> m Bool
-{-# INLINE_FUSED elem #-}
-elem x = S.elem x . sElems
-
-infix 4 `notElem`
--- | Inverse of `elem`
-notElem :: (Monad m, Eq a) => a -> Bundle m v a -> m Bool
-{-# INLINE notElem #-}
-notElem x = S.notElem x . sElems
-
--- | Yield 'Just' the first element that satisfies the predicate or 'Nothing'
--- if no such element exists.
-find :: Monad m => (a -> Bool) -> Bundle m v a -> m (Maybe a)
-{-# INLINE find #-}
-find f = findM (return . f)
-
--- | Yield 'Just' the first element that satisfies the monadic predicate or
--- 'Nothing' if no such element exists.
-findM :: Monad m => (a -> m Bool) -> Bundle m v a -> m (Maybe a)
-{-# INLINE_FUSED findM #-}
-findM f = S.findM f . sElems
-
--- | Yield 'Just' the index of the first element that satisfies the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Monad m => (a -> Bool) -> Bundle m v a -> m (Maybe Int)
-{-# INLINE_FUSED findIndex #-}
-findIndex f = findIndexM (return . f)
-
--- | Yield 'Just' the index of the first element that satisfies the monadic
--- predicate or 'Nothing' if no such element exists.
-findIndexM :: Monad m => (a -> m Bool) -> Bundle m v a -> m (Maybe Int)
-{-# INLINE_FUSED findIndexM #-}
-findIndexM f = S.findIndexM f . sElems
-
--- Folding
--- -------
-
--- | Left fold
-foldl :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> m a
-{-# INLINE foldl #-}
-foldl f = foldlM (\a b -> return (f a b))
-
--- | Left fold with a monadic operator
-foldlM :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> m a
-{-# INLINE_FUSED foldlM #-}
-foldlM m z = S.foldlM m z . sElems
-
--- | Same as 'foldlM'
-foldM :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> m a
-{-# INLINE foldM #-}
-foldM = foldlM
-
--- | Left fold over a non-empty 'Bundle'
-foldl1 :: Monad m => (a -> a -> a) -> Bundle m v a -> m a
-{-# INLINE foldl1 #-}
-foldl1 f = foldl1M (\a b -> return (f a b))
-
--- | Left fold over a non-empty 'Bundle' with a monadic operator
-foldl1M :: Monad m => (a -> a -> m a) -> Bundle m v a -> m a
-{-# INLINE_FUSED foldl1M #-}
-foldl1M f = S.foldl1M f . sElems
-
--- | Same as 'foldl1M'
-fold1M :: Monad m => (a -> a -> m a) -> Bundle m v a -> m a
-{-# INLINE fold1M #-}
-fold1M = foldl1M
-
--- | Left fold with a strict accumulator
-foldl' :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> m a
-{-# INLINE foldl' #-}
-foldl' f = foldlM' (\a b -> return (f a b))
-
--- | Left fold with a strict accumulator and a monadic operator
-foldlM' :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> m a
-{-# INLINE_FUSED foldlM' #-}
-foldlM' m z = S.foldlM' m z . sElems
-
--- | Same as 'foldlM''
-foldM' :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> m a
-{-# INLINE foldM' #-}
-foldM' = foldlM'
-
--- | Left fold over a non-empty 'Bundle' with a strict accumulator
-foldl1' :: Monad m => (a -> a -> a) -> Bundle m v a -> m a
-{-# INLINE foldl1' #-}
-foldl1' f = foldl1M' (\a b -> return (f a b))
-
--- | Left fold over a non-empty 'Bundle' with a strict accumulator and a
--- monadic operator
-foldl1M' :: Monad m => (a -> a -> m a) -> Bundle m v a -> m a
-{-# INLINE_FUSED foldl1M' #-}
-foldl1M' f = S.foldl1M' f . sElems
-
--- | Same as 'foldl1M''
-fold1M' :: Monad m => (a -> a -> m a) -> Bundle m v a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = foldl1M'
-
--- | Right fold
-foldr :: Monad m => (a -> b -> b) -> b -> Bundle m v a -> m b
-{-# INLINE foldr #-}
-foldr f = foldrM (\a b -> return (f a b))
-
--- | Right fold with a monadic operator
-foldrM :: Monad m => (a -> b -> m b) -> b -> Bundle m v a -> m b
-{-# INLINE_FUSED foldrM #-}
-foldrM f z = S.foldrM f z . sElems
-
--- | Right fold over a non-empty stream
-foldr1 :: Monad m => (a -> a -> a) -> Bundle m v a -> m a
-{-# INLINE foldr1 #-}
-foldr1 f = foldr1M (\a b -> return (f a b))
-
--- | Right fold over a non-empty stream with a monadic operator
-foldr1M :: Monad m => (a -> a -> m a) -> Bundle m v a -> m a
-{-# INLINE_FUSED foldr1M #-}
-foldr1M f = S.foldr1M f . sElems
-
--- Specialised folds
--- -----------------
-
-and :: Monad m => Bundle m v Bool -> m Bool
-{-# INLINE_FUSED and #-}
-and = S.and . sElems
-
-or :: Monad m => Bundle m v Bool -> m Bool
-{-# INLINE_FUSED or #-}
-or = S.or . sElems
-
-concatMap :: Monad m => (a -> Bundle m v b) -> Bundle m v a -> Bundle m v b
-{-# INLINE concatMap #-}
-concatMap f = concatMapM (return . f)
-
-concatMapM :: Monad m => (a -> m (Bundle m v b)) -> Bundle m v a -> Bundle m v b
-{-# INLINE_FUSED concatMapM #-}
-concatMapM f Bundle{sElems = s} = fromStream (S.concatMapM (liftM sElems . f) s) Unknown
-
--- | Create a 'Bundle' of values from a 'Bundle' of streamable things
-flatten :: Monad m => (a -> m s) -> (s -> m (Step s b)) -> Size
-                   -> Bundle m v a -> Bundle m v b
-{-# INLINE_FUSED flatten #-}
-flatten mk istep sz Bundle{sElems = s} = fromStream (S.flatten mk istep s) sz
-
--- Unfolding
--- ---------
-
--- | Unfold
-unfoldr :: Monad m => (s -> Maybe (a, s)) -> s -> Bundle m u a
-{-# INLINE_FUSED unfoldr #-}
-unfoldr f = unfoldrM (return . f)
-
--- | Unfold with a monadic function
-unfoldrM :: Monad m => (s -> m (Maybe (a, s))) -> s -> Bundle m u a
-{-# INLINE_FUSED unfoldrM #-}
-unfoldrM f s = fromStream (S.unfoldrM f s) Unknown
-
--- | Unfold at most @n@ elements
-unfoldrN :: Monad m => Int -> (s -> Maybe (a, s)) -> s -> Bundle m u a
-{-# INLINE_FUSED unfoldrN #-}
-unfoldrN n f = unfoldrNM n (return . f)
-
--- | Unfold at most @n@ elements with a monadic functions
-unfoldrNM :: Monad m => Int -> (s -> m (Maybe (a, s))) -> s -> Bundle m u a
-{-# INLINE_FUSED unfoldrNM #-}
-unfoldrNM n f s = fromStream (S.unfoldrNM n f s) (Max (delay_inline max n 0))
-
--- | Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: Monad m => Int -> (a -> m a) -> a -> Bundle m u a
-{-# INLINE_FUSED iterateNM #-}
-iterateNM n f x0 = fromStream (S.iterateNM n f x0) (Exact (delay_inline max n 0))
-
--- | Apply function n times to value. Zeroth element is original value.
-iterateN :: Monad m => Int -> (a -> a) -> a -> Bundle m u a
-{-# INLINE_FUSED iterateN #-}
-iterateN n f x0 = iterateNM n (return . f) x0
-
--- Scans
--- -----
-
--- | Prefix scan
-prescanl :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE prescanl #-}
-prescanl f = prescanlM (\a b -> return (f a b))
-
--- | Prefix scan with a monadic operator
-prescanlM :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE_FUSED prescanlM #-}
-prescanlM f z Bundle{sElems = s, sSize = sz} = fromStream (S.prescanlM f z s) sz
-
--- | Prefix scan with strict accumulator
-prescanl' :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE prescanl' #-}
-prescanl' f = prescanlM' (\a b -> return (f a b))
-
--- | Prefix scan with strict accumulator and a monadic operator
-prescanlM' :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE_FUSED prescanlM' #-}
-prescanlM' f z Bundle{sElems = s, sSize = sz} = fromStream (S.prescanlM' f z s) sz
-
--- | Suffix scan
-postscanl :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE postscanl #-}
-postscanl f = postscanlM (\a b -> return (f a b))
-
--- | Suffix scan with a monadic operator
-postscanlM :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE_FUSED postscanlM #-}
-postscanlM f z Bundle{sElems = s, sSize = sz} = fromStream (S.postscanlM f z s) sz
-
--- | Suffix scan with strict accumulator
-postscanl' :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE postscanl' #-}
-postscanl' f = postscanlM' (\a b -> return (f a b))
-
--- | Suffix scan with strict acccumulator and a monadic operator
-postscanlM' :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE_FUSED postscanlM' #-}
-postscanlM' f z Bundle{sElems = s, sSize = sz} = fromStream (S.postscanlM' f z s) sz
-
--- | Haskell-style scan
-scanl :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE scanl #-}
-scanl f = scanlM (\a b -> return (f a b))
-
--- | Haskell-style scan with a monadic operator
-scanlM :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE scanlM #-}
-scanlM f z s = z `cons` postscanlM f z s
-
--- | Haskell-style scan with strict accumulator
-scanl' :: Monad m => (a -> b -> a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE scanl' #-}
-scanl' f = scanlM' (\a b -> return (f a b))
-
--- | Haskell-style scan with strict accumulator and a monadic operator
-scanlM' :: Monad m => (a -> b -> m a) -> a -> Bundle m v b -> Bundle m v a
-{-# INLINE scanlM' #-}
-scanlM' f z s = z `seq` (z `cons` postscanlM f z s)
-
--- | Scan over a non-empty 'Bundle'
-scanl1 :: Monad m => (a -> a -> a) -> Bundle m v a -> Bundle m v a
-{-# INLINE scanl1 #-}
-scanl1 f = scanl1M (\x y -> return (f x y))
-
--- | Scan over a non-empty 'Bundle' with a monadic operator
-scanl1M :: Monad m => (a -> a -> m a) -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED scanl1M #-}
-scanl1M f Bundle{sElems = s, sSize = sz} = fromStream (S.scanl1M f s) sz
-
--- | Scan over a non-empty 'Bundle' with a strict accumulator
-scanl1' :: Monad m => (a -> a -> a) -> Bundle m v a -> Bundle m v a
-{-# INLINE scanl1' #-}
-scanl1' f = scanl1M' (\x y -> return (f x y))
-
--- | Scan over a non-empty 'Bundle' with a strict accumulator and a monadic
--- operator
-scanl1M' :: Monad m => (a -> a -> m a) -> Bundle m v a -> Bundle m v a
-{-# INLINE_FUSED scanl1M' #-}
-scanl1M' f Bundle{sElems = s, sSize = sz} = fromStream (S.scanl1M' f s) sz
-
--- Enumerations
--- ------------
-
--- The Enum class is broken for this, there just doesn't seem to be a
--- way to implement this generically. We have to specialise for as many types
--- as we can but this doesn't help in polymorphic loops.
-
--- | Yield a 'Bundle' of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc.
-enumFromStepN :: (Num a, Monad m) => a -> a -> Int -> Bundle m v a
-{-# INLINE_FUSED enumFromStepN #-}
-enumFromStepN x y n = fromStream (S.enumFromStepN x y n) (Exact (delay_inline max n 0))
-
--- | Enumerate values
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromTo :: (Enum a, Monad m) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo #-}
-enumFromTo x y = fromList [x .. y]
-
--- NOTE: We use (x+1) instead of (succ x) below because the latter checks for
--- overflow which can't happen here.
-
--- FIXME: add "too large" test for Int
-enumFromTo_small :: (Integral a, Monad m) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo_small #-}
-enumFromTo_small x y = x `seq` y `seq` fromStream (Stream step x) (Exact n)
-  where
-    n = delay_inline max (fromIntegral y - fromIntegral x + 1) 0
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Int8> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Int8 -> Int8 -> Bundle m v Int8
-
-"enumFromTo<Int16> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Int16 -> Int16 -> Bundle m v Int16
-
-"enumFromTo<Word8> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Word8 -> Word8 -> Bundle m v Word8
-
-"enumFromTo<Word16> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Word16 -> Word16 -> Bundle m v Word16   #-}
-
-
-
-#if WORD_SIZE_IN_BITS > 32
-
-{-# RULES
-
-"enumFromTo<Int32> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Int32 -> Int32 -> Bundle m v Int32
-
-"enumFromTo<Word32> [Bundle]"
-  enumFromTo = enumFromTo_small :: Monad m => Word32 -> Word32 -> Bundle m v Word32   #-}
-
-#endif
-
--- NOTE: We could implement a generic "too large" test:
---
--- len x y | x > y = 0
---         | n > 0 && n <= fromIntegral (maxBound :: Int) = fromIntegral n
---         | otherwise = error
---   where
---     n = y-x+1
---
--- Alas, GHC won't eliminate unnecessary comparisons (such as n >= 0 for
--- unsigned types). See http://hackage.haskell.org/trac/ghc/ticket/3744
---
-
-enumFromTo_int :: forall m v. Monad m => Int -> Int -> Bundle m v Int
-{-# INLINE_FUSED enumFromTo_int #-}
-enumFromTo_int x y = x `seq` y `seq` fromStream (Stream step x) (Exact (len x y))
-  where
-    {-# INLINE [0] len #-}
-    len :: Int -> Int -> Int
-    len u v | u > v     = 0
-            | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-                          (n > 0)
-                        $ n
-      where
-        n = v-u+1
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-enumFromTo_intlike :: (Integral a, Monad m) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo_intlike #-}
-enumFromTo_intlike x y = x `seq` y `seq` fromStream (Stream step x) (Exact (len x y))
-  where
-    {-# INLINE [0] len #-}
-    len u v | u > v     = 0
-            | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-                          (n > 0)
-                        $ fromIntegral n
-      where
-        n = v-u+1
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Int> [Bundle]"
-  enumFromTo = enumFromTo_int :: Monad m => Int -> Int -> Bundle m v Int
-
-#if WORD_SIZE_IN_BITS > 32
-
-"enumFromTo<Int64> [Bundle]"
-  enumFromTo = enumFromTo_intlike :: Monad m => Int64 -> Int64 -> Bundle m v Int64    #-}
-
-#else
-
-"enumFromTo<Int32> [Bundle]"
-  enumFromTo = enumFromTo_intlike :: Monad m => Int32 -> Int32 -> Bundle m v Int32    #-}
-
-#endif
-
-
-
-enumFromTo_big_word :: (Integral a, Monad m) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo_big_word #-}
-enumFromTo_big_word x y = x `seq` y `seq` fromStream (Stream step x) (Exact (len x y))
-  where
-    {-# INLINE [0] len #-}
-    len u v | u > v     = 0
-            | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-                          (n < fromIntegral (maxBound :: Int))
-                        $ fromIntegral (n+1)
-      where
-        n = v-u
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Word> [Bundle]"
-  enumFromTo = enumFromTo_big_word :: Monad m => Word -> Word -> Bundle m v Word
-
-"enumFromTo<Word64> [Bundle]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Word64 -> Word64 -> Bundle m v Word64
-
-#if WORD_SIZE_IN_BITS == 32
-
-"enumFromTo<Word32> [Bundle]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Word32 -> Word32 -> Bundle m v Word32
-
-#endif
-
-"enumFromTo<Integer> [Bundle]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Integer -> Integer -> Bundle m v Integer   #-}
-
-
-#if WORD_SIZE_IN_BITS > 32
-
--- FIXME: the "too large" test is totally wrong
-enumFromTo_big_int :: (Integral a, Monad m) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo_big_int #-}
-enumFromTo_big_int x y = x `seq` y `seq` fromStream (Stream step x) (Exact (len x y))
-  where
-    {-# INLINE [0] len #-}
-    len u v | u > v     = 0
-            | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-                          (n > 0 && n <= fromIntegral (maxBound :: Int))
-                        $ fromIntegral n
-      where
-        n = v-u+1
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-
-{-# RULES
-
-"enumFromTo<Int64> [Bundle]"
-  enumFromTo = enumFromTo_big_int :: Monad m => Int64 -> Int64 -> Bundle m v Int64   #-}
-
-
-
-#endif
-
-enumFromTo_char :: Monad m => Char -> Char -> Bundle m v Char
-{-# INLINE_FUSED enumFromTo_char #-}
-enumFromTo_char x y = x `seq` y `seq` fromStream (Stream step xn) (Exact n)
-  where
-    xn = ord x
-    yn = ord y
-
-    n = delay_inline max 0 (yn - xn + 1)
-
-    {-# INLINE_INNER step #-}
-    step zn | zn <= yn  = return $ Yield (unsafeChr zn) (zn+1)
-            | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Char> [Bundle]"
-  enumFromTo = enumFromTo_char   #-}
-
-
-
-------------------------------------------------------------------------
-
--- Specialise enumFromTo for Float and Double.
--- Also, try to do something about pairs?
-
-enumFromTo_double :: (Monad m, Ord a, RealFrac a) => a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromTo_double #-}
-enumFromTo_double n m = n `seq` m `seq` fromStream (Stream step n) (Max (len n lim))
-  where
-    lim = m + 1/2 -- important to float out
-
-    {-# INLINE [0] len #-}
-    len x y | x > y     = 0
-            | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-                          (l > 0)
-                        $ fromIntegral l
-      where
-        l :: Integer
-        l = truncate (y-x)+2
-
-    {-# INLINE_INNER step #-}
-    step x | x <= lim  = return $ Yield x (x+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Double> [Bundle]"
-  enumFromTo = enumFromTo_double :: Monad m => Double -> Double -> Bundle m v Double
-
-"enumFromTo<Float> [Bundle]"
-  enumFromTo = enumFromTo_double :: Monad m => Float -> Float -> Bundle m v Float   #-}
-
-
-
-------------------------------------------------------------------------
-
--- | Enumerate values with a given step.
---
--- /WARNING:/ This operation is very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Enum a, Monad m) => a -> a -> a -> Bundle m v a
-{-# INLINE_FUSED enumFromThenTo #-}
-enumFromThenTo x y z = fromList [x, y .. z]
-
--- FIXME: Specialise enumFromThenTo.
-
--- Conversions
--- -----------
-
--- | Convert a 'Bundle' to a list
-toList :: Monad m => Bundle m v a -> m [a]
-{-# INLINE toList #-}
-toList = foldr (:) []
-
--- | Convert a list to a 'Bundle'
-fromList :: Monad m => [a] -> Bundle m v a
-{-# INLINE fromList #-}
-fromList xs = unsafeFromList Unknown xs
-
--- | Convert the first @n@ elements of a list to a 'Bundle'
-fromListN :: Monad m => Int -> [a] -> Bundle m v a
-{-# INLINE_FUSED fromListN #-}
-fromListN n xs = fromStream (S.fromListN n xs) (Max (delay_inline max n 0))
-
--- | Convert a list to a 'Bundle' with the given 'Size' hint.
-unsafeFromList :: Monad m => Size -> [a] -> Bundle m v a
-{-# INLINE_FUSED unsafeFromList #-}
-unsafeFromList sz xs = fromStream (S.fromList xs) sz
-
-fromVector :: (Monad m, Vector v a) => v a -> Bundle m v a
-{-# INLINE_FUSED fromVector #-}
-fromVector v = v `seq` n `seq` Bundle (Stream step 0)
-                                      (Stream vstep True)
-                                      (Just v)
-                                      (Exact n)
-  where
-    n = basicLength v
-
-    {-# INLINE step #-}
-    step i | i >= n = return Done
-           | otherwise = case basicUnsafeIndexM v i of
-                           Box x -> return $ Yield x (i+1)
-
-
-    {-# INLINE vstep #-}
-    vstep True  = return (Yield (Chunk (basicLength v) (\mv -> basicUnsafeCopy mv v)) False)
-    vstep False = return Done
-
-fromVectors :: forall m v a. (Monad m, Vector v a) => [v a] -> Bundle m v a
-{-# INLINE_FUSED fromVectors #-}
-fromVectors us = Bundle (Stream pstep (Left us))
-                        (Stream vstep us)
-                        Nothing
-                        (Exact n)
-  where
-    n = List.foldl' (\k v -> k + basicLength v) 0 us
-
-    pstep (Left []) = return Done
-    pstep (Left (v:vs)) = basicLength v `seq` return (Skip (Right (v,0,vs)))
-
-    pstep (Right (v,i,vs))
-      | i >= basicLength v = return $ Skip (Left vs)
-      | otherwise          = case basicUnsafeIndexM v i of
-                               Box x -> return $ Yield x (Right (v,i+1,vs))
-
-    -- FIXME: work around bug in GHC 7.6.1
-    vstep :: [v a] -> m (Step [v a] (Chunk v a))
-    vstep [] = return Done
-    vstep (v:vs) = return $ Yield (Chunk (basicLength v)
-                                         (\mv -> INTERNAL_CHECK(check) "concatVectors" "length mismatch"
-                                                                       (M.basicLength mv == basicLength v)
-                                                 $ basicUnsafeCopy mv v)) vs
-
-
-concatVectors :: (Monad m, Vector v a) => Bundle m u (v a) -> Bundle m v a
-{-# INLINE_FUSED concatVectors #-}
-concatVectors Bundle{sElems = Stream step t}
-  = Bundle (Stream pstep (Left t))
-           (Stream vstep t)
-           Nothing
-           Unknown
-  where
-    pstep (Left s) = do
-      r <- step s
-      case r of
-        Yield v s' -> basicLength v `seq` return (Skip (Right (v,0,s')))
-        Skip    s' -> return (Skip (Left s'))
-        Done       -> return Done
-
-    pstep (Right (v,i,s))
-      | i >= basicLength v = return (Skip (Left s))
-      | otherwise          = case basicUnsafeIndexM v i of
-                               Box x -> return (Yield x (Right (v,i+1,s)))
-
-
-    vstep s = do
-      r <- step s
-      case r of
-        Yield v s' -> return (Yield (Chunk (basicLength v)
-                                           (\mv -> INTERNAL_CHECK(check) "concatVectors" "length mismatch"
-                                                                          (M.basicLength mv == basicLength v)
-                                                   $ basicUnsafeCopy mv v)) s')
-        Skip    s' -> return (Skip s')
-        Done       -> return Done
-
-reVector :: Monad m => Bundle m u a -> Bundle m v a
-{-# INLINE_FUSED reVector #-}
-reVector Bundle{sElems = s, sSize = n} = fromStream s n
-
-{-# RULES
-
-"reVector [Vector]"
-  reVector = id
-
-"reVector/reVector [Vector]" forall s.
-  reVector (reVector s) = s   #-}
-
-
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Size.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Size.hs
deleted file mode 100644
index e90cf373202d..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Bundle/Size.hs
+++ /dev/null
@@ -1,121 +0,0 @@
--- |
--- Module      : Data.Vector.Fusion.Bundle.Size
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : portable
---
--- Size hints for streams.
---
-
-module Data.Vector.Fusion.Bundle.Size (
-  Size(..), clampedSubtract, smaller, larger, toMax, upperBound, lowerBound
-) where
-
-import Data.Vector.Fusion.Util ( delay_inline )
-
--- | Size hint
-data Size = Exact Int          -- ^ Exact size
-          | Max   Int          -- ^ Upper bound on the size
-          | Unknown            -- ^ Unknown size
-        deriving( Eq, Show )
-
-instance Num Size where
-  Exact m + Exact n = checkedAdd Exact m n
-  Exact m + Max   n = checkedAdd Max m n
-
-  Max   m + Exact n = checkedAdd Max m n
-  Max   m + Max   n = checkedAdd Max m n
-
-  _       + _       = Unknown
-
-
-  Exact m - Exact n = checkedSubtract Exact m n
-  Exact m - Max   _ = Max   m
-
-  Max   m - Exact n = checkedSubtract Max m n
-  Max   m - Max   _ = Max   m
-  Max   m - Unknown = Max   m
-
-  _       - _       = Unknown
-
-
-  fromInteger n     = Exact (fromInteger n)
-
-  (*)    = error "vector: internal error * for Bundle.size isn't defined"
-  abs    = error "vector: internal error abs for Bundle.size isn't defined"
-  signum = error "vector: internal error signum for Bundle.size isn't defined"
-
-{-# INLINE checkedAdd #-}
-checkedAdd :: (Int -> Size) -> Int -> Int -> Size
-checkedAdd con m n
-    -- Note: we assume m and n are >= 0.
-  | r < m || r < n =
-      error $ "Data.Vector.Fusion.Bundle.Size.checkedAdd: overflow: " ++ show r
-  | otherwise = con r
-  where
-    r = m + n
-
-{-# INLINE checkedSubtract #-}
-checkedSubtract :: (Int -> Size) -> Int -> Int -> Size
-checkedSubtract con m n
-  | r < 0 =
-      error $ "Data.Vector.Fusion.Bundle.Size.checkedSubtract: underflow: " ++ show r
-  | otherwise = con r
-  where
-    r = m - n
-
--- | Subtract two sizes with clamping to 0, for drop-like things
-{-# INLINE clampedSubtract #-}
-clampedSubtract :: Size -> Size -> Size
-clampedSubtract (Exact m) (Exact n) = Exact (max 0 (m - n))
-clampedSubtract (Max   m) (Exact n)
-  | m <= n = Exact 0
-  | otherwise = Max (m - n)
-clampedSubtract (Exact m) (Max   _) = Max m
-clampedSubtract (Max   m) (Max   _) = Max m
-clampedSubtract _         _ = Unknown
-
--- | Minimum of two size hints
-smaller :: Size -> Size -> Size
-{-# INLINE smaller #-}
-smaller (Exact m) (Exact n) = Exact (delay_inline min m n)
-smaller (Exact m) (Max   n) = Max   (delay_inline min m n)
-smaller (Exact m) Unknown   = Max   m
-smaller (Max   m) (Exact n) = Max   (delay_inline min m n)
-smaller (Max   m) (Max   n) = Max   (delay_inline min m n)
-smaller (Max   m) Unknown   = Max   m
-smaller Unknown   (Exact n) = Max   n
-smaller Unknown   (Max   n) = Max   n
-smaller Unknown   Unknown   = Unknown
-
--- | Maximum of two size hints
-larger :: Size -> Size -> Size
-{-# INLINE larger #-}
-larger (Exact m) (Exact n)             = Exact (delay_inline max m n)
-larger (Exact m) (Max   n) | m >= n    = Exact m
-                           | otherwise = Max   n
-larger (Max   m) (Exact n) | n >= m    = Exact n
-                           | otherwise = Max   m
-larger (Max   m) (Max   n)             = Max   (delay_inline max m n)
-larger _         _                     = Unknown
-
--- | Convert a size hint to an upper bound
-toMax :: Size -> Size
-toMax (Exact n) = Max n
-toMax (Max   n) = Max n
-toMax Unknown   = Unknown
-
--- | Compute the minimum size from a size hint
-lowerBound :: Size -> Int
-lowerBound (Exact n) = n
-lowerBound _         = 0
-
--- | Compute the maximum size from a size hint if possible
-upperBound :: Size -> Maybe Int
-upperBound (Exact n) = Just n
-upperBound (Max   n) = Just n
-upperBound Unknown   = Nothing
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Stream/Monadic.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Stream/Monadic.hs
deleted file mode 100644
index cca002ca6f74..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Stream/Monadic.hs
+++ /dev/null
@@ -1,1639 +0,0 @@
-{-# LANGUAGE CPP, ExistentialQuantification, MultiParamTypeClasses, FlexibleInstances, Rank2Types, BangPatterns, KindSignatures, GADTs, ScopedTypeVariables #-}
-
--- |
--- Module      : Data.Vector.Fusion.Stream.Monadic
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Monadic stream combinators.
---
-
-module Data.Vector.Fusion.Stream.Monadic (
-  Stream(..), Step(..), SPEC(..),
-
-  -- * Length
-  length, null,
-
-  -- * Construction
-  empty, singleton, cons, snoc, replicate, replicateM, generate, generateM, (++),
-
-  -- * Accessing elements
-  head, last, (!!), (!?),
-
-  -- * Substreams
-  slice, init, tail, take, drop,
-
-  -- * Mapping
-  map, mapM, mapM_, trans, unbox, concatMap, flatten,
-
-  -- * Zipping
-  indexed, indexedR, zipWithM_,
-  zipWithM, zipWith3M, zipWith4M, zipWith5M, zipWith6M,
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- * Comparisons
-  eqBy, cmpBy,
-
-  -- * Filtering
-  filter, filterM, uniq, mapMaybe, takeWhile, takeWhileM, dropWhile, dropWhileM,
-
-  -- * Searching
-  elem, notElem, find, findM, findIndex, findIndexM,
-
-  -- * Folding
-  foldl, foldlM, foldl1, foldl1M, foldM, fold1M,
-  foldl', foldlM', foldl1', foldl1M', foldM', fold1M',
-  foldr, foldrM, foldr1, foldr1M,
-
-  -- * Specialised folds
-  and, or, concatMapM,
-
-  -- * Unfolding
-  unfoldr, unfoldrM,
-  unfoldrN, unfoldrNM,
-  iterateN, iterateNM,
-
-  -- * Scans
-  prescanl, prescanlM, prescanl', prescanlM',
-  postscanl, postscanlM, postscanl', postscanlM',
-  scanl, scanlM, scanl', scanlM',
-  scanl1, scanl1M, scanl1', scanl1M',
-
-  -- * Enumerations
-  enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- * Conversions
-  toList, fromList, fromListN
-) where
-
-import Data.Vector.Fusion.Util ( Box(..) )
-
-import Data.Char      ( ord )
-import GHC.Base       ( unsafeChr )
-import Control.Monad  ( liftM )
-import Prelude hiding ( length, null,
-                        replicate, (++),
-                        head, last, (!!),
-                        init, tail, take, drop,
-                        map, mapM, mapM_, concatMap,
-                        zipWith, zipWith3, zip, zip3,
-                        filter, takeWhile, dropWhile,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        and, or,
-                        scanl, scanl1,
-                        enumFromTo, enumFromThenTo )
-
-import Data.Int  ( Int8, Int16, Int32 )
-import Data.Word ( Word8, Word16, Word32, Word64 )
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Word ( Word8, Word16, Word32, Word, Word64 )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-import GHC.Types ( SPEC(..) )
-#elif __GLASGOW_HASKELL__ >= 700
-import GHC.Exts ( SpecConstrAnnotation(..) )
-#endif
-
-#include "vector.h"
-#include "MachDeps.h"
-
-#if WORD_SIZE_IN_BITS > 32
-import Data.Int  ( Int64 )
-#endif
-
-#if __GLASGOW_HASKELL__ < 708
-data SPEC = SPEC | SPEC2
-#if __GLASGOW_HASKELL__ >= 700
-{-# ANN type SPEC ForceSpecConstr #-}
-#endif
-#endif
-
-emptyStream :: String
-{-# NOINLINE emptyStream #-}
-emptyStream = "empty stream"
-
-#define EMPTY_STREAM (\state -> ERROR state emptyStream)
-
--- | Result of taking a single step in a stream
-data Step s a where
-  Yield :: a -> s -> Step s a
-  Skip  :: s -> Step s a
-  Done  :: Step s a
-
-instance Functor (Step s) where
-  {-# INLINE fmap #-}
-  fmap f (Yield x s) = Yield (f x) s
-  fmap _ (Skip s) = Skip s
-  fmap _ Done = Done
-
--- | Monadic streams
-data Stream m a = forall s. Stream (s -> m (Step s a)) s
-
--- Length
--- ------
-
--- | Length of a 'Stream'
-length :: Monad m => Stream m a -> m Int
-{-# INLINE_FUSED length #-}
-length = foldl' (\n _ -> n+1) 0
-
--- | Check if a 'Stream' is empty
-null :: Monad m => Stream m a -> m Bool
-{-# INLINE_FUSED null #-}
-null (Stream step t) = null_loop t
-  where
-    null_loop s = do
-      r <- step s
-      case r of
-        Yield _ _ -> return False
-        Skip s'   -> null_loop s'
-        Done      -> return True
-
--- Construction
--- ------------
-
--- | Empty 'Stream'
-empty :: Monad m => Stream m a
-{-# INLINE_FUSED empty #-}
-empty = Stream (const (return Done)) ()
-
--- | Singleton 'Stream'
-singleton :: Monad m => a -> Stream m a
-{-# INLINE_FUSED singleton #-}
-singleton x = Stream (return . step) True
-  where
-    {-# INLINE_INNER step #-}
-    step True  = Yield x False
-    step False = Done
-
--- | Replicate a value to a given length
-replicate :: Monad m => Int -> a -> Stream m a
-{-# INLINE_FUSED replicate #-}
-replicate n x = replicateM n (return x)
-
--- | Yield a 'Stream' of values obtained by performing the monadic action the
--- given number of times
-replicateM :: Monad m => Int -> m a -> Stream m a
-{-# INLINE_FUSED replicateM #-}
-replicateM n p = Stream step n
-  where
-    {-# INLINE_INNER step #-}
-    step i | i <= 0    = return Done
-           | otherwise = do { x <- p; return $ Yield x (i-1) }
-
-generate :: Monad m => Int -> (Int -> a) -> Stream m a
-{-# INLINE generate #-}
-generate n f = generateM n (return . f)
-
--- | Generate a stream from its indices
-generateM :: Monad m => Int -> (Int -> m a) -> Stream m a
-{-# INLINE_FUSED generateM #-}
-generateM n f = n `seq` Stream step 0
-  where
-    {-# INLINE_INNER step #-}
-    step i | i < n     = do
-                           x <- f i
-                           return $ Yield x (i+1)
-           | otherwise = return Done
-
--- | Prepend an element
-cons :: Monad m => a -> Stream m a -> Stream m a
-{-# INLINE cons #-}
-cons x s = singleton x ++ s
-
--- | Append an element
-snoc :: Monad m => Stream m a -> a -> Stream m a
-{-# INLINE snoc #-}
-snoc s x = s ++ singleton x
-
-infixr 5 ++
--- | Concatenate two 'Stream's
-(++) :: Monad m => Stream m a -> Stream m a -> Stream m a
-{-# INLINE_FUSED (++) #-}
-Stream stepa ta ++ Stream stepb tb = Stream step (Left ta)
-  where
-    {-# INLINE_INNER step #-}
-    step (Left  sa) = do
-                        r <- stepa sa
-                        case r of
-                          Yield x sa' -> return $ Yield x (Left  sa')
-                          Skip    sa' -> return $ Skip    (Left  sa')
-                          Done        -> return $ Skip    (Right tb)
-    step (Right sb) = do
-                        r <- stepb sb
-                        case r of
-                          Yield x sb' -> return $ Yield x (Right sb')
-                          Skip    sb' -> return $ Skip    (Right sb')
-                          Done        -> return $ Done
-
--- Accessing elements
--- ------------------
-
--- | First element of the 'Stream' or error if empty
-head :: Monad m => Stream m a -> m a
-{-# INLINE_FUSED head #-}
-head (Stream step t) = head_loop SPEC t
-  where
-    head_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x _  -> return x
-            Skip    s' -> head_loop SPEC s'
-            Done       -> EMPTY_STREAM "head"
-
-
-
--- | Last element of the 'Stream' or error if empty
-last :: Monad m => Stream m a -> m a
-{-# INLINE_FUSED last #-}
-last (Stream step t) = last_loop0 SPEC t
-  where
-    last_loop0 !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> last_loop1 SPEC x s'
-            Skip    s' -> last_loop0 SPEC   s'
-            Done       -> EMPTY_STREAM "last"
-
-    last_loop1 !_ x s
-      = do
-          r <- step s
-          case r of
-            Yield y s' -> last_loop1 SPEC y s'
-            Skip    s' -> last_loop1 SPEC x s'
-            Done       -> return x
-
-infixl 9 !!
--- | Element at the given position
-(!!) :: Monad m => Stream m a -> Int -> m a
-{-# INLINE (!!) #-}
-Stream step t !! j | j < 0     = ERROR "!!" "negative index"
-                   | otherwise = index_loop SPEC t j
-  where
-    index_loop !_ s i
-      = i `seq`
-        do
-          r <- step s
-          case r of
-            Yield x s' | i == 0    -> return x
-                       | otherwise -> index_loop SPEC s' (i-1)
-            Skip    s'             -> index_loop SPEC s' i
-            Done                   -> EMPTY_STREAM "!!"
-
-infixl 9 !?
--- | Element at the given position or 'Nothing' if out of bounds
-(!?) :: Monad m => Stream m a -> Int -> m (Maybe a)
-{-# INLINE (!?) #-}
-Stream step t !? j = index_loop SPEC t j
-  where
-    index_loop !_ s i
-      = i `seq`
-        do
-          r <- step s
-          case r of
-            Yield x s' | i == 0    -> return (Just x)
-                       | otherwise -> index_loop SPEC s' (i-1)
-            Skip    s'             -> index_loop SPEC s' i
-            Done                   -> return Nothing
-
--- Substreams
--- ----------
-
--- | Extract a substream of the given length starting at the given position.
-slice :: Monad m => Int   -- ^ starting index
-                 -> Int   -- ^ length
-                 -> Stream m a
-                 -> Stream m a
-{-# INLINE slice #-}
-slice i n s = take n (drop i s)
-
--- | All but the last element
-init :: Monad m => Stream m a -> Stream m a
-{-# INLINE_FUSED init #-}
-init (Stream step t) = Stream step' (Nothing, t)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (Nothing, s) = liftM (\r ->
-                           case r of
-                             Yield x s' -> Skip (Just x,  s')
-                             Skip    s' -> Skip (Nothing, s')
-                             Done       -> EMPTY_STREAM "init"
-                         ) (step s)
-
-    step' (Just x,  s) = liftM (\r ->
-                           case r of
-                             Yield y s' -> Yield x (Just y, s')
-                             Skip    s' -> Skip    (Just x, s')
-                             Done       -> Done
-                         ) (step s)
-
--- | All but the first element
-tail :: Monad m => Stream m a -> Stream m a
-{-# INLINE_FUSED tail #-}
-tail (Stream step t) = Stream step' (Left t)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (Left  s) = liftM (\r ->
-                        case r of
-                          Yield _ s' -> Skip (Right s')
-                          Skip    s' -> Skip (Left  s')
-                          Done       -> EMPTY_STREAM "tail"
-                      ) (step s)
-
-    step' (Right s) = liftM (\r ->
-                        case r of
-                          Yield x s' -> Yield x (Right s')
-                          Skip    s' -> Skip    (Right s')
-                          Done       -> Done
-                      ) (step s)
-
--- | The first @n@ elements
-take :: Monad m => Int -> Stream m a -> Stream m a
-{-# INLINE_FUSED take #-}
-take n (Stream step t) = n `seq` Stream step' (t, 0)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s, i) | i < n = liftM (\r ->
-                             case r of
-                               Yield x s' -> Yield x (s', i+1)
-                               Skip    s' -> Skip    (s', i)
-                               Done       -> Done
-                           ) (step s)
-    step' (_, _) = return Done
-
--- | All but the first @n@ elements
-drop :: Monad m => Int -> Stream m a -> Stream m a
-{-# INLINE_FUSED drop #-}
-drop n (Stream step t) = Stream step' (t, Just n)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s, Just i) | i > 0 = liftM (\r ->
-                                case r of
-                                   Yield _ s' -> Skip (s', Just (i-1))
-                                   Skip    s' -> Skip (s', Just i)
-                                   Done       -> Done
-                                ) (step s)
-                      | otherwise = return $ Skip (s, Nothing)
-
-    step' (s, Nothing) = liftM (\r ->
-                           case r of
-                             Yield x s' -> Yield x (s', Nothing)
-                             Skip    s' -> Skip    (s', Nothing)
-                             Done       -> Done
-                           ) (step s)
-
--- Mapping
--- -------
-
-instance Monad m => Functor (Stream m) where
-  {-# INLINE fmap #-}
-  fmap = map
-
--- | Map a function over a 'Stream'
-map :: Monad m => (a -> b) -> Stream m a -> Stream m b
-{-# INLINE map #-}
-map f = mapM (return . f)
-
-
--- | Map a monadic function over a 'Stream'
-mapM :: Monad m => (a -> m b) -> Stream m a -> Stream m b
-{-# INLINE_FUSED mapM #-}
-mapM f (Stream step t) = Stream step' t
-  where
-    {-# INLINE_INNER step' #-}
-    step' s = do
-                r <- step s
-                case r of
-                  Yield x s' -> liftM  (`Yield` s') (f x)
-                  Skip    s' -> return (Skip    s')
-                  Done       -> return Done
-
-consume :: Monad m => Stream m a -> m ()
-{-# INLINE_FUSED consume #-}
-consume (Stream step t) = consume_loop SPEC t
-  where
-    consume_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield _ s' -> consume_loop SPEC s'
-            Skip    s' -> consume_loop SPEC s'
-            Done       -> return ()
-
--- | Execute a monadic action for each element of the 'Stream'
-mapM_ :: Monad m => (a -> m b) -> Stream m a -> m ()
-{-# INLINE_FUSED mapM_ #-}
-mapM_ m = consume . mapM m
-
--- | Transform a 'Stream' to use a different monad
-trans :: (Monad m, Monad m')
-      => (forall z. m z -> m' z) -> Stream m a -> Stream m' a
-{-# INLINE_FUSED trans #-}
-trans f (Stream step s) = Stream (f . step) s
-
-unbox :: Monad m => Stream m (Box a) -> Stream m a
-{-# INLINE_FUSED unbox #-}
-unbox (Stream step t) = Stream step' t
-  where
-    {-# INLINE_INNER step' #-}
-    step' s = do
-                r <- step s
-                case r of
-                  Yield (Box x) s' -> return $ Yield x s'
-                  Skip          s' -> return $ Skip    s'
-                  Done             -> return $ Done
-
--- Zipping
--- -------
-
--- | Pair each element in a 'Stream' with its index
-indexed :: Monad m => Stream m a -> Stream m (Int,a)
-{-# INLINE_FUSED indexed #-}
-indexed (Stream step t) = Stream step' (t,0)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,i) = i `seq`
-                  do
-                    r <- step s
-                    case r of
-                      Yield x s' -> return $ Yield (i,x) (s', i+1)
-                      Skip    s' -> return $ Skip        (s', i)
-                      Done       -> return Done
-
--- | Pair each element in a 'Stream' with its index, starting from the right
--- and counting down
-indexedR :: Monad m => Int -> Stream m a -> Stream m (Int,a)
-{-# INLINE_FUSED indexedR #-}
-indexedR m (Stream step t) = Stream step' (t,m)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,i) = i `seq`
-                  do
-                    r <- step s
-                    case r of
-                      Yield x s' -> let i' = i-1
-                                    in
-                                    return $ Yield (i',x) (s', i')
-                      Skip    s' -> return $ Skip         (s', i)
-                      Done       -> return Done
-
--- | Zip two 'Stream's with the given monadic function
-zipWithM :: Monad m => (a -> b -> m c) -> Stream m a -> Stream m b -> Stream m c
-{-# INLINE_FUSED zipWithM #-}
-zipWithM f (Stream stepa ta) (Stream stepb tb) = Stream step (ta, tb, Nothing)
-  where
-    {-# INLINE_INNER step #-}
-    step (sa, sb, Nothing) = liftM (\r ->
-                               case r of
-                                 Yield x sa' -> Skip (sa', sb, Just x)
-                                 Skip    sa' -> Skip (sa', sb, Nothing)
-                                 Done        -> Done
-                             ) (stepa sa)
-
-    step (sa, sb, Just x)  = do
-                               r <- stepb sb
-                               case r of
-                                 Yield y sb' ->
-                                   do
-                                     z <- f x y
-                                     return $ Yield z (sa, sb', Nothing)
-                                 Skip    sb' -> return $ Skip (sa, sb', Just x)
-                                 Done        -> return $ Done
-
--- FIXME: This might expose an opportunity for inplace execution.
-{-# RULES
-
-"zipWithM xs xs [Vector.Stream]" forall f xs.
-  zipWithM f xs xs = mapM (\x -> f x x) xs   #-}
-
-
-zipWithM_ :: Monad m => (a -> b -> m c) -> Stream m a -> Stream m b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ f sa sb = consume (zipWithM f sa sb)
-
-zipWith3M :: Monad m => (a -> b -> c -> m d) -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-{-# INLINE_FUSED zipWith3M #-}
-zipWith3M f (Stream stepa ta)
-            (Stream stepb tb)
-            (Stream stepc tc) = Stream step (ta, tb, tc, Nothing)
-  where
-    {-# INLINE_INNER step #-}
-    step (sa, sb, sc, Nothing) = do
-        r <- stepa sa
-        return $ case r of
-            Yield x sa' -> Skip (sa', sb, sc, Just (x, Nothing))
-            Skip    sa' -> Skip (sa', sb, sc, Nothing)
-            Done        -> Done
-
-    step (sa, sb, sc, Just (x, Nothing)) = do
-        r <- stepb sb
-        return $ case r of
-            Yield y sb' -> Skip (sa, sb', sc, Just (x, Just y))
-            Skip    sb' -> Skip (sa, sb', sc, Just (x, Nothing))
-            Done        -> Done
-
-    step (sa, sb, sc, Just (x, Just y)) = do
-        r <- stepc sc
-        case r of
-            Yield z sc' -> f x y z >>= (\res -> return $ Yield res (sa, sb, sc', Nothing))
-            Skip    sc' -> return $ Skip (sa, sb, sc', Just (x, Just y))
-            Done        -> return $ Done
-
-zipWith4M :: Monad m => (a -> b -> c -> d -> m e)
-                     -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                     -> Stream m e
-{-# INLINE zipWith4M #-}
-zipWith4M f sa sb sc sd
-  = zipWithM (\(a,b) (c,d) -> f a b c d) (zip sa sb) (zip sc sd)
-
-zipWith5M :: Monad m => (a -> b -> c -> d -> e -> m f)
-                     -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                     -> Stream m e -> Stream m f
-{-# INLINE zipWith5M #-}
-zipWith5M f sa sb sc sd se
-  = zipWithM (\(a,b,c) (d,e) -> f a b c d e) (zip3 sa sb sc) (zip sd se)
-
-zipWith6M :: Monad m => (a -> b -> c -> d -> e -> f -> m g)
-                     -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                     -> Stream m e -> Stream m f -> Stream m g
-{-# INLINE zipWith6M #-}
-zipWith6M fn sa sb sc sd se sf
-  = zipWithM (\(a,b,c) (d,e,f) -> fn a b c d e f) (zip3 sa sb sc)
-                                                  (zip3 sd se sf)
-
-zipWith :: Monad m => (a -> b -> c) -> Stream m a -> Stream m b -> Stream m c
-{-# INLINE zipWith #-}
-zipWith f = zipWithM (\a b -> return (f a b))
-
-zipWith3 :: Monad m => (a -> b -> c -> d)
-                    -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-{-# INLINE zipWith3 #-}
-zipWith3 f = zipWith3M (\a b c -> return (f a b c))
-
-zipWith4 :: Monad m => (a -> b -> c -> d -> e)
-                    -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                    -> Stream m e
-{-# INLINE zipWith4 #-}
-zipWith4 f = zipWith4M (\a b c d -> return (f a b c d))
-
-zipWith5 :: Monad m => (a -> b -> c -> d -> e -> f)
-                    -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                    -> Stream m e -> Stream m f
-{-# INLINE zipWith5 #-}
-zipWith5 f = zipWith5M (\a b c d e -> return (f a b c d e))
-
-zipWith6 :: Monad m => (a -> b -> c -> d -> e -> f -> g)
-                    -> Stream m a -> Stream m b -> Stream m c -> Stream m d
-                    -> Stream m e -> Stream m f -> Stream m g
-{-# INLINE zipWith6 #-}
-zipWith6 fn = zipWith6M (\a b c d e f -> return (fn a b c d e f))
-
-zip :: Monad m => Stream m a -> Stream m b -> Stream m (a,b)
-{-# INLINE zip #-}
-zip = zipWith (,)
-
-zip3 :: Monad m => Stream m a -> Stream m b -> Stream m c -> Stream m (a,b,c)
-{-# INLINE zip3 #-}
-zip3 = zipWith3 (,,)
-
-zip4 :: Monad m => Stream m a -> Stream m b -> Stream m c -> Stream m d
-                -> Stream m (a,b,c,d)
-{-# INLINE zip4 #-}
-zip4 = zipWith4 (,,,)
-
-zip5 :: Monad m => Stream m a -> Stream m b -> Stream m c -> Stream m d
-                -> Stream m e -> Stream m (a,b,c,d,e)
-{-# INLINE zip5 #-}
-zip5 = zipWith5 (,,,,)
-
-zip6 :: Monad m => Stream m a -> Stream m b -> Stream m c -> Stream m d
-                -> Stream m e -> Stream m f -> Stream m (a,b,c,d,e,f)
-{-# INLINE zip6 #-}
-zip6 = zipWith6 (,,,,,)
-
--- Comparisons
--- -----------
-
--- | Check if two 'Stream's are equal
-eqBy :: (Monad m) => (a -> b -> Bool) -> Stream m a -> Stream m b -> m Bool
-{-# INLINE_FUSED eqBy #-}
-eqBy eq (Stream step1 t1) (Stream step2 t2) = eq_loop0 SPEC t1 t2
-  where
-    eq_loop0 !_ s1 s2 = do
-      r <- step1 s1
-      case r of
-        Yield x s1' -> eq_loop1 SPEC x s1' s2
-        Skip    s1' -> eq_loop0 SPEC   s1' s2
-        Done        -> eq_null s2
-
-    eq_loop1 !_ x s1 s2 = do
-      r <- step2 s2
-      case r of
-        Yield y s2'
-          | eq x y    -> eq_loop0 SPEC   s1 s2'
-          | otherwise -> return False
-        Skip    s2'   -> eq_loop1 SPEC x s1 s2'
-        Done          -> return False
-
-    eq_null s2 = do
-      r <- step2 s2
-      case r of
-        Yield _ _ -> return False
-        Skip s2'  -> eq_null s2'
-        Done      -> return True
-
--- | Lexicographically compare two 'Stream's
-cmpBy :: (Monad m) => (a -> b -> Ordering) -> Stream m a -> Stream m b -> m Ordering
-{-# INLINE_FUSED cmpBy #-}
-cmpBy cmp (Stream step1 t1) (Stream step2 t2) = cmp_loop0 SPEC t1 t2
-  where
-    cmp_loop0 !_ s1 s2 = do
-      r <- step1 s1
-      case r of
-        Yield x s1' -> cmp_loop1 SPEC x s1' s2
-        Skip    s1' -> cmp_loop0 SPEC   s1' s2
-        Done        -> cmp_null s2
-
-    cmp_loop1 !_ x s1 s2 = do
-      r <- step2 s2
-      case r of
-        Yield y s2' -> case x `cmp` y of
-                         EQ -> cmp_loop0 SPEC s1 s2'
-                         c  -> return c
-        Skip    s2' -> cmp_loop1 SPEC x s1 s2'
-        Done        -> return GT
-
-    cmp_null s2 = do
-      r <- step2 s2
-      case r of
-        Yield _ _ -> return LT
-        Skip s2'  -> cmp_null s2'
-        Done      -> return EQ
-
--- Filtering
--- ---------
-
--- | Drop elements which do not satisfy the predicate
-filter :: Monad m => (a -> Bool) -> Stream m a -> Stream m a
-{-# INLINE filter #-}
-filter f = filterM (return . f)
-
-mapMaybe :: Monad m => (a -> Maybe b) -> Stream m a -> Stream m b
-{-# INLINE_FUSED mapMaybe #-}
-mapMaybe f (Stream step t) = Stream step' t
-  where
-    {-# INLINE_INNER step' #-}
-    step' s = do
-                r <- step s
-                case r of
-                  Yield x s' -> do
-                                  return $ case f x of
-                                    Nothing -> Skip s'
-                                    Just b' -> Yield b' s'
-                  Skip    s' -> return $ Skip s'
-                  Done       -> return $ Done
-
--- | Drop elements which do not satisfy the monadic predicate
-filterM :: Monad m => (a -> m Bool) -> Stream m a -> Stream m a
-{-# INLINE_FUSED filterM #-}
-filterM f (Stream step t) = Stream step' t
-  where
-    {-# INLINE_INNER step' #-}
-    step' s = do
-                r <- step s
-                case r of
-                  Yield x s' -> do
-                                  b <- f x
-                                  return $ if b then Yield x s'
-                                                else Skip    s'
-                  Skip    s' -> return $ Skip s'
-                  Done       -> return $ Done
-
--- | Drop repeated adjacent elements.
-uniq :: (Eq a, Monad m) => Stream m a -> Stream m a
-{-# INLINE_FUSED uniq #-}
-uniq (Stream step st) = Stream step' (Nothing,st)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (Nothing, s) = do r <- step s
-                            case r of
-                              Yield x s' -> return $ Yield x (Just x , s')
-                              Skip  s'   -> return $ Skip  (Nothing, s')
-                              Done       -> return   Done
-    step' (Just x0, s) = do r <- step s
-                            case r of
-                              Yield x s' | x == x0   -> return $ Skip    (Just x0, s')
-                                         | otherwise -> return $ Yield x (Just x , s')
-                              Skip  s'   -> return $ Skip (Just x0, s')
-                              Done       -> return   Done
-
--- | Longest prefix of elements that satisfy the predicate
-takeWhile :: Monad m => (a -> Bool) -> Stream m a -> Stream m a
-{-# INLINE takeWhile #-}
-takeWhile f = takeWhileM (return . f)
-
--- | Longest prefix of elements that satisfy the monadic predicate
-takeWhileM :: Monad m => (a -> m Bool) -> Stream m a -> Stream m a
-{-# INLINE_FUSED takeWhileM #-}
-takeWhileM f (Stream step t) = Stream step' t
-  where
-    {-# INLINE_INNER step' #-}
-    step' s = do
-                r <- step s
-                case r of
-                  Yield x s' -> do
-                                  b <- f x
-                                  return $ if b then Yield x s' else Done
-                  Skip    s' -> return $ Skip s'
-                  Done       -> return $ Done
-
--- | Drop the longest prefix of elements that satisfy the predicate
-dropWhile :: Monad m => (a -> Bool) -> Stream m a -> Stream m a
-{-# INLINE dropWhile #-}
-dropWhile f = dropWhileM (return . f)
-
-data DropWhile s a = DropWhile_Drop s | DropWhile_Yield a s | DropWhile_Next s
-
--- | Drop the longest prefix of elements that satisfy the monadic predicate
-dropWhileM :: Monad m => (a -> m Bool) -> Stream m a -> Stream m a
-{-# INLINE_FUSED dropWhileM #-}
-dropWhileM f (Stream step t) = Stream step' (DropWhile_Drop t)
-  where
-    -- NOTE: we jump through hoops here to have only one Yield; local data
-    -- declarations would be nice!
-
-    {-# INLINE_INNER step' #-}
-    step' (DropWhile_Drop s)
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> do
-                            b <- f x
-                            return $ if b then Skip (DropWhile_Drop    s')
-                                          else Skip (DropWhile_Yield x s')
-            Skip    s' -> return $ Skip (DropWhile_Drop    s')
-            Done       -> return $ Done
-
-    step' (DropWhile_Yield x s) = return $ Yield x (DropWhile_Next s)
-
-    step' (DropWhile_Next s)
-      = liftM (\r ->
-          case r of
-            Yield x s' -> Skip    (DropWhile_Yield x s')
-            Skip    s' -> Skip    (DropWhile_Next    s')
-            Done       -> Done
-        ) (step s)
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | Check whether the 'Stream' contains an element
-elem :: (Monad m, Eq a) => a -> Stream m a -> m Bool
-{-# INLINE_FUSED elem #-}
-elem x (Stream step t) = elem_loop SPEC t
-  where
-    elem_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield y s' | x == y    -> return True
-                       | otherwise -> elem_loop SPEC s'
-            Skip    s'             -> elem_loop SPEC s'
-            Done                   -> return False
-
-infix 4 `notElem`
--- | Inverse of `elem`
-notElem :: (Monad m, Eq a) => a -> Stream m a -> m Bool
-{-# INLINE notElem #-}
-notElem x s = liftM not (elem x s)
-
--- | Yield 'Just' the first element that satisfies the predicate or 'Nothing'
--- if no such element exists.
-find :: Monad m => (a -> Bool) -> Stream m a -> m (Maybe a)
-{-# INLINE find #-}
-find f = findM (return . f)
-
--- | Yield 'Just' the first element that satisfies the monadic predicate or
--- 'Nothing' if no such element exists.
-findM :: Monad m => (a -> m Bool) -> Stream m a -> m (Maybe a)
-{-# INLINE_FUSED findM #-}
-findM f (Stream step t) = find_loop SPEC t
-  where
-    find_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> do
-                            b <- f x
-                            if b then return $ Just x
-                                 else find_loop SPEC s'
-            Skip    s' -> find_loop SPEC s'
-            Done       -> return Nothing
-
--- | Yield 'Just' the index of the first element that satisfies the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Monad m => (a -> Bool) -> Stream m a -> m (Maybe Int)
-{-# INLINE_FUSED findIndex #-}
-findIndex f = findIndexM (return . f)
-
--- | Yield 'Just' the index of the first element that satisfies the monadic
--- predicate or 'Nothing' if no such element exists.
-findIndexM :: Monad m => (a -> m Bool) -> Stream m a -> m (Maybe Int)
-{-# INLINE_FUSED findIndexM #-}
-findIndexM f (Stream step t) = findIndex_loop SPEC t 0
-  where
-    findIndex_loop !_ s i
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> do
-                            b <- f x
-                            if b then return $ Just i
-                                 else findIndex_loop SPEC s' (i+1)
-            Skip    s' -> findIndex_loop SPEC s' i
-            Done       -> return Nothing
-
--- Folding
--- -------
-
--- | Left fold
-foldl :: Monad m => (a -> b -> a) -> a -> Stream m b -> m a
-{-# INLINE foldl #-}
-foldl f = foldlM (\a b -> return (f a b))
-
--- | Left fold with a monadic operator
-foldlM :: Monad m => (a -> b -> m a) -> a -> Stream m b -> m a
-{-# INLINE_FUSED foldlM #-}
-foldlM m w (Stream step t) = foldlM_loop SPEC w t
-  where
-    foldlM_loop !_ z s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> do { z' <- m z x; foldlM_loop SPEC z' s' }
-            Skip    s' -> foldlM_loop SPEC z s'
-            Done       -> return z
-
--- | Same as 'foldlM'
-foldM :: Monad m => (a -> b -> m a) -> a -> Stream m b -> m a
-{-# INLINE foldM #-}
-foldM = foldlM
-
--- | Left fold over a non-empty 'Stream'
-foldl1 :: Monad m => (a -> a -> a) -> Stream m a -> m a
-{-# INLINE foldl1 #-}
-foldl1 f = foldl1M (\a b -> return (f a b))
-
--- | Left fold over a non-empty 'Stream' with a monadic operator
-foldl1M :: Monad m => (a -> a -> m a) -> Stream m a -> m a
-{-# INLINE_FUSED foldl1M #-}
-foldl1M f (Stream step t) = foldl1M_loop SPEC t
-  where
-    foldl1M_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> foldlM f x (Stream step s')
-            Skip    s' -> foldl1M_loop SPEC s'
-            Done       -> EMPTY_STREAM "foldl1M"
-
--- | Same as 'foldl1M'
-fold1M :: Monad m => (a -> a -> m a) -> Stream m a -> m a
-{-# INLINE fold1M #-}
-fold1M = foldl1M
-
--- | Left fold with a strict accumulator
-foldl' :: Monad m => (a -> b -> a) -> a -> Stream m b -> m a
-{-# INLINE foldl' #-}
-foldl' f = foldlM' (\a b -> return (f a b))
-
--- | Left fold with a strict accumulator and a monadic operator
-foldlM' :: Monad m => (a -> b -> m a) -> a -> Stream m b -> m a
-{-# INLINE_FUSED foldlM' #-}
-foldlM' m w (Stream step t) = foldlM'_loop SPEC w t
-  where
-    foldlM'_loop !_ z s
-      = z `seq`
-        do
-          r <- step s
-          case r of
-            Yield x s' -> do { z' <- m z x; foldlM'_loop SPEC z' s' }
-            Skip    s' -> foldlM'_loop SPEC z s'
-            Done       -> return z
-
--- | Same as 'foldlM''
-foldM' :: Monad m => (a -> b -> m a) -> a -> Stream m b -> m a
-{-# INLINE foldM' #-}
-foldM' = foldlM'
-
--- | Left fold over a non-empty 'Stream' with a strict accumulator
-foldl1' :: Monad m => (a -> a -> a) -> Stream m a -> m a
-{-# INLINE foldl1' #-}
-foldl1' f = foldl1M' (\a b -> return (f a b))
-
--- | Left fold over a non-empty 'Stream' with a strict accumulator and a
--- monadic operator
-foldl1M' :: Monad m => (a -> a -> m a) -> Stream m a -> m a
-{-# INLINE_FUSED foldl1M' #-}
-foldl1M' f (Stream step t) = foldl1M'_loop SPEC t
-  where
-    foldl1M'_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> foldlM' f x (Stream step s')
-            Skip    s' -> foldl1M'_loop SPEC s'
-            Done       -> EMPTY_STREAM "foldl1M'"
-
--- | Same as 'foldl1M''
-fold1M' :: Monad m => (a -> a -> m a) -> Stream m a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = foldl1M'
-
--- | Right fold
-foldr :: Monad m => (a -> b -> b) -> b -> Stream m a -> m b
-{-# INLINE foldr #-}
-foldr f = foldrM (\a b -> return (f a b))
-
--- | Right fold with a monadic operator
-foldrM :: Monad m => (a -> b -> m b) -> b -> Stream m a -> m b
-{-# INLINE_FUSED foldrM #-}
-foldrM f z (Stream step t) = foldrM_loop SPEC t
-  where
-    foldrM_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> f x =<< foldrM_loop SPEC s'
-            Skip    s' -> foldrM_loop SPEC s'
-            Done       -> return z
-
--- | Right fold over a non-empty stream
-foldr1 :: Monad m => (a -> a -> a) -> Stream m a -> m a
-{-# INLINE foldr1 #-}
-foldr1 f = foldr1M (\a b -> return (f a b))
-
--- | Right fold over a non-empty stream with a monadic operator
-foldr1M :: Monad m => (a -> a -> m a) -> Stream m a -> m a
-{-# INLINE_FUSED foldr1M #-}
-foldr1M f (Stream step t) = foldr1M_loop0 SPEC t
-  where
-    foldr1M_loop0 !_ s
-      = do
-          r <- step s
-          case r of
-            Yield x s' -> foldr1M_loop1 SPEC x s'
-            Skip    s' -> foldr1M_loop0 SPEC   s'
-            Done       -> EMPTY_STREAM "foldr1M"
-
-    foldr1M_loop1 !_ x s
-      = do
-          r <- step s
-          case r of
-            Yield y s' -> f x =<< foldr1M_loop1 SPEC y s'
-            Skip    s' -> foldr1M_loop1 SPEC x s'
-            Done       -> return x
-
--- Specialised folds
--- -----------------
-
-and :: Monad m => Stream m Bool -> m Bool
-{-# INLINE_FUSED and #-}
-and (Stream step t) = and_loop SPEC t
-  where
-    and_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield False _  -> return False
-            Yield True  s' -> and_loop SPEC s'
-            Skip        s' -> and_loop SPEC s'
-            Done           -> return True
-
-or :: Monad m => Stream m Bool -> m Bool
-{-# INLINE_FUSED or #-}
-or (Stream step t) = or_loop SPEC t
-  where
-    or_loop !_ s
-      = do
-          r <- step s
-          case r of
-            Yield False s' -> or_loop SPEC s'
-            Yield True  _  -> return True
-            Skip        s' -> or_loop SPEC s'
-            Done           -> return False
-
-concatMap :: Monad m => (a -> Stream m b) -> Stream m a -> Stream m b
-{-# INLINE concatMap #-}
-concatMap f = concatMapM (return . f)
-
-concatMapM :: Monad m => (a -> m (Stream m b)) -> Stream m a -> Stream m b
-{-# INLINE_FUSED concatMapM #-}
-concatMapM f (Stream step t) = Stream concatMap_go (Left t)
-  where
-    concatMap_go (Left s) = do
-        r <- step s
-        case r of
-            Yield a s' -> do
-                b_stream <- f a
-                return $ Skip (Right (b_stream, s'))
-            Skip    s' -> return $ Skip (Left s')
-            Done       -> return Done
-    concatMap_go (Right (Stream inner_step inner_s, s)) = do
-        r <- inner_step inner_s
-        case r of
-            Yield b inner_s' -> return $ Yield b (Right (Stream inner_step inner_s', s))
-            Skip    inner_s' -> return $ Skip (Right (Stream inner_step inner_s', s))
-            Done             -> return $ Skip (Left s)
-
--- | Create a 'Stream' of values from a 'Stream' of streamable things
-flatten :: Monad m => (a -> m s) -> (s -> m (Step s b)) -> Stream m a -> Stream m b
-{-# INLINE_FUSED flatten #-}
-flatten mk istep (Stream ostep u) = Stream step (Left u)
-  where
-    {-# INLINE_INNER step #-}
-    step (Left t) = do
-                      r <- ostep t
-                      case r of
-                        Yield a t' -> do
-                                        s <- mk a
-                                        s `seq` return (Skip (Right (s,t')))
-                        Skip    t' -> return $ Skip (Left t')
-                        Done       -> return $ Done
-
-
-    step (Right (s,t)) = do
-                           r <- istep s
-                           case r of
-                             Yield x s' -> return $ Yield x (Right (s',t))
-                             Skip    s' -> return $ Skip    (Right (s',t))
-                             Done       -> return $ Skip    (Left t)
-
--- Unfolding
--- ---------
-
--- | Unfold
-unfoldr :: Monad m => (s -> Maybe (a, s)) -> s -> Stream m a
-{-# INLINE_FUSED unfoldr #-}
-unfoldr f = unfoldrM (return . f)
-
--- | Unfold with a monadic function
-unfoldrM :: Monad m => (s -> m (Maybe (a, s))) -> s -> Stream m a
-{-# INLINE_FUSED unfoldrM #-}
-unfoldrM f t = Stream step t
-  where
-    {-# INLINE_INNER step #-}
-    step s = liftM (\r ->
-               case r of
-                 Just (x, s') -> Yield x s'
-                 Nothing      -> Done
-             ) (f s)
-
-unfoldrN :: Monad m => Int -> (s -> Maybe (a, s)) -> s -> Stream m a
-{-# INLINE_FUSED unfoldrN #-}
-unfoldrN n f = unfoldrNM n (return . f)
-
--- | Unfold at most @n@ elements with a monadic functions
-unfoldrNM :: Monad m => Int -> (s -> m (Maybe (a, s))) -> s -> Stream m a
-{-# INLINE_FUSED unfoldrNM #-}
-unfoldrNM m f t = Stream step (t,m)
-  where
-    {-# INLINE_INNER step #-}
-    step (s,n) | n <= 0    = return Done
-               | otherwise = liftM (\r ->
-                               case r of
-                                 Just (x,s') -> Yield x (s',n-1)
-                                 Nothing     -> Done
-                             ) (f s)
-
--- | Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: Monad m => Int -> (a -> m a) -> a -> Stream m a
-{-# INLINE_FUSED iterateNM #-}
-iterateNM n f x0 = Stream step (x0,n)
-  where
-    {-# INLINE_INNER step #-}
-    step (x,i) | i <= 0    = return Done
-               | i == n    = return $ Yield x (x,i-1)
-               | otherwise = do a <- f x
-                                return $ Yield a (a,i-1)
-
--- | Apply function n times to value. Zeroth element is original value.
-iterateN :: Monad m => Int -> (a -> a) -> a -> Stream m a
-{-# INLINE_FUSED iterateN #-}
-iterateN n f x0 = iterateNM n (return . f) x0
-
--- Scans
--- -----
-
--- | Prefix scan
-prescanl :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE prescanl #-}
-prescanl f = prescanlM (\a b -> return (f a b))
-
--- | Prefix scan with a monadic operator
-prescanlM :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE_FUSED prescanlM #-}
-prescanlM f w (Stream step t) = Stream step' (t,w)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,x) = do
-                    r <- step s
-                    case r of
-                      Yield y s' -> do
-                                      z <- f x y
-                                      return $ Yield x (s', z)
-                      Skip    s' -> return $ Skip (s', x)
-                      Done       -> return Done
-
--- | Prefix scan with strict accumulator
-prescanl' :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE prescanl' #-}
-prescanl' f = prescanlM' (\a b -> return (f a b))
-
--- | Prefix scan with strict accumulator and a monadic operator
-prescanlM' :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE_FUSED prescanlM' #-}
-prescanlM' f w (Stream step t) = Stream step' (t,w)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,x) = x `seq`
-                  do
-                    r <- step s
-                    case r of
-                      Yield y s' -> do
-                                      z <- f x y
-                                      return $ Yield x (s', z)
-                      Skip    s' -> return $ Skip (s', x)
-                      Done       -> return Done
-
--- | Suffix scan
-postscanl :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE postscanl #-}
-postscanl f = postscanlM (\a b -> return (f a b))
-
--- | Suffix scan with a monadic operator
-postscanlM :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE_FUSED postscanlM #-}
-postscanlM f w (Stream step t) = Stream step' (t,w)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,x) = do
-                    r <- step s
-                    case r of
-                      Yield y s' -> do
-                                      z <- f x y
-                                      return $ Yield z (s',z)
-                      Skip    s' -> return $ Skip (s',x)
-                      Done       -> return Done
-
--- | Suffix scan with strict accumulator
-postscanl' :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE postscanl' #-}
-postscanl' f = postscanlM' (\a b -> return (f a b))
-
--- | Suffix scan with strict acccumulator and a monadic operator
-postscanlM' :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE_FUSED postscanlM' #-}
-postscanlM' f w (Stream step t) = w `seq` Stream step' (t,w)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s,x) = x `seq`
-                  do
-                    r <- step s
-                    case r of
-                      Yield y s' -> do
-                                      z <- f x y
-                                      z `seq` return (Yield z (s',z))
-                      Skip    s' -> return $ Skip (s',x)
-                      Done       -> return Done
-
--- | Haskell-style scan
-scanl :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE scanl #-}
-scanl f = scanlM (\a b -> return (f a b))
-
--- | Haskell-style scan with a monadic operator
-scanlM :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE scanlM #-}
-scanlM f z s = z `cons` postscanlM f z s
-
--- | Haskell-style scan with strict accumulator
-scanl' :: Monad m => (a -> b -> a) -> a -> Stream m b -> Stream m a
-{-# INLINE scanl' #-}
-scanl' f = scanlM' (\a b -> return (f a b))
-
--- | Haskell-style scan with strict accumulator and a monadic operator
-scanlM' :: Monad m => (a -> b -> m a) -> a -> Stream m b -> Stream m a
-{-# INLINE scanlM' #-}
-scanlM' f z s = z `seq` (z `cons` postscanlM f z s)
-
--- | Scan over a non-empty 'Stream'
-scanl1 :: Monad m => (a -> a -> a) -> Stream m a -> Stream m a
-{-# INLINE scanl1 #-}
-scanl1 f = scanl1M (\x y -> return (f x y))
-
--- | Scan over a non-empty 'Stream' with a monadic operator
-scanl1M :: Monad m => (a -> a -> m a) -> Stream m a -> Stream m a
-{-# INLINE_FUSED scanl1M #-}
-scanl1M f (Stream step t) = Stream step' (t, Nothing)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s, Nothing) = do
-                           r <- step s
-                           case r of
-                             Yield x s' -> return $ Yield x (s', Just x)
-                             Skip    s' -> return $ Skip (s', Nothing)
-                             Done       -> EMPTY_STREAM "scanl1M"
-
-    step' (s, Just x) = do
-                          r <- step s
-                          case r of
-                            Yield y s' -> do
-                                            z <- f x y
-                                            return $ Yield z (s', Just z)
-                            Skip    s' -> return $ Skip (s', Just x)
-                            Done       -> return Done
-
--- | Scan over a non-empty 'Stream' with a strict accumulator
-scanl1' :: Monad m => (a -> a -> a) -> Stream m a -> Stream m a
-{-# INLINE scanl1' #-}
-scanl1' f = scanl1M' (\x y -> return (f x y))
-
--- | Scan over a non-empty 'Stream' with a strict accumulator and a monadic
--- operator
-scanl1M' :: Monad m => (a -> a -> m a) -> Stream m a -> Stream m a
-{-# INLINE_FUSED scanl1M' #-}
-scanl1M' f (Stream step t) = Stream step' (t, Nothing)
-  where
-    {-# INLINE_INNER step' #-}
-    step' (s, Nothing) = do
-                           r <- step s
-                           case r of
-                             Yield x s' -> x `seq` return (Yield x (s', Just x))
-                             Skip    s' -> return $ Skip (s', Nothing)
-                             Done       -> EMPTY_STREAM "scanl1M"
-
-    step' (s, Just x) = x `seq`
-                        do
-                          r <- step s
-                          case r of
-                            Yield y s' -> do
-                                            z <- f x y
-                                            z `seq` return (Yield z (s', Just z))
-                            Skip    s' -> return $ Skip (s', Just x)
-                            Done       -> return Done
-
--- Enumerations
--- ------------
-
--- The Enum class is broken for this, there just doesn't seem to be a
--- way to implement this generically. We have to specialise for as many types
--- as we can but this doesn't help in polymorphic loops.
-
--- | Yield a 'Stream' of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc.
-enumFromStepN :: (Num a, Monad m) => a -> a -> Int -> Stream m a
-{-# INLINE_FUSED enumFromStepN #-}
-enumFromStepN x y n = x `seq` y `seq` n `seq` Stream step (x,n)
-  where
-    {-# INLINE_INNER step #-}
-    step (w,m) | m > 0     = return $ Yield w (w+y,m-1)
-               | otherwise = return $ Done
-
--- | Enumerate values
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromTo :: (Enum a, Monad m) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo #-}
-enumFromTo x y = fromList [x .. y]
-
--- NOTE: We use (x+1) instead of (succ x) below because the latter checks for
--- overflow which can't happen here.
-
--- FIXME: add "too large" test for Int
-enumFromTo_small :: (Integral a, Monad m) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo_small #-}
-enumFromTo_small x y = x `seq` y `seq` Stream step x
-  where
-    {-# INLINE_INNER step #-}
-    step w | w <= y    = return $ Yield w (w+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Int8> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Int8 -> Int8 -> Stream m Int8
-
-"enumFromTo<Int16> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Int16 -> Int16 -> Stream m Int16
-
-"enumFromTo<Word8> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Word8 -> Word8 -> Stream m Word8
-
-"enumFromTo<Word16> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Word16 -> Word16 -> Stream m Word16   #-}
-
-
-#if WORD_SIZE_IN_BITS > 32
-
-{-# RULES
-
-"enumFromTo<Int32> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Int32 -> Int32 -> Stream m Int32
-
-"enumFromTo<Word32> [Stream]"
-  enumFromTo = enumFromTo_small :: Monad m => Word32 -> Word32 -> Stream m Word32   #-}
-
-
-#endif
-
--- NOTE: We could implement a generic "too large" test:
---
--- len x y | x > y = 0
---         | n > 0 && n <= fromIntegral (maxBound :: Int) = fromIntegral n
---         | otherwise = error
---   where
---     n = y-x+1
---
--- Alas, GHC won't eliminate unnecessary comparisons (such as n >= 0 for
--- unsigned types). See http://hackage.haskell.org/trac/ghc/ticket/3744
---
-
-enumFromTo_int :: forall m. Monad m => Int -> Int -> Stream m Int
-{-# INLINE_FUSED enumFromTo_int #-}
-enumFromTo_int x y = x `seq` y `seq` Stream step x
-  where
-    -- {-# INLINE [0] len #-}
-    -- len :: Int -> Int -> Int
-    -- len u v | u > v     = 0
-    --         | otherwise = BOUNDS_CHECK(check) "enumFromTo" "vector too large"
-    --                       (n > 0)
-    --                     $ n
-    --   where
-    --     n = v-u+1
-
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-enumFromTo_intlike :: (Integral a, Monad m) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo_intlike #-}
-enumFromTo_intlike x y = x `seq` y `seq` Stream step x
-  where
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Int> [Stream]"
-  enumFromTo = enumFromTo_int :: Monad m => Int -> Int -> Stream m Int
-
-#if WORD_SIZE_IN_BITS > 32
-
-"enumFromTo<Int64> [Stream]"
-  enumFromTo = enumFromTo_intlike :: Monad m => Int64 -> Int64 -> Stream m Int64 #-}
-
-#else
-
-"enumFromTo<Int32> [Stream]"
-  enumFromTo = enumFromTo_intlike :: Monad m => Int32 -> Int32 -> Stream m Int32 #-}
-
-#endif
-
-enumFromTo_big_word :: (Integral a, Monad m) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo_big_word #-}
-enumFromTo_big_word x y = x `seq` y `seq` Stream step x
-  where
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Word> [Stream]"
-  enumFromTo = enumFromTo_big_word :: Monad m => Word -> Word -> Stream m Word
-
-"enumFromTo<Word64> [Stream]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Word64 -> Word64 -> Stream m Word64
-
-#if WORD_SIZE_IN_BITS == 32
-
-"enumFromTo<Word32> [Stream]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Word32 -> Word32 -> Stream m Word32
-
-#endif
-
-"enumFromTo<Integer> [Stream]"
-  enumFromTo = enumFromTo_big_word
-                        :: Monad m => Integer -> Integer -> Stream m Integer   #-}
-
-
-
-#if WORD_SIZE_IN_BITS > 32
-
--- FIXME: the "too large" test is totally wrong
-enumFromTo_big_int :: (Integral a, Monad m) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo_big_int #-}
-enumFromTo_big_int x y = x `seq` y `seq` Stream step x
-  where
-    {-# INLINE_INNER step #-}
-    step z | z <= y    = return $ Yield z (z+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Int64> [Stream]"
-  enumFromTo = enumFromTo_big_int :: Monad m => Int64 -> Int64 -> Stream m Int64   #-}
-
-
-
-#endif
-
-enumFromTo_char :: Monad m => Char -> Char -> Stream m Char
-{-# INLINE_FUSED enumFromTo_char #-}
-enumFromTo_char x y = x `seq` y `seq` Stream step xn
-  where
-    xn = ord x
-    yn = ord y
-
-    {-# INLINE_INNER step #-}
-    step zn | zn <= yn  = return $ Yield (unsafeChr zn) (zn+1)
-            | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Char> [Stream]"
-  enumFromTo = enumFromTo_char   #-}
-
-
-
-------------------------------------------------------------------------
-
--- Specialise enumFromTo for Float and Double.
--- Also, try to do something about pairs?
-
-enumFromTo_double :: (Monad m, Ord a, RealFrac a) => a -> a -> Stream m a
-{-# INLINE_FUSED enumFromTo_double #-}
-enumFromTo_double n m = n `seq` m `seq` Stream step n
-  where
-    lim = m + 1/2 -- important to float out
-
-    {-# INLINE_INNER step #-}
-    step x | x <= lim  = return $ Yield x (x+1)
-           | otherwise = return $ Done
-
-{-# RULES
-
-"enumFromTo<Double> [Stream]"
-  enumFromTo = enumFromTo_double :: Monad m => Double -> Double -> Stream m Double
-
-"enumFromTo<Float> [Stream]"
-  enumFromTo = enumFromTo_double :: Monad m => Float -> Float -> Stream m Float   #-}
-
-
-
-------------------------------------------------------------------------
-
--- | Enumerate values with a given step.
---
--- /WARNING:/ This operation is very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Enum a, Monad m) => a -> a -> a -> Stream m a
-{-# INLINE_FUSED enumFromThenTo #-}
-enumFromThenTo x y z = fromList [x, y .. z]
-
--- FIXME: Specialise enumFromThenTo.
-
--- Conversions
--- -----------
-
--- | Convert a 'Stream' to a list
-toList :: Monad m => Stream m a -> m [a]
-{-# INLINE toList #-}
-toList = foldr (:) []
-
--- | Convert a list to a 'Stream'
-fromList :: Monad m => [a] -> Stream m a
-{-# INLINE fromList #-}
-fromList zs = Stream step zs
-  where
-    step (x:xs) = return (Yield x xs)
-    step []     = return Done
-
--- | Convert the first @n@ elements of a list to a 'Bundle'
-fromListN :: Monad m => Int -> [a] -> Stream m a
-{-# INLINE_FUSED fromListN #-}
-fromListN m zs = Stream step (zs,m)
-  where
-    {-# INLINE_INNER step #-}
-    step (_, n) | n <= 0 = return Done
-    step (x:xs,n)        = return (Yield x (xs,n-1))
-    step ([],_)          = return Done
-
-{-
-fromVector :: (Monad m, Vector v a) => v a -> Stream m a
-{-# INLINE_FUSED fromVector #-}
-fromVector v = v `seq` n `seq` Stream (Unf step 0)
-                                      (Unf vstep True)
-                                      (Just v)
-                                      (Exact n)
-  where
-    n = basicLength v
-
-    {-# INLINE step #-}
-    step i | i >= n = return Done
-           | otherwise = case basicUnsafeIndexM v i of
-                           Box x -> return $ Yield x (i+1)
-
-
-    {-# INLINE vstep #-}
-    vstep True  = return (Yield (Chunk (basicLength v) (\mv -> basicUnsafeCopy mv v)) False)
-    vstep False = return Done
-
-fromVectors :: forall m a. (Monad m, Vector v a) => [v a] -> Stream m a
-{-# INLINE_FUSED fromVectors #-}
-fromVectors vs = Stream (Unf pstep (Left vs))
-                        (Unf vstep vs)
-                        Nothing
-                        (Exact n)
-  where
-    n = List.foldl' (\k v -> k + basicLength v) 0 vs
-
-    pstep (Left []) = return Done
-    pstep (Left (v:vs)) = basicLength v `seq` return (Skip (Right (v,0,vs)))
-
-    pstep (Right (v,i,vs))
-      | i >= basicLength v = return $ Skip (Left vs)
-      | otherwise          = case basicUnsafeIndexM v i of
-                               Box x -> return $ Yield x (Right (v,i+1,vs))
-
-    -- FIXME: work around bug in GHC 7.6.1
-    vstep :: [v a] -> m (Step [v a] (Chunk v a))
-    vstep [] = return Done
-    vstep (v:vs) = return $ Yield (Chunk (basicLength v)
-                                         (\mv -> INTERNAL_CHECK(check) "concatVectors" "length mismatch"
-                                                                       (M.basicLength mv == basicLength v)
-                                                 $ basicUnsafeCopy mv v)) vs
-
-
-concatVectors :: (Monad m, Vector v a) => Stream m (v a) -> Stream m a
-{-# INLINE_FUSED concatVectors #-}
-concatVectors (Stream step s}
-  = Stream (Unf pstep (Left s))
-           (Unf vstep s)
-           Nothing
-           Unknown
-  where
-    pstep (Left s) = do
-      r <- step s
-      case r of
-        Yield v s' -> basicLength v `seq` return (Skip (Right (v,0,s')))
-        Skip    s' -> return (Skip (Left s'))
-        Done       -> return Done
-
-    pstep (Right (v,i,s))
-      | i >= basicLength v = return (Skip (Left s))
-      | otherwise          = case basicUnsafeIndexM v i of
-                               Box x -> return (Yield x (Right (v,i+1,s)))
-
-
-    vstep s = do
-      r <- step s
-      case r of
-        Yield v s' -> return (Yield (Chunk (basicLength v)
-                                           (\mv -> INTERNAL_CHECK(check) "concatVectors" "length mismatch"
-                                                                          (M.basicLength mv == basicLength v)
-                                                   $ basicUnsafeCopy mv v)) s')
-        Skip    s' -> return (Skip s')
-        Done       -> return Done
-
-reVector :: Monad m => Stream m a -> Stream m a
-{-# INLINE_FUSED reVector #-}
-reVector (Stream step s, sSize = n} = Stream step s n
-
-{-# RULES
-
-"reVector [Vector]"
-  reVector = id
-
-"reVector/reVector [Vector]" forall s.
-  reVector (reVector s) = s   #-}
-
-
--}
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Util.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Util.hs
deleted file mode 100644
index 855bf5ddd40d..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Fusion/Util.hs
+++ /dev/null
@@ -1,60 +0,0 @@
-{-# LANGUAGE CPP #-}
--- |
--- Module      : Data.Vector.Fusion.Util
--- Copyright   : (c) Roman Leshchinskiy 2009
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : portable
---
--- Fusion-related utility types
---
-
-module Data.Vector.Fusion.Util (
-  Id(..), Box(..),
-
-  delay_inline, delayed_min
-) where
-
-#if !MIN_VERSION_base(4,8,0)
-import Control.Applicative (Applicative(..))
-#endif
-
--- | Identity monad
-newtype Id a = Id { unId :: a }
-
-instance Functor Id where
-  fmap f (Id x) = Id (f x)
-
-instance Applicative Id where
-  pure = Id
-  Id f <*> Id x = Id (f x)
-
-instance Monad Id where
-  return = pure
-  Id x >>= f = f x
-
--- | Box monad
-data Box a = Box { unBox :: a }
-
-instance Functor Box where
-  fmap f (Box x) = Box (f x)
-
-instance Applicative Box where
-  pure = Box
-  Box f <*> Box x = Box (f x)
-
-instance Monad Box where
-  return = pure
-  Box x >>= f = f x
-
--- | Delay inlining a function until late in the game (simplifier phase 0).
-delay_inline :: (a -> b) -> a -> b
-{-# INLINE [0] delay_inline #-}
-delay_inline f = f
-
--- | `min` inlined in phase 0
-delayed_min :: Int -> Int -> Int
-{-# INLINE [0] delayed_min #-}
-delayed_min m n = min m n
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic.hs
deleted file mode 100644
index 066c07fd3d1d..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic.hs
+++ /dev/null
@@ -1,2206 +0,0 @@
-{-# LANGUAGE CPP, Rank2Types, MultiParamTypeClasses, FlexibleContexts,
-             TypeFamilies, ScopedTypeVariables, BangPatterns #-}
--- |
--- Module      : Data.Vector.Generic
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Generic interface to pure vectors.
---
-
-module Data.Vector.Generic (
-  -- * Immutable vectors
-  Vector(..), Mutable,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Indexing
-  (!), (!?), head, last,
-  unsafeIndex, unsafeHead, unsafeLast,
-
-  -- ** Monadic indexing
-  indexM, headM, lastM,
-  unsafeIndexM, unsafeHeadM, unsafeLastM,
-
-  -- ** Extracting subvectors (slicing)
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- * Construction
-
-  -- ** Initialisation
-  empty, singleton, replicate, generate, iterateN,
-
-  -- ** Monadic initialisation
-  replicateM, generateM, iterateNM, create, createT,
-
-  -- ** Unfolding
-  unfoldr, unfoldrN,
-  unfoldrM, unfoldrNM,
-  constructN, constructrN,
-
-  -- ** Enumeration
-  enumFromN, enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- ** Concatenation
-  cons, snoc, (++), concat, concatNE,
-
-  -- ** Restricting memory usage
-  force,
-
-  -- * Modifying vectors
-
-  -- ** Bulk updates
-  (//), update, update_,
-  unsafeUpd, unsafeUpdate, unsafeUpdate_,
-
-  -- ** Accumulations
-  accum, accumulate, accumulate_,
-  unsafeAccum, unsafeAccumulate, unsafeAccumulate_,
-
-  -- ** Permutations
-  reverse, backpermute, unsafeBackpermute,
-
-  -- ** Safe destructive updates
-  modify,
-
-  -- * Elementwise operations
-
-  -- ** Indexing
-  indexed,
-
-  -- ** Mapping
-  map, imap, concatMap,
-
-  -- ** Monadic mapping
-  mapM, imapM, mapM_, imapM_, forM, forM_,
-
-  -- ** Zipping
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  izipWith, izipWith3, izipWith4, izipWith5, izipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- ** Monadic zipping
-  zipWithM, izipWithM, zipWithM_, izipWithM_,
-
-  -- ** Unzipping
-  unzip, unzip3, unzip4, unzip5, unzip6,
-
-  -- * Working with predicates
-
-  -- ** Filtering
-  filter, ifilter, uniq,
-  mapMaybe, imapMaybe,
-  filterM,
-  takeWhile, dropWhile,
-
-  -- ** Partitioning
-  partition, unstablePartition, span, break,
-
-  -- ** Searching
-  elem, notElem, find, findIndex, findIndices, elemIndex, elemIndices,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1, foldr', foldr1',
-  ifoldl, ifoldl', ifoldr, ifoldr',
-
-  -- ** Specialised folds
-  all, any, and, or,
-  sum, product,
-  maximum, maximumBy, minimum, minimumBy,
-  minIndex, minIndexBy, maxIndex, maxIndexBy,
-
-  -- ** Monadic folds
-  foldM, ifoldM, foldM', ifoldM',
-  fold1M, fold1M', foldM_, ifoldM_,
-  foldM'_, ifoldM'_, fold1M_, fold1M'_,
-
-  -- ** Monadic sequencing
-  sequence, sequence_,
-
-  -- * Prefix sums (scans)
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl', scanl1, scanl1',
-  iscanl, iscanl',
-  prescanr, prescanr',
-  postscanr, postscanr',
-  scanr, scanr', scanr1, scanr1',
-  iscanr, iscanr',
-
-  -- * Conversions
-
-  -- ** Lists
-  toList, fromList, fromListN,
-
-  -- ** Different vector types
-  convert,
-
-  -- ** Mutable vectors
-  freeze, thaw, copy, unsafeFreeze, unsafeThaw, unsafeCopy,
-
-  -- * Fusion support
-
-  -- ** Conversion to/from Bundles
-  stream, unstream, streamR, unstreamR,
-
-  -- ** Recycling support
-  new, clone,
-
-  -- * Utilities
-
-  -- ** Comparisons
-  eq, cmp,
-  eqBy, cmpBy,
-
-  -- ** Show and Read
-  showsPrec, readPrec,
-  liftShowsPrec, liftReadsPrec,
-
-  -- ** @Data@ and @Typeable@
-  gfoldl, dataCast, mkType
-) where
-
-import           Data.Vector.Generic.Base
-
-import qualified Data.Vector.Generic.Mutable as M
-
-import qualified Data.Vector.Generic.New as New
-import           Data.Vector.Generic.New ( New )
-
-import qualified Data.Vector.Fusion.Bundle as Bundle
-import           Data.Vector.Fusion.Bundle ( Bundle, MBundle, lift, inplace )
-import qualified Data.Vector.Fusion.Bundle.Monadic as MBundle
-import           Data.Vector.Fusion.Stream.Monadic ( Stream )
-import qualified Data.Vector.Fusion.Stream.Monadic as S
-import           Data.Vector.Fusion.Bundle.Size
-import           Data.Vector.Fusion.Util
-
-import Control.Monad.ST ( ST, runST )
-import Control.Monad.Primitive
-import Prelude hiding ( length, null,
-                        replicate, (++), concat,
-                        head, last,
-                        init, tail, take, drop, splitAt, reverse,
-                        map, concat, concatMap,
-                        zipWith, zipWith3, zip, zip3, unzip, unzip3,
-                        filter, takeWhile, dropWhile, span, break,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        all, any, and, or, sum, product, maximum, minimum,
-                        scanl, scanl1, scanr, scanr1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_, sequence, sequence_,
-                        showsPrec )
-
-import qualified Text.Read as Read
-import qualified Data.List.NonEmpty as NonEmpty
-
-#if __GLASGOW_HASKELL__ >= 707
-import Data.Typeable ( Typeable, gcast1 )
-#else
-import Data.Typeable ( Typeable1, gcast1 )
-#endif
-
-#include "vector.h"
-
-import Data.Data ( Data, DataType )
-#if MIN_VERSION_base(4,2,0)
-import Data.Data ( mkNoRepType )
-#else
-import Data.Data ( mkNorepType )
-mkNoRepType :: String -> DataType
-mkNoRepType = mkNorepType
-#endif
-
-import qualified Data.Traversable as T (Traversable(mapM))
-
--- Length information
--- ------------------
-
--- | /O(1)/ Yield the length of the vector
-length :: Vector v a => v a -> Int
-{-# INLINE length #-}
-length = Bundle.length . stream'
-
--- | /O(1)/ Test whether a vector is empty
-null :: Vector v a => v a -> Bool
-{-# INLINE null #-}
-null = Bundle.null . stream
-
--- Indexing
--- --------
-
-infixl 9 !
--- | O(1) Indexing
-(!) :: Vector v a => v a -> Int -> a
-{-# INLINE_FUSED (!) #-}
-(!) v i = BOUNDS_CHECK(checkIndex) "(!)" i (length v)
-        $ unId (basicUnsafeIndexM v i)
-
-infixl 9 !?
--- | O(1) Safe indexing
-(!?) :: Vector v a => v a -> Int -> Maybe a
-{-# INLINE_FUSED (!?) #-}
-v !? i | i < 0 || i >= length v = Nothing
-       | otherwise              = Just $ unsafeIndex v i
-
--- | /O(1)/ First element
-head :: Vector v a => v a -> a
-{-# INLINE_FUSED head #-}
-head v = v ! 0
-
--- | /O(1)/ Last element
-last :: Vector v a => v a -> a
-{-# INLINE_FUSED last #-}
-last v = v ! (length v - 1)
-
--- | /O(1)/ Unsafe indexing without bounds checking
-unsafeIndex :: Vector v a => v a -> Int -> a
-{-# INLINE_FUSED unsafeIndex #-}
-unsafeIndex v i = UNSAFE_CHECK(checkIndex) "unsafeIndex" i (length v)
-                $ unId (basicUnsafeIndexM v i)
-
--- | /O(1)/ First element without checking if the vector is empty
-unsafeHead :: Vector v a => v a -> a
-{-# INLINE_FUSED unsafeHead #-}
-unsafeHead v = unsafeIndex v 0
-
--- | /O(1)/ Last element without checking if the vector is empty
-unsafeLast :: Vector v a => v a -> a
-{-# INLINE_FUSED unsafeLast #-}
-unsafeLast v = unsafeIndex v (length v - 1)
-
-{-# RULES
-
-"(!)/unstream [Vector]" forall i s.
-  new (New.unstream s) ! i = s Bundle.!! i
-
-"(!?)/unstream [Vector]" forall i s.
-  new (New.unstream s) !? i = s Bundle.!? i
-
-"head/unstream [Vector]" forall s.
-  head (new (New.unstream s)) = Bundle.head s
-
-"last/unstream [Vector]" forall s.
-  last (new (New.unstream s)) = Bundle.last s
-
-"unsafeIndex/unstream [Vector]" forall i s.
-  unsafeIndex (new (New.unstream s)) i = s Bundle.!! i
-
-"unsafeHead/unstream [Vector]" forall s.
-  unsafeHead (new (New.unstream s)) = Bundle.head s
-
-"unsafeLast/unstream [Vector]" forall s.
-  unsafeLast (new (New.unstream s)) = Bundle.last s  #-}
-
-
-
--- Monadic indexing
--- ----------------
-
--- | /O(1)/ Indexing in a monad.
---
--- The monad allows operations to be strict in the vector when necessary.
--- Suppose vector copying is implemented like this:
---
--- > copy mv v = ... write mv i (v ! i) ...
---
--- For lazy vectors, @v ! i@ would not be evaluated which means that @mv@
--- would unnecessarily retain a reference to @v@ in each element written.
---
--- With 'indexM', copying can be implemented like this instead:
---
--- > copy mv v = ... do
--- >                   x <- indexM v i
--- >                   write mv i x
---
--- Here, no references to @v@ are retained because indexing (but /not/ the
--- elements) is evaluated eagerly.
---
-indexM :: (Vector v a, Monad m) => v a -> Int -> m a
-{-# INLINE_FUSED indexM #-}
-indexM v i = BOUNDS_CHECK(checkIndex) "indexM" i (length v)
-           $ basicUnsafeIndexM v i
-
--- | /O(1)/ First element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-headM :: (Vector v a, Monad m) => v a -> m a
-{-# INLINE_FUSED headM #-}
-headM v = indexM v 0
-
--- | /O(1)/ Last element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-lastM :: (Vector v a, Monad m) => v a -> m a
-{-# INLINE_FUSED lastM #-}
-lastM v = indexM v (length v - 1)
-
--- | /O(1)/ Indexing in a monad without bounds checks. See 'indexM' for an
--- explanation of why this is useful.
-unsafeIndexM :: (Vector v a, Monad m) => v a -> Int -> m a
-{-# INLINE_FUSED unsafeIndexM #-}
-unsafeIndexM v i = UNSAFE_CHECK(checkIndex) "unsafeIndexM" i (length v)
-                 $ basicUnsafeIndexM v i
-
--- | /O(1)/ First element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeHeadM :: (Vector v a, Monad m) => v a -> m a
-{-# INLINE_FUSED unsafeHeadM #-}
-unsafeHeadM v = unsafeIndexM v 0
-
--- | /O(1)/ Last element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeLastM :: (Vector v a, Monad m) => v a -> m a
-{-# INLINE_FUSED unsafeLastM #-}
-unsafeLastM v = unsafeIndexM v (length v - 1)
-
-{-# RULES
-
-"indexM/unstream [Vector]" forall s i.
-  indexM (new (New.unstream s)) i = lift s MBundle.!! i
-
-"headM/unstream [Vector]" forall s.
-  headM (new (New.unstream s)) = MBundle.head (lift s)
-
-"lastM/unstream [Vector]" forall s.
-  lastM (new (New.unstream s)) = MBundle.last (lift s)
-
-"unsafeIndexM/unstream [Vector]" forall s i.
-  unsafeIndexM (new (New.unstream s)) i = lift s MBundle.!! i
-
-"unsafeHeadM/unstream [Vector]" forall s.
-  unsafeHeadM (new (New.unstream s)) = MBundle.head (lift s)
-
-"unsafeLastM/unstream [Vector]" forall s.
-  unsafeLastM (new (New.unstream s)) = MBundle.last (lift s)   #-}
-
-
-
--- Extracting subvectors (slicing)
--- -------------------------------
-
--- | /O(1)/ Yield a slice of the vector without copying it. The vector must
--- contain at least @i+n@ elements.
-slice :: Vector v a => Int   -- ^ @i@ starting index
-                    -> Int   -- ^ @n@ length
-                    -> v a
-                    -> v a
-{-# INLINE_FUSED slice #-}
-slice i n v = BOUNDS_CHECK(checkSlice) "slice" i n (length v)
-            $ basicUnsafeSlice i n v
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty.
-init :: Vector v a => v a -> v a
-{-# INLINE_FUSED init #-}
-init v = slice 0 (length v - 1) v
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty.
-tail :: Vector v a => v a -> v a
-{-# INLINE_FUSED tail #-}
-tail v = slice 1 (length v - 1) v
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case it is returned unchanged.
-take :: Vector v a => Int -> v a -> v a
-{-# INLINE_FUSED take #-}
-take n v = unsafeSlice 0 (delay_inline min n' (length v)) v
-  where n' = max n 0
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case an empty vector is returned.
-drop :: Vector v a => Int -> v a -> v a
-{-# INLINE_FUSED drop #-}
-drop n v = unsafeSlice (delay_inline min n' len)
-                       (delay_inline max 0 (len - n')) v
-  where n' = max n 0
-        len = length v
-
--- | /O(1)/ Yield the first @n@ elements paired with the remainder without copying.
---
--- Note that @'splitAt' n v@ is equivalent to @('take' n v, 'drop' n v)@
--- but slightly more efficient.
-{-# INLINE_FUSED splitAt #-}
-splitAt :: Vector v a => Int -> v a -> (v a, v a)
-splitAt n v = ( unsafeSlice 0 m v
-              , unsafeSlice m (delay_inline max 0 (len - n')) v
-              )
-    where
-      m   = delay_inline min n' len
-      n'  = max n 0
-      len = length v
-
--- | /O(1)/ Yield a slice of the vector without copying. The vector must
--- contain at least @i+n@ elements but this is not checked.
-unsafeSlice :: Vector v a => Int   -- ^ @i@ starting index
-                          -> Int   -- ^ @n@ length
-                          -> v a
-                          -> v a
-{-# INLINE_FUSED unsafeSlice #-}
-unsafeSlice i n v = UNSAFE_CHECK(checkSlice) "unsafeSlice" i n (length v)
-                  $ basicUnsafeSlice i n v
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty but this is not checked.
-unsafeInit :: Vector v a => v a -> v a
-{-# INLINE_FUSED unsafeInit #-}
-unsafeInit v = unsafeSlice 0 (length v - 1) v
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty but this is not checked.
-unsafeTail :: Vector v a => v a -> v a
-{-# INLINE_FUSED unsafeTail #-}
-unsafeTail v = unsafeSlice 1 (length v - 1) v
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector must
--- contain at least @n@ elements but this is not checked.
-unsafeTake :: Vector v a => Int -> v a -> v a
-{-# INLINE unsafeTake #-}
-unsafeTake n v = unsafeSlice 0 n v
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector
--- must contain at least @n@ elements but this is not checked.
-unsafeDrop :: Vector v a => Int -> v a -> v a
-{-# INLINE unsafeDrop #-}
-unsafeDrop n v = unsafeSlice n (length v - n) v
-
-{-# RULES
-
-"slice/new [Vector]" forall i n p.
-  slice i n (new p) = new (New.slice i n p)
-
-"init/new [Vector]" forall p.
-  init (new p) = new (New.init p)
-
-"tail/new [Vector]" forall p.
-  tail (new p) = new (New.tail p)
-
-"take/new [Vector]" forall n p.
-  take n (new p) = new (New.take n p)
-
-"drop/new [Vector]" forall n p.
-  drop n (new p) = new (New.drop n p)
-
-"unsafeSlice/new [Vector]" forall i n p.
-  unsafeSlice i n (new p) = new (New.unsafeSlice i n p)
-
-"unsafeInit/new [Vector]" forall p.
-  unsafeInit (new p) = new (New.unsafeInit p)
-
-"unsafeTail/new [Vector]" forall p.
-  unsafeTail (new p) = new (New.unsafeTail p)   #-}
-
-
-
--- Initialisation
--- --------------
-
--- | /O(1)/ Empty vector
-empty :: Vector v a => v a
-{-# INLINE empty #-}
-empty = unstream Bundle.empty
-
--- | /O(1)/ Vector with exactly one element
-singleton :: forall v a. Vector v a => a -> v a
-{-# INLINE singleton #-}
-singleton x = elemseq (undefined :: v a) x
-            $ unstream (Bundle.singleton x)
-
--- | /O(n)/ Vector of the given length with the same value in each position
-replicate :: forall v a. Vector v a => Int -> a -> v a
-{-# INLINE replicate #-}
-replicate n x = elemseq (undefined :: v a) x
-              $ unstream
-              $ Bundle.replicate n x
-
--- | /O(n)/ Construct a vector of the given length by applying the function to
--- each index
-generate :: Vector v a => Int -> (Int -> a) -> v a
-{-# INLINE generate #-}
-generate n f = unstream (Bundle.generate n f)
-
--- | /O(n)/ Apply function n times to value. Zeroth element is original value.
-iterateN :: Vector v a => Int -> (a -> a) -> a -> v a
-{-# INLINE iterateN #-}
-iterateN n f x = unstream (Bundle.iterateN n f x)
-
--- Unfolding
--- ---------
-
--- | /O(n)/ Construct a vector by repeatedly applying the generator function
--- to a seed. The generator function yields 'Just' the next element and the
--- new seed or 'Nothing' if there are no more elements.
---
--- > unfoldr (\n -> if n == 0 then Nothing else Just (n,n-1)) 10
--- >  = <10,9,8,7,6,5,4,3,2,1>
-unfoldr :: Vector v a => (b -> Maybe (a, b)) -> b -> v a
-{-# INLINE unfoldr #-}
-unfoldr f = unstream . Bundle.unfoldr f
-
--- | /O(n)/ Construct a vector with at most @n@ elements by repeatedly applying
--- the generator function to a seed. The generator function yields 'Just' the
--- next element and the new seed or 'Nothing' if there are no more elements.
---
--- > unfoldrN 3 (\n -> Just (n,n-1)) 10 = <10,9,8>
-unfoldrN  :: Vector v a => Int -> (b -> Maybe (a, b)) -> b -> v a
-{-# INLINE unfoldrN #-}
-unfoldrN n f = unstream . Bundle.unfoldrN n f
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrM :: (Monad m, Vector v a) => (b -> m (Maybe (a, b))) -> b -> m (v a)
-{-# INLINE unfoldrM #-}
-unfoldrM f = unstreamM . MBundle.unfoldrM f
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrNM :: (Monad m, Vector v a) => Int -> (b -> m (Maybe (a, b))) -> b -> m (v a)
-{-# INLINE unfoldrNM #-}
-unfoldrNM n f = unstreamM . MBundle.unfoldrNM n f
-
--- | /O(n)/ Construct a vector with @n@ elements by repeatedly applying the
--- generator function to the already constructed part of the vector.
---
--- > constructN 3 f = let a = f <> ; b = f <a> ; c = f <a,b> in f <a,b,c>
---
-constructN :: forall v a. Vector v a => Int -> (v a -> a) -> v a
-{-# INLINE constructN #-}
--- NOTE: We *CANNOT* wrap this in New and then fuse because the elements
--- might contain references to the immutable vector!
-constructN !n f = runST (
-  do
-    v  <- M.new n
-    v' <- unsafeFreeze v
-    fill v' 0
-  )
-  where
-    fill :: forall s. v a -> Int -> ST s (v a)
-    fill !v i | i < n = let x = f (unsafeTake i v)
-                        in
-                        elemseq v x $
-                        do
-                          v'  <- unsafeThaw v
-                          M.unsafeWrite v' i x
-                          v'' <- unsafeFreeze v'
-                          fill v'' (i+1)
-
-    fill v _ = return v
-
--- | /O(n)/ Construct a vector with @n@ elements from right to left by
--- repeatedly applying the generator function to the already constructed part
--- of the vector.
---
--- > constructrN 3 f = let a = f <> ; b = f<a> ; c = f <b,a> in f <c,b,a>
---
-constructrN :: forall v a. Vector v a => Int -> (v a -> a) -> v a
-{-# INLINE constructrN #-}
--- NOTE: We *CANNOT* wrap this in New and then fuse because the elements
--- might contain references to the immutable vector!
-constructrN !n f = runST (
-  do
-    v  <- n `seq` M.new n
-    v' <- unsafeFreeze v
-    fill v' 0
-  )
-  where
-    fill :: forall s. v a -> Int -> ST s (v a)
-    fill !v i | i < n = let x = f (unsafeSlice (n-i) i v)
-                        in
-                        elemseq v x $
-                        do
-                          v'  <- unsafeThaw v
-                          M.unsafeWrite v' (n-i-1) x
-                          v'' <- unsafeFreeze v'
-                          fill v'' (i+1)
-
-    fill v _ = return v
-
-
--- Enumeration
--- -----------
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+1@
--- etc. This operation is usually more efficient than 'enumFromTo'.
---
--- > enumFromN 5 3 = <5,6,7>
-enumFromN :: (Vector v a, Num a) => a -> Int -> v a
-{-# INLINE enumFromN #-}
-enumFromN x n = enumFromStepN x 1 n
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc. This operations is usually more efficient than 'enumFromThenTo'.
---
--- > enumFromStepN 1 0.1 5 = <1,1.1,1.2,1.3,1.4>
-enumFromStepN :: forall v a. (Vector v a, Num a) => a -> a -> Int -> v a
-{-# INLINE enumFromStepN #-}
-enumFromStepN x y n = elemseq (undefined :: v a) x
-                    $ elemseq (undefined :: v a) y
-                    $ unstream
-                    $ Bundle.enumFromStepN  x y n
-
--- | /O(n)/ Enumerate values from @x@ to @y@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromN' instead.
-enumFromTo :: (Vector v a, Enum a) => a -> a -> v a
-{-# INLINE enumFromTo #-}
-enumFromTo x y = unstream (Bundle.enumFromTo x y)
-
--- | /O(n)/ Enumerate values from @x@ to @y@ with a specific step @z@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Vector v a, Enum a) => a -> a -> a -> v a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo x y z = unstream (Bundle.enumFromThenTo x y z)
-
--- Concatenation
--- -------------
-
--- | /O(n)/ Prepend an element
-cons :: forall v a. Vector v a => a -> v a -> v a
-{-# INLINE cons #-}
-cons x v = elemseq (undefined :: v a) x
-         $ unstream
-         $ Bundle.cons x
-         $ stream v
-
--- | /O(n)/ Append an element
-snoc :: forall v a. Vector v a => v a -> a -> v a
-{-# INLINE snoc #-}
-snoc v x = elemseq (undefined :: v a) x
-         $ unstream
-         $ Bundle.snoc (stream v) x
-
-infixr 5 ++
--- | /O(m+n)/ Concatenate two vectors
-(++) :: Vector v a => v a -> v a -> v a
-{-# INLINE (++) #-}
-v ++ w = unstream (stream v Bundle.++ stream w)
-
--- | /O(n)/ Concatenate all vectors in the list
-concat :: Vector v a => [v a] -> v a
-{-# INLINE concat #-}
-concat = unstream . Bundle.fromVectors
-{-
-concat vs = unstream (Bundle.flatten mk step (Exact n) (Bundle.fromList vs))
-  where
-    n = List.foldl' (\k v -> k + length v) 0 vs
-
-    {-# INLINE_INNER step #-}
-    step (v,i,k)
-      | i < k = case unsafeIndexM v i of
-                  Box x -> Bundle.Yield x (v,i+1,k)
-      | otherwise = Bundle.Done
-
-    {-# INLINE mk #-}
-    mk v = let k = length v
-           in
-           k `seq` (v,0,k)
--}
-
--- | /O(n)/ Concatenate all vectors in the non-empty list
-concatNE :: Vector v a => NonEmpty.NonEmpty (v a) -> v a
-concatNE = concat . NonEmpty.toList
-
--- Monadic initialisation
--- ----------------------
-
--- | /O(n)/ Execute the monadic action the given number of times and store the
--- results in a vector.
-replicateM :: (Monad m, Vector v a) => Int -> m a -> m (v a)
-{-# INLINE replicateM #-}
-replicateM n m = unstreamM (MBundle.replicateM n m)
-
--- | /O(n)/ Construct a vector of the given length by applying the monadic
--- action to each index
-generateM :: (Monad m, Vector v a) => Int -> (Int -> m a) -> m (v a)
-{-# INLINE generateM #-}
-generateM n f = unstreamM (MBundle.generateM n f)
-
--- | /O(n)/ Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: (Monad m, Vector v a) => Int -> (a -> m a) -> a -> m (v a)
-{-# INLINE iterateNM #-}
-iterateNM n f x = unstreamM (MBundle.iterateNM n f x)
-
--- | Execute the monadic action and freeze the resulting vector.
---
--- @
--- create (do { v \<- 'M.new' 2; 'M.write' v 0 \'a\'; 'M.write' v 1 \'b\'; return v }) = \<'a','b'\>
--- @
-create :: Vector v a => (forall s. ST s (Mutable v s a)) -> v a
-{-# INLINE create #-}
-create p = new (New.create p)
-
--- | Execute the monadic action and freeze the resulting vectors.
-createT
-  :: (T.Traversable f, Vector v a)
-  => (forall s. ST s (f (Mutable v s a))) -> f (v a)
-{-# INLINE createT #-}
-createT p = runST (p >>= T.mapM unsafeFreeze)
-
--- Restricting memory usage
--- ------------------------
-
--- | /O(n)/ Yield the argument but force it not to retain any extra memory,
--- possibly by copying it.
---
--- This is especially useful when dealing with slices. For example:
---
--- > force (slice 0 2 <huge vector>)
---
--- Here, the slice retains a reference to the huge vector. Forcing it creates
--- a copy of just the elements that belong to the slice and allows the huge
--- vector to be garbage collected.
-force :: Vector v a => v a -> v a
--- FIXME: we probably ought to inline this later as the rules still might fire
--- otherwise
-{-# INLINE_FUSED force #-}
-force v = new (clone v)
-
--- Bulk updates
--- ------------
-
--- | /O(m+n)/ For each pair @(i,a)@ from the list, replace the vector
--- element at position @i@ by @a@.
---
--- > <5,9,2,7> // [(2,1),(0,3),(2,8)] = <3,9,8,7>
---
-(//) :: Vector v a => v a        -- ^ initial vector (of length @m@)
-                   -> [(Int, a)] -- ^ list of index/value pairs (of length @n@)
-                   -> v a
-{-# INLINE (//) #-}
-v // us = update_stream v (Bundle.fromList us)
-
--- | /O(m+n)/ For each pair @(i,a)@ from the vector of index/value pairs,
--- replace the vector element at position @i@ by @a@.
---
--- > update <5,9,2,7> <(2,1),(0,3),(2,8)> = <3,9,8,7>
---
-update :: (Vector v a, Vector v (Int, a))
-        => v a        -- ^ initial vector (of length @m@)
-        -> v (Int, a) -- ^ vector of index/value pairs (of length @n@)
-        -> v a
-{-# INLINE update #-}
-update v w = update_stream v (stream w)
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @a@ from the value vector, replace the element of the
--- initial vector at position @i@ by @a@.
---
--- > update_ <5,9,2,7>  <2,0,2> <1,3,8> = <3,9,8,7>
---
--- This function is useful for instances of 'Vector' that cannot store pairs.
--- Otherwise, 'update' is probably more convenient.
---
--- @
--- update_ xs is ys = 'update' xs ('zip' is ys)
--- @
-update_ :: (Vector v a, Vector v Int)
-        => v a   -- ^ initial vector (of length @m@)
-        -> v Int -- ^ index vector (of length @n1@)
-        -> v a   -- ^ value vector (of length @n2@)
-        -> v a
-{-# INLINE update_ #-}
-update_ v is w = update_stream v (Bundle.zipWith (,) (stream is) (stream w))
-
-update_stream :: Vector v a => v a -> Bundle u (Int,a) -> v a
-{-# INLINE update_stream #-}
-update_stream = modifyWithBundle M.update
-
--- | Same as ('//') but without bounds checking.
-unsafeUpd :: Vector v a => v a -> [(Int, a)] -> v a
-{-# INLINE unsafeUpd #-}
-unsafeUpd v us = unsafeUpdate_stream v (Bundle.fromList us)
-
--- | Same as 'update' but without bounds checking.
-unsafeUpdate :: (Vector v a, Vector v (Int, a)) => v a -> v (Int, a) -> v a
-{-# INLINE unsafeUpdate #-}
-unsafeUpdate v w = unsafeUpdate_stream v (stream w)
-
--- | Same as 'update_' but without bounds checking.
-unsafeUpdate_ :: (Vector v a, Vector v Int) => v a -> v Int -> v a -> v a
-{-# INLINE unsafeUpdate_ #-}
-unsafeUpdate_ v is w
-  = unsafeUpdate_stream v (Bundle.zipWith (,) (stream is) (stream w))
-
-unsafeUpdate_stream :: Vector v a => v a -> Bundle u (Int,a) -> v a
-{-# INLINE unsafeUpdate_stream #-}
-unsafeUpdate_stream = modifyWithBundle M.unsafeUpdate
-
--- Accumulations
--- -------------
-
--- | /O(m+n)/ For each pair @(i,b)@ from the list, replace the vector element
--- @a@ at position @i@ by @f a b@.
---
--- > accum (+) <5,9,2> [(2,4),(1,6),(0,3),(1,7)] = <5+3, 9+6+7, 2+4>
-accum :: Vector v a
-      => (a -> b -> a) -- ^ accumulating function @f@
-      -> v a           -- ^ initial vector (of length @m@)
-      -> [(Int,b)]     -- ^ list of index/value pairs (of length @n@)
-      -> v a
-{-# INLINE accum #-}
-accum f v us = accum_stream f v (Bundle.fromList us)
-
--- | /O(m+n)/ For each pair @(i,b)@ from the vector of pairs, replace the vector
--- element @a@ at position @i@ by @f a b@.
---
--- > accumulate (+) <5,9,2> <(2,4),(1,6),(0,3),(1,7)> = <5+3, 9+6+7, 2+4>
-accumulate :: (Vector v a, Vector v (Int, b))
-           => (a -> b -> a) -- ^ accumulating function @f@
-           -> v a           -- ^ initial vector (of length @m@)
-           -> v (Int,b)     -- ^ vector of index/value pairs (of length @n@)
-           -> v a
-{-# INLINE accumulate #-}
-accumulate f v us = accum_stream f v (stream us)
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @b@ from the the value vector,
--- replace the element of the initial vector at
--- position @i@ by @f a b@.
---
--- > accumulate_ (+) <5,9,2> <2,1,0,1> <4,6,3,7> = <5+3, 9+6+7, 2+4>
---
--- This function is useful for instances of 'Vector' that cannot store pairs.
--- Otherwise, 'accumulate' is probably more convenient:
---
--- @
--- accumulate_ f as is bs = 'accumulate' f as ('zip' is bs)
--- @
-accumulate_ :: (Vector v a, Vector v Int, Vector v b)
-                => (a -> b -> a) -- ^ accumulating function @f@
-                -> v a           -- ^ initial vector (of length @m@)
-                -> v Int         -- ^ index vector (of length @n1@)
-                -> v b           -- ^ value vector (of length @n2@)
-                -> v a
-{-# INLINE accumulate_ #-}
-accumulate_ f v is xs = accum_stream f v (Bundle.zipWith (,) (stream is)
-                                                             (stream xs))
-
-
-accum_stream :: Vector v a => (a -> b -> a) -> v a -> Bundle u (Int,b) -> v a
-{-# INLINE accum_stream #-}
-accum_stream f = modifyWithBundle (M.accum f)
-
--- | Same as 'accum' but without bounds checking.
-unsafeAccum :: Vector v a => (a -> b -> a) -> v a -> [(Int,b)] -> v a
-{-# INLINE unsafeAccum #-}
-unsafeAccum f v us = unsafeAccum_stream f v (Bundle.fromList us)
-
--- | Same as 'accumulate' but without bounds checking.
-unsafeAccumulate :: (Vector v a, Vector v (Int, b))
-                => (a -> b -> a) -> v a -> v (Int,b) -> v a
-{-# INLINE unsafeAccumulate #-}
-unsafeAccumulate f v us = unsafeAccum_stream f v (stream us)
-
--- | Same as 'accumulate_' but without bounds checking.
-unsafeAccumulate_ :: (Vector v a, Vector v Int, Vector v b)
-                => (a -> b -> a) -> v a -> v Int -> v b -> v a
-{-# INLINE unsafeAccumulate_ #-}
-unsafeAccumulate_ f v is xs
-  = unsafeAccum_stream f v (Bundle.zipWith (,) (stream is) (stream xs))
-
-unsafeAccum_stream
-  :: Vector v a => (a -> b -> a) -> v a -> Bundle u (Int,b) -> v a
-{-# INLINE unsafeAccum_stream #-}
-unsafeAccum_stream f = modifyWithBundle (M.unsafeAccum f)
-
--- Permutations
--- ------------
-
--- | /O(n)/ Reverse a vector
-reverse :: (Vector v a) => v a -> v a
-{-# INLINE reverse #-}
--- FIXME: make this fuse better, add support for recycling
-reverse = unstream . streamR
-
--- | /O(n)/ Yield the vector obtained by replacing each element @i@ of the
--- index vector by @xs'!'i@. This is equivalent to @'map' (xs'!') is@ but is
--- often much more efficient.
---
--- > backpermute <a,b,c,d> <0,3,2,3,1,0> = <a,d,c,d,b,a>
-backpermute :: (Vector v a, Vector v Int)
-            => v a   -- ^ @xs@ value vector
-            -> v Int -- ^ @is@ index vector (of length @n@)
-            -> v a
-{-# INLINE backpermute #-}
--- This somewhat non-intuitive definition ensures that the resulting vector
--- does not retain references to the original one even if it is lazy in its
--- elements. This would not be the case if we simply used map (v!)
-backpermute v is = seq v
-                 $ seq n
-                 $ unstream
-                 $ Bundle.unbox
-                 $ Bundle.map index
-                 $ stream is
-  where
-    n = length v
-
-    {-# INLINE index #-}
-    -- NOTE: we do it this way to avoid triggering LiberateCase on n in
-    -- polymorphic code
-    index i = BOUNDS_CHECK(checkIndex) "backpermute" i n
-            $ basicUnsafeIndexM v i
-
--- | Same as 'backpermute' but without bounds checking.
-unsafeBackpermute :: (Vector v a, Vector v Int) => v a -> v Int -> v a
-{-# INLINE unsafeBackpermute #-}
-unsafeBackpermute v is = seq v
-                       $ seq n
-                       $ unstream
-                       $ Bundle.unbox
-                       $ Bundle.map index
-                       $ stream is
-  where
-    n = length v
-
-    {-# INLINE index #-}
-    -- NOTE: we do it this way to avoid triggering LiberateCase on n in
-    -- polymorphic code
-    index i = UNSAFE_CHECK(checkIndex) "unsafeBackpermute" i n
-            $ basicUnsafeIndexM v i
-
--- Safe destructive updates
--- ------------------------
-
--- | Apply a destructive operation to a vector. The operation will be
--- performed in place if it is safe to do so and will modify a copy of the
--- vector otherwise.
---
--- @
--- modify (\\v -> 'M.write' v 0 \'x\') ('replicate' 3 \'a\') = \<\'x\',\'a\',\'a\'\>
--- @
-modify :: Vector v a => (forall s. Mutable v s a -> ST s ()) -> v a -> v a
-{-# INLINE modify #-}
-modify p = new . New.modify p . clone
-
--- We have to make sure that this is strict in the stream but we can't seq on
--- it while fusion is happening. Hence this ugliness.
-modifyWithBundle :: Vector v a
-                 => (forall s. Mutable v s a -> Bundle u b -> ST s ())
-                 -> v a -> Bundle u b -> v a
-{-# INLINE modifyWithBundle #-}
-modifyWithBundle p v s = new (New.modifyWithBundle p (clone v) s)
-
--- Indexing
--- --------
-
--- | /O(n)/ Pair each element in a vector with its index
-indexed :: (Vector v a, Vector v (Int,a)) => v a -> v (Int,a)
-{-# INLINE indexed #-}
-indexed = unstream . Bundle.indexed . stream
-
--- Mapping
--- -------
-
--- | /O(n)/ Map a function over a vector
-map :: (Vector v a, Vector v b) => (a -> b) -> v a -> v b
-{-# INLINE map #-}
-map f = unstream . inplace (S.map f) id . stream
-
--- | /O(n)/ Apply a function to every element of a vector and its index
-imap :: (Vector v a, Vector v b) => (Int -> a -> b) -> v a -> v b
-{-# INLINE imap #-}
-imap f = unstream . inplace (S.map (uncurry f) . S.indexed) id
-                  . stream
-
--- | Map a function over a vector and concatenate the results.
-concatMap :: (Vector v a, Vector v b) => (a -> v b) -> v a -> v b
-{-# INLINE concatMap #-}
--- NOTE: We can't fuse concatMap anyway so don't pretend we do.
--- This seems to be slightly slower
--- concatMap f = concat . Bundle.toList . Bundle.map f . stream
-
--- Slowest
--- concatMap f = unstream . Bundle.concatMap (stream . f) . stream
-
--- Used to be fastest
-{-
-concatMap f = unstream
-            . Bundle.flatten mk step Unknown
-            . stream
-  where
-    {-# INLINE_INNER step #-}
-    step (v,i,k)
-      | i < k = case unsafeIndexM v i of
-                  Box x -> Bundle.Yield x (v,i+1,k)
-      | otherwise = Bundle.Done
-
-    {-# INLINE mk #-}
-    mk x = let v = f x
-               k = length v
-           in
-           k `seq` (v,0,k)
--}
-
--- This seems to be fastest now
-concatMap f = unstream
-            . Bundle.concatVectors
-            . Bundle.map f
-            . stream
-
--- Monadic mapping
--- ---------------
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results
-mapM :: (Monad m, Vector v a, Vector v b) => (a -> m b) -> v a -> m (v b)
-{-# INLINE mapM #-}
-mapM f = unstreamM . Bundle.mapM f . stream
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, yielding a vector of results
-imapM :: (Monad m, Vector v a, Vector v b)
-      => (Int -> a -> m b) -> v a -> m (v b)
-imapM f = unstreamM . Bundle.mapM (uncurry f) . Bundle.indexed . stream
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results
-mapM_ :: (Monad m, Vector v a) => (a -> m b) -> v a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ f = Bundle.mapM_ f . stream
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, ignoring the results
-imapM_ :: (Monad m, Vector v a) => (Int -> a -> m b) -> v a -> m ()
-{-# INLINE imapM_ #-}
-imapM_ f = Bundle.mapM_ (uncurry f) . Bundle.indexed . stream
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results. Equivalent to @flip 'mapM'@.
-forM :: (Monad m, Vector v a, Vector v b) => v a -> (a -> m b) -> m (v b)
-{-# INLINE forM #-}
-forM as f = mapM f as
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results. Equivalent to @flip 'mapM_'@.
-forM_ :: (Monad m, Vector v a) => v a -> (a -> m b) -> m ()
-{-# INLINE forM_ #-}
-forM_ as f = mapM_ f as
-
--- Zipping
--- -------
-
--- | /O(min(m,n))/ Zip two vectors with the given function.
-zipWith :: (Vector v a, Vector v b, Vector v c)
-        => (a -> b -> c) -> v a -> v b -> v c
-{-# INLINE zipWith #-}
-zipWith f = \xs ys -> unstream (Bundle.zipWith f (stream xs) (stream ys))
-
--- | Zip three vectors with the given function.
-zipWith3 :: (Vector v a, Vector v b, Vector v c, Vector v d)
-         => (a -> b -> c -> d) -> v a -> v b -> v c -> v d
-{-# INLINE zipWith3 #-}
-zipWith3 f = \as bs cs -> unstream (Bundle.zipWith3 f (stream as)
-                                                  (stream bs)
-                                                  (stream cs))
-
-zipWith4 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e)
-         => (a -> b -> c -> d -> e) -> v a -> v b -> v c -> v d -> v e
-{-# INLINE zipWith4 #-}
-zipWith4 f = \as bs cs ds ->
-    unstream (Bundle.zipWith4 f (stream as)
-                                (stream bs)
-                                (stream cs)
-                                (stream ds))
-
-zipWith5 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-             Vector v f)
-         => (a -> b -> c -> d -> e -> f) -> v a -> v b -> v c -> v d -> v e
-                                         -> v f
-{-# INLINE zipWith5 #-}
-zipWith5 f = \as bs cs ds es ->
-    unstream (Bundle.zipWith5 f (stream as)
-                                (stream bs)
-                                (stream cs)
-                                (stream ds)
-                                (stream es))
-
-zipWith6 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-             Vector v f, Vector v g)
-         => (a -> b -> c -> d -> e -> f -> g)
-         -> v a -> v b -> v c -> v d -> v e -> v f -> v g
-{-# INLINE zipWith6 #-}
-zipWith6 f = \as bs cs ds es fs ->
-    unstream (Bundle.zipWith6 f (stream as)
-                                (stream bs)
-                                (stream cs)
-                                (stream ds)
-                                (stream es)
-                                (stream fs))
-
--- | /O(min(m,n))/ Zip two vectors with a function that also takes the
--- elements' indices.
-izipWith :: (Vector v a, Vector v b, Vector v c)
-        => (Int -> a -> b -> c) -> v a -> v b -> v c
-{-# INLINE izipWith #-}
-izipWith f = \xs ys ->
-    unstream (Bundle.zipWith (uncurry f) (Bundle.indexed (stream xs))
-                                                         (stream ys))
-
-izipWith3 :: (Vector v a, Vector v b, Vector v c, Vector v d)
-         => (Int -> a -> b -> c -> d) -> v a -> v b -> v c -> v d
-{-# INLINE izipWith3 #-}
-izipWith3 f = \as bs cs ->
-    unstream (Bundle.zipWith3 (uncurry f) (Bundle.indexed (stream as))
-                                                          (stream bs)
-                                                          (stream cs))
-
-izipWith4 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e)
-         => (Int -> a -> b -> c -> d -> e) -> v a -> v b -> v c -> v d -> v e
-{-# INLINE izipWith4 #-}
-izipWith4 f = \as bs cs ds ->
-    unstream (Bundle.zipWith4 (uncurry f) (Bundle.indexed (stream as))
-                                                          (stream bs)
-                                                          (stream cs)
-                                                          (stream ds))
-
-izipWith5 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-             Vector v f)
-         => (Int -> a -> b -> c -> d -> e -> f) -> v a -> v b -> v c -> v d
-                                                -> v e -> v f
-{-# INLINE izipWith5 #-}
-izipWith5 f = \as bs cs ds es ->
-    unstream (Bundle.zipWith5 (uncurry f) (Bundle.indexed (stream as))
-                                                          (stream bs)
-                                                          (stream cs)
-                                                          (stream ds)
-                                                          (stream es))
-
-izipWith6 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-             Vector v f, Vector v g)
-         => (Int -> a -> b -> c -> d -> e -> f -> g)
-         -> v a -> v b -> v c -> v d -> v e -> v f -> v g
-{-# INLINE izipWith6 #-}
-izipWith6 f = \as bs cs ds es fs ->
-    unstream (Bundle.zipWith6 (uncurry f) (Bundle.indexed (stream as))
-                                                          (stream bs)
-                                                          (stream cs)
-                                                          (stream ds)
-                                                          (stream es)
-                                                          (stream fs))
-
--- | /O(min(m,n))/ Zip two vectors
-zip :: (Vector v a, Vector v b, Vector v (a,b)) => v a -> v b -> v (a, b)
-{-# INLINE zip #-}
-zip = zipWith (,)
-
-zip3 :: (Vector v a, Vector v b, Vector v c, Vector v (a, b, c))
-     => v a -> v b -> v c -> v (a, b, c)
-{-# INLINE zip3 #-}
-zip3 = zipWith3 (,,)
-
-zip4 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v (a, b, c, d))
-     => v a -> v b -> v c -> v d -> v (a, b, c, d)
-{-# INLINE zip4 #-}
-zip4 = zipWith4 (,,,)
-
-zip5 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-         Vector v (a, b, c, d, e))
-     => v a -> v b -> v c -> v d -> v e -> v (a, b, c, d, e)
-{-# INLINE zip5 #-}
-zip5 = zipWith5 (,,,,)
-
-zip6 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-         Vector v f, Vector v (a, b, c, d, e, f))
-     => v a -> v b -> v c -> v d -> v e -> v f -> v (a, b, c, d, e, f)
-{-# INLINE zip6 #-}
-zip6 = zipWith6 (,,,,,)
-
--- Monadic zipping
--- ---------------
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and yield a
--- vector of results
-zipWithM :: (Monad m, Vector v a, Vector v b, Vector v c)
-         => (a -> b -> m c) -> v a -> v b -> m (v c)
--- FIXME: specialise for ST and IO?
-{-# INLINE zipWithM #-}
-zipWithM f = \as bs -> unstreamM $ Bundle.zipWithM f (stream as) (stream bs)
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and yield a vector of results
-izipWithM :: (Monad m, Vector v a, Vector v b, Vector v c)
-         => (Int -> a -> b -> m c) -> v a -> v b -> m (v c)
-{-# INLINE izipWithM #-}
-izipWithM m as bs = unstreamM . Bundle.zipWithM (uncurry m)
-                                (Bundle.indexed (stream as))
-                                $ stream bs
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and ignore the
--- results
-zipWithM_ :: (Monad m, Vector v a, Vector v b)
-          => (a -> b -> m c) -> v a -> v b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ f = \as bs -> Bundle.zipWithM_ f (stream as) (stream bs)
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and ignore the results
-izipWithM_ :: (Monad m, Vector v a, Vector v b)
-          => (Int -> a -> b -> m c) -> v a -> v b -> m ()
-{-# INLINE izipWithM_ #-}
-izipWithM_ m as bs = Bundle.zipWithM_ (uncurry m)
-                      (Bundle.indexed (stream as))
-                      $ stream bs
-
--- Unzipping
--- ---------
-
--- | /O(min(m,n))/ Unzip a vector of pairs.
-unzip :: (Vector v a, Vector v b, Vector v (a,b)) => v (a, b) -> (v a, v b)
-{-# INLINE unzip #-}
-unzip xs = (map fst xs, map snd xs)
-
-unzip3 :: (Vector v a, Vector v b, Vector v c, Vector v (a, b, c))
-       => v (a, b, c) -> (v a, v b, v c)
-{-# INLINE unzip3 #-}
-unzip3 xs = (map (\(a, _, _) -> a) xs,
-             map (\(_, b, _) -> b) xs,
-             map (\(_, _, c) -> c) xs)
-
-unzip4 :: (Vector v a, Vector v b, Vector v c, Vector v d,
-           Vector v (a, b, c, d))
-       => v (a, b, c, d) -> (v a, v b, v c, v d)
-{-# INLINE unzip4 #-}
-unzip4 xs = (map (\(a, _, _, _) -> a) xs,
-             map (\(_, b, _, _) -> b) xs,
-             map (\(_, _, c, _) -> c) xs,
-             map (\(_, _, _, d) -> d) xs)
-
-unzip5 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-           Vector v (a, b, c, d, e))
-       => v (a, b, c, d, e) -> (v a, v b, v c, v d, v e)
-{-# INLINE unzip5 #-}
-unzip5 xs = (map (\(a, _, _, _, _) -> a) xs,
-             map (\(_, b, _, _, _) -> b) xs,
-             map (\(_, _, c, _, _) -> c) xs,
-             map (\(_, _, _, d, _) -> d) xs,
-             map (\(_, _, _, _, e) -> e) xs)
-
-unzip6 :: (Vector v a, Vector v b, Vector v c, Vector v d, Vector v e,
-           Vector v f, Vector v (a, b, c, d, e, f))
-       => v (a, b, c, d, e, f) -> (v a, v b, v c, v d, v e, v f)
-{-# INLINE unzip6 #-}
-unzip6 xs = (map (\(a, _, _, _, _, _) -> a) xs,
-             map (\(_, b, _, _, _, _) -> b) xs,
-             map (\(_, _, c, _, _, _) -> c) xs,
-             map (\(_, _, _, d, _, _) -> d) xs,
-             map (\(_, _, _, _, e, _) -> e) xs,
-             map (\(_, _, _, _, _, f) -> f) xs)
-
--- Filtering
--- ---------
-
--- | /O(n)/ Drop elements that do not satisfy the predicate
-filter :: Vector v a => (a -> Bool) -> v a -> v a
-{-# INLINE filter #-}
-filter f = unstream . inplace (S.filter f) toMax . stream
-
--- | /O(n)/ Drop elements that do not satisfy the predicate which is applied to
--- values and their indices
-ifilter :: Vector v a => (Int -> a -> Bool) -> v a -> v a
-{-# INLINE ifilter #-}
-ifilter f = unstream
-          . inplace (S.map snd . S.filter (uncurry f) . S.indexed) toMax
-          . stream
-
--- | /O(n)/ Drop repeated adjacent elements.
-uniq :: (Vector v a, Eq a) => v a -> v a
-{-# INLINE uniq #-}
-uniq = unstream . inplace S.uniq toMax . stream
-
--- | /O(n)/ Drop elements when predicate returns Nothing
-mapMaybe :: (Vector v a, Vector v b) => (a -> Maybe b) -> v a -> v b
-{-# INLINE mapMaybe #-}
-mapMaybe f = unstream . inplace (S.mapMaybe f) toMax . stream
-
--- | /O(n)/ Drop elements when predicate, applied to index and value, returns Nothing
-imapMaybe :: (Vector v a, Vector v b) => (Int -> a -> Maybe b) -> v a -> v b
-{-# INLINE imapMaybe #-}
-imapMaybe f = unstream
-          . inplace (S.mapMaybe (uncurry f) . S.indexed) toMax
-          . stream
-
-
--- | /O(n)/ Drop elements that do not satisfy the monadic predicate
-filterM :: (Monad m, Vector v a) => (a -> m Bool) -> v a -> m (v a)
-{-# INLINE filterM #-}
-filterM f = unstreamM . Bundle.filterM f . stream
-
--- | /O(n)/ Yield the longest prefix of elements satisfying the predicate
--- without copying.
-takeWhile :: Vector v a => (a -> Bool) -> v a -> v a
-{-# INLINE takeWhile #-}
-takeWhile f = unstream . Bundle.takeWhile f . stream
-
--- | /O(n)/ Drop the longest prefix of elements that satisfy the predicate
--- without copying.
-dropWhile :: Vector v a => (a -> Bool) -> v a -> v a
-{-# INLINE dropWhile #-}
-dropWhile f = unstream . Bundle.dropWhile f . stream
-
--- Parititioning
--- -------------
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't. The
--- relative order of the elements is preserved at the cost of a sometimes
--- reduced performance compared to 'unstablePartition'.
-partition :: Vector v a => (a -> Bool) -> v a -> (v a, v a)
-{-# INLINE partition #-}
-partition f = partition_stream f . stream
-
--- FIXME: Make this inplace-fusible (look at how stable_partition is
--- implemented in C++)
-
-partition_stream :: Vector v a => (a -> Bool) -> Bundle u a -> (v a, v a)
-{-# INLINE_FUSED partition_stream #-}
-partition_stream f s = s `seq` runST (
-  do
-    (mv1,mv2) <- M.partitionBundle f s
-    v1 <- unsafeFreeze mv1
-    v2 <- unsafeFreeze mv2
-    return (v1,v2))
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't.
--- The order of the elements is not preserved but the operation is often
--- faster than 'partition'.
-unstablePartition :: Vector v a => (a -> Bool) -> v a -> (v a, v a)
-{-# INLINE unstablePartition #-}
-unstablePartition f = unstablePartition_stream f . stream
-
-unstablePartition_stream
-  :: Vector v a => (a -> Bool) -> Bundle u a -> (v a, v a)
-{-# INLINE_FUSED unstablePartition_stream #-}
-unstablePartition_stream f s = s `seq` runST (
-  do
-    (mv1,mv2) <- M.unstablePartitionBundle f s
-    v1 <- unsafeFreeze mv1
-    v2 <- unsafeFreeze mv2
-    return (v1,v2))
-
-unstablePartition_new :: Vector v a => (a -> Bool) -> New v a -> (v a, v a)
-{-# INLINE_FUSED unstablePartition_new #-}
-unstablePartition_new f (New.New p) = runST (
-  do
-    mv <- p
-    i <- M.unstablePartition f mv
-    v <- unsafeFreeze mv
-    return (unsafeTake i v, unsafeDrop i v))
-
-{-# RULES
-
-"unstablePartition" forall f p.
-  unstablePartition_stream f (stream (new p))
-    = unstablePartition_new f p   #-}
-
-
-
-
--- FIXME: make span and break fusible
-
--- | /O(n)/ Split the vector into the longest prefix of elements that satisfy
--- the predicate and the rest without copying.
-span :: Vector v a => (a -> Bool) -> v a -> (v a, v a)
-{-# INLINE span #-}
-span f = break (not . f)
-
--- | /O(n)/ Split the vector into the longest prefix of elements that do not
--- satisfy the predicate and the rest without copying.
-break :: Vector v a => (a -> Bool) -> v a -> (v a, v a)
-{-# INLINE break #-}
-break f xs = case findIndex f xs of
-               Just i  -> (unsafeSlice 0 i xs, unsafeSlice i (length xs - i) xs)
-               Nothing -> (xs, empty)
-
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | /O(n)/ Check if the vector contains an element
-elem :: (Vector v a, Eq a) => a -> v a -> Bool
-{-# INLINE elem #-}
-elem x = Bundle.elem x . stream
-
-infix 4 `notElem`
--- | /O(n)/ Check if the vector does not contain an element (inverse of 'elem')
-notElem :: (Vector v a, Eq a) => a -> v a -> Bool
-{-# INLINE notElem #-}
-notElem x = Bundle.notElem x . stream
-
--- | /O(n)/ Yield 'Just' the first element matching the predicate or 'Nothing'
--- if no such element exists.
-find :: Vector v a => (a -> Bool) -> v a -> Maybe a
-{-# INLINE find #-}
-find f = Bundle.find f . stream
-
--- | /O(n)/ Yield 'Just' the index of the first element matching the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Vector v a => (a -> Bool) -> v a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex f = Bundle.findIndex f . stream
-
--- | /O(n)/ Yield the indices of elements satisfying the predicate in ascending
--- order.
-findIndices :: (Vector v a, Vector v Int) => (a -> Bool) -> v a -> v Int
-{-# INLINE findIndices #-}
-findIndices f = unstream
-              . inplace (S.map fst . S.filter (f . snd) . S.indexed) toMax
-              . stream
-
--- | /O(n)/ Yield 'Just' the index of the first occurence of the given element or
--- 'Nothing' if the vector does not contain the element. This is a specialised
--- version of 'findIndex'.
-elemIndex :: (Vector v a, Eq a) => a -> v a -> Maybe Int
-{-# INLINE elemIndex #-}
-elemIndex x = findIndex (x==)
-
--- | /O(n)/ Yield the indices of all occurences of the given element in
--- ascending order. This is a specialised version of 'findIndices'.
-elemIndices :: (Vector v a, Vector v Int, Eq a) => a -> v a -> v Int
-{-# INLINE elemIndices #-}
-elemIndices x = findIndices (x==)
-
--- Folding
--- -------
-
--- | /O(n)/ Left fold
-foldl :: Vector v b => (a -> b -> a) -> a -> v b -> a
-{-# INLINE foldl #-}
-foldl f z = Bundle.foldl f z . stream
-
--- | /O(n)/ Left fold on non-empty vectors
-foldl1 :: Vector v a => (a -> a -> a) -> v a -> a
-{-# INLINE foldl1 #-}
-foldl1 f = Bundle.foldl1 f . stream
-
--- | /O(n)/ Left fold with strict accumulator
-foldl' :: Vector v b => (a -> b -> a) -> a -> v b -> a
-{-# INLINE foldl' #-}
-foldl' f z = Bundle.foldl' f z . stream
-
--- | /O(n)/ Left fold on non-empty vectors with strict accumulator
-foldl1' :: Vector v a => (a -> a -> a) -> v a -> a
-{-# INLINE foldl1' #-}
-foldl1' f = Bundle.foldl1' f . stream
-
--- | /O(n)/ Right fold
-foldr :: Vector v a => (a -> b -> b) -> b -> v a -> b
-{-# INLINE foldr #-}
-foldr f z = Bundle.foldr f z . stream
-
--- | /O(n)/ Right fold on non-empty vectors
-foldr1 :: Vector v a => (a -> a -> a) -> v a -> a
-{-# INLINE foldr1 #-}
-foldr1 f = Bundle.foldr1 f . stream
-
--- | /O(n)/ Right fold with a strict accumulator
-foldr' :: Vector v a => (a -> b -> b) -> b -> v a -> b
-{-# INLINE foldr' #-}
-foldr' f z = Bundle.foldl' (flip f) z . streamR
-
--- | /O(n)/ Right fold on non-empty vectors with strict accumulator
-foldr1' :: Vector v a => (a -> a -> a) -> v a -> a
-{-# INLINE foldr1' #-}
-foldr1' f = Bundle.foldl1' (flip f) . streamR
-
--- | /O(n)/ Left fold (function applied to each element and its index)
-ifoldl :: Vector v b => (a -> Int -> b -> a) -> a -> v b -> a
-{-# INLINE ifoldl #-}
-ifoldl f z = Bundle.foldl (uncurry . f) z . Bundle.indexed . stream
-
--- | /O(n)/ Left fold with strict accumulator (function applied to each element
--- and its index)
-ifoldl' :: Vector v b => (a -> Int -> b -> a) -> a -> v b -> a
-{-# INLINE ifoldl' #-}
-ifoldl' f z = Bundle.foldl' (uncurry . f) z . Bundle.indexed . stream
-
--- | /O(n)/ Right fold (function applied to each element and its index)
-ifoldr :: Vector v a => (Int -> a -> b -> b) -> b -> v a -> b
-{-# INLINE ifoldr #-}
-ifoldr f z = Bundle.foldr (uncurry f) z . Bundle.indexed . stream
-
--- | /O(n)/ Right fold with strict accumulator (function applied to each
--- element and its index)
-ifoldr' :: Vector v a => (Int -> a -> b -> b) -> b -> v a -> b
-{-# INLINE ifoldr' #-}
-ifoldr' f z xs = Bundle.foldl' (flip (uncurry f)) z
-               $ Bundle.indexedR (length xs) $ streamR xs
-
--- Specialised folds
--- -----------------
-
--- | /O(n)/ Check if all elements satisfy the predicate.
-all :: Vector v a => (a -> Bool) -> v a -> Bool
-{-# INLINE all #-}
-all f = Bundle.and . Bundle.map f . stream
-
--- | /O(n)/ Check if any element satisfies the predicate.
-any :: Vector v a => (a -> Bool) -> v a -> Bool
-{-# INLINE any #-}
-any f = Bundle.or . Bundle.map f . stream
-
--- | /O(n)/ Check if all elements are 'True'
-and :: Vector v Bool => v Bool -> Bool
-{-# INLINE and #-}
-and = Bundle.and . stream
-
--- | /O(n)/ Check if any element is 'True'
-or :: Vector v Bool => v Bool -> Bool
-{-# INLINE or #-}
-or = Bundle.or . stream
-
--- | /O(n)/ Compute the sum of the elements
-sum :: (Vector v a, Num a) => v a -> a
-{-# INLINE sum #-}
-sum = Bundle.foldl' (+) 0 . stream
-
--- | /O(n)/ Compute the produce of the elements
-product :: (Vector v a, Num a) => v a -> a
-{-# INLINE product #-}
-product = Bundle.foldl' (*) 1 . stream
-
--- | /O(n)/ Yield the maximum element of the vector. The vector may not be
--- empty.
-maximum :: (Vector v a, Ord a) => v a -> a
-{-# INLINE maximum #-}
-maximum = Bundle.foldl1' max . stream
-
--- | /O(n)/ Yield the maximum element of the vector according to the given
--- comparison function. The vector may not be empty.
-maximumBy :: Vector v a => (a -> a -> Ordering) -> v a -> a
-{-# INLINE maximumBy #-}
-maximumBy cmpr = Bundle.foldl1' maxBy . stream
-  where
-    {-# INLINE maxBy #-}
-    maxBy x y = case cmpr x y of
-                  LT -> y
-                  _  -> x
-
--- | /O(n)/ Yield the minimum element of the vector. The vector may not be
--- empty.
-minimum :: (Vector v a, Ord a) => v a -> a
-{-# INLINE minimum #-}
-minimum = Bundle.foldl1' min . stream
-
--- | /O(n)/ Yield the minimum element of the vector according to the given
--- comparison function. The vector may not be empty.
-minimumBy :: Vector v a => (a -> a -> Ordering) -> v a -> a
-{-# INLINE minimumBy #-}
-minimumBy cmpr = Bundle.foldl1' minBy . stream
-  where
-    {-# INLINE minBy #-}
-    minBy x y = case cmpr x y of
-                  GT -> y
-                  _  -> x
-
--- | /O(n)/ Yield the index of the maximum element of the vector. The vector
--- may not be empty.
-maxIndex :: (Vector v a, Ord a) => v a -> Int
-{-# INLINE maxIndex #-}
-maxIndex = maxIndexBy compare
-
--- | /O(n)/ Yield the index of the maximum element of the vector according to
--- the given comparison function. The vector may not be empty.
-maxIndexBy :: Vector v a => (a -> a -> Ordering) -> v a -> Int
-{-# INLINE maxIndexBy #-}
-maxIndexBy cmpr = fst . Bundle.foldl1' imax . Bundle.indexed . stream
-  where
-    imax (i,x) (j,y) = i `seq` j `seq`
-                       case cmpr x y of
-                         LT -> (j,y)
-                         _  -> (i,x)
-
--- | /O(n)/ Yield the index of the minimum element of the vector. The vector
--- may not be empty.
-minIndex :: (Vector v a, Ord a) => v a -> Int
-{-# INLINE minIndex #-}
-minIndex = minIndexBy compare
-
--- | /O(n)/ Yield the index of the minimum element of the vector according to
--- the given comparison function. The vector may not be empty.
-minIndexBy :: Vector v a => (a -> a -> Ordering) -> v a -> Int
-{-# INLINE minIndexBy #-}
-minIndexBy cmpr = fst . Bundle.foldl1' imin . Bundle.indexed . stream
-  where
-    imin (i,x) (j,y) = i `seq` j `seq`
-                       case cmpr x y of
-                         GT -> (j,y)
-                         _  -> (i,x)
-
--- Monadic folds
--- -------------
-
--- | /O(n)/ Monadic fold
-foldM :: (Monad m, Vector v b) => (a -> b -> m a) -> a -> v b -> m a
-{-# INLINE foldM #-}
-foldM m z = Bundle.foldM m z . stream
-
--- | /O(n)/ Monadic fold (action applied to each element and its index)
-ifoldM :: (Monad m, Vector v b) => (a -> Int -> b -> m a) -> a -> v b -> m a
-{-# INLINE ifoldM #-}
-ifoldM m z = Bundle.foldM (uncurry . m) z . Bundle.indexed . stream
-
--- | /O(n)/ Monadic fold over non-empty vectors
-fold1M :: (Monad m, Vector v a) => (a -> a -> m a) -> v a -> m a
-{-# INLINE fold1M #-}
-fold1M m = Bundle.fold1M m . stream
-
--- | /O(n)/ Monadic fold with strict accumulator
-foldM' :: (Monad m, Vector v b) => (a -> b -> m a) -> a -> v b -> m a
-{-# INLINE foldM' #-}
-foldM' m z = Bundle.foldM' m z . stream
-
--- | /O(n)/ Monadic fold with strict accumulator (action applied to each
--- element and its index)
-ifoldM' :: (Monad m, Vector v b) => (a -> Int -> b -> m a) -> a -> v b -> m a
-{-# INLINE ifoldM' #-}
-ifoldM' m z = Bundle.foldM' (uncurry . m) z . Bundle.indexed . stream
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
-fold1M' :: (Monad m, Vector v a) => (a -> a -> m a) -> v a -> m a
-{-# INLINE fold1M' #-}
-fold1M' m = Bundle.fold1M' m . stream
-
-discard :: Monad m => m a -> m ()
-{-# INLINE discard #-}
-discard m = m >> return ()
-
--- | /O(n)/ Monadic fold that discards the result
-foldM_ :: (Monad m, Vector v b) => (a -> b -> m a) -> a -> v b -> m ()
-{-# INLINE foldM_ #-}
-foldM_ m z = discard . Bundle.foldM m z . stream
-
--- | /O(n)/ Monadic fold that discards the result (action applied to
--- each element and its index)
-ifoldM_ :: (Monad m, Vector v b) => (a -> Int -> b -> m a) -> a -> v b -> m ()
-{-# INLINE ifoldM_ #-}
-ifoldM_ m z = discard . Bundle.foldM (uncurry . m) z . Bundle.indexed . stream
-
--- | /O(n)/ Monadic fold over non-empty vectors that discards the result
-fold1M_ :: (Monad m, Vector v a) => (a -> a -> m a) -> v a -> m ()
-{-# INLINE fold1M_ #-}
-fold1M_ m = discard . Bundle.fold1M m . stream
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
-foldM'_ :: (Monad m, Vector v b) => (a -> b -> m a) -> a -> v b -> m ()
-{-# INLINE foldM'_ #-}
-foldM'_ m z = discard . Bundle.foldM' m z . stream
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
--- (action applied to each element and its index)
-ifoldM'_ :: (Monad m, Vector v b) => (a -> Int -> b -> m a) -> a -> v b -> m ()
-{-# INLINE ifoldM'_ #-}
-ifoldM'_ m z = discard . Bundle.foldM' (uncurry . m) z . Bundle.indexed . stream
-
--- | /O(n)/ Monad fold over non-empty vectors with strict accumulator
--- that discards the result
-fold1M'_ :: (Monad m, Vector v a) => (a -> a -> m a) -> v a -> m ()
-{-# INLINE fold1M'_ #-}
-fold1M'_ m = discard . Bundle.fold1M' m . stream
-
--- Monadic sequencing
--- ------------------
-
--- | Evaluate each action and collect the results
-sequence :: (Monad m, Vector v a, Vector v (m a)) => v (m a) -> m (v a)
-{-# INLINE sequence #-}
-sequence = mapM id
-
--- | Evaluate each action and discard the results
-sequence_ :: (Monad m, Vector v (m a)) => v (m a) -> m ()
-{-# INLINE sequence_ #-}
-sequence_ = mapM_ id
-
--- Prefix sums (scans)
--- -------------------
-
--- | /O(n)/ Prescan
---
--- @
--- prescanl f z = 'init' . 'scanl' f z
--- @
---
--- Example: @prescanl (+) 0 \<1,2,3,4\> = \<0,1,3,6\>@
---
-prescanl :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE prescanl #-}
-prescanl f z = unstream . inplace (S.prescanl f z) id . stream
-
--- | /O(n)/ Prescan with strict accumulator
-prescanl' :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE prescanl' #-}
-prescanl' f z = unstream . inplace (S.prescanl' f z) id . stream
-
--- | /O(n)/ Scan
---
--- @
--- postscanl f z = 'tail' . 'scanl' f z
--- @
---
--- Example: @postscanl (+) 0 \<1,2,3,4\> = \<1,3,6,10\>@
---
-postscanl :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE postscanl #-}
-postscanl f z = unstream . inplace (S.postscanl f z) id . stream
-
--- | /O(n)/ Scan with strict accumulator
-postscanl' :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE postscanl' #-}
-postscanl' f z = unstream . inplace (S.postscanl' f z) id . stream
-
--- | /O(n)/ Haskell-style scan
---
--- > scanl f z <x1,...,xn> = <y1,...,y(n+1)>
--- >   where y1 = z
--- >         yi = f y(i-1) x(i-1)
---
--- Example: @scanl (+) 0 \<1,2,3,4\> = \<0,1,3,6,10\>@
---
-scanl :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE scanl #-}
-scanl f z = unstream . Bundle.scanl f z . stream
-
--- | /O(n)/ Haskell-style scan with strict accumulator
-scanl' :: (Vector v a, Vector v b) => (a -> b -> a) -> a -> v b -> v a
-{-# INLINE scanl' #-}
-scanl' f z = unstream . Bundle.scanl' f z . stream
-
--- | /O(n)/ Scan over a vector with its index
-iscanl :: (Vector v a, Vector v b) => (Int -> a -> b -> a) -> a -> v b -> v a
-{-# INLINE iscanl #-}
-iscanl f z =
-    unstream
-  . inplace (S.scanl (\a (i, b) -> f i a b) z . S.indexed) (+1)
-  . stream
-
--- | /O(n)/ Scan over a vector (strictly) with its index
-iscanl' :: (Vector v a, Vector v b) => (Int -> a -> b -> a) -> a -> v b -> v a
-{-# INLINE iscanl' #-}
-iscanl' f z =
-    unstream
-  . inplace (S.scanl' (\a (i, b) -> f i a b) z . S.indexed) (+1)
-  . stream
-
-
--- | /O(n)/ Scan over a non-empty vector
---
--- > scanl f <x1,...,xn> = <y1,...,yn>
--- >   where y1 = x1
--- >         yi = f y(i-1) xi
---
-scanl1 :: Vector v a => (a -> a -> a) -> v a -> v a
-{-# INLINE scanl1 #-}
-scanl1 f = unstream . inplace (S.scanl1 f) id . stream
-
--- | /O(n)/ Scan over a non-empty vector with a strict accumulator
-scanl1' :: Vector v a => (a -> a -> a) -> v a -> v a
-{-# INLINE scanl1' #-}
-scanl1' f = unstream . inplace (S.scanl1' f) id . stream
-
--- | /O(n)/ Right-to-left prescan
---
--- @
--- prescanr f z = 'reverse' . 'prescanl' (flip f) z . 'reverse'
--- @
---
-prescanr :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE prescanr #-}
-prescanr f z = unstreamR . inplace (S.prescanl (flip f) z) id . streamR
-
--- | /O(n)/ Right-to-left prescan with strict accumulator
-prescanr' :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE prescanr' #-}
-prescanr' f z = unstreamR . inplace (S.prescanl' (flip f) z) id . streamR
-
--- | /O(n)/ Right-to-left scan
-postscanr :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE postscanr #-}
-postscanr f z = unstreamR . inplace (S.postscanl (flip f) z) id . streamR
-
--- | /O(n)/ Right-to-left scan with strict accumulator
-postscanr' :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE postscanr' #-}
-postscanr' f z = unstreamR . inplace (S.postscanl' (flip f) z) id . streamR
-
--- | /O(n)/ Right-to-left Haskell-style scan
-scanr :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE scanr #-}
-scanr f z = unstreamR . Bundle.scanl (flip f) z . streamR
-
--- | /O(n)/ Right-to-left Haskell-style scan with strict accumulator
-scanr' :: (Vector v a, Vector v b) => (a -> b -> b) -> b -> v a -> v b
-{-# INLINE scanr' #-}
-scanr' f z = unstreamR . Bundle.scanl' (flip f) z . streamR
-
--- | /O(n)/ Right-to-left scan over a vector with its index
-iscanr :: (Vector v a, Vector v b) => (Int -> a -> b -> b) -> b -> v a -> v b
-{-# INLINE iscanr #-}
-iscanr f z v =
-    unstreamR
-  . inplace (S.scanl (flip $ uncurry f) z . S.indexedR n) (+1)
-  . streamR
-  $ v
- where n = length v
-
--- | /O(n)/ Right-to-left scan over a vector (strictly) with its index
-iscanr' :: (Vector v a, Vector v b) => (Int -> a -> b -> b) -> b -> v a -> v b
-{-# INLINE iscanr' #-}
-iscanr' f z v =
-    unstreamR
-  . inplace (S.scanl' (flip $ uncurry f) z . S.indexedR n) (+1)
-  . streamR
-  $ v
- where n = length v
-
--- | /O(n)/ Right-to-left scan over a non-empty vector
-scanr1 :: Vector v a => (a -> a -> a) -> v a -> v a
-{-# INLINE scanr1 #-}
-scanr1 f = unstreamR . inplace (S.scanl1 (flip f)) id . streamR
-
--- | /O(n)/ Right-to-left scan over a non-empty vector with a strict
--- accumulator
-scanr1' :: Vector v a => (a -> a -> a) -> v a -> v a
-{-# INLINE scanr1' #-}
-scanr1' f = unstreamR . inplace (S.scanl1' (flip f)) id . streamR
-
--- Conversions - Lists
--- ------------------------
-
--- | /O(n)/ Convert a vector to a list
-toList :: Vector v a => v a -> [a]
-{-# INLINE toList #-}
-toList = Bundle.toList . stream
-
--- | /O(n)/ Convert a list to a vector
-fromList :: Vector v a => [a] -> v a
-{-# INLINE fromList #-}
-fromList = unstream . Bundle.fromList
-
--- | /O(n)/ Convert the first @n@ elements of a list to a vector
---
--- @
--- fromListN n xs = 'fromList' ('take' n xs)
--- @
-fromListN :: Vector v a => Int -> [a] -> v a
-{-# INLINE fromListN #-}
-fromListN n = unstream . Bundle.fromListN n
-
--- Conversions - Immutable vectors
--- -------------------------------
-
--- | /O(n)/ Convert different vector types
-convert :: (Vector v a, Vector w a) => v a -> w a
-{-# INLINE convert #-}
-convert = unstream . Bundle.reVector . stream
-
--- Conversions - Mutable vectors
--- -----------------------------
-
--- | /O(1)/ Unsafe convert a mutable vector to an immutable one without
--- copying. The mutable vector may not be used after this operation.
-unsafeFreeze
-  :: (PrimMonad m, Vector v a) => Mutable v (PrimState m) a -> m (v a)
-{-# INLINE unsafeFreeze #-}
-unsafeFreeze = basicUnsafeFreeze
-
--- | /O(n)/ Yield an immutable copy of the mutable vector.
-freeze :: (PrimMonad m, Vector v a) => Mutable v (PrimState m) a -> m (v a)
-{-# INLINE freeze #-}
-freeze mv = unsafeFreeze =<< M.clone mv
-
--- | /O(1)/ Unsafely convert an immutable vector to a mutable one without
--- copying. The immutable vector may not be used after this operation.
-unsafeThaw :: (PrimMonad m, Vector v a) => v a -> m (Mutable v (PrimState m) a)
-{-# INLINE_FUSED unsafeThaw #-}
-unsafeThaw = basicUnsafeThaw
-
--- | /O(n)/ Yield a mutable copy of the immutable vector.
-thaw :: (PrimMonad m, Vector v a) => v a -> m (Mutable v (PrimState m) a)
-{-# INLINE_FUSED thaw #-}
-thaw v = do
-           mv <- M.unsafeNew (length v)
-           unsafeCopy mv v
-           return mv
-
-{-# RULES
-
-"unsafeThaw/new [Vector]" forall p.
-  unsafeThaw (new p) = New.runPrim p
-
-"thaw/new [Vector]" forall p.
-  thaw (new p) = New.runPrim p   #-}
-
-
-
-{-
--- | /O(n)/ Yield a mutable vector containing copies of each vector in the
--- list.
-thawMany :: (PrimMonad m, Vector v a) => [v a] -> m (Mutable v (PrimState m) a)
-{-# INLINE_FUSED thawMany #-}
--- FIXME: add rule for (stream (new (New.create (thawMany vs))))
--- NOTE: We don't try to consume the list lazily as this wouldn't significantly
--- change the space requirements anyway.
-thawMany vs = do
-                mv <- M.new n
-                thaw_loop mv vs
-                return mv
-  where
-    n = List.foldl' (\k v -> k + length v) 0 vs
-
-    thaw_loop mv [] = mv `seq` return ()
-    thaw_loop mv (v:vs)
-      = do
-          let n = length v
-          unsafeCopy (M.unsafeTake n mv) v
-          thaw_loop (M.unsafeDrop n mv) vs
--}
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length.
-copy
-  :: (PrimMonad m, Vector v a) => Mutable v (PrimState m) a -> v a -> m ()
-{-# INLINE copy #-}
-copy dst src = BOUNDS_CHECK(check) "copy" "length mismatch"
-                                          (M.length dst == length src)
-             $ unsafeCopy dst src
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length. This is not checked.
-unsafeCopy
-  :: (PrimMonad m, Vector v a) => Mutable v (PrimState m) a -> v a -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy dst src = UNSAFE_CHECK(check) "unsafeCopy" "length mismatch"
-                                         (M.length dst == length src)
-                   $ (dst `seq` src `seq` basicUnsafeCopy dst src)
-
--- Conversions to/from Bundles
--- ---------------------------
-
--- | /O(1)/ Convert a vector to a 'Bundle'
-stream :: Vector v a => v a -> Bundle v a
-{-# INLINE_FUSED stream #-}
-stream v = stream' v
-
--- Same as 'stream', but can be used to avoid having a cycle in the dependency
--- graph of functions, which forces GHC to create a loop breaker.
-stream' :: Vector v a => v a -> Bundle v a
-{-# INLINE stream' #-}
-stream' v = Bundle.fromVector v
-
-{-
-stream v = v `seq` n `seq` (Bundle.unfoldr get 0 `Bundle.sized` Exact n)
-  where
-    n = length v
-
-    -- NOTE: the False case comes first in Core so making it the recursive one
-    -- makes the code easier to read
-    {-# INLINE get #-}
-    get i | i >= n    = Nothing
-          | otherwise = case basicUnsafeIndexM v i of Box x -> Just (x, i+1)
--}
-
--- | /O(n)/ Construct a vector from a 'Bundle'
-unstream :: Vector v a => Bundle v a -> v a
-{-# INLINE unstream #-}
-unstream s = new (New.unstream s)
-
-{-# RULES
-
-"stream/unstream [Vector]" forall s.
-  stream (new (New.unstream s)) = s
-
-"New.unstream/stream [Vector]" forall v.
-  New.unstream (stream v) = clone v
-
-"clone/new [Vector]" forall p.
-  clone (new p) = p
-
-"inplace [Vector]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a) g m.
-  New.unstream (inplace f g (stream (new m))) = New.transform f g m
-
-"uninplace [Vector]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a) g m.
-  stream (new (New.transform f g m)) = inplace f g (stream (new m))  #-}
-
-
-
--- | /O(1)/ Convert a vector to a 'Bundle', proceeding from right to left
-streamR :: Vector v a => v a -> Bundle u a
-{-# INLINE_FUSED streamR #-}
-streamR v = v `seq` n `seq` (Bundle.unfoldr get n `Bundle.sized` Exact n)
-  where
-    n = length v
-
-    {-# INLINE get #-}
-    get 0 = Nothing
-    get i = let i' = i-1
-            in
-            case basicUnsafeIndexM v i' of Box x -> Just (x, i')
-
--- | /O(n)/ Construct a vector from a 'Bundle', proceeding from right to left
-unstreamR :: Vector v a => Bundle v a -> v a
-{-# INLINE unstreamR #-}
-unstreamR s = new (New.unstreamR s)
-
-{-# RULES
-
-"streamR/unstreamR [Vector]" forall s.
-  streamR (new (New.unstreamR s)) = s
-
-"New.unstreamR/streamR/new [Vector]" forall p.
-  New.unstreamR (streamR (new p)) = p
-
-"New.unstream/streamR/new [Vector]" forall p.
-  New.unstream (streamR (new p)) = New.modify M.reverse p
-
-"New.unstreamR/stream/new [Vector]" forall p.
-  New.unstreamR (stream (new p)) = New.modify M.reverse p
-
-"inplace right [Vector]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a) g m.
-  New.unstreamR (inplace f g (streamR (new m))) = New.transformR f g m
-
-"uninplace right [Vector]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a) g m.
-  streamR (new (New.transformR f g m)) = inplace f g (streamR (new m))  #-}
-
-
-
-unstreamM :: (Monad m, Vector v a) => MBundle m u a -> m (v a)
-{-# INLINE_FUSED unstreamM #-}
-unstreamM s = do
-                xs <- MBundle.toList s
-                return $ unstream $ Bundle.unsafeFromList (MBundle.size s) xs
-
-unstreamPrimM :: (PrimMonad m, Vector v a) => MBundle m u a -> m (v a)
-{-# INLINE_FUSED unstreamPrimM #-}
-unstreamPrimM s = M.munstream s >>= unsafeFreeze
-
--- FIXME: the next two functions are only necessary for the specialisations
-unstreamPrimM_IO :: Vector v a => MBundle IO u a -> IO (v a)
-{-# INLINE unstreamPrimM_IO #-}
-unstreamPrimM_IO = unstreamPrimM
-
-unstreamPrimM_ST :: Vector v a => MBundle (ST s) u a -> ST s (v a)
-{-# INLINE unstreamPrimM_ST #-}
-unstreamPrimM_ST = unstreamPrimM
-
-{-# RULES
-
-"unstreamM[IO]" unstreamM = unstreamPrimM_IO
-"unstreamM[ST]" unstreamM = unstreamPrimM_ST  #-}
-
-
-
-
--- Recycling support
--- -----------------
-
--- | Construct a vector from a monadic initialiser.
-new :: Vector v a => New v a -> v a
-{-# INLINE_FUSED new #-}
-new m = m `seq` runST (unsafeFreeze =<< New.run m)
-
--- | Convert a vector to an initialiser which, when run, produces a copy of
--- the vector.
-clone :: Vector v a => v a -> New v a
-{-# INLINE_FUSED clone #-}
-clone v = v `seq` New.create (
-  do
-    mv <- M.new (length v)
-    unsafeCopy mv v
-    return mv)
-
--- Comparisons
--- -----------
-
--- | /O(n)/ Check if two vectors are equal. All 'Vector' instances are also
--- instances of 'Eq' and it is usually more appropriate to use those. This
--- function is primarily intended for implementing 'Eq' instances for new
--- vector types.
-eq :: (Vector v a, Eq a) => v a -> v a -> Bool
-{-# INLINE eq #-}
-xs `eq` ys = stream xs == stream ys
-
--- | /O(n)/
-eqBy :: (Vector v a, Vector v b) => (a -> b -> Bool) -> v a -> v b -> Bool
-{-# INLINE eqBy #-}
-eqBy e xs ys = Bundle.eqBy e (stream xs) (stream ys)
-
--- | /O(n)/ Compare two vectors lexicographically. All 'Vector' instances are
--- also instances of 'Ord' and it is usually more appropriate to use those. This
--- function is primarily intended for implementing 'Ord' instances for new
--- vector types.
-cmp :: (Vector v a, Ord a) => v a -> v a -> Ordering
-{-# INLINE cmp #-}
-cmp xs ys = compare (stream xs) (stream ys)
-
--- | /O(n)/
-cmpBy :: (Vector v a, Vector v b) => (a -> b -> Ordering) -> v a -> v b -> Ordering
-cmpBy c xs ys = Bundle.cmpBy c (stream xs) (stream ys)
-
--- Show
--- ----
-
--- | Generic definition of 'Prelude.showsPrec'
-showsPrec :: (Vector v a, Show a) => Int -> v a -> ShowS
-{-# INLINE showsPrec #-}
-showsPrec _ = shows . toList
-
-liftShowsPrec :: (Vector v a) => (Int -> a -> ShowS) -> ([a] -> ShowS) -> Int -> v a -> ShowS
-{-# INLINE liftShowsPrec #-}
-liftShowsPrec _ s _ = s . toList
-
--- | Generic definition of 'Text.Read.readPrec'
-readPrec :: (Vector v a, Read a) => Read.ReadPrec (v a)
-{-# INLINE readPrec #-}
-readPrec = do
-  xs <- Read.readPrec
-  return (fromList xs)
-
--- | /Note:/ uses 'ReadS'
-liftReadsPrec :: (Vector v a) => (Int -> Read.ReadS a) -> ReadS [a] -> Int -> Read.ReadS (v a)
-liftReadsPrec _ r _ s = [ (fromList v, s') | (v, s') <- r s ]
-
--- Data and Typeable
--- -----------------
-
--- | Generic definion of 'Data.Data.gfoldl' that views a 'Vector' as a
--- list.
-gfoldl :: (Vector v a, Data a)
-       => (forall d b. Data d => c (d -> b) -> d -> c b)
-       -> (forall g. g -> c g)
-       -> v a
-       -> c (v a)
-{-# INLINE gfoldl #-}
-gfoldl f z v = z fromList `f` toList v
-
-mkType :: String -> DataType
-{-# INLINE mkType #-}
-mkType = mkNoRepType
-
-#if __GLASGOW_HASKELL__ >= 707
-dataCast :: (Vector v a, Data a, Typeable v, Typeable t)
-#else
-dataCast :: (Vector v a, Data a, Typeable1 v, Typeable1 t)
-#endif
-         => (forall d. Data  d => c (t d)) -> Maybe  (c (v a))
-{-# INLINE dataCast #-}
-dataCast f = gcast1 f
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Base.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Base.hs
deleted file mode 100644
index a760329c599f..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Base.hs
+++ /dev/null
@@ -1,140 +0,0 @@
-{-# LANGUAGE Rank2Types, MultiParamTypeClasses, FlexibleContexts,
-             TypeFamilies, ScopedTypeVariables, BangPatterns #-}
-{-# OPTIONS_HADDOCK hide #-}
-
--- |
--- Module      : Data.Vector.Generic.Base
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Class of pure vectors
---
-
-module Data.Vector.Generic.Base (
-  Vector(..), Mutable
-) where
-
-import           Data.Vector.Generic.Mutable.Base ( MVector )
-import qualified Data.Vector.Generic.Mutable.Base as M
-
-import Control.Monad.Primitive
-
--- | @Mutable v s a@ is the mutable version of the pure vector type @v a@ with
--- the state token @s@
---
-type family Mutable (v :: * -> *) :: * -> * -> *
-
--- | Class of immutable vectors. Every immutable vector is associated with its
--- mutable version through the 'Mutable' type family. Methods of this class
--- should not be used directly. Instead, "Data.Vector.Generic" and other
--- Data.Vector modules provide safe and fusible wrappers.
---
--- Minimum complete implementation:
---
---   * 'basicUnsafeFreeze'
---
---   * 'basicUnsafeThaw'
---
---   * 'basicLength'
---
---   * 'basicUnsafeSlice'
---
---   * 'basicUnsafeIndexM'
---
-class MVector (Mutable v) a => Vector v a where
-  -- | /Assumed complexity: O(1)/
-  --
-  -- Unsafely convert a mutable vector to its immutable version
-  -- without copying. The mutable vector may not be used after
-  -- this operation.
-  basicUnsafeFreeze :: PrimMonad m => Mutable v (PrimState m) a -> m (v a)
-
-  -- | /Assumed complexity: O(1)/
-  --
-  -- Unsafely convert an immutable vector to its mutable version without
-  -- copying. The immutable vector may not be used after this operation.
-  basicUnsafeThaw :: PrimMonad m => v a -> m (Mutable v (PrimState m) a)
-
-  -- | /Assumed complexity: O(1)/
-  --
-  -- Yield the length of the vector.
-  basicLength      :: v a -> Int
-
-  -- | /Assumed complexity: O(1)/
-  --
-  -- Yield a slice of the vector without copying it. No range checks are
-  -- performed.
-  basicUnsafeSlice  :: Int -- ^ starting index
-                    -> Int -- ^ length
-                    -> v a -> v a
-
-  -- | /Assumed complexity: O(1)/
-  --
-  -- Yield the element at the given position in a monad. No range checks are
-  -- performed.
-  --
-  -- The monad allows us to be strict in the vector if we want. Suppose we had
-  --
-  -- > unsafeIndex :: v a -> Int -> a
-  --
-  -- instead. Now, if we wanted to copy a vector, we'd do something like
-  --
-  -- > copy mv v ... = ... unsafeWrite mv i (unsafeIndex v i) ...
-  --
-  -- For lazy vectors, the indexing would not be evaluated which means that we
-  -- would retain a reference to the original vector in each element we write.
-  -- This is not what we want!
-  --
-  -- With 'basicUnsafeIndexM', we can do
-  --
-  -- > copy mv v ... = ... case basicUnsafeIndexM v i of
-  -- >                       Box x -> unsafeWrite mv i x ...
-  --
-  -- which does not have this problem because indexing (but not the returned
-  -- element!) is evaluated immediately.
-  --
-  basicUnsafeIndexM  :: Monad m => v a -> Int -> m a
-
-  -- |  /Assumed complexity: O(n)/
-  --
-  -- Copy an immutable vector into a mutable one. The two vectors must have
-  -- the same length but this is not checked.
-  --
-  -- Instances of 'Vector' should redefine this method if they wish to support
-  -- an efficient block copy operation.
-  --
-  -- Default definition: copying basic on 'basicUnsafeIndexM' and
-  -- 'basicUnsafeWrite'.
-  basicUnsafeCopy :: PrimMonad m => Mutable v (PrimState m) a -> v a -> m ()
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy !dst !src = do_copy 0
-    where
-      !n = basicLength src
-
-      do_copy i | i < n = do
-                            x <- basicUnsafeIndexM src i
-                            M.basicUnsafeWrite dst i x
-                            do_copy (i+1)
-                | otherwise = return ()
-
-  -- | Evaluate @a@ as far as storing it in a vector would and yield @b@.
-  -- The @v a@ argument only fixes the type and is not touched. The method is
-  -- only used for optimisation purposes. Thus, it is safe for instances of
-  -- 'Vector' to evaluate @a@ less than it would be when stored in a vector
-  -- although this might result in suboptimal code.
-  --
-  -- > elemseq v x y = (singleton x `asTypeOf` v) `seq` y
-  --
-  -- Default defintion: @a@ is not evaluated at all
-  --
-  elemseq :: v a -> a -> b -> b
-
-  {-# INLINE elemseq #-}
-  elemseq _ = \_ x -> x
-
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable.hs
deleted file mode 100644
index 89bebf360765..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable.hs
+++ /dev/null
@@ -1,1034 +0,0 @@
-{-# LANGUAGE CPP, MultiParamTypeClasses, FlexibleContexts, BangPatterns, TypeFamilies, ScopedTypeVariables #-}
--- |
--- Module      : Data.Vector.Generic.Mutable
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Generic interface to mutable vectors
---
-
-module Data.Vector.Generic.Mutable (
-  -- * Class of mutable vector types
-  MVector(..),
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Extracting subvectors
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- ** Overlapping
-  overlaps,
-
-  -- * Construction
-
-  -- ** Initialisation
-  new, unsafeNew, replicate, replicateM, clone,
-
-  -- ** Growing
-  grow, unsafeGrow,
-  growFront, unsafeGrowFront,
-
-  -- ** Restricting memory usage
-  clear,
-
-  -- * Accessing individual elements
-  read, write, modify, swap, exchange,
-  unsafeRead, unsafeWrite, unsafeModify, unsafeSwap, unsafeExchange,
-
-  -- * Modifying vectors
-  nextPermutation,
-
-  -- ** Filling and copying
-  set, copy, move, unsafeCopy, unsafeMove,
-
-  -- * Internal operations
-  mstream, mstreamR,
-  unstream, unstreamR, vunstream,
-  munstream, munstreamR,
-  transform, transformR,
-  fill, fillR,
-  unsafeAccum, accum, unsafeUpdate, update, reverse,
-  unstablePartition, unstablePartitionBundle, partitionBundle
-) where
-
-import           Data.Vector.Generic.Mutable.Base
-import qualified Data.Vector.Generic.Base as V
-
-import qualified Data.Vector.Fusion.Bundle      as Bundle
-import           Data.Vector.Fusion.Bundle      ( Bundle, MBundle, Chunk(..) )
-import qualified Data.Vector.Fusion.Bundle.Monadic as MBundle
-import           Data.Vector.Fusion.Stream.Monadic ( Stream )
-import qualified Data.Vector.Fusion.Stream.Monadic as Stream
-import           Data.Vector.Fusion.Bundle.Size
-import           Data.Vector.Fusion.Util        ( delay_inline )
-
-import Control.Monad.Primitive ( PrimMonad, PrimState )
-
-import Prelude hiding ( length, null, replicate, reverse, map, read,
-                        take, drop, splitAt, init, tail )
-
-#include "vector.h"
-
-{-
-type family Immutable (v :: * -> * -> *) :: * -> *
-
--- | Class of mutable vectors parametrised with a primitive state token.
---
-class MBundle.Pointer u a => MVector v a where
-  -- | Length of the mutable vector. This method should not be
-  -- called directly, use 'length' instead.
-  basicLength       :: v s a -> Int
-
-  -- | Yield a part of the mutable vector without copying it. This method
-  -- should not be called directly, use 'unsafeSlice' instead.
-  basicUnsafeSlice :: Int  -- ^ starting index
-                   -> Int  -- ^ length of the slice
-                   -> v s a
-                   -> v s a
-
-  -- Check whether two vectors overlap. This method should not be
-  -- called directly, use 'overlaps' instead.
-  basicOverlaps    :: v s a -> v s a -> Bool
-
-  -- | Create a mutable vector of the given length. This method should not be
-  -- called directly, use 'unsafeNew' instead.
-  basicUnsafeNew   :: PrimMonad m => Int -> m (v (PrimState m) a)
-
-  -- | Create a mutable vector of the given length and fill it with an
-  -- initial value. This method should not be called directly, use
-  -- 'replicate' instead.
-  basicUnsafeReplicate :: PrimMonad m => Int -> a -> m (v (PrimState m) a)
-
-  -- | Yield the element at the given position. This method should not be
-  -- called directly, use 'unsafeRead' instead.
-  basicUnsafeRead  :: PrimMonad m => v (PrimState m) a -> Int -> m a
-
-  -- | Replace the element at the given position. This method should not be
-  -- called directly, use 'unsafeWrite' instead.
-  basicUnsafeWrite :: PrimMonad m => v (PrimState m) a -> Int -> a -> m ()
-
-  -- | Reset all elements of the vector to some undefined value, clearing all
-  -- references to external objects. This is usually a noop for unboxed
-  -- vectors. This method should not be called directly, use 'clear' instead.
-  basicClear       :: PrimMonad m => v (PrimState m) a -> m ()
-
-  -- | Set all elements of the vector to the given value. This method should
-  -- not be called directly, use 'set' instead.
-  basicSet         :: PrimMonad m => v (PrimState m) a -> a -> m ()
-
-  basicUnsafeCopyPointer :: PrimMonad m => v (PrimState m) a
-                                        -> Immutable v a
-                                        -> m ()
-
-  -- | Copy a vector. The two vectors may not overlap. This method should not
-  -- be called directly, use 'unsafeCopy' instead.
-  basicUnsafeCopy  :: PrimMonad m => v (PrimState m) a   -- ^ target
-                                  -> v (PrimState m) a   -- ^ source
-                                  -> m ()
-
-  -- | Move the contents of a vector. The two vectors may overlap. This method
-  -- should not be called directly, use 'unsafeMove' instead.
-  basicUnsafeMove  :: PrimMonad m => v (PrimState m) a   -- ^ target
-                                  -> v (PrimState m) a   -- ^ source
-                                  -> m ()
-
-  -- | Grow a vector by the given number of elements. This method should not be
-  -- called directly, use 'unsafeGrow' instead.
-  basicUnsafeGrow  :: PrimMonad m => v (PrimState m) a -> Int
-                                                       -> m (v (PrimState m) a)
-
-  {-# INLINE basicUnsafeReplicate #-}
-  basicUnsafeReplicate n x
-    = do
-        v <- basicUnsafeNew n
-        basicSet v x
-        return v
-
-  {-# INLINE basicClear #-}
-  basicClear _ = return ()
-
-  {-# INLINE basicSet #-}
-  basicSet !v x
-    | n == 0    = return ()
-    | otherwise = do
-                    basicUnsafeWrite v 0 x
-                    do_set 1
-    where
-      !n = basicLength v
-
-      do_set i | 2*i < n = do basicUnsafeCopy (basicUnsafeSlice i i v)
-                                              (basicUnsafeSlice 0 i v)
-                              do_set (2*i)
-               | otherwise = basicUnsafeCopy (basicUnsafeSlice i (n-i) v)
-                                             (basicUnsafeSlice 0 (n-i) v)
-
-  {-# INLINE basicUnsafeCopyPointer #-}
-  basicUnsafeCopyPointer !dst !src = do_copy 0 src
-    where
-      do_copy !i p | Just (x,q) <- MBundle.pget p = do
-                                                      basicUnsafeWrite dst i x
-                                                      do_copy (i+1) q
-                   | otherwise = return ()
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy !dst !src = do_copy 0
-    where
-      !n = basicLength src
-
-      do_copy i | i < n = do
-                            x <- basicUnsafeRead src i
-                            basicUnsafeWrite dst i x
-                            do_copy (i+1)
-                | otherwise = return ()
-
-  {-# INLINE basicUnsafeMove #-}
-  basicUnsafeMove !dst !src
-    | basicOverlaps dst src = do
-        srcCopy <- clone src
-        basicUnsafeCopy dst srcCopy
-    | otherwise = basicUnsafeCopy dst src
-
-  {-# INLINE basicUnsafeGrow #-}
-  basicUnsafeGrow v by
-    = do
-        v' <- basicUnsafeNew (n+by)
-        basicUnsafeCopy (basicUnsafeSlice 0 n v') v
-        return v'
-    where
-      n = basicLength v
--}
-
--- ------------------
--- Internal functions
--- ------------------
-
-unsafeAppend1 :: (PrimMonad m, MVector v a)
-        => v (PrimState m) a -> Int -> a -> m (v (PrimState m) a)
-{-# INLINE_INNER unsafeAppend1 #-}
-    -- NOTE: The case distinction has to be on the outside because
-    -- GHC creates a join point for the unsafeWrite even when everything
-    -- is inlined. This is bad because with the join point, v isn't getting
-    -- unboxed.
-unsafeAppend1 v i x
-  | i < length v = do
-                     unsafeWrite v i x
-                     return v
-  | otherwise    = do
-                     v' <- enlarge v
-                     INTERNAL_CHECK(checkIndex) "unsafeAppend1" i (length v')
-                       $ unsafeWrite v' i x
-                     return v'
-
-unsafePrepend1 :: (PrimMonad m, MVector v a)
-        => v (PrimState m) a -> Int -> a -> m (v (PrimState m) a, Int)
-{-# INLINE_INNER unsafePrepend1 #-}
-unsafePrepend1 v i x
-  | i /= 0    = do
-                  let i' = i-1
-                  unsafeWrite v i' x
-                  return (v, i')
-  | otherwise = do
-                  (v', j) <- enlargeFront v
-                  let i' = j-1
-                  INTERNAL_CHECK(checkIndex) "unsafePrepend1" i' (length v')
-                    $ unsafeWrite v' i' x
-                  return (v', i')
-
-mstream :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Stream m a
-{-# INLINE mstream #-}
-mstream v = v `seq` n `seq` (Stream.unfoldrM get 0)
-  where
-    n = length v
-
-    {-# INLINE_INNER get #-}
-    get i | i < n     = do x <- unsafeRead v i
-                           return $ Just (x, i+1)
-          | otherwise = return $ Nothing
-
-fill :: (PrimMonad m, MVector v a)
-     => v (PrimState m) a -> Stream m a -> m (v (PrimState m) a)
-{-# INLINE fill #-}
-fill v s = v `seq` do
-                     n' <- Stream.foldM put 0 s
-                     return $ unsafeSlice 0 n' v
-  where
-    {-# INLINE_INNER put #-}
-    put i x = do
-                INTERNAL_CHECK(checkIndex) "fill" i (length v)
-                  $ unsafeWrite v i x
-                return (i+1)
-
-transform
-  :: (PrimMonad m, MVector v a)
-  => (Stream m a -> Stream m a) -> v (PrimState m) a -> m (v (PrimState m) a)
-{-# INLINE_FUSED transform #-}
-transform f v = fill v (f (mstream v))
-
-mstreamR :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Stream m a
-{-# INLINE mstreamR #-}
-mstreamR v = v `seq` n `seq` (Stream.unfoldrM get n)
-  where
-    n = length v
-
-    {-# INLINE_INNER get #-}
-    get i | j >= 0    = do x <- unsafeRead v j
-                           return $ Just (x,j)
-          | otherwise = return Nothing
-      where
-        j = i-1
-
-fillR :: (PrimMonad m, MVector v a)
-      => v (PrimState m) a -> Stream m a -> m (v (PrimState m) a)
-{-# INLINE fillR #-}
-fillR v s = v `seq` do
-                      i <- Stream.foldM put n s
-                      return $ unsafeSlice i (n-i) v
-  where
-    n = length v
-
-    {-# INLINE_INNER put #-}
-    put i x = do
-                unsafeWrite v j x
-                return j
-      where
-        j = i-1
-
-transformR
-  :: (PrimMonad m, MVector v a)
-  => (Stream m a -> Stream m a) -> v (PrimState m) a -> m (v (PrimState m) a)
-{-# INLINE_FUSED transformR #-}
-transformR f v = fillR v (f (mstreamR v))
-
--- | Create a new mutable vector and fill it with elements from the 'Bundle'.
--- The vector will grow exponentially if the maximum size of the 'Bundle' is
--- unknown.
-unstream :: (PrimMonad m, MVector v a)
-         => Bundle u a -> m (v (PrimState m) a)
--- NOTE: replace INLINE_FUSED by INLINE? (also in unstreamR)
-{-# INLINE_FUSED unstream #-}
-unstream s = munstream (Bundle.lift s)
-
--- | Create a new mutable vector and fill it with elements from the monadic
--- stream. The vector will grow exponentially if the maximum size of the stream
--- is unknown.
-munstream :: (PrimMonad m, MVector v a)
-          => MBundle m u a -> m (v (PrimState m) a)
-{-# INLINE_FUSED munstream #-}
-munstream s = case upperBound (MBundle.size s) of
-               Just n  -> munstreamMax     s n
-               Nothing -> munstreamUnknown s
-
--- FIXME: I can't think of how to prevent GHC from floating out
--- unstreamUnknown. That is bad because SpecConstr then generates two
--- specialisations: one for when it is called from unstream (it doesn't know
--- the shape of the vector) and one for when the vector has grown. To see the
--- problem simply compile this:
---
--- fromList = Data.Vector.Unboxed.unstream . Bundle.fromList
---
--- I'm not sure this still applies (19/04/2010)
-
-munstreamMax :: (PrimMonad m, MVector v a)
-             => MBundle m u a -> Int -> m (v (PrimState m) a)
-{-# INLINE munstreamMax #-}
-munstreamMax s n
-  = do
-      v <- INTERNAL_CHECK(checkLength) "munstreamMax" n
-           $ unsafeNew n
-      let put i x = do
-                       INTERNAL_CHECK(checkIndex) "munstreamMax" i n
-                         $ unsafeWrite v i x
-                       return (i+1)
-      n' <- MBundle.foldM' put 0 s
-      return $ INTERNAL_CHECK(checkSlice) "munstreamMax" 0 n' n
-             $ unsafeSlice 0 n' v
-
-munstreamUnknown :: (PrimMonad m, MVector v a)
-                 => MBundle m u a -> m (v (PrimState m) a)
-{-# INLINE munstreamUnknown #-}
-munstreamUnknown s
-  = do
-      v <- unsafeNew 0
-      (v', n) <- MBundle.foldM put (v, 0) s
-      return $ INTERNAL_CHECK(checkSlice) "munstreamUnknown" 0 n (length v')
-             $ unsafeSlice 0 n v'
-  where
-    {-# INLINE_INNER put #-}
-    put (v,i) x = do
-                    v' <- unsafeAppend1 v i x
-                    return (v',i+1)
-
-
-
-
-
-
-
--- | Create a new mutable vector and fill it with elements from the 'Bundle'.
--- The vector will grow exponentially if the maximum size of the 'Bundle' is
--- unknown.
-vunstream :: (PrimMonad m, V.Vector v a)
-         => Bundle v a -> m (V.Mutable v (PrimState m) a)
--- NOTE: replace INLINE_FUSED by INLINE? (also in unstreamR)
-{-# INLINE_FUSED vunstream #-}
-vunstream s = vmunstream (Bundle.lift s)
-
--- | Create a new mutable vector and fill it with elements from the monadic
--- stream. The vector will grow exponentially if the maximum size of the stream
--- is unknown.
-vmunstream :: (PrimMonad m, V.Vector v a)
-           => MBundle m v a -> m (V.Mutable v (PrimState m) a)
-{-# INLINE_FUSED vmunstream #-}
-vmunstream s = case upperBound (MBundle.size s) of
-               Just n  -> vmunstreamMax     s n
-               Nothing -> vmunstreamUnknown s
-
--- FIXME: I can't think of how to prevent GHC from floating out
--- unstreamUnknown. That is bad because SpecConstr then generates two
--- specialisations: one for when it is called from unstream (it doesn't know
--- the shape of the vector) and one for when the vector has grown. To see the
--- problem simply compile this:
---
--- fromList = Data.Vector.Unboxed.unstream . Bundle.fromList
---
--- I'm not sure this still applies (19/04/2010)
-
-vmunstreamMax :: (PrimMonad m, V.Vector v a)
-              => MBundle m v a -> Int -> m (V.Mutable v (PrimState m) a)
-{-# INLINE vmunstreamMax #-}
-vmunstreamMax s n
-  = do
-      v <- INTERNAL_CHECK(checkLength) "munstreamMax" n
-           $ unsafeNew n
-      let {-# INLINE_INNER copyChunk #-}
-          copyChunk i (Chunk m f) =
-            INTERNAL_CHECK(checkSlice) "munstreamMax.copyChunk" i m (length v) $ do
-              f (basicUnsafeSlice i m v)
-              return (i+m)
-
-      n' <- Stream.foldlM' copyChunk 0 (MBundle.chunks s)
-      return $ INTERNAL_CHECK(checkSlice) "munstreamMax" 0 n' n
-             $ unsafeSlice 0 n' v
-
-vmunstreamUnknown :: (PrimMonad m, V.Vector v a)
-                 => MBundle m v a -> m (V.Mutable v (PrimState m) a)
-{-# INLINE vmunstreamUnknown #-}
-vmunstreamUnknown s
-  = do
-      v <- unsafeNew 0
-      (v', n) <- Stream.foldlM copyChunk (v,0) (MBundle.chunks s)
-      return $ INTERNAL_CHECK(checkSlice) "munstreamUnknown" 0 n (length v')
-             $ unsafeSlice 0 n v'
-  where
-    {-# INLINE_INNER copyChunk #-}
-    copyChunk (v,i) (Chunk n f)
-      = do
-          let j = i+n
-          v' <- if basicLength v < j
-                  then unsafeGrow v (delay_inline max (enlarge_delta v) (j - basicLength v))
-                  else return v
-          INTERNAL_CHECK(checkSlice) "munstreamUnknown.copyChunk" i n (length v')
-            $ f (basicUnsafeSlice i n v')
-          return (v',j)
-
-
-
-
--- | Create a new mutable vector and fill it with elements from the 'Bundle'
--- from right to left. The vector will grow exponentially if the maximum size
--- of the 'Bundle' is unknown.
-unstreamR :: (PrimMonad m, MVector v a)
-          => Bundle u a -> m (v (PrimState m) a)
--- NOTE: replace INLINE_FUSED by INLINE? (also in unstream)
-{-# INLINE_FUSED unstreamR #-}
-unstreamR s = munstreamR (Bundle.lift s)
-
--- | Create a new mutable vector and fill it with elements from the monadic
--- stream from right to left. The vector will grow exponentially if the maximum
--- size of the stream is unknown.
-munstreamR :: (PrimMonad m, MVector v a)
-           => MBundle m u a -> m (v (PrimState m) a)
-{-# INLINE_FUSED munstreamR #-}
-munstreamR s = case upperBound (MBundle.size s) of
-               Just n  -> munstreamRMax     s n
-               Nothing -> munstreamRUnknown s
-
-munstreamRMax :: (PrimMonad m, MVector v a)
-              => MBundle m u a -> Int -> m (v (PrimState m) a)
-{-# INLINE munstreamRMax #-}
-munstreamRMax s n
-  = do
-      v <- INTERNAL_CHECK(checkLength) "munstreamRMax" n
-           $ unsafeNew n
-      let put i x = do
-                      let i' = i-1
-                      INTERNAL_CHECK(checkIndex) "munstreamRMax" i' n
-                        $ unsafeWrite v i' x
-                      return i'
-      i <- MBundle.foldM' put n s
-      return $ INTERNAL_CHECK(checkSlice) "munstreamRMax" i (n-i) n
-             $ unsafeSlice i (n-i) v
-
-munstreamRUnknown :: (PrimMonad m, MVector v a)
-                  => MBundle m u a -> m (v (PrimState m) a)
-{-# INLINE munstreamRUnknown #-}
-munstreamRUnknown s
-  = do
-      v <- unsafeNew 0
-      (v', i) <- MBundle.foldM put (v, 0) s
-      let n = length v'
-      return $ INTERNAL_CHECK(checkSlice) "unstreamRUnknown" i (n-i) n
-             $ unsafeSlice i (n-i) v'
-  where
-    {-# INLINE_INNER put #-}
-    put (v,i) x = unsafePrepend1 v i x
-
--- Length
--- ------
-
--- | Length of the mutable vector.
-length :: MVector v a => v s a -> Int
-{-# INLINE length #-}
-length = basicLength
-
--- | Check whether the vector is empty
-null :: MVector v a => v s a -> Bool
-{-# INLINE null #-}
-null v = length v == 0
-
--- Extracting subvectors
--- ---------------------
-
--- | Yield a part of the mutable vector without copying it.
-slice :: MVector v a => Int -> Int -> v s a -> v s a
-{-# INLINE slice #-}
-slice i n v = BOUNDS_CHECK(checkSlice) "slice" i n (length v)
-            $ unsafeSlice i n v
-
-take :: MVector v a => Int -> v s a -> v s a
-{-# INLINE take #-}
-take n v = unsafeSlice 0 (min (max n 0) (length v)) v
-
-drop :: MVector v a => Int -> v s a -> v s a
-{-# INLINE drop #-}
-drop n v = unsafeSlice (min m n') (max 0 (m - n')) v
-  where
-    n' = max n 0
-    m  = length v
-
-{-# INLINE splitAt #-}
-splitAt :: MVector v a => Int -> v s a -> (v s a, v s a)
-splitAt n v = ( unsafeSlice 0 m v
-              , unsafeSlice m (max 0 (len - n')) v
-              )
-    where
-      m   = min n' len
-      n'  = max n 0
-      len = length v
-
-init :: MVector v a => v s a -> v s a
-{-# INLINE init #-}
-init v = slice 0 (length v - 1) v
-
-tail :: MVector v a => v s a -> v s a
-{-# INLINE tail #-}
-tail v = slice 1 (length v - 1) v
-
--- | Yield a part of the mutable vector without copying it. No bounds checks
--- are performed.
-unsafeSlice :: MVector v a => Int  -- ^ starting index
-                           -> Int  -- ^ length of the slice
-                           -> v s a
-                           -> v s a
-{-# INLINE unsafeSlice #-}
-unsafeSlice i n v = UNSAFE_CHECK(checkSlice) "unsafeSlice" i n (length v)
-                  $ basicUnsafeSlice i n v
-
-unsafeInit :: MVector v a => v s a -> v s a
-{-# INLINE unsafeInit #-}
-unsafeInit v = unsafeSlice 0 (length v - 1) v
-
-unsafeTail :: MVector v a => v s a -> v s a
-{-# INLINE unsafeTail #-}
-unsafeTail v = unsafeSlice 1 (length v - 1) v
-
-unsafeTake :: MVector v a => Int -> v s a -> v s a
-{-# INLINE unsafeTake #-}
-unsafeTake n v = unsafeSlice 0 n v
-
-unsafeDrop :: MVector v a => Int -> v s a -> v s a
-{-# INLINE unsafeDrop #-}
-unsafeDrop n v = unsafeSlice n (length v - n) v
-
--- Overlapping
--- -----------
-
--- | Check whether two vectors overlap.
-overlaps :: MVector v a => v s a -> v s a -> Bool
-{-# INLINE overlaps #-}
-overlaps = basicOverlaps
-
--- Initialisation
--- --------------
-
--- | Create a mutable vector of the given length.
-new :: (PrimMonad m, MVector v a) => Int -> m (v (PrimState m) a)
-{-# INLINE new #-}
-new n = BOUNDS_CHECK(checkLength) "new" n
-      $ unsafeNew n >>= \v -> basicInitialize v >> return v
-
--- | Create a mutable vector of the given length. The memory is not initialized.
-unsafeNew :: (PrimMonad m, MVector v a) => Int -> m (v (PrimState m) a)
-{-# INLINE unsafeNew #-}
-unsafeNew n = UNSAFE_CHECK(checkLength) "unsafeNew" n
-            $ basicUnsafeNew n
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with an initial value.
-replicate :: (PrimMonad m, MVector v a) => Int -> a -> m (v (PrimState m) a)
-{-# INLINE replicate #-}
-replicate n x = basicUnsafeReplicate (delay_inline max 0 n) x
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with values produced by repeatedly executing the monadic action.
-replicateM :: (PrimMonad m, MVector v a) => Int -> m a -> m (v (PrimState m) a)
-{-# INLINE replicateM #-}
-replicateM n m = munstream (MBundle.replicateM n m)
-
--- | Create a copy of a mutable vector.
-clone :: (PrimMonad m, MVector v a) => v (PrimState m) a -> m (v (PrimState m) a)
-{-# INLINE clone #-}
-clone v = do
-            v' <- unsafeNew (length v)
-            unsafeCopy v' v
-            return v'
-
--- Growing
--- -------
-
--- | Grow a vector by the given number of elements. The number must be
--- positive.
-grow :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> Int -> m (v (PrimState m) a)
-{-# INLINE grow #-}
-grow v by = BOUNDS_CHECK(checkLength) "grow" by
-          $ do vnew <- unsafeGrow v by
-               basicInitialize $ basicUnsafeSlice (length v) by vnew
-               return vnew
-
-growFront :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> Int -> m (v (PrimState m) a)
-{-# INLINE growFront #-}
-growFront v by = BOUNDS_CHECK(checkLength) "growFront" by
-               $ do vnew <- unsafeGrowFront v by
-                    basicInitialize $ basicUnsafeSlice 0 by vnew
-                    return vnew
-
-enlarge_delta :: MVector v a => v s a -> Int
-enlarge_delta v = max (length v) 1
-
--- | Grow a vector logarithmically
-enlarge :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> m (v (PrimState m) a)
-{-# INLINE enlarge #-}
-enlarge v = do vnew <- unsafeGrow v by
-               basicInitialize $ basicUnsafeSlice (length v) by vnew
-               return vnew
-  where
-    by = enlarge_delta v
-
-enlargeFront :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> m (v (PrimState m) a, Int)
-{-# INLINE enlargeFront #-}
-enlargeFront v = do
-                   v' <- unsafeGrowFront v by
-                   basicInitialize $ basicUnsafeSlice 0 by v'
-                   return (v', by)
-  where
-    by = enlarge_delta v
-
--- | Grow a vector by the given number of elements. The number must be
--- positive but this is not checked.
-unsafeGrow :: (PrimMonad m, MVector v a)
-                        => v (PrimState m) a -> Int -> m (v (PrimState m) a)
-{-# INLINE unsafeGrow #-}
-unsafeGrow v n = UNSAFE_CHECK(checkLength) "unsafeGrow" n
-               $ basicUnsafeGrow v n
-
-unsafeGrowFront :: (PrimMonad m, MVector v a)
-                        => v (PrimState m) a -> Int -> m (v (PrimState m) a)
-{-# INLINE unsafeGrowFront #-}
-unsafeGrowFront v by = UNSAFE_CHECK(checkLength) "unsafeGrowFront" by
-                     $ do
-                         let n = length v
-                         v' <- basicUnsafeNew (by+n)
-                         basicUnsafeCopy (basicUnsafeSlice by n v') v
-                         return v'
-
--- Restricting memory usage
--- ------------------------
-
--- | Reset all elements of the vector to some undefined value, clearing all
--- references to external objects. This is usually a noop for unboxed vectors.
-clear :: (PrimMonad m, MVector v a) => v (PrimState m) a -> m ()
-{-# INLINE clear #-}
-clear = basicClear
-
--- Accessing individual elements
--- -----------------------------
-
--- | Yield the element at the given position.
-read :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Int -> m a
-{-# INLINE read #-}
-read v i = BOUNDS_CHECK(checkIndex) "read" i (length v)
-         $ unsafeRead v i
-
--- | Replace the element at the given position.
-write :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Int -> a -> m ()
-{-# INLINE write #-}
-write v i x = BOUNDS_CHECK(checkIndex) "write" i (length v)
-            $ unsafeWrite v i x
-
--- | Modify the element at the given position.
-modify :: (PrimMonad m, MVector v a) => v (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE modify #-}
-modify v f i = BOUNDS_CHECK(checkIndex) "modify" i (length v)
-             $ unsafeModify v f i
-
--- | Swap the elements at the given positions.
-swap :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE swap #-}
-swap v i j = BOUNDS_CHECK(checkIndex) "swap" i (length v)
-           $ BOUNDS_CHECK(checkIndex) "swap" j (length v)
-           $ unsafeSwap v i j
-
--- | Replace the element at the give position and return the old element.
-exchange :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Int -> a -> m a
-{-# INLINE exchange #-}
-exchange v i x = BOUNDS_CHECK(checkIndex) "exchange" i (length v)
-               $ unsafeExchange v i x
-
--- | Yield the element at the given position. No bounds checks are performed.
-unsafeRead :: (PrimMonad m, MVector v a) => v (PrimState m) a -> Int -> m a
-{-# INLINE unsafeRead #-}
-unsafeRead v i = UNSAFE_CHECK(checkIndex) "unsafeRead" i (length v)
-               $ basicUnsafeRead v i
-
--- | Replace the element at the given position. No bounds checks are performed.
-unsafeWrite :: (PrimMonad m, MVector v a)
-                                => v (PrimState m) a -> Int -> a -> m ()
-{-# INLINE unsafeWrite #-}
-unsafeWrite v i x = UNSAFE_CHECK(checkIndex) "unsafeWrite" i (length v)
-                  $ basicUnsafeWrite v i x
-
--- | Modify the element at the given position. No bounds checks are performed.
-unsafeModify :: (PrimMonad m, MVector v a) => v (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE unsafeModify #-}
-unsafeModify v f i = UNSAFE_CHECK(checkIndex) "unsafeModify" i (length v)
-                   $ basicUnsafeRead v i >>= \x ->
-                     basicUnsafeWrite v i (f x)
-
--- | Swap the elements at the given positions. No bounds checks are performed.
-unsafeSwap :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE unsafeSwap #-}
-unsafeSwap v i j = UNSAFE_CHECK(checkIndex) "unsafeSwap" i (length v)
-                 $ UNSAFE_CHECK(checkIndex) "unsafeSwap" j (length v)
-                 $ do
-                     x <- unsafeRead v i
-                     y <- unsafeRead v j
-                     unsafeWrite v i y
-                     unsafeWrite v j x
-
--- | Replace the element at the give position and return the old element. No
--- bounds checks are performed.
-unsafeExchange :: (PrimMonad m, MVector v a)
-                                => v (PrimState m) a -> Int -> a -> m a
-{-# INLINE unsafeExchange #-}
-unsafeExchange v i x = UNSAFE_CHECK(checkIndex) "unsafeExchange" i (length v)
-                     $ do
-                         y <- unsafeRead v i
-                         unsafeWrite v i x
-                         return y
-
--- Filling and copying
--- -------------------
-
--- | Set all elements of the vector to the given value.
-set :: (PrimMonad m, MVector v a) => v (PrimState m) a -> a -> m ()
-{-# INLINE set #-}
-set = basicSet
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap.
-copy :: (PrimMonad m, MVector v a) => v (PrimState m) a   -- ^ target
-                                   -> v (PrimState m) a   -- ^ source
-                                   -> m ()
-{-# INLINE copy #-}
-copy dst src = BOUNDS_CHECK(check) "copy" "overlapping vectors"
-                                          (not (dst `overlaps` src))
-             $ BOUNDS_CHECK(check) "copy" "length mismatch"
-                                          (length dst == length src)
-             $ unsafeCopy dst src
-
--- | Move the contents of a vector. The two vectors must have the same
--- length.
---
--- If the vectors do not overlap, then this is equivalent to 'copy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-move :: (PrimMonad m, MVector v a)
-                => v (PrimState m) a -> v (PrimState m) a -> m ()
-{-# INLINE move #-}
-move dst src = BOUNDS_CHECK(check) "move" "length mismatch"
-                                          (length dst == length src)
-             $ unsafeMove dst src
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap. This is not checked.
-unsafeCopy :: (PrimMonad m, MVector v a) => v (PrimState m) a   -- ^ target
-                                         -> v (PrimState m) a   -- ^ source
-                                         -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy dst src = UNSAFE_CHECK(check) "unsafeCopy" "length mismatch"
-                                         (length dst == length src)
-                   $ UNSAFE_CHECK(check) "unsafeCopy" "overlapping vectors"
-                                         (not (dst `overlaps` src))
-                   $ (dst `seq` src `seq` basicUnsafeCopy dst src)
-
--- | Move the contents of a vector. The two vectors must have the same
--- length, but this is not checked.
---
--- If the vectors do not overlap, then this is equivalent to 'unsafeCopy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-unsafeMove :: (PrimMonad m, MVector v a) => v (PrimState m) a   -- ^ target
-                                         -> v (PrimState m) a   -- ^ source
-                                         -> m ()
-{-# INLINE unsafeMove #-}
-unsafeMove dst src = UNSAFE_CHECK(check) "unsafeMove" "length mismatch"
-                                         (length dst == length src)
-                   $ (dst `seq` src `seq` basicUnsafeMove dst src)
-
--- Permutations
--- ------------
-
-accum :: (PrimMonad m, MVector v a)
-      => (a -> b -> a) -> v (PrimState m) a -> Bundle u (Int, b) -> m ()
-{-# INLINE accum #-}
-accum f !v s = Bundle.mapM_ upd s
-  where
-    {-# INLINE_INNER upd #-}
-    upd (i,b) = do
-                  a <- BOUNDS_CHECK(checkIndex) "accum" i n
-                     $ unsafeRead v i
-                  unsafeWrite v i (f a b)
-
-    !n = length v
-
-update :: (PrimMonad m, MVector v a)
-                        => v (PrimState m) a -> Bundle u (Int, a) -> m ()
-{-# INLINE update #-}
-update !v s = Bundle.mapM_ upd s
-  where
-    {-# INLINE_INNER upd #-}
-    upd (i,b) = BOUNDS_CHECK(checkIndex) "update" i n
-              $ unsafeWrite v i b
-
-    !n = length v
-
-unsafeAccum :: (PrimMonad m, MVector v a)
-            => (a -> b -> a) -> v (PrimState m) a -> Bundle u (Int, b) -> m ()
-{-# INLINE unsafeAccum #-}
-unsafeAccum f !v s = Bundle.mapM_ upd s
-  where
-    {-# INLINE_INNER upd #-}
-    upd (i,b) = do
-                  a <- UNSAFE_CHECK(checkIndex) "accum" i n
-                     $ unsafeRead v i
-                  unsafeWrite v i (f a b)
-
-    !n = length v
-
-unsafeUpdate :: (PrimMonad m, MVector v a)
-                        => v (PrimState m) a -> Bundle u (Int, a) -> m ()
-{-# INLINE unsafeUpdate #-}
-unsafeUpdate !v s = Bundle.mapM_ upd s
-  where
-    {-# INLINE_INNER upd #-}
-    upd (i,b) = UNSAFE_CHECK(checkIndex) "accum" i n
-                  $ unsafeWrite v i b
-
-    !n = length v
-
-reverse :: (PrimMonad m, MVector v a) => v (PrimState m) a -> m ()
-{-# INLINE reverse #-}
-reverse !v = reverse_loop 0 (length v - 1)
-  where
-    reverse_loop i j | i < j = do
-                                 unsafeSwap v i j
-                                 reverse_loop (i + 1) (j - 1)
-    reverse_loop _ _ = return ()
-
-unstablePartition :: forall m v a. (PrimMonad m, MVector v a)
-                  => (a -> Bool) -> v (PrimState m) a -> m Int
-{-# INLINE unstablePartition #-}
-unstablePartition f !v = from_left 0 (length v)
-  where
-    -- NOTE: GHC 6.10.4 panics without the signatures on from_left and
-    -- from_right
-    from_left :: Int -> Int -> m Int
-    from_left i j
-      | i == j    = return i
-      | otherwise = do
-                      x <- unsafeRead v i
-                      if f x
-                        then from_left (i+1) j
-                        else from_right i (j-1)
-
-    from_right :: Int -> Int -> m Int
-    from_right i j
-      | i == j    = return i
-      | otherwise = do
-                      x <- unsafeRead v j
-                      if f x
-                        then do
-                               y <- unsafeRead v i
-                               unsafeWrite v i x
-                               unsafeWrite v j y
-                               from_left (i+1) j
-                        else from_right i (j-1)
-
-unstablePartitionBundle :: (PrimMonad m, MVector v a)
-        => (a -> Bool) -> Bundle u a -> m (v (PrimState m) a, v (PrimState m) a)
-{-# INLINE unstablePartitionBundle #-}
-unstablePartitionBundle f s
-  = case upperBound (Bundle.size s) of
-      Just n  -> unstablePartitionMax f s n
-      Nothing -> partitionUnknown f s
-
-unstablePartitionMax :: (PrimMonad m, MVector v a)
-        => (a -> Bool) -> Bundle u a -> Int
-        -> m (v (PrimState m) a, v (PrimState m) a)
-{-# INLINE unstablePartitionMax #-}
-unstablePartitionMax f s n
-  = do
-      v <- INTERNAL_CHECK(checkLength) "unstablePartitionMax" n
-           $ unsafeNew n
-      let {-# INLINE_INNER put #-}
-          put (i, j) x
-            | f x       = do
-                            unsafeWrite v i x
-                            return (i+1, j)
-            | otherwise = do
-                            unsafeWrite v (j-1) x
-                            return (i, j-1)
-
-      (i,j) <- Bundle.foldM' put (0, n) s
-      return (unsafeSlice 0 i v, unsafeSlice j (n-j) v)
-
-partitionBundle :: (PrimMonad m, MVector v a)
-        => (a -> Bool) -> Bundle u a -> m (v (PrimState m) a, v (PrimState m) a)
-{-# INLINE partitionBundle #-}
-partitionBundle f s
-  = case upperBound (Bundle.size s) of
-      Just n  -> partitionMax f s n
-      Nothing -> partitionUnknown f s
-
-partitionMax :: (PrimMonad m, MVector v a)
-  => (a -> Bool) -> Bundle u a -> Int -> m (v (PrimState m) a, v (PrimState m) a)
-{-# INLINE partitionMax #-}
-partitionMax f s n
-  = do
-      v <- INTERNAL_CHECK(checkLength) "unstablePartitionMax" n
-         $ unsafeNew n
-
-      let {-# INLINE_INNER put #-}
-          put (i,j) x
-            | f x       = do
-                            unsafeWrite v i x
-                            return (i+1,j)
-
-            | otherwise = let j' = j-1 in
-                          do
-                            unsafeWrite v j' x
-                            return (i,j')
-
-      (i,j) <- Bundle.foldM' put (0,n) s
-      INTERNAL_CHECK(check) "partitionMax" "invalid indices" (i <= j)
-        $ return ()
-      let l = unsafeSlice 0 i v
-          r = unsafeSlice j (n-j) v
-      reverse r
-      return (l,r)
-
-partitionUnknown :: (PrimMonad m, MVector v a)
-        => (a -> Bool) -> Bundle u a -> m (v (PrimState m) a, v (PrimState m) a)
-{-# INLINE partitionUnknown #-}
-partitionUnknown f s
-  = do
-      v1 <- unsafeNew 0
-      v2 <- unsafeNew 0
-      (v1', n1, v2', n2) <- Bundle.foldM' put (v1, 0, v2, 0) s
-      INTERNAL_CHECK(checkSlice) "partitionUnknown" 0 n1 (length v1')
-        $ INTERNAL_CHECK(checkSlice) "partitionUnknown" 0 n2 (length v2')
-        $ return (unsafeSlice 0 n1 v1', unsafeSlice 0 n2 v2')
-  where
-    -- NOTE: The case distinction has to be on the outside because
-    -- GHC creates a join point for the unsafeWrite even when everything
-    -- is inlined. This is bad because with the join point, v isn't getting
-    -- unboxed.
-    {-# INLINE_INNER put #-}
-    put (v1, i1, v2, i2) x
-      | f x       = do
-                      v1' <- unsafeAppend1 v1 i1 x
-                      return (v1', i1+1, v2, i2)
-      | otherwise = do
-                      v2' <- unsafeAppend1 v2 i2 x
-                      return (v1, i1, v2', i2+1)
-
-{-
-http://en.wikipedia.org/wiki/Permutation#Algorithms_to_generate_permutations
-
-The following algorithm generates the next permutation lexicographically after
-a given permutation. It changes the given permutation in-place.
-
-1. Find the largest index k such that a[k] < a[k + 1]. If no such index exists,
-   the permutation is the last permutation.
-2. Find the largest index l greater than k such that a[k] < a[l].
-3. Swap the value of a[k] with that of a[l].
-4. Reverse the sequence from a[k + 1] up to and including the final element a[n]
--}
-
--- | Compute the next (lexicographically) permutation of given vector in-place.
---   Returns False when input is the last permtuation
-nextPermutation :: (PrimMonad m,Ord e,MVector v e) => v (PrimState m) e -> m Bool
-nextPermutation v
-    | dim < 2 = return False
-    | otherwise = do
-        val <- unsafeRead v 0
-        (k,l) <- loop val (-1) 0 val 1
-        if k < 0
-         then return False
-         else unsafeSwap v k l >>
-              reverse (unsafeSlice (k+1) (dim-k-1) v) >>
-              return True
-    where loop !kval !k !l !prev !i
-              | i == dim = return (k,l)
-              | otherwise  = do
-                  cur <- unsafeRead v i
-                  -- TODO: make tuple unboxed
-                  let (kval',k') = if prev < cur then (prev,i-1) else (kval,k)
-                      l' = if kval' < cur then i else l
-                  loop kval' k' l' cur (i+1)
-          dim = length v
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable/Base.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable/Base.hs
deleted file mode 100644
index ce931eec9b41..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/Mutable/Base.hs
+++ /dev/null
@@ -1,145 +0,0 @@
-{-# LANGUAGE CPP, MultiParamTypeClasses, BangPatterns, TypeFamilies #-}
--- |
--- Module      : Data.Vector.Generic.Mutable.Base
--- Copyright   : (c) Roman Leshchinskiy 2008-2011
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Class of mutable vectors
---
-
-module Data.Vector.Generic.Mutable.Base (
-  MVector(..)
-) where
-
-import Control.Monad.Primitive ( PrimMonad, PrimState )
-
--- Data.Vector.Internal.Check is unused
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- | Class of mutable vectors parametrised with a primitive state token.
---
-class MVector v a where
-  -- | Length of the mutable vector. This method should not be
-  -- called directly, use 'length' instead.
-  basicLength       :: v s a -> Int
-
-  -- | Yield a part of the mutable vector without copying it. This method
-  -- should not be called directly, use 'unsafeSlice' instead.
-  basicUnsafeSlice :: Int  -- ^ starting index
-                   -> Int  -- ^ length of the slice
-                   -> v s a
-                   -> v s a
-
-  -- | Check whether two vectors overlap. This method should not be
-  -- called directly, use 'overlaps' instead.
-  basicOverlaps    :: v s a -> v s a -> Bool
-
-  -- | Create a mutable vector of the given length. This method should not be
-  -- called directly, use 'unsafeNew' instead.
-  basicUnsafeNew   :: PrimMonad m => Int -> m (v (PrimState m) a)
-
-  -- | Initialize a vector to a standard value. This is intended to be called as
-  -- part of the safe new operation (and similar operations), to properly blank
-  -- the newly allocated memory if necessary.
-  --
-  -- Vectors that are necessarily initialized as part of creation may implement
-  -- this as a no-op.
-  basicInitialize :: PrimMonad m => v (PrimState m) a -> m ()
-
-  -- | Create a mutable vector of the given length and fill it with an
-  -- initial value. This method should not be called directly, use
-  -- 'replicate' instead.
-  basicUnsafeReplicate :: PrimMonad m => Int -> a -> m (v (PrimState m) a)
-
-  -- | Yield the element at the given position. This method should not be
-  -- called directly, use 'unsafeRead' instead.
-  basicUnsafeRead  :: PrimMonad m => v (PrimState m) a -> Int -> m a
-
-  -- | Replace the element at the given position. This method should not be
-  -- called directly, use 'unsafeWrite' instead.
-  basicUnsafeWrite :: PrimMonad m => v (PrimState m) a -> Int -> a -> m ()
-
-  -- | Reset all elements of the vector to some undefined value, clearing all
-  -- references to external objects. This is usually a noop for unboxed
-  -- vectors. This method should not be called directly, use 'clear' instead.
-  basicClear       :: PrimMonad m => v (PrimState m) a -> m ()
-
-  -- | Set all elements of the vector to the given value. This method should
-  -- not be called directly, use 'set' instead.
-  basicSet         :: PrimMonad m => v (PrimState m) a -> a -> m ()
-
-  -- | Copy a vector. The two vectors may not overlap. This method should not
-  -- be called directly, use 'unsafeCopy' instead.
-  basicUnsafeCopy  :: PrimMonad m => v (PrimState m) a   -- ^ target
-                                  -> v (PrimState m) a   -- ^ source
-                                  -> m ()
-
-  -- | Move the contents of a vector. The two vectors may overlap. This method
-  -- should not be called directly, use 'unsafeMove' instead.
-  basicUnsafeMove  :: PrimMonad m => v (PrimState m) a   -- ^ target
-                                  -> v (PrimState m) a   -- ^ source
-                                  -> m ()
-
-  -- | Grow a vector by the given number of elements. This method should not be
-  -- called directly, use 'unsafeGrow' instead.
-  basicUnsafeGrow  :: PrimMonad m => v (PrimState m) a -> Int
-                                                       -> m (v (PrimState m) a)
-
-  {-# INLINE basicUnsafeReplicate #-}
-  basicUnsafeReplicate n x
-    = do
-        v <- basicUnsafeNew n
-        basicSet v x
-        return v
-
-  {-# INLINE basicClear #-}
-  basicClear _ = return ()
-
-  {-# INLINE basicSet #-}
-  basicSet !v x
-    | n == 0    = return ()
-    | otherwise = do
-                    basicUnsafeWrite v 0 x
-                    do_set 1
-    where
-      !n = basicLength v
-
-      do_set i | 2*i < n = do basicUnsafeCopy (basicUnsafeSlice i i v)
-                                              (basicUnsafeSlice 0 i v)
-                              do_set (2*i)
-               | otherwise = basicUnsafeCopy (basicUnsafeSlice i (n-i) v)
-                                             (basicUnsafeSlice 0 (n-i) v)
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy !dst !src = do_copy 0
-    where
-      !n = basicLength src
-
-      do_copy i | i < n = do
-                            x <- basicUnsafeRead src i
-                            basicUnsafeWrite dst i x
-                            do_copy (i+1)
-                | otherwise = return ()
-
-  {-# INLINE basicUnsafeMove #-}
-  basicUnsafeMove !dst !src
-    | basicOverlaps dst src = do
-        srcCopy <- basicUnsafeNew (basicLength src)
-        basicUnsafeCopy srcCopy src
-        basicUnsafeCopy dst srcCopy
-    | otherwise = basicUnsafeCopy dst src
-
-  {-# INLINE basicUnsafeGrow #-}
-  basicUnsafeGrow v by
-    = do
-        v' <- basicUnsafeNew (n+by)
-        basicUnsafeCopy (basicUnsafeSlice 0 n v') v
-        return v'
-    where
-      n = basicLength v
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/New.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/New.hs
deleted file mode 100644
index e94ce19e1669..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Generic/New.hs
+++ /dev/null
@@ -1,178 +0,0 @@
-{-# LANGUAGE CPP, Rank2Types, FlexibleContexts, MultiParamTypeClasses #-}
-
--- |
--- Module      : Data.Vector.Generic.New
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Purely functional interface to initialisation of mutable vectors
---
-
-module Data.Vector.Generic.New (
-  New(..), create, run, runPrim, apply, modify, modifyWithBundle,
-  unstream, transform, unstreamR, transformR,
-  slice, init, tail, take, drop,
-  unsafeSlice, unsafeInit, unsafeTail
-) where
-
-import qualified Data.Vector.Generic.Mutable as MVector
-
-import           Data.Vector.Generic.Base ( Vector, Mutable )
-
-import           Data.Vector.Fusion.Bundle ( Bundle )
-import qualified Data.Vector.Fusion.Bundle as Bundle
-import           Data.Vector.Fusion.Stream.Monadic ( Stream )
-import           Data.Vector.Fusion.Bundle.Size
-
-import Control.Monad.Primitive
-import Control.Monad.ST ( ST )
-import Control.Monad  ( liftM )
-import Prelude hiding ( init, tail, take, drop, reverse, map, filter )
-
--- Data.Vector.Internal.Check is unused
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
-data New v a = New (forall s. ST s (Mutable v s a))
-
-create :: (forall s. ST s (Mutable v s a)) -> New v a
-{-# INLINE create #-}
-create p = New p
-
-run :: New v a -> ST s (Mutable v s a)
-{-# INLINE run #-}
-run (New p) = p
-
-runPrim :: PrimMonad m => New v a -> m (Mutable v (PrimState m) a)
-{-# INLINE runPrim #-}
-runPrim (New p) = primToPrim p
-
-apply :: (forall s. Mutable v s a -> Mutable v s a) -> New v a -> New v a
-{-# INLINE apply #-}
-apply f (New p) = New (liftM f p)
-
-modify :: (forall s. Mutable v s a -> ST s ()) -> New v a -> New v a
-{-# INLINE modify #-}
-modify f (New p) = New (do { v <- p; f v; return v })
-
-modifyWithBundle :: (forall s. Mutable v s a -> Bundle u b -> ST s ())
-                 -> New v a -> Bundle u b -> New v a
-{-# INLINE_FUSED modifyWithBundle #-}
-modifyWithBundle f (New p) s = s `seq` New (do { v <- p; f v s; return v })
-
-unstream :: Vector v a => Bundle v a -> New v a
-{-# INLINE_FUSED unstream #-}
-unstream s = s `seq` New (MVector.vunstream s)
-
-transform
-  :: Vector v a => (forall m. Monad m => Stream m a -> Stream m a)
-                -> (Size -> Size) -> New v a -> New v a
-{-# INLINE_FUSED transform #-}
-transform f _ (New p) = New (MVector.transform f =<< p)
-
-{-# RULES
-
-"transform/transform [New]"
-  forall (f1 :: forall m. Monad m => Stream m a -> Stream m a)
-         (f2 :: forall m. Monad m => Stream m a -> Stream m a)
-         g1 g2 p .
-  transform f1 g1 (transform f2 g2 p) = transform (f1 . f2) (g1 . g2) p
-
-"transform/unstream [New]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a)
-         g s.
-  transform f g (unstream s) = unstream (Bundle.inplace f g s)  #-}
-
-
-
-
-unstreamR :: Vector v a => Bundle v a -> New v a
-{-# INLINE_FUSED unstreamR #-}
-unstreamR s = s `seq` New (MVector.unstreamR s)
-
-transformR
-  :: Vector v a => (forall m. Monad m => Stream m a -> Stream m a)
-                -> (Size -> Size) -> New v a -> New v a
-{-# INLINE_FUSED transformR #-}
-transformR f _ (New p) = New (MVector.transformR f =<< p)
-
-{-# RULES
-
-"transformR/transformR [New]"
-  forall (f1 :: forall m. Monad m => Stream m a -> Stream m a)
-         (f2 :: forall m. Monad m => Stream m a -> Stream m a)
-         g1 g2
-         p .
-  transformR f1 g1 (transformR f2 g2 p) = transformR (f1 . f2) (g1 . g2) p
-
-"transformR/unstreamR [New]"
-  forall (f :: forall m. Monad m => Stream m a -> Stream m a)
-         g s.
-  transformR f g (unstreamR s) = unstreamR (Bundle.inplace f g s)  #-}
-
-
-
-slice :: Vector v a => Int -> Int -> New v a -> New v a
-{-# INLINE_FUSED slice #-}
-slice i n m = apply (MVector.slice i n) m
-
-init :: Vector v a => New v a -> New v a
-{-# INLINE_FUSED init #-}
-init m = apply MVector.init m
-
-tail :: Vector v a => New v a -> New v a
-{-# INLINE_FUSED tail #-}
-tail m = apply MVector.tail m
-
-take :: Vector v a => Int -> New v a -> New v a
-{-# INLINE_FUSED take #-}
-take n m = apply (MVector.take n) m
-
-drop :: Vector v a => Int -> New v a -> New v a
-{-# INLINE_FUSED drop #-}
-drop n m = apply (MVector.drop n) m
-
-unsafeSlice :: Vector v a => Int -> Int -> New v a -> New v a
-{-# INLINE_FUSED unsafeSlice #-}
-unsafeSlice i n m = apply (MVector.unsafeSlice i n) m
-
-unsafeInit :: Vector v a => New v a -> New v a
-{-# INLINE_FUSED unsafeInit #-}
-unsafeInit m = apply MVector.unsafeInit m
-
-unsafeTail :: Vector v a => New v a -> New v a
-{-# INLINE_FUSED unsafeTail #-}
-unsafeTail m = apply MVector.unsafeTail m
-
-{-# RULES
-
-"slice/unstream [New]" forall i n s.
-  slice i n (unstream s) = unstream (Bundle.slice i n s)
-
-"init/unstream [New]" forall s.
-  init (unstream s) = unstream (Bundle.init s)
-
-"tail/unstream [New]" forall s.
-  tail (unstream s) = unstream (Bundle.tail s)
-
-"take/unstream [New]" forall n s.
-  take n (unstream s) = unstream (Bundle.take n s)
-
-"drop/unstream [New]" forall n s.
-  drop n (unstream s) = unstream (Bundle.drop n s)
-
-"unsafeSlice/unstream [New]" forall i n s.
-  unsafeSlice i n (unstream s) = unstream (Bundle.slice i n s)
-
-"unsafeInit/unstream [New]" forall s.
-  unsafeInit (unstream s) = unstream (Bundle.init s)
-
-"unsafeTail/unstream [New]" forall s.
-  unsafeTail (unstream s) = unstream (Bundle.tail s)   #-}
-
-
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Internal/Check.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Internal/Check.hs
deleted file mode 100644
index 4a4ef80fe172..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Internal/Check.hs
+++ /dev/null
@@ -1,152 +0,0 @@
-{-# LANGUAGE CPP #-}
-
--- |
--- Module      : Data.Vector.Internal.Check
--- Copyright   : (c) Roman Leshchinskiy 2009
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Bounds checking infrastructure
---
-
-{-# LANGUAGE MagicHash #-}
-
-module Data.Vector.Internal.Check (
-  Checks(..), doChecks,
-
-  error, internalError,
-  check, checkIndex, checkLength, checkSlice
-) where
-
-import GHC.Base( Int(..) )
-import GHC.Prim( Int# )
-import Prelude hiding( error, (&&), (||), not )
-import qualified Prelude as P
-
--- NOTE: This is a workaround for GHC's weird behaviour where it doesn't inline
--- these functions into unfoldings which makes the intermediate code size
--- explode. See http://hackage.haskell.org/trac/ghc/ticket/5539.
-infixr 2 ||
-infixr 3 &&
-
-not :: Bool -> Bool
-{-# INLINE not #-}
-not True = False
-not False = True
-
-(&&) :: Bool -> Bool -> Bool
-{-# INLINE (&&) #-}
-False && _ = False
-True && x = x
-
-(||) :: Bool -> Bool -> Bool
-{-# INLINE (||) #-}
-True || _ = True
-False || x = x
-
-
-data Checks = Bounds | Unsafe | Internal deriving( Eq )
-
-doBoundsChecks :: Bool
-#ifdef VECTOR_BOUNDS_CHECKS
-doBoundsChecks = True
-#else
-doBoundsChecks = False
-#endif
-
-doUnsafeChecks :: Bool
-#ifdef VECTOR_UNSAFE_CHECKS
-doUnsafeChecks = True
-#else
-doUnsafeChecks = False
-#endif
-
-doInternalChecks :: Bool
-#ifdef VECTOR_INTERNAL_CHECKS
-doInternalChecks = True
-#else
-doInternalChecks = False
-#endif
-
-
-doChecks :: Checks -> Bool
-{-# INLINE doChecks #-}
-doChecks Bounds   = doBoundsChecks
-doChecks Unsafe   = doUnsafeChecks
-doChecks Internal = doInternalChecks
-
-error_msg :: String -> Int -> String -> String -> String
-error_msg file line loc msg = file ++ ":" ++ show line ++ " (" ++ loc ++ "): " ++ msg
-
-error :: String -> Int -> String -> String -> a
-{-# NOINLINE error #-}
-error file line loc msg
-  = P.error $ error_msg file line loc msg
-
-internalError :: String -> Int -> String -> String -> a
-{-# NOINLINE internalError #-}
-internalError file line loc msg
-  = P.error $ unlines
-        ["*** Internal error in package vector ***"
-        ,"*** Please submit a bug report at http://trac.haskell.org/vector"
-        ,error_msg file line loc msg]
-
-
-checkError :: String -> Int -> Checks -> String -> String -> a
-{-# NOINLINE checkError #-}
-checkError file line kind loc msg
-  = case kind of
-      Internal -> internalError file line loc msg
-      _ -> error file line loc msg
-
-check :: String -> Int -> Checks -> String -> String -> Bool -> a -> a
-{-# INLINE check #-}
-check file line kind loc msg cond x
-  | not (doChecks kind) || cond = x
-  | otherwise = checkError file line kind loc msg
-
-checkIndex_msg :: Int -> Int -> String
-{-# INLINE checkIndex_msg #-}
-checkIndex_msg (I# i#) (I# n#) = checkIndex_msg# i# n#
-
-checkIndex_msg# :: Int# -> Int# -> String
-{-# NOINLINE checkIndex_msg# #-}
-checkIndex_msg# i# n# = "index out of bounds " ++ show (I# i#, I# n#)
-
-checkIndex :: String -> Int -> Checks -> String -> Int -> Int -> a -> a
-{-# INLINE checkIndex #-}
-checkIndex file line kind loc i n x
-  = check file line kind loc (checkIndex_msg i n) (i >= 0 && i<n) x
-
-
-checkLength_msg :: Int -> String
-{-# INLINE checkLength_msg #-}
-checkLength_msg (I# n#) = checkLength_msg# n#
-
-checkLength_msg# :: Int# -> String
-{-# NOINLINE checkLength_msg# #-}
-checkLength_msg# n# = "negative length " ++ show (I# n#)
-
-checkLength :: String -> Int -> Checks -> String -> Int -> a -> a
-{-# INLINE checkLength #-}
-checkLength file line kind loc n x
-  = check file line kind loc (checkLength_msg n) (n >= 0) x
-
-
-checkSlice_msg :: Int -> Int -> Int -> String
-{-# INLINE checkSlice_msg #-}
-checkSlice_msg (I# i#) (I# m#) (I# n#) = checkSlice_msg# i# m# n#
-
-checkSlice_msg# :: Int# -> Int# -> Int# -> String
-{-# NOINLINE checkSlice_msg# #-}
-checkSlice_msg# i# m# n# = "invalid slice " ++ show (I# i#, I# m#, I# n#)
-
-checkSlice :: String -> Int -> Checks -> String -> Int -> Int -> Int -> a -> a
-{-# INLINE checkSlice #-}
-checkSlice file line kind loc i m n x
-  = check file line kind loc (checkSlice_msg i m n)
-                             (i >= 0 && m >= 0 && i+m <= n) x
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Mutable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Mutable.hs
deleted file mode 100644
index ba701afb6a19..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Mutable.hs
+++ /dev/null
@@ -1,416 +0,0 @@
-{-# LANGUAGE CPP, DeriveDataTypeable, MultiParamTypeClasses, FlexibleInstances, BangPatterns, TypeFamilies #-}
-
--- |
--- Module      : Data.Vector.Mutable
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Mutable boxed vectors.
---
-
-module Data.Vector.Mutable (
-  -- * Mutable boxed vectors
-  MVector(..), IOVector, STVector,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Extracting subvectors
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- ** Overlapping
-  overlaps,
-
-  -- * Construction
-
-  -- ** Initialisation
-  new, unsafeNew, replicate, replicateM, clone,
-
-  -- ** Growing
-  grow, unsafeGrow,
-
-  -- ** Restricting memory usage
-  clear,
-
-  -- * Accessing individual elements
-  read, write, modify, swap,
-  unsafeRead, unsafeWrite, unsafeModify, unsafeSwap,
-
-  -- * Modifying vectors
-  nextPermutation,
-
-  -- ** Filling and copying
-  set, copy, move, unsafeCopy, unsafeMove
-) where
-
-import           Control.Monad (when)
-import qualified Data.Vector.Generic.Mutable as G
-import           Data.Primitive.Array
-import           Control.Monad.Primitive
-
-import Prelude hiding ( length, null, replicate, reverse, read,
-                        take, drop, splitAt, init, tail )
-
-import Data.Typeable ( Typeable )
-
-#include "vector.h"
-
--- | Mutable boxed vectors keyed on the monad they live in ('IO' or @'ST' s@).
-data MVector s a = MVector {-# UNPACK #-} !Int
-                           {-# UNPACK #-} !Int
-                           {-# UNPACK #-} !(MutableArray s a)
-        deriving ( Typeable )
-
-type IOVector = MVector RealWorld
-type STVector s = MVector s
-
--- NOTE: This seems unsafe, see http://trac.haskell.org/vector/ticket/54
-{-
-instance NFData a => NFData (MVector s a) where
-    rnf (MVector i n arr) = unsafeInlineST $ force i
-        where
-          force !ix | ix < n    = do x <- readArray arr ix
-                                     rnf x `seq` force (ix+1)
-                    | otherwise = return ()
--}
-
-instance G.MVector MVector a where
-  {-# INLINE basicLength #-}
-  basicLength (MVector _ n _) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice j m (MVector i _ arr) = MVector (i+j) m arr
-
-  {-# INLINE basicOverlaps #-}
-  basicOverlaps (MVector i m arr1) (MVector j n arr2)
-    = sameMutableArray arr1 arr2
-      && (between i j (j+n) || between j i (i+m))
-    where
-      between x y z = x >= y && x < z
-
-  {-# INLINE basicUnsafeNew #-}
-  basicUnsafeNew n
-    = do
-        arr <- newArray n uninitialised
-        return (MVector 0 n arr)
-
-  {-# INLINE basicInitialize #-}
-  -- initialization is unnecessary for boxed vectors
-  basicInitialize _ = return ()
-
-  {-# INLINE basicUnsafeReplicate #-}
-  basicUnsafeReplicate n x
-    = do
-        arr <- newArray n x
-        return (MVector 0 n arr)
-
-  {-# INLINE basicUnsafeRead #-}
-  basicUnsafeRead (MVector i _ arr) j = readArray arr (i+j)
-
-  {-# INLINE basicUnsafeWrite #-}
-  basicUnsafeWrite (MVector i _ arr) j x = writeArray arr (i+j) x
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector i n dst) (MVector j _ src)
-    = copyMutableArray dst i src j n
-
-  basicUnsafeMove dst@(MVector iDst n arrDst) src@(MVector iSrc _ arrSrc)
-    = case n of
-        0 -> return ()
-        1 -> readArray arrSrc iSrc >>= writeArray arrDst iDst
-        2 -> do
-               x <- readArray arrSrc iSrc
-               y <- readArray arrSrc (iSrc + 1)
-               writeArray arrDst iDst x
-               writeArray arrDst (iDst + 1) y
-        _
-          | overlaps dst src
-             -> case compare iDst iSrc of
-                  LT -> moveBackwards arrDst iDst iSrc n
-                  EQ -> return ()
-                  GT | (iDst - iSrc) * 2 < n
-                        -> moveForwardsLargeOverlap arrDst iDst iSrc n
-                     | otherwise
-                        -> moveForwardsSmallOverlap arrDst iDst iSrc n
-          | otherwise -> G.basicUnsafeCopy dst src
-
-  {-# INLINE basicClear #-}
-  basicClear v = G.set v uninitialised
-
-{-# INLINE moveBackwards #-}
-moveBackwards :: PrimMonad m => MutableArray (PrimState m) a -> Int -> Int -> Int -> m ()
-moveBackwards !arr !dstOff !srcOff !len =
-  INTERNAL_CHECK(check) "moveBackwards" "not a backwards move" (dstOff < srcOff)
-  $ loopM len $ \ i -> readArray arr (srcOff + i) >>= writeArray arr (dstOff + i)
-
-{-# INLINE moveForwardsSmallOverlap #-}
--- Performs a move when dstOff > srcOff, optimized for when the overlap of the intervals is small.
-moveForwardsSmallOverlap :: PrimMonad m => MutableArray (PrimState m) a -> Int -> Int -> Int -> m ()
-moveForwardsSmallOverlap !arr !dstOff !srcOff !len =
-  INTERNAL_CHECK(check) "moveForwardsSmallOverlap" "not a forward move" (dstOff > srcOff)
-  $ do
-      tmp <- newArray overlap uninitialised
-      loopM overlap $ \ i -> readArray arr (dstOff + i) >>= writeArray tmp i
-      loopM nonOverlap $ \ i -> readArray arr (srcOff + i) >>= writeArray arr (dstOff + i)
-      loopM overlap $ \ i -> readArray tmp i >>= writeArray arr (dstOff + nonOverlap + i)
-  where nonOverlap = dstOff - srcOff; overlap = len - nonOverlap
-
--- Performs a move when dstOff > srcOff, optimized for when the overlap of the intervals is large.
-moveForwardsLargeOverlap :: PrimMonad m => MutableArray (PrimState m) a -> Int -> Int -> Int -> m ()
-moveForwardsLargeOverlap !arr !dstOff !srcOff !len =
-  INTERNAL_CHECK(check) "moveForwardsLargeOverlap" "not a forward move" (dstOff > srcOff)
-  $ do
-      queue <- newArray nonOverlap uninitialised
-      loopM nonOverlap $ \ i -> readArray arr (srcOff + i) >>= writeArray queue i
-      let mov !i !qTop = when (i < dstOff + len) $ do
-            x <- readArray arr i
-            y <- readArray queue qTop
-            writeArray arr i y
-            writeArray queue qTop x
-            mov (i+1) (if qTop + 1 >= nonOverlap then 0 else qTop + 1)
-      mov dstOff 0
-  where nonOverlap = dstOff - srcOff
-
-{-# INLINE loopM #-}
-loopM :: Monad m => Int -> (Int -> m a) -> m ()
-loopM !n k = let
-  go i = when (i < n) (k i >> go (i+1))
-  in go 0
-
-uninitialised :: a
-uninitialised = error "Data.Vector.Mutable: uninitialised element"
-
--- Length information
--- ------------------
-
--- | Length of the mutable vector.
-length :: MVector s a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | Check whether the vector is empty
-null :: MVector s a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Extracting subvectors
--- ---------------------
-
--- | Yield a part of the mutable vector without copying it.
-slice :: Int -> Int -> MVector s a -> MVector s a
-{-# INLINE slice #-}
-slice = G.slice
-
-take :: Int -> MVector s a -> MVector s a
-{-# INLINE take #-}
-take = G.take
-
-drop :: Int -> MVector s a -> MVector s a
-{-# INLINE drop #-}
-drop = G.drop
-
-{-# INLINE splitAt #-}
-splitAt :: Int -> MVector s a -> (MVector s a, MVector s a)
-splitAt = G.splitAt
-
-init :: MVector s a -> MVector s a
-{-# INLINE init #-}
-init = G.init
-
-tail :: MVector s a -> MVector s a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | Yield a part of the mutable vector without copying it. No bounds checks
--- are performed.
-unsafeSlice :: Int  -- ^ starting index
-            -> Int  -- ^ length of the slice
-            -> MVector s a
-            -> MVector s a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
-unsafeTake :: Int -> MVector s a -> MVector s a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
-unsafeDrop :: Int -> MVector s a -> MVector s a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
-unsafeInit :: MVector s a -> MVector s a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
-unsafeTail :: MVector s a -> MVector s a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- Overlapping
--- -----------
-
--- | Check whether two vectors overlap.
-overlaps :: MVector s a -> MVector s a -> Bool
-{-# INLINE overlaps #-}
-overlaps = G.overlaps
-
--- Initialisation
--- --------------
-
--- | Create a mutable vector of the given length.
-new :: PrimMonad m => Int -> m (MVector (PrimState m) a)
-{-# INLINE new #-}
-new = G.new
-
--- | Create a mutable vector of the given length. The memory is not initialized.
-unsafeNew :: PrimMonad m => Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeNew #-}
-unsafeNew = G.unsafeNew
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with an initial value.
-replicate :: PrimMonad m => Int -> a -> m (MVector (PrimState m) a)
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with values produced by repeatedly executing the monadic action.
-replicateM :: PrimMonad m => Int -> m a -> m (MVector (PrimState m) a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | Create a copy of a mutable vector.
-clone :: PrimMonad m => MVector (PrimState m) a -> m (MVector (PrimState m) a)
-{-# INLINE clone #-}
-clone = G.clone
-
--- Growing
--- -------
-
--- | Grow a vector by the given number of elements. The number must be
--- positive.
-grow :: PrimMonad m
-              => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE grow #-}
-grow = G.grow
-
--- | Grow a vector by the given number of elements. The number must be
--- positive but this is not checked.
-unsafeGrow :: PrimMonad m
-               => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeGrow #-}
-unsafeGrow = G.unsafeGrow
-
--- Restricting memory usage
--- ------------------------
-
--- | Reset all elements of the vector to some undefined value, clearing all
--- references to external objects. This is usually a noop for unboxed vectors.
-clear :: PrimMonad m => MVector (PrimState m) a -> m ()
-{-# INLINE clear #-}
-clear = G.clear
-
--- Accessing individual elements
--- -----------------------------
-
--- | Yield the element at the given position.
-read :: PrimMonad m => MVector (PrimState m) a -> Int -> m a
-{-# INLINE read #-}
-read = G.read
-
--- | Replace the element at the given position.
-write :: PrimMonad m => MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE write #-}
-write = G.write
-
--- | Modify the element at the given position.
-modify :: PrimMonad m => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE modify #-}
-modify = G.modify
-
--- | Swap the elements at the given positions.
-swap :: PrimMonad m => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE swap #-}
-swap = G.swap
-
-
--- | Yield the element at the given position. No bounds checks are performed.
-unsafeRead :: PrimMonad m => MVector (PrimState m) a -> Int -> m a
-{-# INLINE unsafeRead #-}
-unsafeRead = G.unsafeRead
-
--- | Replace the element at the given position. No bounds checks are performed.
-unsafeWrite :: PrimMonad m => MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE unsafeWrite #-}
-unsafeWrite = G.unsafeWrite
-
--- | Modify the element at the given position. No bounds checks are performed.
-unsafeModify :: PrimMonad m => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE unsafeModify #-}
-unsafeModify = G.unsafeModify
-
--- | Swap the elements at the given positions. No bounds checks are performed.
-unsafeSwap :: PrimMonad m => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE unsafeSwap #-}
-unsafeSwap = G.unsafeSwap
-
--- Filling and copying
--- -------------------
-
--- | Set all elements of the vector to the given value.
-set :: PrimMonad m => MVector (PrimState m) a -> a -> m ()
-{-# INLINE set #-}
-set = G.set
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap.
-copy :: PrimMonad m
-                 => MVector (PrimState m) a -> MVector (PrimState m) a -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap. This is not checked.
-unsafeCopy :: PrimMonad m => MVector (PrimState m) a   -- ^ target
-                          -> MVector (PrimState m) a   -- ^ source
-                          -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | Move the contents of a vector. The two vectors must have the same
--- length.
---
--- If the vectors do not overlap, then this is equivalent to 'copy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-move :: PrimMonad m
-                 => MVector (PrimState m) a -> MVector (PrimState m) a -> m ()
-{-# INLINE move #-}
-move = G.move
-
--- | Move the contents of a vector. The two vectors must have the same
--- length, but this is not checked.
---
--- If the vectors do not overlap, then this is equivalent to 'unsafeCopy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-unsafeMove :: PrimMonad m => MVector (PrimState m) a   -- ^ target
-                          -> MVector (PrimState m) a   -- ^ source
-                          -> m ()
-{-# INLINE unsafeMove #-}
-unsafeMove = G.unsafeMove
-
--- | Compute the next (lexicographically) permutation of given vector in-place.
---   Returns False when input is the last permtuation
-nextPermutation :: (PrimMonad m,Ord e) => MVector (PrimState m) e -> m Bool
-{-# INLINE nextPermutation #-}
-nextPermutation = G.nextPermutation
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive.hs
deleted file mode 100644
index ba18f9ba957f..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive.hs
+++ /dev/null
@@ -1,1393 +0,0 @@
-{-# LANGUAGE CPP, DeriveDataTypeable, FlexibleInstances, MultiParamTypeClasses, TypeFamilies, ScopedTypeVariables, Rank2Types #-}
-
--- |
--- Module      : Data.Vector.Primitive
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Unboxed vectors of primitive types. The use of this module is not
--- recommended except in very special cases. Adaptive unboxed vectors defined
--- in "Data.Vector.Unboxed" are significantly more flexible at no performance
--- cost.
---
-
-module Data.Vector.Primitive (
-  -- * Primitive vectors
-  Vector(..), MVector(..), Prim,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Indexing
-  (!), (!?), head, last,
-  unsafeIndex, unsafeHead, unsafeLast,
-
-  -- ** Monadic indexing
-  indexM, headM, lastM,
-  unsafeIndexM, unsafeHeadM, unsafeLastM,
-
-  -- ** Extracting subvectors (slicing)
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- * Construction
-
-  -- ** Initialisation
-  empty, singleton, replicate, generate, iterateN,
-
-  -- ** Monadic initialisation
-  replicateM, generateM, iterateNM, create, createT,
-
-  -- ** Unfolding
-  unfoldr, unfoldrN,
-  unfoldrM, unfoldrNM,
-  constructN, constructrN,
-
-  -- ** Enumeration
-  enumFromN, enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- ** Concatenation
-  cons, snoc, (++), concat,
-
-  -- ** Restricting memory usage
-  force,
-
-  -- * Modifying vectors
-
-  -- ** Bulk updates
-  (//), update_,
-  unsafeUpd, unsafeUpdate_,
-
-  -- ** Accumulations
-  accum, accumulate_,
-  unsafeAccum, unsafeAccumulate_,
-
-  -- ** Permutations
-  reverse, backpermute, unsafeBackpermute,
-
-  -- ** Safe destructive updates
-  modify,
-
-  -- * Elementwise operations
-
-  -- ** Mapping
-  map, imap, concatMap,
-
-  -- ** Monadic mapping
-  mapM, mapM_, forM, forM_,
-
-  -- ** Zipping
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  izipWith, izipWith3, izipWith4, izipWith5, izipWith6,
-
-  -- ** Monadic zipping
-  zipWithM, zipWithM_,
-
-  -- * Working with predicates
-
-  -- ** Filtering
-  filter, ifilter, uniq,
-  mapMaybe, imapMaybe,
-  filterM,
-  takeWhile, dropWhile,
-
-  -- ** Partitioning
-  partition, unstablePartition, span, break,
-
-  -- ** Searching
-  elem, notElem, find, findIndex, findIndices, elemIndex, elemIndices,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1, foldr', foldr1',
-  ifoldl, ifoldl', ifoldr, ifoldr',
-
-  -- ** Specialised folds
-  all, any,
-  sum, product,
-  maximum, maximumBy, minimum, minimumBy,
-  minIndex, minIndexBy, maxIndex, maxIndexBy,
-
-  -- ** Monadic folds
-  foldM, foldM', fold1M, fold1M',
-  foldM_, foldM'_, fold1M_, fold1M'_,
-
-  -- * Prefix sums (scans)
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl', scanl1, scanl1',
-  prescanr, prescanr',
-  postscanr, postscanr',
-  scanr, scanr', scanr1, scanr1',
-
-  -- * Conversions
-
-  -- ** Lists
-  toList, fromList, fromListN,
-
-  -- ** Other vector types
-  G.convert,
-
-  -- ** Mutable vectors
-  freeze, thaw, copy, unsafeFreeze, unsafeThaw, unsafeCopy
-) where
-
-import qualified Data.Vector.Generic           as G
-import           Data.Vector.Primitive.Mutable ( MVector(..) )
-import qualified Data.Vector.Fusion.Bundle as Bundle
-import           Data.Primitive.ByteArray
-import           Data.Primitive ( Prim, sizeOf )
-
-import Control.DeepSeq ( NFData(rnf) )
-
-import Control.Monad ( liftM )
-import Control.Monad.ST ( ST )
-import Control.Monad.Primitive
-
-import Prelude hiding ( length, null,
-                        replicate, (++), concat,
-                        head, last,
-                        init, tail, take, drop, splitAt, reverse,
-                        map, concatMap,
-                        zipWith, zipWith3, zip, zip3, unzip, unzip3,
-                        filter, takeWhile, dropWhile, span, break,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        all, any, sum, product, minimum, maximum,
-                        scanl, scanl1, scanr, scanr1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_ )
-
-import Data.Typeable  ( Typeable )
-import Data.Data      ( Data(..) )
-import Text.Read      ( Read(..), readListPrecDefault )
-import Data.Semigroup ( Semigroup(..) )
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Monoid   ( Monoid(..) )
-import Data.Traversable ( Traversable )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-import qualified GHC.Exts as Exts
-#endif
-
--- | Unboxed vectors of primitive types
-data Vector a = Vector {-# UNPACK #-} !Int
-                       {-# UNPACK #-} !Int
-                       {-# UNPACK #-} !ByteArray -- ^ offset, length, underlying byte array
-  deriving ( Typeable )
-
-instance NFData (Vector a) where
-  rnf (Vector _ _ _) = ()
-
-instance (Show a, Prim a) => Show (Vector a) where
-  showsPrec = G.showsPrec
-
-instance (Read a, Prim a) => Read (Vector a) where
-  readPrec = G.readPrec
-  readListPrec = readListPrecDefault
-
-instance (Data a, Prim a) => Data (Vector a) where
-  gfoldl       = G.gfoldl
-  toConstr _   = error "toConstr"
-  gunfold _ _  = error "gunfold"
-  dataTypeOf _ = G.mkType "Data.Vector.Primitive.Vector"
-  dataCast1    = G.dataCast
-
-
-type instance G.Mutable Vector = MVector
-
-instance Prim a => G.Vector Vector a where
-  {-# INLINE basicUnsafeFreeze #-}
-  basicUnsafeFreeze (MVector i n marr)
-    = Vector i n `liftM` unsafeFreezeByteArray marr
-
-  {-# INLINE basicUnsafeThaw #-}
-  basicUnsafeThaw (Vector i n arr)
-    = MVector i n `liftM` unsafeThawByteArray arr
-
-  {-# INLINE basicLength #-}
-  basicLength (Vector _ n _) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice j n (Vector i _ arr) = Vector (i+j) n arr
-
-  {-# INLINE basicUnsafeIndexM #-}
-  basicUnsafeIndexM (Vector i _ arr) j = return $! indexByteArray arr (i+j)
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector i n dst) (Vector j _ src)
-    = copyByteArray dst (i*sz) src (j*sz) (n*sz)
-    where
-      sz = sizeOf (undefined :: a)
-
-  {-# INLINE elemseq #-}
-  elemseq _ = seq
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Prim a, Eq a) => Eq (Vector a) where
-  {-# INLINE (==) #-}
-  xs == ys = Bundle.eq (G.stream xs) (G.stream ys)
-
-  {-# INLINE (/=) #-}
-  xs /= ys = not (Bundle.eq (G.stream xs) (G.stream ys))
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Prim a, Ord a) => Ord (Vector a) where
-  {-# INLINE compare #-}
-  compare xs ys = Bundle.cmp (G.stream xs) (G.stream ys)
-
-  {-# INLINE (<) #-}
-  xs < ys = Bundle.cmp (G.stream xs) (G.stream ys) == LT
-
-  {-# INLINE (<=) #-}
-  xs <= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= GT
-
-  {-# INLINE (>) #-}
-  xs > ys = Bundle.cmp (G.stream xs) (G.stream ys) == GT
-
-  {-# INLINE (>=) #-}
-  xs >= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= LT
-
-instance Prim a => Semigroup (Vector a) where
-  {-# INLINE (<>) #-}
-  (<>) = (++)
-
-  {-# INLINE sconcat #-}
-  sconcat = G.concatNE
-
-instance Prim a => Monoid (Vector a) where
-  {-# INLINE mempty #-}
-  mempty = empty
-
-  {-# INLINE mappend #-}
-  mappend = (++)
-
-  {-# INLINE mconcat #-}
-  mconcat = concat
-
-#if __GLASGOW_HASKELL__ >= 708
-
-instance Prim a => Exts.IsList (Vector a) where
-  type Item (Vector a) = a
-  fromList = fromList
-  fromListN = fromListN
-  toList = toList
-
-#endif
--- Length
--- ------
-
--- | /O(1)/ Yield the length of the vector
-length :: Prim a => Vector a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | /O(1)/ Test whether a vector is empty
-null :: Prim a => Vector a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Indexing
--- --------
-
--- | O(1) Indexing
-(!) :: Prim a => Vector a -> Int -> a
-{-# INLINE (!) #-}
-(!) = (G.!)
-
--- | O(1) Safe indexing
-(!?) :: Prim a => Vector a -> Int -> Maybe a
-{-# INLINE (!?) #-}
-(!?) = (G.!?)
-
--- | /O(1)/ First element
-head :: Prim a => Vector a -> a
-{-# INLINE head #-}
-head = G.head
-
--- | /O(1)/ Last element
-last :: Prim a => Vector a -> a
-{-# INLINE last #-}
-last = G.last
-
--- | /O(1)/ Unsafe indexing without bounds checking
-unsafeIndex :: Prim a => Vector a -> Int -> a
-{-# INLINE unsafeIndex #-}
-unsafeIndex = G.unsafeIndex
-
--- | /O(1)/ First element without checking if the vector is empty
-unsafeHead :: Prim a => Vector a -> a
-{-# INLINE unsafeHead #-}
-unsafeHead = G.unsafeHead
-
--- | /O(1)/ Last element without checking if the vector is empty
-unsafeLast :: Prim a => Vector a -> a
-{-# INLINE unsafeLast #-}
-unsafeLast = G.unsafeLast
-
--- Monadic indexing
--- ----------------
-
--- | /O(1)/ Indexing in a monad.
---
--- The monad allows operations to be strict in the vector when necessary.
--- Suppose vector copying is implemented like this:
---
--- > copy mv v = ... write mv i (v ! i) ...
---
--- For lazy vectors, @v ! i@ would not be evaluated which means that @mv@
--- would unnecessarily retain a reference to @v@ in each element written.
---
--- With 'indexM', copying can be implemented like this instead:
---
--- > copy mv v = ... do
--- >                   x <- indexM v i
--- >                   write mv i x
---
--- Here, no references to @v@ are retained because indexing (but /not/ the
--- elements) is evaluated eagerly.
---
-indexM :: (Prim a, Monad m) => Vector a -> Int -> m a
-{-# INLINE indexM #-}
-indexM = G.indexM
-
--- | /O(1)/ First element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-headM :: (Prim a, Monad m) => Vector a -> m a
-{-# INLINE headM #-}
-headM = G.headM
-
--- | /O(1)/ Last element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-lastM :: (Prim a, Monad m) => Vector a -> m a
-{-# INLINE lastM #-}
-lastM = G.lastM
-
--- | /O(1)/ Indexing in a monad without bounds checks. See 'indexM' for an
--- explanation of why this is useful.
-unsafeIndexM :: (Prim a, Monad m) => Vector a -> Int -> m a
-{-# INLINE unsafeIndexM #-}
-unsafeIndexM = G.unsafeIndexM
-
--- | /O(1)/ First element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeHeadM :: (Prim a, Monad m) => Vector a -> m a
-{-# INLINE unsafeHeadM #-}
-unsafeHeadM = G.unsafeHeadM
-
--- | /O(1)/ Last element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeLastM :: (Prim a, Monad m) => Vector a -> m a
-{-# INLINE unsafeLastM #-}
-unsafeLastM = G.unsafeLastM
-
--- Extracting subvectors (slicing)
--- -------------------------------
-
--- | /O(1)/ Yield a slice of the vector without copying it. The vector must
--- contain at least @i+n@ elements.
-slice :: Prim a
-      => Int   -- ^ @i@ starting index
-      -> Int   -- ^ @n@ length
-      -> Vector a
-      -> Vector a
-{-# INLINE slice #-}
-slice = G.slice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty.
-init :: Prim a => Vector a -> Vector a
-{-# INLINE init #-}
-init = G.init
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty.
-tail :: Prim a => Vector a -> Vector a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | /O(1)/ Yield at the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case it is returned unchanged.
-take :: Prim a => Int -> Vector a -> Vector a
-{-# INLINE take #-}
-take = G.take
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case an empty vector is returned.
-drop :: Prim a => Int -> Vector a -> Vector a
-{-# INLINE drop #-}
-drop = G.drop
-
--- | /O(1)/ Yield the first @n@ elements paired with the remainder without copying.
---
--- Note that @'splitAt' n v@ is equivalent to @('take' n v, 'drop' n v)@
--- but slightly more efficient.
-{-# INLINE splitAt #-}
-splitAt :: Prim a => Int -> Vector a -> (Vector a, Vector a)
-splitAt = G.splitAt
-
--- | /O(1)/ Yield a slice of the vector without copying. The vector must
--- contain at least @i+n@ elements but this is not checked.
-unsafeSlice :: Prim a => Int   -- ^ @i@ starting index
-                       -> Int   -- ^ @n@ length
-                       -> Vector a
-                       -> Vector a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty but this is not checked.
-unsafeInit :: Prim a => Vector a -> Vector a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty but this is not checked.
-unsafeTail :: Prim a => Vector a -> Vector a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector must
--- contain at least @n@ elements but this is not checked.
-unsafeTake :: Prim a => Int -> Vector a -> Vector a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector
--- must contain at least @n@ elements but this is not checked.
-unsafeDrop :: Prim a => Int -> Vector a -> Vector a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
--- Initialisation
--- --------------
-
--- | /O(1)/ Empty vector
-empty :: Prim a => Vector a
-{-# INLINE empty #-}
-empty = G.empty
-
--- | /O(1)/ Vector with exactly one element
-singleton :: Prim a => a -> Vector a
-{-# INLINE singleton #-}
-singleton = G.singleton
-
--- | /O(n)/ Vector of the given length with the same value in each position
-replicate :: Prim a => Int -> a -> Vector a
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | /O(n)/ Construct a vector of the given length by applying the function to
--- each index
-generate :: Prim a => Int -> (Int -> a) -> Vector a
-{-# INLINE generate #-}
-generate = G.generate
-
--- | /O(n)/ Apply function n times to value. Zeroth element is original value.
-iterateN :: Prim a => Int -> (a -> a) -> a -> Vector a
-{-# INLINE iterateN #-}
-iterateN = G.iterateN
-
--- Unfolding
--- ---------
-
--- | /O(n)/ Construct a vector by repeatedly applying the generator function
--- to a seed. The generator function yields 'Just' the next element and the
--- new seed or 'Nothing' if there are no more elements.
---
--- > unfoldr (\n -> if n == 0 then Nothing else Just (n,n-1)) 10
--- >  = <10,9,8,7,6,5,4,3,2,1>
-unfoldr :: Prim a => (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldr #-}
-unfoldr = G.unfoldr
-
--- | /O(n)/ Construct a vector with at most @n@ elements by repeatedly applying
--- the generator function to a seed. The generator function yields 'Just' the
--- next element and the new seed or 'Nothing' if there are no more elements.
---
--- > unfoldrN 3 (\n -> Just (n,n-1)) 10 = <10,9,8>
-unfoldrN :: Prim a => Int -> (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldrN #-}
-unfoldrN = G.unfoldrN
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrM :: (Monad m, Prim a) => (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrM #-}
-unfoldrM = G.unfoldrM
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrNM :: (Monad m, Prim a) => Int -> (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrNM #-}
-unfoldrNM = G.unfoldrNM
-
--- | /O(n)/ Construct a vector with @n@ elements by repeatedly applying the
--- generator function to the already constructed part of the vector.
---
--- > constructN 3 f = let a = f <> ; b = f <a> ; c = f <a,b> in f <a,b,c>
---
-constructN :: Prim a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructN #-}
-constructN = G.constructN
-
--- | /O(n)/ Construct a vector with @n@ elements from right to left by
--- repeatedly applying the generator function to the already constructed part
--- of the vector.
---
--- > constructrN 3 f = let a = f <> ; b = f<a> ; c = f <b,a> in f <c,b,a>
---
-constructrN :: Prim a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructrN #-}
-constructrN = G.constructrN
-
--- Enumeration
--- -----------
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+1@
--- etc. This operation is usually more efficient than 'enumFromTo'.
---
--- > enumFromN 5 3 = <5,6,7>
-enumFromN :: (Prim a, Num a) => a -> Int -> Vector a
-{-# INLINE enumFromN #-}
-enumFromN = G.enumFromN
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc. This operations is usually more efficient than 'enumFromThenTo'.
---
--- > enumFromStepN 1 0.1 5 = <1,1.1,1.2,1.3,1.4>
-enumFromStepN :: (Prim a, Num a) => a -> a -> Int -> Vector a
-{-# INLINE enumFromStepN #-}
-enumFromStepN = G.enumFromStepN
-
--- | /O(n)/ Enumerate values from @x@ to @y@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromN' instead.
-enumFromTo :: (Prim a, Enum a) => a -> a -> Vector a
-{-# INLINE enumFromTo #-}
-enumFromTo = G.enumFromTo
-
--- | /O(n)/ Enumerate values from @x@ to @y@ with a specific step @z@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Prim a, Enum a) => a -> a -> a -> Vector a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo = G.enumFromThenTo
-
--- Concatenation
--- -------------
-
--- | /O(n)/ Prepend an element
-cons :: Prim a => a -> Vector a -> Vector a
-{-# INLINE cons #-}
-cons = G.cons
-
--- | /O(n)/ Append an element
-snoc :: Prim a => Vector a -> a -> Vector a
-{-# INLINE snoc #-}
-snoc = G.snoc
-
-infixr 5 ++
--- | /O(m+n)/ Concatenate two vectors
-(++) :: Prim a => Vector a -> Vector a -> Vector a
-{-# INLINE (++) #-}
-(++) = (G.++)
-
--- | /O(n)/ Concatenate all vectors in the list
-concat :: Prim a => [Vector a] -> Vector a
-{-# INLINE concat #-}
-concat = G.concat
-
--- Monadic initialisation
--- ----------------------
-
--- | /O(n)/ Execute the monadic action the given number of times and store the
--- results in a vector.
-replicateM :: (Monad m, Prim a) => Int -> m a -> m (Vector a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | /O(n)/ Construct a vector of the given length by applying the monadic
--- action to each index
-generateM :: (Monad m, Prim a) => Int -> (Int -> m a) -> m (Vector a)
-{-# INLINE generateM #-}
-generateM = G.generateM
-
--- | /O(n)/ Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: (Monad m, Prim a) => Int -> (a -> m a) -> a -> m (Vector a)
-{-# INLINE iterateNM #-}
-iterateNM = G.iterateNM
-
--- | Execute the monadic action and freeze the resulting vector.
---
--- @
--- create (do { v \<- new 2; write v 0 \'a\'; write v 1 \'b\'; return v }) = \<'a','b'\>
--- @
-create :: Prim a => (forall s. ST s (MVector s a)) -> Vector a
-{-# INLINE create #-}
--- NOTE: eta-expanded due to http://hackage.haskell.org/trac/ghc/ticket/4120
-create p = G.create p
-
--- | Execute the monadic action and freeze the resulting vectors.
-createT :: (Traversable f, Prim a) => (forall s. ST s (f (MVector s a))) -> f (Vector a)
-{-# INLINE createT #-}
-createT p = G.createT p
-
--- Restricting memory usage
--- ------------------------
-
--- | /O(n)/ Yield the argument but force it not to retain any extra memory,
--- possibly by copying it.
---
--- This is especially useful when dealing with slices. For example:
---
--- > force (slice 0 2 <huge vector>)
---
--- Here, the slice retains a reference to the huge vector. Forcing it creates
--- a copy of just the elements that belong to the slice and allows the huge
--- vector to be garbage collected.
-force :: Prim a => Vector a -> Vector a
-{-# INLINE force #-}
-force = G.force
-
--- Bulk updates
--- ------------
-
--- | /O(m+n)/ For each pair @(i,a)@ from the list, replace the vector
--- element at position @i@ by @a@.
---
--- > <5,9,2,7> // [(2,1),(0,3),(2,8)] = <3,9,8,7>
---
-(//) :: Prim a => Vector a   -- ^ initial vector (of length @m@)
-                -> [(Int, a)] -- ^ list of index/value pairs (of length @n@)
-                -> Vector a
-{-# INLINE (//) #-}
-(//) = (G.//)
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @a@ from the value vector, replace the element of the
--- initial vector at position @i@ by @a@.
---
--- > update_ <5,9,2,7>  <2,0,2> <1,3,8> = <3,9,8,7>
---
-update_ :: Prim a
-        => Vector a   -- ^ initial vector (of length @m@)
-        -> Vector Int -- ^ index vector (of length @n1@)
-        -> Vector a   -- ^ value vector (of length @n2@)
-        -> Vector a
-{-# INLINE update_ #-}
-update_ = G.update_
-
--- | Same as ('//') but without bounds checking.
-unsafeUpd :: Prim a => Vector a -> [(Int, a)] -> Vector a
-{-# INLINE unsafeUpd #-}
-unsafeUpd = G.unsafeUpd
-
--- | Same as 'update_' but without bounds checking.
-unsafeUpdate_ :: Prim a => Vector a -> Vector Int -> Vector a -> Vector a
-{-# INLINE unsafeUpdate_ #-}
-unsafeUpdate_ = G.unsafeUpdate_
-
--- Accumulations
--- -------------
-
--- | /O(m+n)/ For each pair @(i,b)@ from the list, replace the vector element
--- @a@ at position @i@ by @f a b@.
---
--- > accum (+) <5,9,2> [(2,4),(1,6),(0,3),(1,7)] = <5+3, 9+6+7, 2+4>
-accum :: Prim a
-      => (a -> b -> a) -- ^ accumulating function @f@
-      -> Vector a      -- ^ initial vector (of length @m@)
-      -> [(Int,b)]     -- ^ list of index/value pairs (of length @n@)
-      -> Vector a
-{-# INLINE accum #-}
-accum = G.accum
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @b@ from the the value vector,
--- replace the element of the initial vector at
--- position @i@ by @f a b@.
---
--- > accumulate_ (+) <5,9,2> <2,1,0,1> <4,6,3,7> = <5+3, 9+6+7, 2+4>
---
-accumulate_ :: (Prim a, Prim b)
-            => (a -> b -> a) -- ^ accumulating function @f@
-            -> Vector a      -- ^ initial vector (of length @m@)
-            -> Vector Int    -- ^ index vector (of length @n1@)
-            -> Vector b      -- ^ value vector (of length @n2@)
-            -> Vector a
-{-# INLINE accumulate_ #-}
-accumulate_ = G.accumulate_
-
--- | Same as 'accum' but without bounds checking.
-unsafeAccum :: Prim a => (a -> b -> a) -> Vector a -> [(Int,b)] -> Vector a
-{-# INLINE unsafeAccum #-}
-unsafeAccum = G.unsafeAccum
-
--- | Same as 'accumulate_' but without bounds checking.
-unsafeAccumulate_ :: (Prim a, Prim b) =>
-               (a -> b -> a) -> Vector a -> Vector Int -> Vector b -> Vector a
-{-# INLINE unsafeAccumulate_ #-}
-unsafeAccumulate_ = G.unsafeAccumulate_
-
--- Permutations
--- ------------
-
--- | /O(n)/ Reverse a vector
-reverse :: Prim a => Vector a -> Vector a
-{-# INLINE reverse #-}
-reverse = G.reverse
-
--- | /O(n)/ Yield the vector obtained by replacing each element @i@ of the
--- index vector by @xs'!'i@. This is equivalent to @'map' (xs'!') is@ but is
--- often much more efficient.
---
--- > backpermute <a,b,c,d> <0,3,2,3,1,0> = <a,d,c,d,b,a>
-backpermute :: Prim a => Vector a -> Vector Int -> Vector a
-{-# INLINE backpermute #-}
-backpermute = G.backpermute
-
--- | Same as 'backpermute' but without bounds checking.
-unsafeBackpermute :: Prim a => Vector a -> Vector Int -> Vector a
-{-# INLINE unsafeBackpermute #-}
-unsafeBackpermute = G.unsafeBackpermute
-
--- Safe destructive updates
--- ------------------------
-
--- | Apply a destructive operation to a vector. The operation will be
--- performed in place if it is safe to do so and will modify a copy of the
--- vector otherwise.
---
--- @
--- modify (\\v -> write v 0 \'x\') ('replicate' 3 \'a\') = \<\'x\',\'a\',\'a\'\>
--- @
-modify :: Prim a => (forall s. MVector s a -> ST s ()) -> Vector a -> Vector a
-{-# INLINE modify #-}
-modify p = G.modify p
-
--- Mapping
--- -------
-
--- | /O(n)/ Map a function over a vector
-map :: (Prim a, Prim b) => (a -> b) -> Vector a -> Vector b
-{-# INLINE map #-}
-map = G.map
-
--- | /O(n)/ Apply a function to every element of a vector and its index
-imap :: (Prim a, Prim b) => (Int -> a -> b) -> Vector a -> Vector b
-{-# INLINE imap #-}
-imap = G.imap
-
--- | Map a function over a vector and concatenate the results.
-concatMap :: (Prim a, Prim b) => (a -> Vector b) -> Vector a -> Vector b
-{-# INLINE concatMap #-}
-concatMap = G.concatMap
-
--- Monadic mapping
--- ---------------
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results
-mapM :: (Monad m, Prim a, Prim b) => (a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE mapM #-}
-mapM = G.mapM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results
-mapM_ :: (Monad m, Prim a) => (a -> m b) -> Vector a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ = G.mapM_
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results. Equivalent to @flip 'mapM'@.
-forM :: (Monad m, Prim a, Prim b) => Vector a -> (a -> m b) -> m (Vector b)
-{-# INLINE forM #-}
-forM = G.forM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results. Equivalent to @flip 'mapM_'@.
-forM_ :: (Monad m, Prim a) => Vector a -> (a -> m b) -> m ()
-{-# INLINE forM_ #-}
-forM_ = G.forM_
-
--- Zipping
--- -------
-
--- | /O(min(m,n))/ Zip two vectors with the given function.
-zipWith :: (Prim a, Prim b, Prim c)
-        => (a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE zipWith #-}
-zipWith = G.zipWith
-
--- | Zip three vectors with the given function.
-zipWith3 :: (Prim a, Prim b, Prim c, Prim d)
-         => (a -> b -> c -> d) -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE zipWith3 #-}
-zipWith3 = G.zipWith3
-
-zipWith4 :: (Prim a, Prim b, Prim c, Prim d, Prim e)
-         => (a -> b -> c -> d -> e)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE zipWith4 #-}
-zipWith4 = G.zipWith4
-
-zipWith5 :: (Prim a, Prim b, Prim c, Prim d, Prim e,
-             Prim f)
-         => (a -> b -> c -> d -> e -> f)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f
-{-# INLINE zipWith5 #-}
-zipWith5 = G.zipWith5
-
-zipWith6 :: (Prim a, Prim b, Prim c, Prim d, Prim e,
-             Prim f, Prim g)
-         => (a -> b -> c -> d -> e -> f -> g)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f -> Vector g
-{-# INLINE zipWith6 #-}
-zipWith6 = G.zipWith6
-
--- | /O(min(m,n))/ Zip two vectors with a function that also takes the
--- elements' indices.
-izipWith :: (Prim a, Prim b, Prim c)
-         => (Int -> a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE izipWith #-}
-izipWith = G.izipWith
-
--- | Zip three vectors and their indices with the given function.
-izipWith3 :: (Prim a, Prim b, Prim c, Prim d)
-          => (Int -> a -> b -> c -> d)
-          -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE izipWith3 #-}
-izipWith3 = G.izipWith3
-
-izipWith4 :: (Prim a, Prim b, Prim c, Prim d, Prim e)
-          => (Int -> a -> b -> c -> d -> e)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE izipWith4 #-}
-izipWith4 = G.izipWith4
-
-izipWith5 :: (Prim a, Prim b, Prim c, Prim d, Prim e,
-              Prim f)
-          => (Int -> a -> b -> c -> d -> e -> f)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f
-{-# INLINE izipWith5 #-}
-izipWith5 = G.izipWith5
-
-izipWith6 :: (Prim a, Prim b, Prim c, Prim d, Prim e,
-              Prim f, Prim g)
-          => (Int -> a -> b -> c -> d -> e -> f -> g)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f -> Vector g
-{-# INLINE izipWith6 #-}
-izipWith6 = G.izipWith6
-
--- Monadic zipping
--- ---------------
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and yield a
--- vector of results
-zipWithM :: (Monad m, Prim a, Prim b, Prim c)
-         => (a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE zipWithM #-}
-zipWithM = G.zipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and ignore the
--- results
-zipWithM_ :: (Monad m, Prim a, Prim b)
-          => (a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ = G.zipWithM_
-
--- Filtering
--- ---------
-
--- | /O(n)/ Drop elements that do not satisfy the predicate
-filter :: Prim a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE filter #-}
-filter = G.filter
-
--- | /O(n)/ Drop elements that do not satisfy the predicate which is applied to
--- values and their indices
-ifilter :: Prim a => (Int -> a -> Bool) -> Vector a -> Vector a
-{-# INLINE ifilter #-}
-ifilter = G.ifilter
-
--- | /O(n)/ Drop repeated adjacent elements.
-uniq :: (Prim a, Eq a) => Vector a -> Vector a
-{-# INLINE uniq #-}
-uniq = G.uniq
-
--- | /O(n)/ Drop elements when predicate returns Nothing
-mapMaybe :: (Prim a, Prim b) => (a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE mapMaybe #-}
-mapMaybe = G.mapMaybe
-
--- | /O(n)/ Drop elements when predicate, applied to index and value, returns Nothing
-imapMaybe :: (Prim a, Prim b) => (Int -> a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE imapMaybe #-}
-imapMaybe = G.imapMaybe
-
--- | /O(n)/ Drop elements that do not satisfy the monadic predicate
-filterM :: (Monad m, Prim a) => (a -> m Bool) -> Vector a -> m (Vector a)
-{-# INLINE filterM #-}
-filterM = G.filterM
-
--- | /O(n)/ Yield the longest prefix of elements satisfying the predicate
--- without copying.
-takeWhile :: Prim a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE takeWhile #-}
-takeWhile = G.takeWhile
-
--- | /O(n)/ Drop the longest prefix of elements that satisfy the predicate
--- without copying.
-dropWhile :: Prim a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE dropWhile #-}
-dropWhile = G.dropWhile
-
--- Parititioning
--- -------------
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't. The
--- relative order of the elements is preserved at the cost of a sometimes
--- reduced performance compared to 'unstablePartition'.
-partition :: Prim a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE partition #-}
-partition = G.partition
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't.
--- The order of the elements is not preserved but the operation is often
--- faster than 'partition'.
-unstablePartition :: Prim a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE unstablePartition #-}
-unstablePartition = G.unstablePartition
-
--- | /O(n)/ Split the vector into the longest prefix of elements that satisfy
--- the predicate and the rest without copying.
-span :: Prim a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE span #-}
-span = G.span
-
--- | /O(n)/ Split the vector into the longest prefix of elements that do not
--- satisfy the predicate and the rest without copying.
-break :: Prim a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE break #-}
-break = G.break
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | /O(n)/ Check if the vector contains an element
-elem :: (Prim a, Eq a) => a -> Vector a -> Bool
-{-# INLINE elem #-}
-elem = G.elem
-
-infix 4 `notElem`
--- | /O(n)/ Check if the vector does not contain an element (inverse of 'elem')
-notElem :: (Prim a, Eq a) => a -> Vector a -> Bool
-{-# INLINE notElem #-}
-notElem = G.notElem
-
--- | /O(n)/ Yield 'Just' the first element matching the predicate or 'Nothing'
--- if no such element exists.
-find :: Prim a => (a -> Bool) -> Vector a -> Maybe a
-{-# INLINE find #-}
-find = G.find
-
--- | /O(n)/ Yield 'Just' the index of the first element matching the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Prim a => (a -> Bool) -> Vector a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex = G.findIndex
-
--- | /O(n)/ Yield the indices of elements satisfying the predicate in ascending
--- order.
-findIndices :: Prim a => (a -> Bool) -> Vector a -> Vector Int
-{-# INLINE findIndices #-}
-findIndices = G.findIndices
-
--- | /O(n)/ Yield 'Just' the index of the first occurence of the given element or
--- 'Nothing' if the vector does not contain the element. This is a specialised
--- version of 'findIndex'.
-elemIndex :: (Prim a, Eq a) => a -> Vector a -> Maybe Int
-{-# INLINE elemIndex #-}
-elemIndex = G.elemIndex
-
--- | /O(n)/ Yield the indices of all occurences of the given element in
--- ascending order. This is a specialised version of 'findIndices'.
-elemIndices :: (Prim a, Eq a) => a -> Vector a -> Vector Int
-{-# INLINE elemIndices #-}
-elemIndices = G.elemIndices
-
--- Folding
--- -------
-
--- | /O(n)/ Left fold
-foldl :: Prim b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl #-}
-foldl = G.foldl
-
--- | /O(n)/ Left fold on non-empty vectors
-foldl1 :: Prim a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1 #-}
-foldl1 = G.foldl1
-
--- | /O(n)/ Left fold with strict accumulator
-foldl' :: Prim b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl' #-}
-foldl' = G.foldl'
-
--- | /O(n)/ Left fold on non-empty vectors with strict accumulator
-foldl1' :: Prim a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1' #-}
-foldl1' = G.foldl1'
-
--- | /O(n)/ Right fold
-foldr :: Prim a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr #-}
-foldr = G.foldr
-
--- | /O(n)/ Right fold on non-empty vectors
-foldr1 :: Prim a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1 #-}
-foldr1 = G.foldr1
-
--- | /O(n)/ Right fold with a strict accumulator
-foldr' :: Prim a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr' #-}
-foldr' = G.foldr'
-
--- | /O(n)/ Right fold on non-empty vectors with strict accumulator
-foldr1' :: Prim a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1' #-}
-foldr1' = G.foldr1'
-
--- | /O(n)/ Left fold (function applied to each element and its index)
-ifoldl :: Prim b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl #-}
-ifoldl = G.ifoldl
-
--- | /O(n)/ Left fold with strict accumulator (function applied to each element
--- and its index)
-ifoldl' :: Prim b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl' #-}
-ifoldl' = G.ifoldl'
-
--- | /O(n)/ Right fold (function applied to each element and its index)
-ifoldr :: Prim a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr #-}
-ifoldr = G.ifoldr
-
--- | /O(n)/ Right fold with strict accumulator (function applied to each
--- element and its index)
-ifoldr' :: Prim a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr' #-}
-ifoldr' = G.ifoldr'
-
--- Specialised folds
--- -----------------
-
--- | /O(n)/ Check if all elements satisfy the predicate.
-all :: Prim a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE all #-}
-all = G.all
-
--- | /O(n)/ Check if any element satisfies the predicate.
-any :: Prim a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE any #-}
-any = G.any
-
--- | /O(n)/ Compute the sum of the elements
-sum :: (Prim a, Num a) => Vector a -> a
-{-# INLINE sum #-}
-sum = G.sum
-
--- | /O(n)/ Compute the produce of the elements
-product :: (Prim a, Num a) => Vector a -> a
-{-# INLINE product #-}
-product = G.product
-
--- | /O(n)/ Yield the maximum element of the vector. The vector may not be
--- empty.
-maximum :: (Prim a, Ord a) => Vector a -> a
-{-# INLINE maximum #-}
-maximum = G.maximum
-
--- | /O(n)/ Yield the maximum element of the vector according to the given
--- comparison function. The vector may not be empty.
-maximumBy :: Prim a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE maximumBy #-}
-maximumBy = G.maximumBy
-
--- | /O(n)/ Yield the minimum element of the vector. The vector may not be
--- empty.
-minimum :: (Prim a, Ord a) => Vector a -> a
-{-# INLINE minimum #-}
-minimum = G.minimum
-
--- | /O(n)/ Yield the minimum element of the vector according to the given
--- comparison function. The vector may not be empty.
-minimumBy :: Prim a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE minimumBy #-}
-minimumBy = G.minimumBy
-
--- | /O(n)/ Yield the index of the maximum element of the vector. The vector
--- may not be empty.
-maxIndex :: (Prim a, Ord a) => Vector a -> Int
-{-# INLINE maxIndex #-}
-maxIndex = G.maxIndex
-
--- | /O(n)/ Yield the index of the maximum element of the vector according to
--- the given comparison function. The vector may not be empty.
-maxIndexBy :: Prim a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE maxIndexBy #-}
-maxIndexBy = G.maxIndexBy
-
--- | /O(n)/ Yield the index of the minimum element of the vector. The vector
--- may not be empty.
-minIndex :: (Prim a, Ord a) => Vector a -> Int
-{-# INLINE minIndex #-}
-minIndex = G.minIndex
-
--- | /O(n)/ Yield the index of the minimum element of the vector according to
--- the given comparison function. The vector may not be empty.
-minIndexBy :: Prim a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE minIndexBy #-}
-minIndexBy = G.minIndexBy
-
--- Monadic folds
--- -------------
-
--- | /O(n)/ Monadic fold
-foldM :: (Monad m, Prim b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM #-}
-foldM = G.foldM
-
--- | /O(n)/ Monadic fold over non-empty vectors
-fold1M :: (Monad m, Prim a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M #-}
-fold1M = G.fold1M
-
--- | /O(n)/ Monadic fold with strict accumulator
-foldM' :: (Monad m, Prim b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM' #-}
-foldM' = G.foldM'
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
-fold1M' :: (Monad m, Prim a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = G.fold1M'
-
--- | /O(n)/ Monadic fold that discards the result
-foldM_ :: (Monad m, Prim b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM_ #-}
-foldM_ = G.foldM_
-
--- | /O(n)/ Monadic fold over non-empty vectors that discards the result
-fold1M_ :: (Monad m, Prim a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M_ #-}
-fold1M_ = G.fold1M_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
-foldM'_ :: (Monad m, Prim b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM'_ #-}
-foldM'_ = G.foldM'_
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
--- that discards the result
-fold1M'_ :: (Monad m, Prim a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M'_ #-}
-fold1M'_ = G.fold1M'_
-
--- Prefix sums (scans)
--- -------------------
-
--- | /O(n)/ Prescan
---
--- @
--- prescanl f z = 'init' . 'scanl' f z
--- @
---
--- Example: @prescanl (+) 0 \<1,2,3,4\> = \<0,1,3,6\>@
---
-prescanl :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl #-}
-prescanl = G.prescanl
-
--- | /O(n)/ Prescan with strict accumulator
-prescanl' :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl' #-}
-prescanl' = G.prescanl'
-
--- | /O(n)/ Scan
---
--- @
--- postscanl f z = 'tail' . 'scanl' f z
--- @
---
--- Example: @postscanl (+) 0 \<1,2,3,4\> = \<1,3,6,10\>@
---
-postscanl :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl #-}
-postscanl = G.postscanl
-
--- | /O(n)/ Scan with strict accumulator
-postscanl' :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl' #-}
-postscanl' = G.postscanl'
-
--- | /O(n)/ Haskell-style scan
---
--- > scanl f z <x1,...,xn> = <y1,...,y(n+1)>
--- >   where y1 = z
--- >         yi = f y(i-1) x(i-1)
---
--- Example: @scanl (+) 0 \<1,2,3,4\> = \<0,1,3,6,10\>@
---
-scanl :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl #-}
-scanl = G.scanl
-
--- | /O(n)/ Haskell-style scan with strict accumulator
-scanl' :: (Prim a, Prim b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl' #-}
-scanl' = G.scanl'
-
--- | /O(n)/ Scan over a non-empty vector
---
--- > scanl f <x1,...,xn> = <y1,...,yn>
--- >   where y1 = x1
--- >         yi = f y(i-1) xi
---
-scanl1 :: Prim a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1 #-}
-scanl1 = G.scanl1
-
--- | /O(n)/ Scan over a non-empty vector with a strict accumulator
-scanl1' :: Prim a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1' #-}
-scanl1' = G.scanl1'
-
--- | /O(n)/ Right-to-left prescan
---
--- @
--- prescanr f z = 'reverse' . 'prescanl' (flip f) z . 'reverse'
--- @
---
-prescanr :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr #-}
-prescanr = G.prescanr
-
--- | /O(n)/ Right-to-left prescan with strict accumulator
-prescanr' :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr' #-}
-prescanr' = G.prescanr'
-
--- | /O(n)/ Right-to-left scan
-postscanr :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr #-}
-postscanr = G.postscanr
-
--- | /O(n)/ Right-to-left scan with strict accumulator
-postscanr' :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr' #-}
-postscanr' = G.postscanr'
-
--- | /O(n)/ Right-to-left Haskell-style scan
-scanr :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr #-}
-scanr = G.scanr
-
--- | /O(n)/ Right-to-left Haskell-style scan with strict accumulator
-scanr' :: (Prim a, Prim b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr' #-}
-scanr' = G.scanr'
-
--- | /O(n)/ Right-to-left scan over a non-empty vector
-scanr1 :: Prim a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1 #-}
-scanr1 = G.scanr1
-
--- | /O(n)/ Right-to-left scan over a non-empty vector with a strict
--- accumulator
-scanr1' :: Prim a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1' #-}
-scanr1' = G.scanr1'
-
--- Conversions - Lists
--- ------------------------
-
--- | /O(n)/ Convert a vector to a list
-toList :: Prim a => Vector a -> [a]
-{-# INLINE toList #-}
-toList = G.toList
-
--- | /O(n)/ Convert a list to a vector
-fromList :: Prim a => [a] -> Vector a
-{-# INLINE fromList #-}
-fromList = G.fromList
-
--- | /O(n)/ Convert the first @n@ elements of a list to a vector
---
--- @
--- fromListN n xs = 'fromList' ('take' n xs)
--- @
-fromListN :: Prim a => Int -> [a] -> Vector a
-{-# INLINE fromListN #-}
-fromListN = G.fromListN
-
--- Conversions - Mutable vectors
--- -----------------------------
-
--- | /O(1)/ Unsafe convert a mutable vector to an immutable one without
--- copying. The mutable vector may not be used after this operation.
-unsafeFreeze :: (Prim a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE unsafeFreeze #-}
-unsafeFreeze = G.unsafeFreeze
-
--- | /O(1)/ Unsafely convert an immutable vector to a mutable one without
--- copying. The immutable vector may not be used after this operation.
-unsafeThaw :: (Prim a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE unsafeThaw #-}
-unsafeThaw = G.unsafeThaw
-
--- | /O(n)/ Yield a mutable copy of the immutable vector.
-thaw :: (Prim a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE thaw #-}
-thaw = G.thaw
-
--- | /O(n)/ Yield an immutable copy of the mutable vector.
-freeze :: (Prim a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE freeze #-}
-freeze = G.freeze
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length. This is not checked.
-unsafeCopy
-  :: (Prim a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length.
-copy :: (Prim a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE copy #-}
-copy = G.copy
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive/Mutable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive/Mutable.hs
deleted file mode 100644
index 33aca812e208..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Primitive/Mutable.hs
+++ /dev/null
@@ -1,366 +0,0 @@
-{-# LANGUAGE CPP, DeriveDataTypeable, MultiParamTypeClasses, FlexibleInstances, ScopedTypeVariables #-}
-
--- |
--- Module      : Data.Vector.Primitive.Mutable
--- Copyright   : (c) Roman Leshchinskiy 2008-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Mutable primitive vectors.
---
-
-module Data.Vector.Primitive.Mutable (
-  -- * Mutable vectors of primitive types
-  MVector(..), IOVector, STVector, Prim,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Extracting subvectors
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- ** Overlapping
-  overlaps,
-
-  -- * Construction
-
-  -- ** Initialisation
-  new, unsafeNew, replicate, replicateM, clone,
-
-  -- ** Growing
-  grow, unsafeGrow,
-
-  -- ** Restricting memory usage
-  clear,
-
-  -- * Accessing individual elements
-  read, write, modify, swap,
-  unsafeRead, unsafeWrite, unsafeModify, unsafeSwap,
-
-  -- * Modifying vectors
-  nextPermutation,
-
-  -- ** Filling and copying
-  set, copy, move, unsafeCopy, unsafeMove
-) where
-
-import qualified Data.Vector.Generic.Mutable as G
-import           Data.Primitive.ByteArray
-import           Data.Primitive ( Prim, sizeOf )
-import           Data.Word ( Word8 )
-import           Control.Monad.Primitive
-import           Control.Monad ( liftM )
-
-import Control.DeepSeq ( NFData(rnf) )
-
-import Prelude hiding ( length, null, replicate, reverse, map, read,
-                        take, drop, splitAt, init, tail )
-
-import Data.Typeable ( Typeable )
-
--- Data.Vector.Internal.Check is unnecessary
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- | Mutable vectors of primitive types.
-data MVector s a = MVector {-# UNPACK #-} !Int
-                           {-# UNPACK #-} !Int
-                           {-# UNPACK #-} !(MutableByteArray s) -- ^ offset, length, underlying mutable byte array
-        deriving ( Typeable )
-
-type IOVector = MVector RealWorld
-type STVector s = MVector s
-
-instance NFData (MVector s a) where
-  rnf (MVector _ _ _) = ()
-
-instance Prim a => G.MVector MVector a where
-  basicLength (MVector _ n _) = n
-  basicUnsafeSlice j m (MVector i _ arr)
-    = MVector (i+j) m arr
-
-  {-# INLINE basicOverlaps #-}
-  basicOverlaps (MVector i m arr1) (MVector j n arr2)
-    = sameMutableByteArray arr1 arr2
-      && (between i j (j+n) || between j i (i+m))
-    where
-      between x y z = x >= y && x < z
-
-  {-# INLINE basicUnsafeNew #-}
-  basicUnsafeNew n
-    | n < 0 = error $ "Primitive.basicUnsafeNew: negative length: " ++ show n
-    | n > mx = error $ "Primitive.basicUnsafeNew: length to large: " ++ show n
-    | otherwise = MVector 0 n `liftM` newByteArray (n * size)
-    where
-      size = sizeOf (undefined :: a)
-      mx = maxBound `div` size :: Int
-
-  {-# INLINE basicInitialize #-}
-  basicInitialize (MVector off n v) =
-      setByteArray v (off * size) (n * size) (0 :: Word8)
-    where
-      size = sizeOf (undefined :: a)
-
-
-  {-# INLINE basicUnsafeRead #-}
-  basicUnsafeRead (MVector i _ arr) j = readByteArray arr (i+j)
-
-  {-# INLINE basicUnsafeWrite #-}
-  basicUnsafeWrite (MVector i _ arr) j x = writeByteArray arr (i+j) x
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector i n dst) (MVector j _ src)
-    = copyMutableByteArray dst (i*sz) src (j*sz) (n*sz)
-    where
-      sz = sizeOf (undefined :: a)
-
-  {-# INLINE basicUnsafeMove #-}
-  basicUnsafeMove (MVector i n dst) (MVector j _ src)
-    = moveByteArray dst (i*sz) src (j*sz) (n * sz)
-    where
-      sz = sizeOf (undefined :: a)
-
-  {-# INLINE basicSet #-}
-  basicSet (MVector i n arr) x = setByteArray arr i n x
-
--- Length information
--- ------------------
-
--- | Length of the mutable vector.
-length :: Prim a => MVector s a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | Check whether the vector is empty
-null :: Prim a => MVector s a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Extracting subvectors
--- ---------------------
-
--- | Yield a part of the mutable vector without copying it.
-slice :: Prim a => Int -> Int -> MVector s a -> MVector s a
-{-# INLINE slice #-}
-slice = G.slice
-
-take :: Prim a => Int -> MVector s a -> MVector s a
-{-# INLINE take #-}
-take = G.take
-
-drop :: Prim a => Int -> MVector s a -> MVector s a
-{-# INLINE drop #-}
-drop = G.drop
-
-splitAt :: Prim a => Int -> MVector s a -> (MVector s a, MVector s a)
-{-# INLINE splitAt #-}
-splitAt = G.splitAt
-
-init :: Prim a => MVector s a -> MVector s a
-{-# INLINE init #-}
-init = G.init
-
-tail :: Prim a => MVector s a -> MVector s a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | Yield a part of the mutable vector without copying it. No bounds checks
--- are performed.
-unsafeSlice :: Prim a
-            => Int  -- ^ starting index
-            -> Int  -- ^ length of the slice
-            -> MVector s a
-            -> MVector s a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
-unsafeTake :: Prim a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
-unsafeDrop :: Prim a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
-unsafeInit :: Prim a => MVector s a -> MVector s a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
-unsafeTail :: Prim a => MVector s a -> MVector s a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- Overlapping
--- -----------
-
--- | Check whether two vectors overlap.
-overlaps :: Prim a => MVector s a -> MVector s a -> Bool
-{-# INLINE overlaps #-}
-overlaps = G.overlaps
-
--- Initialisation
--- --------------
-
--- | Create a mutable vector of the given length.
-new :: (PrimMonad m, Prim a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE new #-}
-new = G.new
-
--- | Create a mutable vector of the given length. The memory is not initialized.
-unsafeNew :: (PrimMonad m, Prim a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeNew #-}
-unsafeNew = G.unsafeNew
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with an initial value.
-replicate :: (PrimMonad m, Prim a) => Int -> a -> m (MVector (PrimState m) a)
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with values produced by repeatedly executing the monadic action.
-replicateM :: (PrimMonad m, Prim a) => Int -> m a -> m (MVector (PrimState m) a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | Create a copy of a mutable vector.
-clone :: (PrimMonad m, Prim a)
-      => MVector (PrimState m) a -> m (MVector (PrimState m) a)
-{-# INLINE clone #-}
-clone = G.clone
-
--- Growing
--- -------
-
--- | Grow a vector by the given number of elements. The number must be
--- positive.
-grow :: (PrimMonad m, Prim a)
-              => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE grow #-}
-grow = G.grow
-
--- | Grow a vector by the given number of elements. The number must be
--- positive but this is not checked.
-unsafeGrow :: (PrimMonad m, Prim a)
-               => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeGrow #-}
-unsafeGrow = G.unsafeGrow
-
--- Restricting memory usage
--- ------------------------
-
--- | Reset all elements of the vector to some undefined value, clearing all
--- references to external objects. This is usually a noop for unboxed vectors.
-clear :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> m ()
-{-# INLINE clear #-}
-clear = G.clear
-
--- Accessing individual elements
--- -----------------------------
-
--- | Yield the element at the given position.
-read :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE read #-}
-read = G.read
-
--- | Replace the element at the given position.
-write :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE write #-}
-write = G.write
-
--- | Modify the element at the given position.
-modify :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE modify #-}
-modify = G.modify
-
--- | Swap the elements at the given positions.
-swap :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE swap #-}
-swap = G.swap
-
-
--- | Yield the element at the given position. No bounds checks are performed.
-unsafeRead :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE unsafeRead #-}
-unsafeRead = G.unsafeRead
-
--- | Replace the element at the given position. No bounds checks are performed.
-unsafeWrite
-    :: (PrimMonad m, Prim a) =>  MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE unsafeWrite #-}
-unsafeWrite = G.unsafeWrite
-
--- | Modify the element at the given position. No bounds checks are performed.
-unsafeModify :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE unsafeModify #-}
-unsafeModify = G.unsafeModify
-
--- | Swap the elements at the given positions. No bounds checks are performed.
-unsafeSwap
-    :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE unsafeSwap #-}
-unsafeSwap = G.unsafeSwap
-
--- Filling and copying
--- -------------------
-
--- | Set all elements of the vector to the given value.
-set :: (PrimMonad m, Prim a) => MVector (PrimState m) a -> a -> m ()
-{-# INLINE set #-}
-set = G.set
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap.
-copy :: (PrimMonad m, Prim a)
-     => MVector (PrimState m) a   -- ^ target
-     -> MVector (PrimState m) a   -- ^ source
-     -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap. This is not checked.
-unsafeCopy :: (PrimMonad m, Prim a)
-           => MVector (PrimState m) a   -- ^ target
-           -> MVector (PrimState m) a   -- ^ source
-           -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | Move the contents of a vector. The two vectors must have the same
--- length.
---
--- If the vectors do not overlap, then this is equivalent to 'copy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-move :: (PrimMonad m, Prim a)
-                 => MVector (PrimState m) a -> MVector (PrimState m) a -> m ()
-{-# INLINE move #-}
-move = G.move
-
--- | Move the contents of a vector. The two vectors must have the same
--- length, but this is not checked.
---
--- If the vectors do not overlap, then this is equivalent to 'unsafeCopy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-unsafeMove :: (PrimMonad m, Prim a)
-                          => MVector (PrimState m) a   -- ^ target
-                          -> MVector (PrimState m) a   -- ^ source
-                          -> m ()
-{-# INLINE unsafeMove #-}
-unsafeMove = G.unsafeMove
-
--- | Compute the next (lexicographically) permutation of given vector in-place.
---   Returns False when input is the last permtuation
-nextPermutation :: (PrimMonad m,Ord e,Prim e) => MVector (PrimState m) e -> m Bool
-{-# INLINE nextPermutation #-}
-nextPermutation = G.nextPermutation
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable.hs
deleted file mode 100644
index 30c9a4615c60..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable.hs
+++ /dev/null
@@ -1,1489 +0,0 @@
-{-# LANGUAGE CPP, DeriveDataTypeable, MultiParamTypeClasses, FlexibleInstances, TypeFamilies, Rank2Types, ScopedTypeVariables #-}
-
--- |
--- Module      : Data.Vector.Storable
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- 'Storable'-based vectors.
---
-
-module Data.Vector.Storable (
-  -- * Storable vectors
-  Vector, MVector(..), Storable,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Indexing
-  (!), (!?), head, last,
-  unsafeIndex, unsafeHead, unsafeLast,
-
-  -- ** Monadic indexing
-  indexM, headM, lastM,
-  unsafeIndexM, unsafeHeadM, unsafeLastM,
-
-  -- ** Extracting subvectors (slicing)
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- * Construction
-
-  -- ** Initialisation
-  empty, singleton, replicate, generate, iterateN,
-
-  -- ** Monadic initialisation
-  replicateM, generateM, iterateNM, create, createT,
-
-  -- ** Unfolding
-  unfoldr, unfoldrN,
-  unfoldrM, unfoldrNM,
-  constructN, constructrN,
-
-  -- ** Enumeration
-  enumFromN, enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- ** Concatenation
-  cons, snoc, (++), concat,
-
-  -- ** Restricting memory usage
-  force,
-
-  -- * Modifying vectors
-
-  -- ** Bulk updates
-  (//), update_,
-  unsafeUpd, unsafeUpdate_,
-
-  -- ** Accumulations
-  accum, accumulate_,
-  unsafeAccum, unsafeAccumulate_,
-
-  -- ** Permutations
-  reverse, backpermute, unsafeBackpermute,
-
-  -- ** Safe destructive updates
-  modify,
-
-  -- * Elementwise operations
-
-  -- ** Mapping
-  map, imap, concatMap,
-
-  -- ** Monadic mapping
-  mapM, mapM_, forM, forM_,
-
-  -- ** Zipping
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  izipWith, izipWith3, izipWith4, izipWith5, izipWith6,
-
-  -- ** Monadic zipping
-  zipWithM, zipWithM_,
-
-  -- * Working with predicates
-
-  -- ** Filtering
-  filter, ifilter, uniq,
-  mapMaybe, imapMaybe,
-  filterM,
-  takeWhile, dropWhile,
-
-  -- ** Partitioning
-  partition, unstablePartition, span, break,
-
-  -- ** Searching
-  elem, notElem, find, findIndex, findIndices, elemIndex, elemIndices,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1, foldr', foldr1',
-  ifoldl, ifoldl', ifoldr, ifoldr',
-
-  -- ** Specialised folds
-  all, any, and, or,
-  sum, product,
-  maximum, maximumBy, minimum, minimumBy,
-  minIndex, minIndexBy, maxIndex, maxIndexBy,
-
-  -- ** Monadic folds
-  foldM, foldM', fold1M, fold1M',
-  foldM_, foldM'_, fold1M_, fold1M'_,
-
-  -- * Prefix sums (scans)
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl', scanl1, scanl1',
-  prescanr, prescanr',
-  postscanr, postscanr',
-  scanr, scanr', scanr1, scanr1',
-
-  -- * Conversions
-
-  -- ** Lists
-  toList, fromList, fromListN,
-
-  -- ** Other vector types
-  G.convert, unsafeCast,
-
-  -- ** Mutable vectors
-  freeze, thaw, copy, unsafeFreeze, unsafeThaw, unsafeCopy,
-
-  -- * Raw pointers
-  unsafeFromForeignPtr, unsafeFromForeignPtr0,
-  unsafeToForeignPtr,   unsafeToForeignPtr0,
-  unsafeWith
-) where
-
-import qualified Data.Vector.Generic          as G
-import           Data.Vector.Storable.Mutable ( MVector(..) )
-import Data.Vector.Storable.Internal
-import qualified Data.Vector.Fusion.Bundle as Bundle
-
-import Foreign.Storable
-import Foreign.ForeignPtr
-import Foreign.Ptr
-import Foreign.Marshal.Array ( advancePtr, copyArray )
-
-import Control.DeepSeq ( NFData(rnf) )
-
-import Control.Monad.ST ( ST )
-import Control.Monad.Primitive
-
-import Prelude hiding ( length, null,
-                        replicate, (++), concat,
-                        head, last,
-                        init, tail, take, drop, splitAt, reverse,
-                        map, concatMap,
-                        zipWith, zipWith3, zip, zip3, unzip, unzip3,
-                        filter, takeWhile, dropWhile, span, break,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        all, any, and, or, sum, product, minimum, maximum,
-                        scanl, scanl1, scanr, scanr1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_ )
-
-import Data.Typeable  ( Typeable )
-import Data.Data      ( Data(..) )
-import Text.Read      ( Read(..), readListPrecDefault )
-import Data.Semigroup ( Semigroup(..) )
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Monoid   ( Monoid(..) )
-import Data.Traversable ( Traversable )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-import qualified GHC.Exts as Exts
-#endif
-
--- Data.Vector.Internal.Check is unused
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- | 'Storable'-based vectors
-data Vector a = Vector {-# UNPACK #-} !Int
-                       {-# UNPACK #-} !(ForeignPtr a)
-        deriving ( Typeable )
-
-instance NFData (Vector a) where
-  rnf (Vector _ _) = ()
-
-instance (Show a, Storable a) => Show (Vector a) where
-  showsPrec = G.showsPrec
-
-instance (Read a, Storable a) => Read (Vector a) where
-  readPrec = G.readPrec
-  readListPrec = readListPrecDefault
-
-instance (Data a, Storable a) => Data (Vector a) where
-  gfoldl       = G.gfoldl
-  toConstr _   = error "toConstr"
-  gunfold _ _  = error "gunfold"
-  dataTypeOf _ = G.mkType "Data.Vector.Storable.Vector"
-  dataCast1    = G.dataCast
-
-type instance G.Mutable Vector = MVector
-
-instance Storable a => G.Vector Vector a where
-  {-# INLINE basicUnsafeFreeze #-}
-  basicUnsafeFreeze (MVector n fp) = return $ Vector n fp
-
-  {-# INLINE basicUnsafeThaw #-}
-  basicUnsafeThaw (Vector n fp) = return $ MVector n fp
-
-  {-# INLINE basicLength #-}
-  basicLength (Vector n _) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice i n (Vector _ fp) = Vector n (updPtr (`advancePtr` i) fp)
-
-  {-# INLINE basicUnsafeIndexM #-}
-  basicUnsafeIndexM (Vector _ fp) i = return
-                                    . unsafeInlineIO
-                                    $ withForeignPtr fp $ \p ->
-                                      peekElemOff p i
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector n fp) (Vector _ fq)
-    = unsafePrimToPrim
-    $ withForeignPtr fp $ \p ->
-      withForeignPtr fq $ \q ->
-      copyArray p q n
-
-  {-# INLINE elemseq #-}
-  elemseq _ = seq
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Storable a, Eq a) => Eq (Vector a) where
-  {-# INLINE (==) #-}
-  xs == ys = Bundle.eq (G.stream xs) (G.stream ys)
-
-  {-# INLINE (/=) #-}
-  xs /= ys = not (Bundle.eq (G.stream xs) (G.stream ys))
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Storable a, Ord a) => Ord (Vector a) where
-  {-# INLINE compare #-}
-  compare xs ys = Bundle.cmp (G.stream xs) (G.stream ys)
-
-  {-# INLINE (<) #-}
-  xs < ys = Bundle.cmp (G.stream xs) (G.stream ys) == LT
-
-  {-# INLINE (<=) #-}
-  xs <= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= GT
-
-  {-# INLINE (>) #-}
-  xs > ys = Bundle.cmp (G.stream xs) (G.stream ys) == GT
-
-  {-# INLINE (>=) #-}
-  xs >= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= LT
-
-instance Storable a => Semigroup (Vector a) where
-  {-# INLINE (<>) #-}
-  (<>) = (++)
-
-  {-# INLINE sconcat #-}
-  sconcat = G.concatNE
-
-instance Storable a => Monoid (Vector a) where
-  {-# INLINE mempty #-}
-  mempty = empty
-
-  {-# INLINE mappend #-}
-  mappend = (++)
-
-  {-# INLINE mconcat #-}
-  mconcat = concat
-
-#if __GLASGOW_HASKELL__ >= 708
-
-instance Storable a => Exts.IsList (Vector a) where
-  type Item (Vector a) = a
-  fromList = fromList
-  fromListN = fromListN
-  toList = toList
-
-#endif
-
--- Length
--- ------
-
--- | /O(1)/ Yield the length of the vector
-length :: Storable a => Vector a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | /O(1)/ Test whether a vector is empty
-null :: Storable a => Vector a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Indexing
--- --------
-
--- | O(1) Indexing
-(!) :: Storable a => Vector a -> Int -> a
-{-# INLINE (!) #-}
-(!) = (G.!)
-
--- | O(1) Safe indexing
-(!?) :: Storable a => Vector a -> Int -> Maybe a
-{-# INLINE (!?) #-}
-(!?) = (G.!?)
-
--- | /O(1)/ First element
-head :: Storable a => Vector a -> a
-{-# INLINE head #-}
-head = G.head
-
--- | /O(1)/ Last element
-last :: Storable a => Vector a -> a
-{-# INLINE last #-}
-last = G.last
-
--- | /O(1)/ Unsafe indexing without bounds checking
-unsafeIndex :: Storable a => Vector a -> Int -> a
-{-# INLINE unsafeIndex #-}
-unsafeIndex = G.unsafeIndex
-
--- | /O(1)/ First element without checking if the vector is empty
-unsafeHead :: Storable a => Vector a -> a
-{-# INLINE unsafeHead #-}
-unsafeHead = G.unsafeHead
-
--- | /O(1)/ Last element without checking if the vector is empty
-unsafeLast :: Storable a => Vector a -> a
-{-# INLINE unsafeLast #-}
-unsafeLast = G.unsafeLast
-
--- Monadic indexing
--- ----------------
-
--- | /O(1)/ Indexing in a monad.
---
--- The monad allows operations to be strict in the vector when necessary.
--- Suppose vector copying is implemented like this:
---
--- > copy mv v = ... write mv i (v ! i) ...
---
--- For lazy vectors, @v ! i@ would not be evaluated which means that @mv@
--- would unnecessarily retain a reference to @v@ in each element written.
---
--- With 'indexM', copying can be implemented like this instead:
---
--- > copy mv v = ... do
--- >                   x <- indexM v i
--- >                   write mv i x
---
--- Here, no references to @v@ are retained because indexing (but /not/ the
--- elements) is evaluated eagerly.
---
-indexM :: (Storable a, Monad m) => Vector a -> Int -> m a
-{-# INLINE indexM #-}
-indexM = G.indexM
-
--- | /O(1)/ First element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-headM :: (Storable a, Monad m) => Vector a -> m a
-{-# INLINE headM #-}
-headM = G.headM
-
--- | /O(1)/ Last element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-lastM :: (Storable a, Monad m) => Vector a -> m a
-{-# INLINE lastM #-}
-lastM = G.lastM
-
--- | /O(1)/ Indexing in a monad without bounds checks. See 'indexM' for an
--- explanation of why this is useful.
-unsafeIndexM :: (Storable a, Monad m) => Vector a -> Int -> m a
-{-# INLINE unsafeIndexM #-}
-unsafeIndexM = G.unsafeIndexM
-
--- | /O(1)/ First element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeHeadM :: (Storable a, Monad m) => Vector a -> m a
-{-# INLINE unsafeHeadM #-}
-unsafeHeadM = G.unsafeHeadM
-
--- | /O(1)/ Last element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeLastM :: (Storable a, Monad m) => Vector a -> m a
-{-# INLINE unsafeLastM #-}
-unsafeLastM = G.unsafeLastM
-
--- Extracting subvectors (slicing)
--- -------------------------------
-
--- | /O(1)/ Yield a slice of the vector without copying it. The vector must
--- contain at least @i+n@ elements.
-slice :: Storable a
-      => Int   -- ^ @i@ starting index
-      -> Int   -- ^ @n@ length
-      -> Vector a
-      -> Vector a
-{-# INLINE slice #-}
-slice = G.slice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty.
-init :: Storable a => Vector a -> Vector a
-{-# INLINE init #-}
-init = G.init
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty.
-tail :: Storable a => Vector a -> Vector a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | /O(1)/ Yield at the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case it is returned unchanged.
-take :: Storable a => Int -> Vector a -> Vector a
-{-# INLINE take #-}
-take = G.take
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case an empty vector is returned.
-drop :: Storable a => Int -> Vector a -> Vector a
-{-# INLINE drop #-}
-drop = G.drop
-
--- | /O(1)/ Yield the first @n@ elements paired with the remainder without copying.
---
--- Note that @'splitAt' n v@ is equivalent to @('take' n v, 'drop' n v)@
--- but slightly more efficient.
-{-# INLINE splitAt #-}
-splitAt :: Storable a => Int -> Vector a -> (Vector a, Vector a)
-splitAt = G.splitAt
-
--- | /O(1)/ Yield a slice of the vector without copying. The vector must
--- contain at least @i+n@ elements but this is not checked.
-unsafeSlice :: Storable a => Int   -- ^ @i@ starting index
-                       -> Int   -- ^ @n@ length
-                       -> Vector a
-                       -> Vector a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty but this is not checked.
-unsafeInit :: Storable a => Vector a -> Vector a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty but this is not checked.
-unsafeTail :: Storable a => Vector a -> Vector a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector must
--- contain at least @n@ elements but this is not checked.
-unsafeTake :: Storable a => Int -> Vector a -> Vector a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector
--- must contain at least @n@ elements but this is not checked.
-unsafeDrop :: Storable a => Int -> Vector a -> Vector a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
--- Initialisation
--- --------------
-
--- | /O(1)/ Empty vector
-empty :: Storable a => Vector a
-{-# INLINE empty #-}
-empty = G.empty
-
--- | /O(1)/ Vector with exactly one element
-singleton :: Storable a => a -> Vector a
-{-# INLINE singleton #-}
-singleton = G.singleton
-
--- | /O(n)/ Vector of the given length with the same value in each position
-replicate :: Storable a => Int -> a -> Vector a
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | /O(n)/ Construct a vector of the given length by applying the function to
--- each index
-generate :: Storable a => Int -> (Int -> a) -> Vector a
-{-# INLINE generate #-}
-generate = G.generate
-
--- | /O(n)/ Apply function n times to value. Zeroth element is original value.
-iterateN :: Storable a => Int -> (a -> a) -> a -> Vector a
-{-# INLINE iterateN #-}
-iterateN = G.iterateN
-
--- Unfolding
--- ---------
-
--- | /O(n)/ Construct a vector by repeatedly applying the generator function
--- to a seed. The generator function yields 'Just' the next element and the
--- new seed or 'Nothing' if there are no more elements.
---
--- > unfoldr (\n -> if n == 0 then Nothing else Just (n,n-1)) 10
--- >  = <10,9,8,7,6,5,4,3,2,1>
-unfoldr :: Storable a => (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldr #-}
-unfoldr = G.unfoldr
-
--- | /O(n)/ Construct a vector with at most @n@ elements by repeatedly applying
--- the generator function to a seed. The generator function yields 'Just' the
--- next element and the new seed or 'Nothing' if there are no more elements.
---
--- > unfoldrN 3 (\n -> Just (n,n-1)) 10 = <10,9,8>
-unfoldrN :: Storable a => Int -> (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldrN #-}
-unfoldrN = G.unfoldrN
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrM :: (Monad m, Storable a) => (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrM #-}
-unfoldrM = G.unfoldrM
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrNM :: (Monad m, Storable a) => Int -> (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrNM #-}
-unfoldrNM = G.unfoldrNM
-
--- | /O(n)/ Construct a vector with @n@ elements by repeatedly applying the
--- generator function to the already constructed part of the vector.
---
--- > constructN 3 f = let a = f <> ; b = f <a> ; c = f <a,b> in f <a,b,c>
---
-constructN :: Storable a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructN #-}
-constructN = G.constructN
-
--- | /O(n)/ Construct a vector with @n@ elements from right to left by
--- repeatedly applying the generator function to the already constructed part
--- of the vector.
---
--- > constructrN 3 f = let a = f <> ; b = f<a> ; c = f <b,a> in f <c,b,a>
---
-constructrN :: Storable a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructrN #-}
-constructrN = G.constructrN
-
--- Enumeration
--- -----------
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+1@
--- etc. This operation is usually more efficient than 'enumFromTo'.
---
--- > enumFromN 5 3 = <5,6,7>
-enumFromN :: (Storable a, Num a) => a -> Int -> Vector a
-{-# INLINE enumFromN #-}
-enumFromN = G.enumFromN
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc. This operations is usually more efficient than 'enumFromThenTo'.
---
--- > enumFromStepN 1 0.1 5 = <1,1.1,1.2,1.3,1.4>
-enumFromStepN :: (Storable a, Num a) => a -> a -> Int -> Vector a
-{-# INLINE enumFromStepN #-}
-enumFromStepN = G.enumFromStepN
-
--- | /O(n)/ Enumerate values from @x@ to @y@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromN' instead.
-enumFromTo :: (Storable a, Enum a) => a -> a -> Vector a
-{-# INLINE enumFromTo #-}
-enumFromTo = G.enumFromTo
-
--- | /O(n)/ Enumerate values from @x@ to @y@ with a specific step @z@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Storable a, Enum a) => a -> a -> a -> Vector a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo = G.enumFromThenTo
-
--- Concatenation
--- -------------
-
--- | /O(n)/ Prepend an element
-cons :: Storable a => a -> Vector a -> Vector a
-{-# INLINE cons #-}
-cons = G.cons
-
--- | /O(n)/ Append an element
-snoc :: Storable a => Vector a -> a -> Vector a
-{-# INLINE snoc #-}
-snoc = G.snoc
-
-infixr 5 ++
--- | /O(m+n)/ Concatenate two vectors
-(++) :: Storable a => Vector a -> Vector a -> Vector a
-{-# INLINE (++) #-}
-(++) = (G.++)
-
--- | /O(n)/ Concatenate all vectors in the list
-concat :: Storable a => [Vector a] -> Vector a
-{-# INLINE concat #-}
-concat = G.concat
-
--- Monadic initialisation
--- ----------------------
-
--- | /O(n)/ Execute the monadic action the given number of times and store the
--- results in a vector.
-replicateM :: (Monad m, Storable a) => Int -> m a -> m (Vector a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | /O(n)/ Construct a vector of the given length by applying the monadic
--- action to each index
-generateM :: (Monad m, Storable a) => Int -> (Int -> m a) -> m (Vector a)
-{-# INLINE generateM #-}
-generateM = G.generateM
-
--- | /O(n)/ Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: (Monad m, Storable a) => Int -> (a -> m a) -> a -> m (Vector a)
-{-# INLINE iterateNM #-}
-iterateNM = G.iterateNM
-
--- | Execute the monadic action and freeze the resulting vector.
---
--- @
--- create (do { v \<- new 2; write v 0 \'a\'; write v 1 \'b\'; return v }) = \<'a','b'\>
--- @
-create :: Storable a => (forall s. ST s (MVector s a)) -> Vector a
-{-# INLINE create #-}
--- NOTE: eta-expanded due to http://hackage.haskell.org/trac/ghc/ticket/4120
-create p = G.create p
-
--- | Execute the monadic action and freeze the resulting vectors.
-createT :: (Traversable f, Storable a) => (forall s. ST s (f (MVector s a))) -> f (Vector a)
-{-# INLINE createT #-}
-createT p = G.createT p
-
--- Restricting memory usage
--- ------------------------
-
--- | /O(n)/ Yield the argument but force it not to retain any extra memory,
--- possibly by copying it.
---
--- This is especially useful when dealing with slices. For example:
---
--- > force (slice 0 2 <huge vector>)
---
--- Here, the slice retains a reference to the huge vector. Forcing it creates
--- a copy of just the elements that belong to the slice and allows the huge
--- vector to be garbage collected.
-force :: Storable a => Vector a -> Vector a
-{-# INLINE force #-}
-force = G.force
-
--- Bulk updates
--- ------------
-
--- | /O(m+n)/ For each pair @(i,a)@ from the list, replace the vector
--- element at position @i@ by @a@.
---
--- > <5,9,2,7> // [(2,1),(0,3),(2,8)] = <3,9,8,7>
---
-(//) :: Storable a => Vector a   -- ^ initial vector (of length @m@)
-                -> [(Int, a)] -- ^ list of index/value pairs (of length @n@)
-                -> Vector a
-{-# INLINE (//) #-}
-(//) = (G.//)
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @a@ from the value vector, replace the element of the
--- initial vector at position @i@ by @a@.
---
--- > update_ <5,9,2,7>  <2,0,2> <1,3,8> = <3,9,8,7>
---
-update_ :: Storable a
-        => Vector a   -- ^ initial vector (of length @m@)
-        -> Vector Int -- ^ index vector (of length @n1@)
-        -> Vector a   -- ^ value vector (of length @n2@)
-        -> Vector a
-{-# INLINE update_ #-}
-update_ = G.update_
-
--- | Same as ('//') but without bounds checking.
-unsafeUpd :: Storable a => Vector a -> [(Int, a)] -> Vector a
-{-# INLINE unsafeUpd #-}
-unsafeUpd = G.unsafeUpd
-
--- | Same as 'update_' but without bounds checking.
-unsafeUpdate_ :: Storable a => Vector a -> Vector Int -> Vector a -> Vector a
-{-# INLINE unsafeUpdate_ #-}
-unsafeUpdate_ = G.unsafeUpdate_
-
--- Accumulations
--- -------------
-
--- | /O(m+n)/ For each pair @(i,b)@ from the list, replace the vector element
--- @a@ at position @i@ by @f a b@.
---
--- > accum (+) <5,9,2> [(2,4),(1,6),(0,3),(1,7)] = <5+3, 9+6+7, 2+4>
-accum :: Storable a
-      => (a -> b -> a) -- ^ accumulating function @f@
-      -> Vector a      -- ^ initial vector (of length @m@)
-      -> [(Int,b)]     -- ^ list of index/value pairs (of length @n@)
-      -> Vector a
-{-# INLINE accum #-}
-accum = G.accum
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @b@ from the the value vector,
--- replace the element of the initial vector at
--- position @i@ by @f a b@.
---
--- > accumulate_ (+) <5,9,2> <2,1,0,1> <4,6,3,7> = <5+3, 9+6+7, 2+4>
---
-accumulate_ :: (Storable a, Storable b)
-            => (a -> b -> a) -- ^ accumulating function @f@
-            -> Vector a      -- ^ initial vector (of length @m@)
-            -> Vector Int    -- ^ index vector (of length @n1@)
-            -> Vector b      -- ^ value vector (of length @n2@)
-            -> Vector a
-{-# INLINE accumulate_ #-}
-accumulate_ = G.accumulate_
-
--- | Same as 'accum' but without bounds checking.
-unsafeAccum :: Storable a => (a -> b -> a) -> Vector a -> [(Int,b)] -> Vector a
-{-# INLINE unsafeAccum #-}
-unsafeAccum = G.unsafeAccum
-
--- | Same as 'accumulate_' but without bounds checking.
-unsafeAccumulate_ :: (Storable a, Storable b) =>
-               (a -> b -> a) -> Vector a -> Vector Int -> Vector b -> Vector a
-{-# INLINE unsafeAccumulate_ #-}
-unsafeAccumulate_ = G.unsafeAccumulate_
-
--- Permutations
--- ------------
-
--- | /O(n)/ Reverse a vector
-reverse :: Storable a => Vector a -> Vector a
-{-# INLINE reverse #-}
-reverse = G.reverse
-
--- | /O(n)/ Yield the vector obtained by replacing each element @i@ of the
--- index vector by @xs'!'i@. This is equivalent to @'map' (xs'!') is@ but is
--- often much more efficient.
---
--- > backpermute <a,b,c,d> <0,3,2,3,1,0> = <a,d,c,d,b,a>
-backpermute :: Storable a => Vector a -> Vector Int -> Vector a
-{-# INLINE backpermute #-}
-backpermute = G.backpermute
-
--- | Same as 'backpermute' but without bounds checking.
-unsafeBackpermute :: Storable a => Vector a -> Vector Int -> Vector a
-{-# INLINE unsafeBackpermute #-}
-unsafeBackpermute = G.unsafeBackpermute
-
--- Safe destructive updates
--- ------------------------
-
--- | Apply a destructive operation to a vector. The operation will be
--- performed in place if it is safe to do so and will modify a copy of the
--- vector otherwise.
---
--- @
--- modify (\\v -> write v 0 \'x\') ('replicate' 3 \'a\') = \<\'x\',\'a\',\'a\'\>
--- @
-modify :: Storable a => (forall s. MVector s a -> ST s ()) -> Vector a -> Vector a
-{-# INLINE modify #-}
-modify p = G.modify p
-
--- Mapping
--- -------
-
--- | /O(n)/ Map a function over a vector
-map :: (Storable a, Storable b) => (a -> b) -> Vector a -> Vector b
-{-# INLINE map #-}
-map = G.map
-
--- | /O(n)/ Apply a function to every element of a vector and its index
-imap :: (Storable a, Storable b) => (Int -> a -> b) -> Vector a -> Vector b
-{-# INLINE imap #-}
-imap = G.imap
-
--- | Map a function over a vector and concatenate the results.
-concatMap :: (Storable a, Storable b) => (a -> Vector b) -> Vector a -> Vector b
-{-# INLINE concatMap #-}
-concatMap = G.concatMap
-
--- Monadic mapping
--- ---------------
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results
-mapM :: (Monad m, Storable a, Storable b) => (a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE mapM #-}
-mapM = G.mapM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results
-mapM_ :: (Monad m, Storable a) => (a -> m b) -> Vector a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ = G.mapM_
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results. Equivalent to @flip 'mapM'@.
-forM :: (Monad m, Storable a, Storable b) => Vector a -> (a -> m b) -> m (Vector b)
-{-# INLINE forM #-}
-forM = G.forM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results. Equivalent to @flip 'mapM_'@.
-forM_ :: (Monad m, Storable a) => Vector a -> (a -> m b) -> m ()
-{-# INLINE forM_ #-}
-forM_ = G.forM_
-
--- Zipping
--- -------
-
--- | /O(min(m,n))/ Zip two vectors with the given function.
-zipWith :: (Storable a, Storable b, Storable c)
-        => (a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE zipWith #-}
-zipWith = G.zipWith
-
--- | Zip three vectors with the given function.
-zipWith3 :: (Storable a, Storable b, Storable c, Storable d)
-         => (a -> b -> c -> d) -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE zipWith3 #-}
-zipWith3 = G.zipWith3
-
-zipWith4 :: (Storable a, Storable b, Storable c, Storable d, Storable e)
-         => (a -> b -> c -> d -> e)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE zipWith4 #-}
-zipWith4 = G.zipWith4
-
-zipWith5 :: (Storable a, Storable b, Storable c, Storable d, Storable e,
-             Storable f)
-         => (a -> b -> c -> d -> e -> f)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f
-{-# INLINE zipWith5 #-}
-zipWith5 = G.zipWith5
-
-zipWith6 :: (Storable a, Storable b, Storable c, Storable d, Storable e,
-             Storable f, Storable g)
-         => (a -> b -> c -> d -> e -> f -> g)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f -> Vector g
-{-# INLINE zipWith6 #-}
-zipWith6 = G.zipWith6
-
--- | /O(min(m,n))/ Zip two vectors with a function that also takes the
--- elements' indices.
-izipWith :: (Storable a, Storable b, Storable c)
-         => (Int -> a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE izipWith #-}
-izipWith = G.izipWith
-
--- | Zip three vectors and their indices with the given function.
-izipWith3 :: (Storable a, Storable b, Storable c, Storable d)
-          => (Int -> a -> b -> c -> d)
-          -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE izipWith3 #-}
-izipWith3 = G.izipWith3
-
-izipWith4 :: (Storable a, Storable b, Storable c, Storable d, Storable e)
-          => (Int -> a -> b -> c -> d -> e)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE izipWith4 #-}
-izipWith4 = G.izipWith4
-
-izipWith5 :: (Storable a, Storable b, Storable c, Storable d, Storable e,
-              Storable f)
-          => (Int -> a -> b -> c -> d -> e -> f)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f
-{-# INLINE izipWith5 #-}
-izipWith5 = G.izipWith5
-
-izipWith6 :: (Storable a, Storable b, Storable c, Storable d, Storable e,
-              Storable f, Storable g)
-          => (Int -> a -> b -> c -> d -> e -> f -> g)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f -> Vector g
-{-# INLINE izipWith6 #-}
-izipWith6 = G.izipWith6
-
--- Monadic zipping
--- ---------------
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and yield a
--- vector of results
-zipWithM :: (Monad m, Storable a, Storable b, Storable c)
-         => (a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE zipWithM #-}
-zipWithM = G.zipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and ignore the
--- results
-zipWithM_ :: (Monad m, Storable a, Storable b)
-          => (a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ = G.zipWithM_
-
--- Filtering
--- ---------
-
--- | /O(n)/ Drop elements that do not satisfy the predicate
-filter :: Storable a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE filter #-}
-filter = G.filter
-
--- | /O(n)/ Drop elements that do not satisfy the predicate which is applied to
--- values and their indices
-ifilter :: Storable a => (Int -> a -> Bool) -> Vector a -> Vector a
-{-# INLINE ifilter #-}
-ifilter = G.ifilter
-
--- | /O(n)/ Drop repeated adjacent elements.
-uniq :: (Storable a, Eq a) => Vector a -> Vector a
-{-# INLINE uniq #-}
-uniq = G.uniq
-
--- | /O(n)/ Drop elements when predicate returns Nothing
-mapMaybe :: (Storable a, Storable b) => (a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE mapMaybe #-}
-mapMaybe = G.mapMaybe
-
--- | /O(n)/ Drop elements when predicate, applied to index and value, returns Nothing
-imapMaybe :: (Storable a, Storable b) => (Int -> a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE imapMaybe #-}
-imapMaybe = G.imapMaybe
-
--- | /O(n)/ Drop elements that do not satisfy the monadic predicate
-filterM :: (Monad m, Storable a) => (a -> m Bool) -> Vector a -> m (Vector a)
-{-# INLINE filterM #-}
-filterM = G.filterM
-
--- | /O(n)/ Yield the longest prefix of elements satisfying the predicate
--- without copying.
-takeWhile :: Storable a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE takeWhile #-}
-takeWhile = G.takeWhile
-
--- | /O(n)/ Drop the longest prefix of elements that satisfy the predicate
--- without copying.
-dropWhile :: Storable a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE dropWhile #-}
-dropWhile = G.dropWhile
-
--- Parititioning
--- -------------
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't. The
--- relative order of the elements is preserved at the cost of a sometimes
--- reduced performance compared to 'unstablePartition'.
-partition :: Storable a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE partition #-}
-partition = G.partition
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't.
--- The order of the elements is not preserved but the operation is often
--- faster than 'partition'.
-unstablePartition :: Storable a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE unstablePartition #-}
-unstablePartition = G.unstablePartition
-
--- | /O(n)/ Split the vector into the longest prefix of elements that satisfy
--- the predicate and the rest without copying.
-span :: Storable a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE span #-}
-span = G.span
-
--- | /O(n)/ Split the vector into the longest prefix of elements that do not
--- satisfy the predicate and the rest without copying.
-break :: Storable a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE break #-}
-break = G.break
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | /O(n)/ Check if the vector contains an element
-elem :: (Storable a, Eq a) => a -> Vector a -> Bool
-{-# INLINE elem #-}
-elem = G.elem
-
-infix 4 `notElem`
--- | /O(n)/ Check if the vector does not contain an element (inverse of 'elem')
-notElem :: (Storable a, Eq a) => a -> Vector a -> Bool
-{-# INLINE notElem #-}
-notElem = G.notElem
-
--- | /O(n)/ Yield 'Just' the first element matching the predicate or 'Nothing'
--- if no such element exists.
-find :: Storable a => (a -> Bool) -> Vector a -> Maybe a
-{-# INLINE find #-}
-find = G.find
-
--- | /O(n)/ Yield 'Just' the index of the first element matching the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Storable a => (a -> Bool) -> Vector a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex = G.findIndex
-
--- | /O(n)/ Yield the indices of elements satisfying the predicate in ascending
--- order.
-findIndices :: Storable a => (a -> Bool) -> Vector a -> Vector Int
-{-# INLINE findIndices #-}
-findIndices = G.findIndices
-
--- | /O(n)/ Yield 'Just' the index of the first occurence of the given element or
--- 'Nothing' if the vector does not contain the element. This is a specialised
--- version of 'findIndex'.
-elemIndex :: (Storable a, Eq a) => a -> Vector a -> Maybe Int
-{-# INLINE elemIndex #-}
-elemIndex = G.elemIndex
-
--- | /O(n)/ Yield the indices of all occurences of the given element in
--- ascending order. This is a specialised version of 'findIndices'.
-elemIndices :: (Storable a, Eq a) => a -> Vector a -> Vector Int
-{-# INLINE elemIndices #-}
-elemIndices = G.elemIndices
-
--- Folding
--- -------
-
--- | /O(n)/ Left fold
-foldl :: Storable b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl #-}
-foldl = G.foldl
-
--- | /O(n)/ Left fold on non-empty vectors
-foldl1 :: Storable a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1 #-}
-foldl1 = G.foldl1
-
--- | /O(n)/ Left fold with strict accumulator
-foldl' :: Storable b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl' #-}
-foldl' = G.foldl'
-
--- | /O(n)/ Left fold on non-empty vectors with strict accumulator
-foldl1' :: Storable a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1' #-}
-foldl1' = G.foldl1'
-
--- | /O(n)/ Right fold
-foldr :: Storable a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr #-}
-foldr = G.foldr
-
--- | /O(n)/ Right fold on non-empty vectors
-foldr1 :: Storable a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1 #-}
-foldr1 = G.foldr1
-
--- | /O(n)/ Right fold with a strict accumulator
-foldr' :: Storable a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr' #-}
-foldr' = G.foldr'
-
--- | /O(n)/ Right fold on non-empty vectors with strict accumulator
-foldr1' :: Storable a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1' #-}
-foldr1' = G.foldr1'
-
--- | /O(n)/ Left fold (function applied to each element and its index)
-ifoldl :: Storable b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl #-}
-ifoldl = G.ifoldl
-
--- | /O(n)/ Left fold with strict accumulator (function applied to each element
--- and its index)
-ifoldl' :: Storable b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl' #-}
-ifoldl' = G.ifoldl'
-
--- | /O(n)/ Right fold (function applied to each element and its index)
-ifoldr :: Storable a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr #-}
-ifoldr = G.ifoldr
-
--- | /O(n)/ Right fold with strict accumulator (function applied to each
--- element and its index)
-ifoldr' :: Storable a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr' #-}
-ifoldr' = G.ifoldr'
-
--- Specialised folds
--- -----------------
-
--- | /O(n)/ Check if all elements satisfy the predicate.
-all :: Storable a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE all #-}
-all = G.all
-
--- | /O(n)/ Check if any element satisfies the predicate.
-any :: Storable a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE any #-}
-any = G.any
-
--- | /O(n)/ Check if all elements are 'True'
-and :: Vector Bool -> Bool
-{-# INLINE and #-}
-and = G.and
-
--- | /O(n)/ Check if any element is 'True'
-or :: Vector Bool -> Bool
-{-# INLINE or #-}
-or = G.or
-
--- | /O(n)/ Compute the sum of the elements
-sum :: (Storable a, Num a) => Vector a -> a
-{-# INLINE sum #-}
-sum = G.sum
-
--- | /O(n)/ Compute the produce of the elements
-product :: (Storable a, Num a) => Vector a -> a
-{-# INLINE product #-}
-product = G.product
-
--- | /O(n)/ Yield the maximum element of the vector. The vector may not be
--- empty.
-maximum :: (Storable a, Ord a) => Vector a -> a
-{-# INLINE maximum #-}
-maximum = G.maximum
-
--- | /O(n)/ Yield the maximum element of the vector according to the given
--- comparison function. The vector may not be empty.
-maximumBy :: Storable a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE maximumBy #-}
-maximumBy = G.maximumBy
-
--- | /O(n)/ Yield the minimum element of the vector. The vector may not be
--- empty.
-minimum :: (Storable a, Ord a) => Vector a -> a
-{-# INLINE minimum #-}
-minimum = G.minimum
-
--- | /O(n)/ Yield the minimum element of the vector according to the given
--- comparison function. The vector may not be empty.
-minimumBy :: Storable a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE minimumBy #-}
-minimumBy = G.minimumBy
-
--- | /O(n)/ Yield the index of the maximum element of the vector. The vector
--- may not be empty.
-maxIndex :: (Storable a, Ord a) => Vector a -> Int
-{-# INLINE maxIndex #-}
-maxIndex = G.maxIndex
-
--- | /O(n)/ Yield the index of the maximum element of the vector according to
--- the given comparison function. The vector may not be empty.
-maxIndexBy :: Storable a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE maxIndexBy #-}
-maxIndexBy = G.maxIndexBy
-
--- | /O(n)/ Yield the index of the minimum element of the vector. The vector
--- may not be empty.
-minIndex :: (Storable a, Ord a) => Vector a -> Int
-{-# INLINE minIndex #-}
-minIndex = G.minIndex
-
--- | /O(n)/ Yield the index of the minimum element of the vector according to
--- the given comparison function. The vector may not be empty.
-minIndexBy :: Storable a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE minIndexBy #-}
-minIndexBy = G.minIndexBy
-
--- Monadic folds
--- -------------
-
--- | /O(n)/ Monadic fold
-foldM :: (Monad m, Storable b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM #-}
-foldM = G.foldM
-
--- | /O(n)/ Monadic fold over non-empty vectors
-fold1M :: (Monad m, Storable a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M #-}
-fold1M = G.fold1M
-
--- | /O(n)/ Monadic fold with strict accumulator
-foldM' :: (Monad m, Storable b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM' #-}
-foldM' = G.foldM'
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
-fold1M' :: (Monad m, Storable a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = G.fold1M'
-
--- | /O(n)/ Monadic fold that discards the result
-foldM_ :: (Monad m, Storable b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM_ #-}
-foldM_ = G.foldM_
-
--- | /O(n)/ Monadic fold over non-empty vectors that discards the result
-fold1M_ :: (Monad m, Storable a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M_ #-}
-fold1M_ = G.fold1M_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
-foldM'_ :: (Monad m, Storable b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM'_ #-}
-foldM'_ = G.foldM'_
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
--- that discards the result
-fold1M'_ :: (Monad m, Storable a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M'_ #-}
-fold1M'_ = G.fold1M'_
-
--- Prefix sums (scans)
--- -------------------
-
--- | /O(n)/ Prescan
---
--- @
--- prescanl f z = 'init' . 'scanl' f z
--- @
---
--- Example: @prescanl (+) 0 \<1,2,3,4\> = \<0,1,3,6\>@
---
-prescanl :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl #-}
-prescanl = G.prescanl
-
--- | /O(n)/ Prescan with strict accumulator
-prescanl' :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl' #-}
-prescanl' = G.prescanl'
-
--- | /O(n)/ Scan
---
--- @
--- postscanl f z = 'tail' . 'scanl' f z
--- @
---
--- Example: @postscanl (+) 0 \<1,2,3,4\> = \<1,3,6,10\>@
---
-postscanl :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl #-}
-postscanl = G.postscanl
-
--- | /O(n)/ Scan with strict accumulator
-postscanl' :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl' #-}
-postscanl' = G.postscanl'
-
--- | /O(n)/ Haskell-style scan
---
--- > scanl f z <x1,...,xn> = <y1,...,y(n+1)>
--- >   where y1 = z
--- >         yi = f y(i-1) x(i-1)
---
--- Example: @scanl (+) 0 \<1,2,3,4\> = \<0,1,3,6,10\>@
---
-scanl :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl #-}
-scanl = G.scanl
-
--- | /O(n)/ Haskell-style scan with strict accumulator
-scanl' :: (Storable a, Storable b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl' #-}
-scanl' = G.scanl'
-
--- | /O(n)/ Scan over a non-empty vector
---
--- > scanl f <x1,...,xn> = <y1,...,yn>
--- >   where y1 = x1
--- >         yi = f y(i-1) xi
---
-scanl1 :: Storable a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1 #-}
-scanl1 = G.scanl1
-
--- | /O(n)/ Scan over a non-empty vector with a strict accumulator
-scanl1' :: Storable a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1' #-}
-scanl1' = G.scanl1'
-
--- | /O(n)/ Right-to-left prescan
---
--- @
--- prescanr f z = 'reverse' . 'prescanl' (flip f) z . 'reverse'
--- @
---
-prescanr :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr #-}
-prescanr = G.prescanr
-
--- | /O(n)/ Right-to-left prescan with strict accumulator
-prescanr' :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr' #-}
-prescanr' = G.prescanr'
-
--- | /O(n)/ Right-to-left scan
-postscanr :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr #-}
-postscanr = G.postscanr
-
--- | /O(n)/ Right-to-left scan with strict accumulator
-postscanr' :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr' #-}
-postscanr' = G.postscanr'
-
--- | /O(n)/ Right-to-left Haskell-style scan
-scanr :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr #-}
-scanr = G.scanr
-
--- | /O(n)/ Right-to-left Haskell-style scan with strict accumulator
-scanr' :: (Storable a, Storable b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr' #-}
-scanr' = G.scanr'
-
--- | /O(n)/ Right-to-left scan over a non-empty vector
-scanr1 :: Storable a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1 #-}
-scanr1 = G.scanr1
-
--- | /O(n)/ Right-to-left scan over a non-empty vector with a strict
--- accumulator
-scanr1' :: Storable a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1' #-}
-scanr1' = G.scanr1'
-
--- Conversions - Lists
--- ------------------------
-
--- | /O(n)/ Convert a vector to a list
-toList :: Storable a => Vector a -> [a]
-{-# INLINE toList #-}
-toList = G.toList
-
--- | /O(n)/ Convert a list to a vector
-fromList :: Storable a => [a] -> Vector a
-{-# INLINE fromList #-}
-fromList = G.fromList
-
--- | /O(n)/ Convert the first @n@ elements of a list to a vector
---
--- @
--- fromListN n xs = 'fromList' ('take' n xs)
--- @
-fromListN :: Storable a => Int -> [a] -> Vector a
-{-# INLINE fromListN #-}
-fromListN = G.fromListN
-
--- Conversions - Unsafe casts
--- --------------------------
-
--- | /O(1)/ Unsafely cast a vector from one element type to another.
--- The operation just changes the type of the underlying pointer and does not
--- modify the elements.
---
--- The resulting vector contains as many elements as can fit into the
--- underlying memory block.
---
-unsafeCast :: forall a b. (Storable a, Storable b) => Vector a -> Vector b
-{-# INLINE unsafeCast #-}
-unsafeCast (Vector n fp)
-  = Vector ((n * sizeOf (undefined :: a)) `div` sizeOf (undefined :: b))
-           (castForeignPtr fp)
-
-
--- Conversions - Mutable vectors
--- -----------------------------
-
--- | /O(1)/ Unsafe convert a mutable vector to an immutable one without
--- copying. The mutable vector may not be used after this operation.
-unsafeFreeze
-        :: (Storable a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE unsafeFreeze #-}
-unsafeFreeze = G.unsafeFreeze
-
--- | /O(1)/ Unsafely convert an immutable vector to a mutable one without
--- copying. The immutable vector may not be used after this operation.
-unsafeThaw
-        :: (Storable a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE unsafeThaw #-}
-unsafeThaw = G.unsafeThaw
-
--- | /O(n)/ Yield a mutable copy of the immutable vector.
-thaw :: (Storable a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE thaw #-}
-thaw = G.thaw
-
--- | /O(n)/ Yield an immutable copy of the mutable vector.
-freeze :: (Storable a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE freeze #-}
-freeze = G.freeze
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length. This is not checked.
-unsafeCopy
-  :: (Storable a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length.
-copy :: (Storable a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
--- Conversions - Raw pointers
--- --------------------------
-
--- | /O(1)/ Create a vector from a 'ForeignPtr' with an offset and a length.
---
--- The data may not be modified through the 'ForeignPtr' afterwards.
---
--- If your offset is 0 it is more efficient to use 'unsafeFromForeignPtr0'.
-unsafeFromForeignPtr :: Storable a
-                     => ForeignPtr a    -- ^ pointer
-                     -> Int             -- ^ offset
-                     -> Int             -- ^ length
-                     -> Vector a
-{-# INLINE_FUSED unsafeFromForeignPtr #-}
-unsafeFromForeignPtr fp i n = unsafeFromForeignPtr0 fp' n
-    where
-      fp' = updPtr (`advancePtr` i) fp
-
-{-# RULES
-"unsafeFromForeignPtr fp 0 n -> unsafeFromForeignPtr0 fp n " forall fp n.
-  unsafeFromForeignPtr fp 0 n = unsafeFromForeignPtr0 fp n   #-}
-
-
--- | /O(1)/ Create a vector from a 'ForeignPtr' and a length.
---
--- It is assumed the pointer points directly to the data (no offset).
--- Use `unsafeFromForeignPtr` if you need to specify an offset.
---
--- The data may not be modified through the 'ForeignPtr' afterwards.
-unsafeFromForeignPtr0 :: Storable a
-                      => ForeignPtr a    -- ^ pointer
-                      -> Int             -- ^ length
-                      -> Vector a
-{-# INLINE unsafeFromForeignPtr0 #-}
-unsafeFromForeignPtr0 fp n = Vector n fp
-
--- | /O(1)/ Yield the underlying 'ForeignPtr' together with the offset to the
--- data and its length. The data may not be modified through the 'ForeignPtr'.
-unsafeToForeignPtr :: Storable a => Vector a -> (ForeignPtr a, Int, Int)
-{-# INLINE unsafeToForeignPtr #-}
-unsafeToForeignPtr (Vector n fp) = (fp, 0, n)
-
--- | /O(1)/ Yield the underlying 'ForeignPtr' together with its length.
---
--- You can assume the pointer points directly to the data (no offset).
---
--- The data may not be modified through the 'ForeignPtr'.
-unsafeToForeignPtr0 :: Storable a => Vector a -> (ForeignPtr a, Int)
-{-# INLINE unsafeToForeignPtr0 #-}
-unsafeToForeignPtr0 (Vector n fp) = (fp, n)
-
--- | Pass a pointer to the vector's data to the IO action. The data may not be
--- modified through the 'Ptr.
-unsafeWith :: Storable a => Vector a -> (Ptr a -> IO b) -> IO b
-{-# INLINE unsafeWith #-}
-unsafeWith (Vector _ fp) = withForeignPtr fp
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Internal.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Internal.hs
deleted file mode 100644
index 69a46d84215b..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Internal.hs
+++ /dev/null
@@ -1,33 +0,0 @@
--- |
--- Module      : Data.Vector.Storable.Internal
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Ugly internal utility functions for implementing 'Storable'-based vectors.
---
-
-module Data.Vector.Storable.Internal (
-  getPtr, setPtr, updPtr
-) where
-
-import Foreign.ForeignPtr
-import Foreign.Ptr
-import GHC.ForeignPtr   ( ForeignPtr(..) )
-import GHC.Ptr          ( Ptr(..) )
-
-getPtr :: ForeignPtr a -> Ptr a
-{-# INLINE getPtr #-}
-getPtr (ForeignPtr addr _) = Ptr addr
-
-setPtr :: ForeignPtr a -> Ptr a -> ForeignPtr a
-{-# INLINE setPtr #-}
-setPtr (ForeignPtr _ c) (Ptr addr) = ForeignPtr addr c
-
-updPtr :: (Ptr a -> Ptr a) -> ForeignPtr a -> ForeignPtr a
-{-# INLINE updPtr #-}
-updPtr f (ForeignPtr p c) = case f (Ptr p) of { Ptr q -> ForeignPtr q c }
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Mutable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Mutable.hs
deleted file mode 100644
index 29eb2fbfa31e..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Storable/Mutable.hs
+++ /dev/null
@@ -1,543 +0,0 @@
-{-# LANGUAGE CPP, DeriveDataTypeable, FlexibleInstances, MagicHash, MultiParamTypeClasses, ScopedTypeVariables #-}
-
--- |
--- Module      : Data.Vector.Storable.Mutable
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Mutable vectors based on Storable.
---
-
-module Data.Vector.Storable.Mutable(
-  -- * Mutable vectors of 'Storable' types
-  MVector(..), IOVector, STVector, Storable,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Extracting subvectors
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- ** Overlapping
-  overlaps,
-
-  -- * Construction
-
-  -- ** Initialisation
-  new, unsafeNew, replicate, replicateM, clone,
-
-  -- ** Growing
-  grow, unsafeGrow,
-
-  -- ** Restricting memory usage
-  clear,
-
-  -- * Accessing individual elements
-  read, write, modify, swap,
-  unsafeRead, unsafeWrite, unsafeModify, unsafeSwap,
-
-  -- * Modifying vectors
-
-  -- ** Filling and copying
-  set, copy, move, unsafeCopy, unsafeMove,
-
-  -- * Unsafe conversions
-  unsafeCast,
-
-  -- * Raw pointers
-  unsafeFromForeignPtr, unsafeFromForeignPtr0,
-  unsafeToForeignPtr,   unsafeToForeignPtr0,
-  unsafeWith
-) where
-
-import Control.DeepSeq ( NFData(rnf) )
-
-import qualified Data.Vector.Generic.Mutable as G
-import Data.Vector.Storable.Internal
-
-import Foreign.Storable
-import Foreign.ForeignPtr
-
-#if __GLASGOW_HASKELL__ >= 706
-import GHC.ForeignPtr (mallocPlainForeignPtrAlignedBytes)
-#elif __GLASGOW_HASKELL__ >= 700
-import Data.Primitive.ByteArray (MutableByteArray(..), newAlignedPinnedByteArray,
-                                 unsafeFreezeByteArray)
-import GHC.Prim (byteArrayContents#, unsafeCoerce#)
-import GHC.ForeignPtr
-#endif
-
-import Foreign.Ptr
-import Foreign.Marshal.Array ( advancePtr, copyArray, moveArray )
-
-import Control.Monad.Primitive
-import Data.Primitive.Addr
-import Data.Primitive.Types (Prim)
-
-import GHC.Word (Word8, Word16, Word32, Word64)
-import GHC.Ptr (Ptr(..))
-
-import Prelude hiding ( length, null, replicate, reverse, map, read,
-                        take, drop, splitAt, init, tail )
-
-import Data.Typeable ( Typeable )
-
--- Data.Vector.Internal.Check is not needed
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- | Mutable 'Storable'-based vectors
-data MVector s a = MVector {-# UNPACK #-} !Int
-                           {-# UNPACK #-} !(ForeignPtr a)
-        deriving ( Typeable )
-
-type IOVector = MVector RealWorld
-type STVector s = MVector s
-
-instance NFData (MVector s a) where
-  rnf (MVector _ _) = ()
-
-instance Storable a => G.MVector MVector a where
-  {-# INLINE basicLength #-}
-  basicLength (MVector n _) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice j m (MVector _ fp) = MVector m (updPtr (`advancePtr` j) fp)
-
-  -- FIXME: this relies on non-portable pointer comparisons
-  {-# INLINE basicOverlaps #-}
-  basicOverlaps (MVector m fp) (MVector n fq)
-    = between p q (q `advancePtr` n) || between q p (p `advancePtr` m)
-    where
-      between x y z = x >= y && x < z
-      p = getPtr fp
-      q = getPtr fq
-
-  {-# INLINE basicUnsafeNew #-}
-  basicUnsafeNew n
-    | n < 0 = error $ "Storable.basicUnsafeNew: negative length: " ++ show n
-    | n > mx = error $ "Storable.basicUnsafeNew: length too large: " ++ show n
-    | otherwise = unsafePrimToPrim $ do
-        fp <- mallocVector n
-        return $ MVector n fp
-    where
-      size = sizeOf (undefined :: a)
-      mx = maxBound `quot` size :: Int
-
-  {-# INLINE basicInitialize #-}
-  basicInitialize = storableZero
-
-  {-# INLINE basicUnsafeRead #-}
-  basicUnsafeRead (MVector _ fp) i
-    = unsafePrimToPrim
-    $ withForeignPtr fp (`peekElemOff` i)
-
-  {-# INLINE basicUnsafeWrite #-}
-  basicUnsafeWrite (MVector _ fp) i x
-    = unsafePrimToPrim
-    $ withForeignPtr fp $ \p -> pokeElemOff p i x
-
-  {-# INLINE basicSet #-}
-  basicSet = storableSet
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MVector n fp) (MVector _ fq)
-    = unsafePrimToPrim
-    $ withForeignPtr fp $ \p ->
-      withForeignPtr fq $ \q ->
-      copyArray p q n
-
-  {-# INLINE basicUnsafeMove #-}
-  basicUnsafeMove (MVector n fp) (MVector _ fq)
-    = unsafePrimToPrim
-    $ withForeignPtr fp $ \p ->
-      withForeignPtr fq $ \q ->
-      moveArray p q n
-
-storableZero :: forall a m. (Storable a, PrimMonad m) => MVector (PrimState m) a -> m ()
-{-# INLINE storableZero #-}
-storableZero (MVector n fp) = unsafePrimToPrim . withForeignPtr fp $ \(Ptr p) -> do
-  let q = Addr p
-  setAddr q byteSize (0 :: Word8)
- where
- x :: a
- x = undefined
-
- byteSize :: Int
- byteSize = n * sizeOf x
-
-storableSet :: (Storable a, PrimMonad m) => MVector (PrimState m) a -> a -> m ()
-{-# INLINE storableSet #-}
-storableSet (MVector n fp) x
-  | n == 0 = return ()
-  | otherwise = unsafePrimToPrim $
-                case sizeOf x of
-                  1 -> storableSetAsPrim n fp x (undefined :: Word8)
-                  2 -> storableSetAsPrim n fp x (undefined :: Word16)
-                  4 -> storableSetAsPrim n fp x (undefined :: Word32)
-                  8 -> storableSetAsPrim n fp x (undefined :: Word64)
-                  _ -> withForeignPtr fp $ \p -> do
-                       poke p x
-
-                       let do_set i
-                             | 2*i < n = do
-                                 copyArray (p `advancePtr` i) p i
-                                 do_set (2*i)
-                             | otherwise = copyArray (p `advancePtr` i) p (n-i)
-
-                       do_set 1
-
-storableSetAsPrim
-  :: (Storable a, Prim b) => Int -> ForeignPtr a -> a -> b -> IO ()
-{-# INLINE [0] storableSetAsPrim #-}
-storableSetAsPrim n fp x y = withForeignPtr fp $ \(Ptr p) -> do
-  poke (Ptr p) x
-  let q = Addr p
-  w <- readOffAddr q 0
-  setAddr (q `plusAddr` sizeOf x) (n-1) (w `asTypeOf` y)
-
-{-# INLINE mallocVector #-}
-mallocVector :: Storable a => Int -> IO (ForeignPtr a)
-mallocVector =
-#if __GLASGOW_HASKELL__ >= 706
-  doMalloc undefined
-  where
-    doMalloc :: Storable b => b -> Int -> IO (ForeignPtr b)
-    doMalloc dummy size =
-      mallocPlainForeignPtrAlignedBytes (size * sizeOf dummy) (alignment dummy)
-#elif __GLASGOW_HASKELL__ >= 700
-  doMalloc undefined
-  where
-    doMalloc :: Storable b => b -> Int -> IO (ForeignPtr b)
-    doMalloc dummy size = do
-      arr@(MutableByteArray arr#) <- newAlignedPinnedByteArray arrSize arrAlign
-      newConcForeignPtr
-        (Ptr (byteArrayContents# (unsafeCoerce# arr#)))
-        -- Keep reference to mutable byte array until whole ForeignPtr goes out
-        -- of scope.
-        (touch arr)
-      where
-        arrSize  = size * sizeOf dummy
-        arrAlign = alignment dummy
-#else
-    mallocForeignPtrArray
-#endif
-
--- Length information
--- ------------------
-
--- | Length of the mutable vector.
-length :: Storable a => MVector s a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | Check whether the vector is empty
-null :: Storable a => MVector s a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Extracting subvectors
--- ---------------------
-
--- | Yield a part of the mutable vector without copying it.
-slice :: Storable a => Int -> Int -> MVector s a -> MVector s a
-{-# INLINE slice #-}
-slice = G.slice
-
-take :: Storable a => Int -> MVector s a -> MVector s a
-{-# INLINE take #-}
-take = G.take
-
-drop :: Storable a => Int -> MVector s a -> MVector s a
-{-# INLINE drop #-}
-drop = G.drop
-
-splitAt :: Storable a => Int -> MVector s a -> (MVector s a, MVector s a)
-{-# INLINE splitAt #-}
-splitAt = G.splitAt
-
-init :: Storable a => MVector s a -> MVector s a
-{-# INLINE init #-}
-init = G.init
-
-tail :: Storable a => MVector s a -> MVector s a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | Yield a part of the mutable vector without copying it. No bounds checks
--- are performed.
-unsafeSlice :: Storable a
-            => Int  -- ^ starting index
-            -> Int  -- ^ length of the slice
-            -> MVector s a
-            -> MVector s a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
-unsafeTake :: Storable a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
-unsafeDrop :: Storable a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
-unsafeInit :: Storable a => MVector s a -> MVector s a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
-unsafeTail :: Storable a => MVector s a -> MVector s a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- Overlapping
--- -----------
-
--- | Check whether two vectors overlap.
-overlaps :: Storable a => MVector s a -> MVector s a -> Bool
-{-# INLINE overlaps #-}
-overlaps = G.overlaps
-
--- Initialisation
--- --------------
-
--- | Create a mutable vector of the given length.
-new :: (PrimMonad m, Storable a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE new #-}
-new = G.new
-
--- | Create a mutable vector of the given length. The memory is not initialized.
-unsafeNew :: (PrimMonad m, Storable a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeNew #-}
-unsafeNew = G.unsafeNew
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with an initial value.
-replicate :: (PrimMonad m, Storable a) => Int -> a -> m (MVector (PrimState m) a)
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with values produced by repeatedly executing the monadic action.
-replicateM :: (PrimMonad m, Storable a) => Int -> m a -> m (MVector (PrimState m) a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | Create a copy of a mutable vector.
-clone :: (PrimMonad m, Storable a)
-      => MVector (PrimState m) a -> m (MVector (PrimState m) a)
-{-# INLINE clone #-}
-clone = G.clone
-
--- Growing
--- -------
-
--- | Grow a vector by the given number of elements. The number must be
--- positive.
-grow :: (PrimMonad m, Storable a)
-     => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE grow #-}
-grow = G.grow
-
--- | Grow a vector by the given number of elements. The number must be
--- positive but this is not checked.
-unsafeGrow :: (PrimMonad m, Storable a)
-           => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeGrow #-}
-unsafeGrow = G.unsafeGrow
-
--- Restricting memory usage
--- ------------------------
-
--- | Reset all elements of the vector to some undefined value, clearing all
--- references to external objects. This is usually a noop for unboxed vectors.
-clear :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> m ()
-{-# INLINE clear #-}
-clear = G.clear
-
--- Accessing individual elements
--- -----------------------------
-
--- | Yield the element at the given position.
-read :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE read #-}
-read = G.read
-
--- | Replace the element at the given position.
-write
-    :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE write #-}
-write = G.write
-
--- | Modify the element at the given position.
-modify :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE modify #-}
-modify = G.modify
-
--- | Swap the elements at the given positions.
-swap
-    :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE swap #-}
-swap = G.swap
-
-
--- | Yield the element at the given position. No bounds checks are performed.
-unsafeRead :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE unsafeRead #-}
-unsafeRead = G.unsafeRead
-
--- | Replace the element at the given position. No bounds checks are performed.
-unsafeWrite
-    :: (PrimMonad m, Storable a) =>  MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE unsafeWrite #-}
-unsafeWrite = G.unsafeWrite
-
--- | Modify the element at the given position. No bounds checks are performed.
-unsafeModify :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE unsafeModify #-}
-unsafeModify = G.unsafeModify
-
--- | Swap the elements at the given positions. No bounds checks are performed.
-unsafeSwap
-    :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE unsafeSwap #-}
-unsafeSwap = G.unsafeSwap
-
--- Filling and copying
--- -------------------
-
--- | Set all elements of the vector to the given value.
-set :: (PrimMonad m, Storable a) => MVector (PrimState m) a -> a -> m ()
-{-# INLINE set #-}
-set = G.set
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap.
-copy :: (PrimMonad m, Storable a)
-     => MVector (PrimState m) a   -- ^ target
-     -> MVector (PrimState m) a   -- ^ source
-     -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap. This is not checked.
-unsafeCopy :: (PrimMonad m, Storable a)
-           => MVector (PrimState m) a   -- ^ target
-           -> MVector (PrimState m) a   -- ^ source
-           -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | Move the contents of a vector. The two vectors must have the same
--- length.
---
--- If the vectors do not overlap, then this is equivalent to 'copy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-move :: (PrimMonad m, Storable a)
-     => MVector (PrimState m) a -> MVector (PrimState m) a -> m ()
-{-# INLINE move #-}
-move = G.move
-
--- | Move the contents of a vector. The two vectors must have the same
--- length, but this is not checked.
---
--- If the vectors do not overlap, then this is equivalent to 'unsafeCopy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-unsafeMove :: (PrimMonad m, Storable a)
-           => MVector (PrimState m) a   -- ^ target
-           -> MVector (PrimState m) a   -- ^ source
-           -> m ()
-{-# INLINE unsafeMove #-}
-unsafeMove = G.unsafeMove
-
--- Unsafe conversions
--- ------------------
-
--- | /O(1)/ Unsafely cast a mutable vector from one element type to another.
--- The operation just changes the type of the underlying pointer and does not
--- modify the elements.
---
--- The resulting vector contains as many elements as can fit into the
--- underlying memory block.
---
-unsafeCast :: forall a b s.
-              (Storable a, Storable b) => MVector s a -> MVector s b
-{-# INLINE unsafeCast #-}
-unsafeCast (MVector n fp)
-  = MVector ((n * sizeOf (undefined :: a)) `div` sizeOf (undefined :: b))
-            (castForeignPtr fp)
-
--- Raw pointers
--- ------------
-
--- | Create a mutable vector from a 'ForeignPtr' with an offset and a length.
---
--- Modifying data through the 'ForeignPtr' afterwards is unsafe if the vector
--- could have been frozen before the modification.
---
---  If your offset is 0 it is more efficient to use 'unsafeFromForeignPtr0'.
-unsafeFromForeignPtr :: Storable a
-                     => ForeignPtr a    -- ^ pointer
-                     -> Int             -- ^ offset
-                     -> Int             -- ^ length
-                     -> MVector s a
-{-# INLINE_FUSED unsafeFromForeignPtr #-}
-unsafeFromForeignPtr fp i n = unsafeFromForeignPtr0 fp' n
-    where
-      fp' = updPtr (`advancePtr` i) fp
-
-{-# RULES
-"unsafeFromForeignPtr fp 0 n -> unsafeFromForeignPtr0 fp n " forall fp n.
-  unsafeFromForeignPtr fp 0 n = unsafeFromForeignPtr0 fp n   #-}
-
-
--- | /O(1)/ Create a mutable vector from a 'ForeignPtr' and a length.
---
--- It is assumed the pointer points directly to the data (no offset).
--- Use `unsafeFromForeignPtr` if you need to specify an offset.
---
--- Modifying data through the 'ForeignPtr' afterwards is unsafe if the vector
--- could have been frozen before the modification.
-unsafeFromForeignPtr0 :: Storable a
-                      => ForeignPtr a    -- ^ pointer
-                      -> Int             -- ^ length
-                      -> MVector s a
-{-# INLINE unsafeFromForeignPtr0 #-}
-unsafeFromForeignPtr0 fp n = MVector n fp
-
--- | Yield the underlying 'ForeignPtr' together with the offset to the data
--- and its length. Modifying the data through the 'ForeignPtr' is
--- unsafe if the vector could have frozen before the modification.
-unsafeToForeignPtr :: Storable a => MVector s a -> (ForeignPtr a, Int, Int)
-{-# INLINE unsafeToForeignPtr #-}
-unsafeToForeignPtr (MVector n fp) = (fp, 0, n)
-
--- | /O(1)/ Yield the underlying 'ForeignPtr' together with its length.
---
--- You can assume the pointer points directly to the data (no offset).
---
--- Modifying the data through the 'ForeignPtr' is unsafe if the vector could
--- have frozen before the modification.
-unsafeToForeignPtr0 :: Storable a => MVector s a -> (ForeignPtr a, Int)
-{-# INLINE unsafeToForeignPtr0 #-}
-unsafeToForeignPtr0 (MVector n fp) = (fp, n)
-
--- | Pass a pointer to the vector's data to the IO action. Modifying data
--- through the pointer is unsafe if the vector could have been frozen before
--- the modification.
-unsafeWith :: Storable a => IOVector a -> (Ptr a -> IO b) -> IO b
-{-# INLINE unsafeWith #-}
-unsafeWith (MVector _ fp) = withForeignPtr fp
-
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed.hs
deleted file mode 100644
index 72dd109fb3b4..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed.hs
+++ /dev/null
@@ -1,1488 +0,0 @@
-{-# LANGUAGE CPP, Rank2Types, TypeFamilies #-}
-
--- |
--- Module      : Data.Vector.Unboxed
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Adaptive unboxed vectors. The implementation is based on type families
--- and picks an efficient, specialised representation for every element type.
--- In particular, unboxed vectors of pairs are represented as pairs of unboxed
--- vectors.
---
--- Implementing unboxed vectors for new data types can be very easy. Here is
--- how the library does this for 'Complex' by simply wrapping vectors of
--- pairs.
---
--- @
--- newtype instance 'MVector' s ('Complex' a) = MV_Complex ('MVector' s (a,a))
--- newtype instance 'Vector'    ('Complex' a) = V_Complex  ('Vector'    (a,a))
---
--- instance ('RealFloat' a, 'Unbox' a) => 'Data.Vector.Generic.Mutable.MVector' 'MVector' ('Complex' a) where
---   {-\# INLINE basicLength \#-}
---   basicLength (MV_Complex v) = 'Data.Vector.Generic.Mutable.basicLength' v
---   ...
---
--- instance ('RealFloat' a, 'Unbox' a) => Data.Vector.Generic.Vector 'Vector' ('Complex' a) where
---   {-\# INLINE basicLength \#-}
---   basicLength (V_Complex v) = Data.Vector.Generic.basicLength v
---   ...
---
--- instance ('RealFloat' a, 'Unbox' a) => 'Unbox' ('Complex' a)
--- @
-
-module Data.Vector.Unboxed (
-  -- * Unboxed vectors
-  Vector, MVector(..), Unbox,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Indexing
-  (!), (!?), head, last,
-  unsafeIndex, unsafeHead, unsafeLast,
-
-  -- ** Monadic indexing
-  indexM, headM, lastM,
-  unsafeIndexM, unsafeHeadM, unsafeLastM,
-
-  -- ** Extracting subvectors (slicing)
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- * Construction
-
-  -- ** Initialisation
-  empty, singleton, replicate, generate, iterateN,
-
-  -- ** Monadic initialisation
-  replicateM, generateM, iterateNM, create, createT,
-
-  -- ** Unfolding
-  unfoldr, unfoldrN,
-  unfoldrM, unfoldrNM,
-  constructN, constructrN,
-
-  -- ** Enumeration
-  enumFromN, enumFromStepN, enumFromTo, enumFromThenTo,
-
-  -- ** Concatenation
-  cons, snoc, (++), concat,
-
-  -- ** Restricting memory usage
-  force,
-
-  -- * Modifying vectors
-
-  -- ** Bulk updates
-  (//), update, update_,
-  unsafeUpd, unsafeUpdate, unsafeUpdate_,
-
-  -- ** Accumulations
-  accum, accumulate, accumulate_,
-  unsafeAccum, unsafeAccumulate, unsafeAccumulate_,
-
-  -- ** Permutations
-  reverse, backpermute, unsafeBackpermute,
-
-  -- ** Safe destructive updates
-  modify,
-
-  -- * Elementwise operations
-
-  -- ** Indexing
-  indexed,
-
-  -- ** Mapping
-  map, imap, concatMap,
-
-  -- ** Monadic mapping
-  mapM, imapM, mapM_, imapM_, forM, forM_,
-
-  -- ** Zipping
-  zipWith, zipWith3, zipWith4, zipWith5, zipWith6,
-  izipWith, izipWith3, izipWith4, izipWith5, izipWith6,
-  zip, zip3, zip4, zip5, zip6,
-
-  -- ** Monadic zipping
-  zipWithM, izipWithM, zipWithM_, izipWithM_,
-
-  -- ** Unzipping
-  unzip, unzip3, unzip4, unzip5, unzip6,
-
-  -- * Working with predicates
-
-  -- ** Filtering
-  filter, ifilter, uniq,
-  mapMaybe, imapMaybe,
-  filterM,
-  takeWhile, dropWhile,
-
-  -- ** Partitioning
-  partition, unstablePartition, span, break,
-
-  -- ** Searching
-  elem, notElem, find, findIndex, findIndices, elemIndex, elemIndices,
-
-  -- * Folding
-  foldl, foldl1, foldl', foldl1', foldr, foldr1, foldr', foldr1',
-  ifoldl, ifoldl', ifoldr, ifoldr',
-
-  -- ** Specialised folds
-  all, any, and, or,
-  sum, product,
-  maximum, maximumBy, minimum, minimumBy,
-  minIndex, minIndexBy, maxIndex, maxIndexBy,
-
-  -- ** Monadic folds
-  foldM, ifoldM, foldM', ifoldM',
-  fold1M, fold1M', foldM_, ifoldM_,
-  foldM'_, ifoldM'_, fold1M_, fold1M'_,
-
-  -- * Prefix sums (scans)
-  prescanl, prescanl',
-  postscanl, postscanl',
-  scanl, scanl', scanl1, scanl1',
-  prescanr, prescanr',
-  postscanr, postscanr',
-  scanr, scanr', scanr1, scanr1',
-
-  -- * Conversions
-
-  -- ** Lists
-  toList, fromList, fromListN,
-
-  -- ** Other vector types
-  G.convert,
-
-  -- ** Mutable vectors
-  freeze, thaw, copy, unsafeFreeze, unsafeThaw, unsafeCopy
-) where
-
-import Data.Vector.Unboxed.Base
-import qualified Data.Vector.Generic as G
-import qualified Data.Vector.Fusion.Bundle as Bundle
-import Data.Vector.Fusion.Util ( delayed_min )
-
-import Control.Monad.ST ( ST )
-import Control.Monad.Primitive
-
-import Prelude hiding ( length, null,
-                        replicate, (++), concat,
-                        head, last,
-                        init, tail, take, drop, splitAt, reverse,
-                        map, concatMap,
-                        zipWith, zipWith3, zip, zip3, unzip, unzip3,
-                        filter, takeWhile, dropWhile, span, break,
-                        elem, notElem,
-                        foldl, foldl1, foldr, foldr1,
-                        all, any, and, or, sum, product, minimum, maximum,
-                        scanl, scanl1, scanr, scanr1,
-                        enumFromTo, enumFromThenTo,
-                        mapM, mapM_ )
-
-import Text.Read      ( Read(..), readListPrecDefault )
-import Data.Semigroup ( Semigroup(..) )
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Monoid   ( Monoid(..) )
-import Data.Traversable ( Traversable )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 708
-import qualified GHC.Exts as Exts (IsList(..))
-#endif
-
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Unbox a, Eq a) => Eq (Vector a) where
-  {-# INLINE (==) #-}
-  xs == ys = Bundle.eq (G.stream xs) (G.stream ys)
-
-  {-# INLINE (/=) #-}
-  xs /= ys = not (Bundle.eq (G.stream xs) (G.stream ys))
-
--- See http://trac.haskell.org/vector/ticket/12
-instance (Unbox a, Ord a) => Ord (Vector a) where
-  {-# INLINE compare #-}
-  compare xs ys = Bundle.cmp (G.stream xs) (G.stream ys)
-
-  {-# INLINE (<) #-}
-  xs < ys = Bundle.cmp (G.stream xs) (G.stream ys) == LT
-
-  {-# INLINE (<=) #-}
-  xs <= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= GT
-
-  {-# INLINE (>) #-}
-  xs > ys = Bundle.cmp (G.stream xs) (G.stream ys) == GT
-
-  {-# INLINE (>=) #-}
-  xs >= ys = Bundle.cmp (G.stream xs) (G.stream ys) /= LT
-
-instance Unbox a => Semigroup (Vector a) where
-  {-# INLINE (<>) #-}
-  (<>) = (++)
-
-  {-# INLINE sconcat #-}
-  sconcat = G.concatNE
-
-instance Unbox a => Monoid (Vector a) where
-  {-# INLINE mempty #-}
-  mempty = empty
-
-  {-# INLINE mappend #-}
-  mappend = (++)
-
-  {-# INLINE mconcat #-}
-  mconcat = concat
-
-instance (Show a, Unbox a) => Show (Vector a) where
-  showsPrec = G.showsPrec
-
-instance (Read a, Unbox a) => Read (Vector a) where
-  readPrec = G.readPrec
-  readListPrec = readListPrecDefault
-
-#if __GLASGOW_HASKELL__ >= 708
-
-instance (Unbox e) => Exts.IsList (Vector e) where
-  type Item (Vector e) = e
-  fromList = fromList
-  fromListN = fromListN
-  toList = toList
-
-#endif
-
--- Length information
--- ------------------
-
--- | /O(1)/ Yield the length of the vector
-length :: Unbox a => Vector a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | /O(1)/ Test whether a vector is empty
-null :: Unbox a => Vector a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Indexing
--- --------
-
--- | O(1) Indexing
-(!) :: Unbox a => Vector a -> Int -> a
-{-# INLINE (!) #-}
-(!) = (G.!)
-
--- | O(1) Safe indexing
-(!?) :: Unbox a => Vector a -> Int -> Maybe a
-{-# INLINE (!?) #-}
-(!?) = (G.!?)
-
--- | /O(1)/ First element
-head :: Unbox a => Vector a -> a
-{-# INLINE head #-}
-head = G.head
-
--- | /O(1)/ Last element
-last :: Unbox a => Vector a -> a
-{-# INLINE last #-}
-last = G.last
-
--- | /O(1)/ Unsafe indexing without bounds checking
-unsafeIndex :: Unbox a => Vector a -> Int -> a
-{-# INLINE unsafeIndex #-}
-unsafeIndex = G.unsafeIndex
-
--- | /O(1)/ First element without checking if the vector is empty
-unsafeHead :: Unbox a => Vector a -> a
-{-# INLINE unsafeHead #-}
-unsafeHead = G.unsafeHead
-
--- | /O(1)/ Last element without checking if the vector is empty
-unsafeLast :: Unbox a => Vector a -> a
-{-# INLINE unsafeLast #-}
-unsafeLast = G.unsafeLast
-
--- Monadic indexing
--- ----------------
-
--- | /O(1)/ Indexing in a monad.
---
--- The monad allows operations to be strict in the vector when necessary.
--- Suppose vector copying is implemented like this:
---
--- > copy mv v = ... write mv i (v ! i) ...
---
--- For lazy vectors, @v ! i@ would not be evaluated which means that @mv@
--- would unnecessarily retain a reference to @v@ in each element written.
---
--- With 'indexM', copying can be implemented like this instead:
---
--- > copy mv v = ... do
--- >                   x <- indexM v i
--- >                   write mv i x
---
--- Here, no references to @v@ are retained because indexing (but /not/ the
--- elements) is evaluated eagerly.
---
-indexM :: (Unbox a, Monad m) => Vector a -> Int -> m a
-{-# INLINE indexM #-}
-indexM = G.indexM
-
--- | /O(1)/ First element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-headM :: (Unbox a, Monad m) => Vector a -> m a
-{-# INLINE headM #-}
-headM = G.headM
-
--- | /O(1)/ Last element of a vector in a monad. See 'indexM' for an
--- explanation of why this is useful.
-lastM :: (Unbox a, Monad m) => Vector a -> m a
-{-# INLINE lastM #-}
-lastM = G.lastM
-
--- | /O(1)/ Indexing in a monad without bounds checks. See 'indexM' for an
--- explanation of why this is useful.
-unsafeIndexM :: (Unbox a, Monad m) => Vector a -> Int -> m a
-{-# INLINE unsafeIndexM #-}
-unsafeIndexM = G.unsafeIndexM
-
--- | /O(1)/ First element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeHeadM :: (Unbox a, Monad m) => Vector a -> m a
-{-# INLINE unsafeHeadM #-}
-unsafeHeadM = G.unsafeHeadM
-
--- | /O(1)/ Last element in a monad without checking for empty vectors.
--- See 'indexM' for an explanation of why this is useful.
-unsafeLastM :: (Unbox a, Monad m) => Vector a -> m a
-{-# INLINE unsafeLastM #-}
-unsafeLastM = G.unsafeLastM
-
--- Extracting subvectors (slicing)
--- -------------------------------
-
--- | /O(1)/ Yield a slice of the vector without copying it. The vector must
--- contain at least @i+n@ elements.
-slice :: Unbox a => Int   -- ^ @i@ starting index
-                 -> Int   -- ^ @n@ length
-                 -> Vector a
-                 -> Vector a
-{-# INLINE slice #-}
-slice = G.slice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty.
-init :: Unbox a => Vector a -> Vector a
-{-# INLINE init #-}
-init = G.init
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty.
-tail :: Unbox a => Vector a -> Vector a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | /O(1)/ Yield at the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case it is returned unchanged.
-take :: Unbox a => Int -> Vector a -> Vector a
-{-# INLINE take #-}
-take = G.take
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector may
--- contain less than @n@ elements in which case an empty vector is returned.
-drop :: Unbox a => Int -> Vector a -> Vector a
-{-# INLINE drop #-}
-drop = G.drop
-
--- | /O(1)/ Yield the first @n@ elements paired with the remainder without copying.
---
--- Note that @'splitAt' n v@ is equivalent to @('take' n v, 'drop' n v)@
--- but slightly more efficient.
-{-# INLINE splitAt #-}
-splitAt :: Unbox a => Int -> Vector a -> (Vector a, Vector a)
-splitAt = G.splitAt
-
--- | /O(1)/ Yield a slice of the vector without copying. The vector must
--- contain at least @i+n@ elements but this is not checked.
-unsafeSlice :: Unbox a => Int   -- ^ @i@ starting index
-                       -> Int   -- ^ @n@ length
-                       -> Vector a
-                       -> Vector a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
--- | /O(1)/ Yield all but the last element without copying. The vector may not
--- be empty but this is not checked.
-unsafeInit :: Unbox a => Vector a -> Vector a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
--- | /O(1)/ Yield all but the first element without copying. The vector may not
--- be empty but this is not checked.
-unsafeTail :: Unbox a => Vector a -> Vector a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- | /O(1)/ Yield the first @n@ elements without copying. The vector must
--- contain at least @n@ elements but this is not checked.
-unsafeTake :: Unbox a => Int -> Vector a -> Vector a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
--- | /O(1)/ Yield all but the first @n@ elements without copying. The vector
--- must contain at least @n@ elements but this is not checked.
-unsafeDrop :: Unbox a => Int -> Vector a -> Vector a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
--- Initialisation
--- --------------
-
--- | /O(1)/ Empty vector
-empty :: Unbox a => Vector a
-{-# INLINE empty #-}
-empty = G.empty
-
--- | /O(1)/ Vector with exactly one element
-singleton :: Unbox a => a -> Vector a
-{-# INLINE singleton #-}
-singleton = G.singleton
-
--- | /O(n)/ Vector of the given length with the same value in each position
-replicate :: Unbox a => Int -> a -> Vector a
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | /O(n)/ Construct a vector of the given length by applying the function to
--- each index
-generate :: Unbox a => Int -> (Int -> a) -> Vector a
-{-# INLINE generate #-}
-generate = G.generate
-
--- | /O(n)/ Apply function n times to value. Zeroth element is original value.
-iterateN :: Unbox a => Int -> (a -> a) -> a -> Vector a
-{-# INLINE iterateN #-}
-iterateN = G.iterateN
-
--- Unfolding
--- ---------
-
--- | /O(n)/ Construct a vector by repeatedly applying the generator function
--- to a seed. The generator function yields 'Just' the next element and the
--- new seed or 'Nothing' if there are no more elements.
---
--- > unfoldr (\n -> if n == 0 then Nothing else Just (n,n-1)) 10
--- >  = <10,9,8,7,6,5,4,3,2,1>
-unfoldr :: Unbox a => (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldr #-}
-unfoldr = G.unfoldr
-
--- | /O(n)/ Construct a vector with at most @n@ elements by repeatedly applying
--- the generator function to a seed. The generator function yields 'Just' the
--- next element and the new seed or 'Nothing' if there are no more elements.
---
--- > unfoldrN 3 (\n -> Just (n,n-1)) 10 = <10,9,8>
-unfoldrN :: Unbox a => Int -> (b -> Maybe (a, b)) -> b -> Vector a
-{-# INLINE unfoldrN #-}
-unfoldrN = G.unfoldrN
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrM :: (Monad m, Unbox a) => (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrM #-}
-unfoldrM = G.unfoldrM
-
--- | /O(n)/ Construct a vector by repeatedly applying the monadic
--- generator function to a seed. The generator function yields 'Just'
--- the next element and the new seed or 'Nothing' if there are no more
--- elements.
-unfoldrNM :: (Monad m, Unbox a) => Int -> (b -> m (Maybe (a, b))) -> b -> m (Vector a)
-{-# INLINE unfoldrNM #-}
-unfoldrNM = G.unfoldrNM
-
--- | /O(n)/ Construct a vector with @n@ elements by repeatedly applying the
--- generator function to the already constructed part of the vector.
---
--- > constructN 3 f = let a = f <> ; b = f <a> ; c = f <a,b> in f <a,b,c>
---
-constructN :: Unbox a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructN #-}
-constructN = G.constructN
-
--- | /O(n)/ Construct a vector with @n@ elements from right to left by
--- repeatedly applying the generator function to the already constructed part
--- of the vector.
---
--- > constructrN 3 f = let a = f <> ; b = f<a> ; c = f <b,a> in f <c,b,a>
---
-constructrN :: Unbox a => Int -> (Vector a -> a) -> Vector a
-{-# INLINE constructrN #-}
-constructrN = G.constructrN
-
--- Enumeration
--- -----------
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+1@
--- etc. This operation is usually more efficient than 'enumFromTo'.
---
--- > enumFromN 5 3 = <5,6,7>
-enumFromN :: (Unbox a, Num a) => a -> Int -> Vector a
-{-# INLINE enumFromN #-}
-enumFromN = G.enumFromN
-
--- | /O(n)/ Yield a vector of the given length containing the values @x@, @x+y@,
--- @x+y+y@ etc. This operations is usually more efficient than 'enumFromThenTo'.
---
--- > enumFromStepN 1 0.1 5 = <1,1.1,1.2,1.3,1.4>
-enumFromStepN :: (Unbox a, Num a) => a -> a -> Int -> Vector a
-{-# INLINE enumFromStepN #-}
-enumFromStepN = G.enumFromStepN
-
--- | /O(n)/ Enumerate values from @x@ to @y@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromN' instead.
-enumFromTo :: (Unbox a, Enum a) => a -> a -> Vector a
-{-# INLINE enumFromTo #-}
-enumFromTo = G.enumFromTo
-
--- | /O(n)/ Enumerate values from @x@ to @y@ with a specific step @z@.
---
--- /WARNING:/ This operation can be very inefficient. If at all possible, use
--- 'enumFromStepN' instead.
-enumFromThenTo :: (Unbox a, Enum a) => a -> a -> a -> Vector a
-{-# INLINE enumFromThenTo #-}
-enumFromThenTo = G.enumFromThenTo
-
--- Concatenation
--- -------------
-
--- | /O(n)/ Prepend an element
-cons :: Unbox a => a -> Vector a -> Vector a
-{-# INLINE cons #-}
-cons = G.cons
-
--- | /O(n)/ Append an element
-snoc :: Unbox a => Vector a -> a -> Vector a
-{-# INLINE snoc #-}
-snoc = G.snoc
-
-infixr 5 ++
--- | /O(m+n)/ Concatenate two vectors
-(++) :: Unbox a => Vector a -> Vector a -> Vector a
-{-# INLINE (++) #-}
-(++) = (G.++)
-
--- | /O(n)/ Concatenate all vectors in the list
-concat :: Unbox a => [Vector a] -> Vector a
-{-# INLINE concat #-}
-concat = G.concat
-
--- Monadic initialisation
--- ----------------------
-
--- | /O(n)/ Execute the monadic action the given number of times and store the
--- results in a vector.
-replicateM :: (Monad m, Unbox a) => Int -> m a -> m (Vector a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | /O(n)/ Construct a vector of the given length by applying the monadic
--- action to each index
-generateM :: (Monad m, Unbox a) => Int -> (Int -> m a) -> m (Vector a)
-{-# INLINE generateM #-}
-generateM = G.generateM
-
--- | /O(n)/ Apply monadic function n times to value. Zeroth element is original value.
-iterateNM :: (Monad m, Unbox a) => Int -> (a -> m a) -> a -> m (Vector a)
-{-# INLINE iterateNM #-}
-iterateNM = G.iterateNM
-
--- | Execute the monadic action and freeze the resulting vector.
---
--- @
--- create (do { v \<- new 2; write v 0 \'a\'; write v 1 \'b\'; return v }) = \<'a','b'\>
--- @
-create :: Unbox a => (forall s. ST s (MVector s a)) -> Vector a
-{-# INLINE create #-}
--- NOTE: eta-expanded due to http://hackage.haskell.org/trac/ghc/ticket/4120
-create p = G.create p
-
--- | Execute the monadic action and freeze the resulting vectors.
-createT :: (Traversable f, Unbox a) => (forall s. ST s (f (MVector s a))) -> f (Vector a)
-{-# INLINE createT #-}
-createT p = G.createT p
-
--- Restricting memory usage
--- ------------------------
-
--- | /O(n)/ Yield the argument but force it not to retain any extra memory,
--- possibly by copying it.
---
--- This is especially useful when dealing with slices. For example:
---
--- > force (slice 0 2 <huge vector>)
---
--- Here, the slice retains a reference to the huge vector. Forcing it creates
--- a copy of just the elements that belong to the slice and allows the huge
--- vector to be garbage collected.
-force :: Unbox a => Vector a -> Vector a
-{-# INLINE force #-}
-force = G.force
-
--- Bulk updates
--- ------------
-
--- | /O(m+n)/ For each pair @(i,a)@ from the list, replace the vector
--- element at position @i@ by @a@.
---
--- > <5,9,2,7> // [(2,1),(0,3),(2,8)] = <3,9,8,7>
---
-(//) :: Unbox a => Vector a   -- ^ initial vector (of length @m@)
-                -> [(Int, a)] -- ^ list of index/value pairs (of length @n@)
-                -> Vector a
-{-# INLINE (//) #-}
-(//) = (G.//)
-
--- | /O(m+n)/ For each pair @(i,a)@ from the vector of index/value pairs,
--- replace the vector element at position @i@ by @a@.
---
--- > update <5,9,2,7> <(2,1),(0,3),(2,8)> = <3,9,8,7>
---
-update :: Unbox a
-       => Vector a        -- ^ initial vector (of length @m@)
-       -> Vector (Int, a) -- ^ vector of index/value pairs (of length @n@)
-       -> Vector a
-{-# INLINE update #-}
-update = G.update
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @a@ from the value vector, replace the element of the
--- initial vector at position @i@ by @a@.
---
--- > update_ <5,9,2,7>  <2,0,2> <1,3,8> = <3,9,8,7>
---
--- The function 'update' provides the same functionality and is usually more
--- convenient.
---
--- @
--- update_ xs is ys = 'update' xs ('zip' is ys)
--- @
-update_ :: Unbox a
-        => Vector a   -- ^ initial vector (of length @m@)
-        -> Vector Int -- ^ index vector (of length @n1@)
-        -> Vector a   -- ^ value vector (of length @n2@)
-        -> Vector a
-{-# INLINE update_ #-}
-update_ = G.update_
-
--- | Same as ('//') but without bounds checking.
-unsafeUpd :: Unbox a => Vector a -> [(Int, a)] -> Vector a
-{-# INLINE unsafeUpd #-}
-unsafeUpd = G.unsafeUpd
-
--- | Same as 'update' but without bounds checking.
-unsafeUpdate :: Unbox a => Vector a -> Vector (Int, a) -> Vector a
-{-# INLINE unsafeUpdate #-}
-unsafeUpdate = G.unsafeUpdate
-
--- | Same as 'update_' but without bounds checking.
-unsafeUpdate_ :: Unbox a => Vector a -> Vector Int -> Vector a -> Vector a
-{-# INLINE unsafeUpdate_ #-}
-unsafeUpdate_ = G.unsafeUpdate_
-
--- Accumulations
--- -------------
-
--- | /O(m+n)/ For each pair @(i,b)@ from the list, replace the vector element
--- @a@ at position @i@ by @f a b@.
---
--- > accum (+) <5,9,2> [(2,4),(1,6),(0,3),(1,7)] = <5+3, 9+6+7, 2+4>
-accum :: Unbox a
-      => (a -> b -> a) -- ^ accumulating function @f@
-      -> Vector a      -- ^ initial vector (of length @m@)
-      -> [(Int,b)]     -- ^ list of index/value pairs (of length @n@)
-      -> Vector a
-{-# INLINE accum #-}
-accum = G.accum
-
--- | /O(m+n)/ For each pair @(i,b)@ from the vector of pairs, replace the vector
--- element @a@ at position @i@ by @f a b@.
---
--- > accumulate (+) <5,9,2> <(2,4),(1,6),(0,3),(1,7)> = <5+3, 9+6+7, 2+4>
-accumulate :: (Unbox a, Unbox b)
-            => (a -> b -> a)  -- ^ accumulating function @f@
-            -> Vector a       -- ^ initial vector (of length @m@)
-            -> Vector (Int,b) -- ^ vector of index/value pairs (of length @n@)
-            -> Vector a
-{-# INLINE accumulate #-}
-accumulate = G.accumulate
-
--- | /O(m+min(n1,n2))/ For each index @i@ from the index vector and the
--- corresponding value @b@ from the the value vector,
--- replace the element of the initial vector at
--- position @i@ by @f a b@.
---
--- > accumulate_ (+) <5,9,2> <2,1,0,1> <4,6,3,7> = <5+3, 9+6+7, 2+4>
---
--- The function 'accumulate' provides the same functionality and is usually more
--- convenient.
---
--- @
--- accumulate_ f as is bs = 'accumulate' f as ('zip' is bs)
--- @
-accumulate_ :: (Unbox a, Unbox b)
-            => (a -> b -> a) -- ^ accumulating function @f@
-            -> Vector a      -- ^ initial vector (of length @m@)
-            -> Vector Int    -- ^ index vector (of length @n1@)
-            -> Vector b      -- ^ value vector (of length @n2@)
-            -> Vector a
-{-# INLINE accumulate_ #-}
-accumulate_ = G.accumulate_
-
--- | Same as 'accum' but without bounds checking.
-unsafeAccum :: Unbox a => (a -> b -> a) -> Vector a -> [(Int,b)] -> Vector a
-{-# INLINE unsafeAccum #-}
-unsafeAccum = G.unsafeAccum
-
--- | Same as 'accumulate' but without bounds checking.
-unsafeAccumulate :: (Unbox a, Unbox b)
-                => (a -> b -> a) -> Vector a -> Vector (Int,b) -> Vector a
-{-# INLINE unsafeAccumulate #-}
-unsafeAccumulate = G.unsafeAccumulate
-
--- | Same as 'accumulate_' but without bounds checking.
-unsafeAccumulate_ :: (Unbox a, Unbox b) =>
-               (a -> b -> a) -> Vector a -> Vector Int -> Vector b -> Vector a
-{-# INLINE unsafeAccumulate_ #-}
-unsafeAccumulate_ = G.unsafeAccumulate_
-
--- Permutations
--- ------------
-
--- | /O(n)/ Reverse a vector
-reverse :: Unbox a => Vector a -> Vector a
-{-# INLINE reverse #-}
-reverse = G.reverse
-
--- | /O(n)/ Yield the vector obtained by replacing each element @i@ of the
--- index vector by @xs'!'i@. This is equivalent to @'map' (xs'!') is@ but is
--- often much more efficient.
---
--- > backpermute <a,b,c,d> <0,3,2,3,1,0> = <a,d,c,d,b,a>
-backpermute :: Unbox a => Vector a -> Vector Int -> Vector a
-{-# INLINE backpermute #-}
-backpermute = G.backpermute
-
--- | Same as 'backpermute' but without bounds checking.
-unsafeBackpermute :: Unbox a => Vector a -> Vector Int -> Vector a
-{-# INLINE unsafeBackpermute #-}
-unsafeBackpermute = G.unsafeBackpermute
-
--- Safe destructive updates
--- ------------------------
-
--- | Apply a destructive operation to a vector. The operation will be
--- performed in place if it is safe to do so and will modify a copy of the
--- vector otherwise.
---
--- @
--- modify (\\v -> write v 0 \'x\') ('replicate' 3 \'a\') = \<\'x\',\'a\',\'a\'\>
--- @
-modify :: Unbox a => (forall s. MVector s a -> ST s ()) -> Vector a -> Vector a
-{-# INLINE modify #-}
-modify p = G.modify p
-
--- Indexing
--- --------
-
--- | /O(n)/ Pair each element in a vector with its index
-indexed :: Unbox a => Vector a -> Vector (Int,a)
-{-# INLINE indexed #-}
-indexed = G.indexed
-
--- Mapping
--- -------
-
--- | /O(n)/ Map a function over a vector
-map :: (Unbox a, Unbox b) => (a -> b) -> Vector a -> Vector b
-{-# INLINE map #-}
-map = G.map
-
--- | /O(n)/ Apply a function to every element of a vector and its index
-imap :: (Unbox a, Unbox b) => (Int -> a -> b) -> Vector a -> Vector b
-{-# INLINE imap #-}
-imap = G.imap
-
--- | Map a function over a vector and concatenate the results.
-concatMap :: (Unbox a, Unbox b) => (a -> Vector b) -> Vector a -> Vector b
-{-# INLINE concatMap #-}
-concatMap = G.concatMap
-
--- Monadic mapping
--- ---------------
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results
-mapM :: (Monad m, Unbox a, Unbox b) => (a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE mapM #-}
-mapM = G.mapM
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, yielding a vector of results
-imapM :: (Monad m, Unbox a, Unbox b)
-      => (Int -> a -> m b) -> Vector a -> m (Vector b)
-{-# INLINE imapM #-}
-imapM = G.imapM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results
-mapM_ :: (Monad m, Unbox a) => (a -> m b) -> Vector a -> m ()
-{-# INLINE mapM_ #-}
-mapM_ = G.mapM_
-
--- | /O(n)/ Apply the monadic action to every element of a vector and its
--- index, ignoring the results
-imapM_ :: (Monad m, Unbox a) => (Int -> a -> m b) -> Vector a -> m ()
-{-# INLINE imapM_ #-}
-imapM_ = G.imapM_
-
--- | /O(n)/ Apply the monadic action to all elements of the vector, yielding a
--- vector of results. Equivalent to @flip 'mapM'@.
-forM :: (Monad m, Unbox a, Unbox b) => Vector a -> (a -> m b) -> m (Vector b)
-{-# INLINE forM #-}
-forM = G.forM
-
--- | /O(n)/ Apply the monadic action to all elements of a vector and ignore the
--- results. Equivalent to @flip 'mapM_'@.
-forM_ :: (Monad m, Unbox a) => Vector a -> (a -> m b) -> m ()
-{-# INLINE forM_ #-}
-forM_ = G.forM_
-
--- Zipping
--- -------
-
--- | /O(min(m,n))/ Zip two vectors with the given function.
-zipWith :: (Unbox a, Unbox b, Unbox c)
-        => (a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE zipWith #-}
-zipWith = G.zipWith
-
--- | Zip three vectors with the given function.
-zipWith3 :: (Unbox a, Unbox b, Unbox c, Unbox d)
-         => (a -> b -> c -> d) -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE zipWith3 #-}
-zipWith3 = G.zipWith3
-
-zipWith4 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e)
-         => (a -> b -> c -> d -> e)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE zipWith4 #-}
-zipWith4 = G.zipWith4
-
-zipWith5 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e, Unbox f)
-         => (a -> b -> c -> d -> e -> f)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f
-{-# INLINE zipWith5 #-}
-zipWith5 = G.zipWith5
-
-zipWith6 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e, Unbox f, Unbox g)
-         => (a -> b -> c -> d -> e -> f -> g)
-         -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-         -> Vector f -> Vector g
-{-# INLINE zipWith6 #-}
-zipWith6 = G.zipWith6
-
--- | /O(min(m,n))/ Zip two vectors with a function that also takes the
--- elements' indices.
-izipWith :: (Unbox a, Unbox b, Unbox c)
-         => (Int -> a -> b -> c) -> Vector a -> Vector b -> Vector c
-{-# INLINE izipWith #-}
-izipWith = G.izipWith
-
--- | Zip three vectors and their indices with the given function.
-izipWith3 :: (Unbox a, Unbox b, Unbox c, Unbox d)
-          => (Int -> a -> b -> c -> d)
-          -> Vector a -> Vector b -> Vector c -> Vector d
-{-# INLINE izipWith3 #-}
-izipWith3 = G.izipWith3
-
-izipWith4 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e)
-          => (Int -> a -> b -> c -> d -> e)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-{-# INLINE izipWith4 #-}
-izipWith4 = G.izipWith4
-
-izipWith5 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e, Unbox f)
-          => (Int -> a -> b -> c -> d -> e -> f)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f
-{-# INLINE izipWith5 #-}
-izipWith5 = G.izipWith5
-
-izipWith6 :: (Unbox a, Unbox b, Unbox c, Unbox d, Unbox e, Unbox f, Unbox g)
-          => (Int -> a -> b -> c -> d -> e -> f -> g)
-          -> Vector a -> Vector b -> Vector c -> Vector d -> Vector e
-          -> Vector f -> Vector g
-{-# INLINE izipWith6 #-}
-izipWith6 = G.izipWith6
-
--- Monadic zipping
--- ---------------
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and yield a
--- vector of results
-zipWithM :: (Monad m, Unbox a, Unbox b, Unbox c)
-         => (a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE zipWithM #-}
-zipWithM = G.zipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and yield a vector of results
-izipWithM :: (Monad m, Unbox a, Unbox b, Unbox c)
-         => (Int -> a -> b -> m c) -> Vector a -> Vector b -> m (Vector c)
-{-# INLINE izipWithM #-}
-izipWithM = G.izipWithM
-
--- | /O(min(m,n))/ Zip the two vectors with the monadic action and ignore the
--- results
-zipWithM_ :: (Monad m, Unbox a, Unbox b)
-          => (a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE zipWithM_ #-}
-zipWithM_ = G.zipWithM_
-
--- | /O(min(m,n))/ Zip the two vectors with a monadic action that also takes
--- the element index and ignore the results
-izipWithM_ :: (Monad m, Unbox a, Unbox b)
-          => (Int -> a -> b -> m c) -> Vector a -> Vector b -> m ()
-{-# INLINE izipWithM_ #-}
-izipWithM_ = G.izipWithM_
-
--- Filtering
--- ---------
-
--- | /O(n)/ Drop elements that do not satisfy the predicate
-filter :: Unbox a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE filter #-}
-filter = G.filter
-
--- | /O(n)/ Drop repeated adjacent elements.
-uniq :: (Unbox a, Eq a) => Vector a -> Vector a
-{-# INLINE uniq #-}
-uniq = G.uniq
-
--- | /O(n)/ Drop elements that do not satisfy the predicate which is applied to
--- values and their indices
-ifilter :: Unbox a => (Int -> a -> Bool) -> Vector a -> Vector a
-{-# INLINE ifilter #-}
-ifilter = G.ifilter
-
--- | /O(n)/ Drop elements when predicate returns Nothing
-mapMaybe :: (Unbox a, Unbox b) => (a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE mapMaybe #-}
-mapMaybe = G.mapMaybe
-
--- | /O(n)/ Drop elements when predicate, applied to index and value, returns Nothing
-imapMaybe :: (Unbox a, Unbox b) => (Int -> a -> Maybe b) -> Vector a -> Vector b
-{-# INLINE imapMaybe #-}
-imapMaybe = G.imapMaybe
-
--- | /O(n)/ Drop elements that do not satisfy the monadic predicate
-filterM :: (Monad m, Unbox a) => (a -> m Bool) -> Vector a -> m (Vector a)
-{-# INLINE filterM #-}
-filterM = G.filterM
-
--- | /O(n)/ Yield the longest prefix of elements satisfying the predicate
--- without copying.
-takeWhile :: Unbox a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE takeWhile #-}
-takeWhile = G.takeWhile
-
--- | /O(n)/ Drop the longest prefix of elements that satisfy the predicate
--- without copying.
-dropWhile :: Unbox a => (a -> Bool) -> Vector a -> Vector a
-{-# INLINE dropWhile #-}
-dropWhile = G.dropWhile
-
--- Parititioning
--- -------------
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't. The
--- relative order of the elements is preserved at the cost of a sometimes
--- reduced performance compared to 'unstablePartition'.
-partition :: Unbox a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE partition #-}
-partition = G.partition
-
--- | /O(n)/ Split the vector in two parts, the first one containing those
--- elements that satisfy the predicate and the second one those that don't.
--- The order of the elements is not preserved but the operation is often
--- faster than 'partition'.
-unstablePartition :: Unbox a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE unstablePartition #-}
-unstablePartition = G.unstablePartition
-
--- | /O(n)/ Split the vector into the longest prefix of elements that satisfy
--- the predicate and the rest without copying.
-span :: Unbox a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE span #-}
-span = G.span
-
--- | /O(n)/ Split the vector into the longest prefix of elements that do not
--- satisfy the predicate and the rest without copying.
-break :: Unbox a => (a -> Bool) -> Vector a -> (Vector a, Vector a)
-{-# INLINE break #-}
-break = G.break
-
--- Searching
--- ---------
-
-infix 4 `elem`
--- | /O(n)/ Check if the vector contains an element
-elem :: (Unbox a, Eq a) => a -> Vector a -> Bool
-{-# INLINE elem #-}
-elem = G.elem
-
-infix 4 `notElem`
--- | /O(n)/ Check if the vector does not contain an element (inverse of 'elem')
-notElem :: (Unbox a, Eq a) => a -> Vector a -> Bool
-{-# INLINE notElem #-}
-notElem = G.notElem
-
--- | /O(n)/ Yield 'Just' the first element matching the predicate or 'Nothing'
--- if no such element exists.
-find :: Unbox a => (a -> Bool) -> Vector a -> Maybe a
-{-# INLINE find #-}
-find = G.find
-
--- | /O(n)/ Yield 'Just' the index of the first element matching the predicate
--- or 'Nothing' if no such element exists.
-findIndex :: Unbox a => (a -> Bool) -> Vector a -> Maybe Int
-{-# INLINE findIndex #-}
-findIndex = G.findIndex
-
--- | /O(n)/ Yield the indices of elements satisfying the predicate in ascending
--- order.
-findIndices :: Unbox a => (a -> Bool) -> Vector a -> Vector Int
-{-# INLINE findIndices #-}
-findIndices = G.findIndices
-
--- | /O(n)/ Yield 'Just' the index of the first occurence of the given element or
--- 'Nothing' if the vector does not contain the element. This is a specialised
--- version of 'findIndex'.
-elemIndex :: (Unbox a, Eq a) => a -> Vector a -> Maybe Int
-{-# INLINE elemIndex #-}
-elemIndex = G.elemIndex
-
--- | /O(n)/ Yield the indices of all occurences of the given element in
--- ascending order. This is a specialised version of 'findIndices'.
-elemIndices :: (Unbox a, Eq a) => a -> Vector a -> Vector Int
-{-# INLINE elemIndices #-}
-elemIndices = G.elemIndices
-
--- Folding
--- -------
-
--- | /O(n)/ Left fold
-foldl :: Unbox b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl #-}
-foldl = G.foldl
-
--- | /O(n)/ Left fold on non-empty vectors
-foldl1 :: Unbox a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1 #-}
-foldl1 = G.foldl1
-
--- | /O(n)/ Left fold with strict accumulator
-foldl' :: Unbox b => (a -> b -> a) -> a -> Vector b -> a
-{-# INLINE foldl' #-}
-foldl' = G.foldl'
-
--- | /O(n)/ Left fold on non-empty vectors with strict accumulator
-foldl1' :: Unbox a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldl1' #-}
-foldl1' = G.foldl1'
-
--- | /O(n)/ Right fold
-foldr :: Unbox a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr #-}
-foldr = G.foldr
-
--- | /O(n)/ Right fold on non-empty vectors
-foldr1 :: Unbox a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1 #-}
-foldr1 = G.foldr1
-
--- | /O(n)/ Right fold with a strict accumulator
-foldr' :: Unbox a => (a -> b -> b) -> b -> Vector a -> b
-{-# INLINE foldr' #-}
-foldr' = G.foldr'
-
--- | /O(n)/ Right fold on non-empty vectors with strict accumulator
-foldr1' :: Unbox a => (a -> a -> a) -> Vector a -> a
-{-# INLINE foldr1' #-}
-foldr1' = G.foldr1'
-
--- | /O(n)/ Left fold (function applied to each element and its index)
-ifoldl :: Unbox b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl #-}
-ifoldl = G.ifoldl
-
--- | /O(n)/ Left fold with strict accumulator (function applied to each element
--- and its index)
-ifoldl' :: Unbox b => (a -> Int -> b -> a) -> a -> Vector b -> a
-{-# INLINE ifoldl' #-}
-ifoldl' = G.ifoldl'
-
--- | /O(n)/ Right fold (function applied to each element and its index)
-ifoldr :: Unbox a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr #-}
-ifoldr = G.ifoldr
-
--- | /O(n)/ Right fold with strict accumulator (function applied to each
--- element and its index)
-ifoldr' :: Unbox a => (Int -> a -> b -> b) -> b -> Vector a -> b
-{-# INLINE ifoldr' #-}
-ifoldr' = G.ifoldr'
-
--- Specialised folds
--- -----------------
-
--- | /O(n)/ Check if all elements satisfy the predicate.
-all :: Unbox a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE all #-}
-all = G.all
-
--- | /O(n)/ Check if any element satisfies the predicate.
-any :: Unbox a => (a -> Bool) -> Vector a -> Bool
-{-# INLINE any #-}
-any = G.any
-
--- | /O(n)/ Check if all elements are 'True'
-and :: Vector Bool -> Bool
-{-# INLINE and #-}
-and = G.and
-
--- | /O(n)/ Check if any element is 'True'
-or :: Vector Bool -> Bool
-{-# INLINE or #-}
-or = G.or
-
--- | /O(n)/ Compute the sum of the elements
-sum :: (Unbox a, Num a) => Vector a -> a
-{-# INLINE sum #-}
-sum = G.sum
-
--- | /O(n)/ Compute the produce of the elements
-product :: (Unbox a, Num a) => Vector a -> a
-{-# INLINE product #-}
-product = G.product
-
--- | /O(n)/ Yield the maximum element of the vector. The vector may not be
--- empty.
-maximum :: (Unbox a, Ord a) => Vector a -> a
-{-# INLINE maximum #-}
-maximum = G.maximum
-
--- | /O(n)/ Yield the maximum element of the vector according to the given
--- comparison function. The vector may not be empty.
-maximumBy :: Unbox a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE maximumBy #-}
-maximumBy = G.maximumBy
-
--- | /O(n)/ Yield the minimum element of the vector. The vector may not be
--- empty.
-minimum :: (Unbox a, Ord a) => Vector a -> a
-{-# INLINE minimum #-}
-minimum = G.minimum
-
--- | /O(n)/ Yield the minimum element of the vector according to the given
--- comparison function. The vector may not be empty.
-minimumBy :: Unbox a => (a -> a -> Ordering) -> Vector a -> a
-{-# INLINE minimumBy #-}
-minimumBy = G.minimumBy
-
--- | /O(n)/ Yield the index of the maximum element of the vector. The vector
--- may not be empty.
-maxIndex :: (Unbox a, Ord a) => Vector a -> Int
-{-# INLINE maxIndex #-}
-maxIndex = G.maxIndex
-
--- | /O(n)/ Yield the index of the maximum element of the vector according to
--- the given comparison function. The vector may not be empty.
-maxIndexBy :: Unbox a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE maxIndexBy #-}
-maxIndexBy = G.maxIndexBy
-
--- | /O(n)/ Yield the index of the minimum element of the vector. The vector
--- may not be empty.
-minIndex :: (Unbox a, Ord a) => Vector a -> Int
-{-# INLINE minIndex #-}
-minIndex = G.minIndex
-
--- | /O(n)/ Yield the index of the minimum element of the vector according to
--- the given comparison function. The vector may not be empty.
-minIndexBy :: Unbox a => (a -> a -> Ordering) -> Vector a -> Int
-{-# INLINE minIndexBy #-}
-minIndexBy = G.minIndexBy
-
--- Monadic folds
--- -------------
-
--- | /O(n)/ Monadic fold
-foldM :: (Monad m, Unbox b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM #-}
-foldM = G.foldM
-
--- | /O(n)/ Monadic fold (action applied to each element and its index)
-ifoldM :: (Monad m, Unbox b) => (a -> Int -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE ifoldM #-}
-ifoldM = G.ifoldM
-
--- | /O(n)/ Monadic fold over non-empty vectors
-fold1M :: (Monad m, Unbox a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M #-}
-fold1M = G.fold1M
-
--- | /O(n)/ Monadic fold with strict accumulator
-foldM' :: (Monad m, Unbox b) => (a -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE foldM' #-}
-foldM' = G.foldM'
-
--- | /O(n)/ Monadic fold with strict accumulator (action applied to each
--- element and its index)
-ifoldM' :: (Monad m, Unbox b) => (a -> Int -> b -> m a) -> a -> Vector b -> m a
-{-# INLINE ifoldM' #-}
-ifoldM' = G.ifoldM'
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
-fold1M' :: (Monad m, Unbox a) => (a -> a -> m a) -> Vector a -> m a
-{-# INLINE fold1M' #-}
-fold1M' = G.fold1M'
-
--- | /O(n)/ Monadic fold that discards the result
-foldM_ :: (Monad m, Unbox b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM_ #-}
-foldM_ = G.foldM_
-
--- | /O(n)/ Monadic fold that discards the result (action applied to each
--- element and its index)
-ifoldM_ :: (Monad m, Unbox b) => (a -> Int -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE ifoldM_ #-}
-ifoldM_ = G.ifoldM_
-
--- | /O(n)/ Monadic fold over non-empty vectors that discards the result
-fold1M_ :: (Monad m, Unbox a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M_ #-}
-fold1M_ = G.fold1M_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
-foldM'_ :: (Monad m, Unbox b) => (a -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE foldM'_ #-}
-foldM'_ = G.foldM'_
-
--- | /O(n)/ Monadic fold with strict accumulator that discards the result
--- (action applied to each element and its index)
-ifoldM'_ :: (Monad m, Unbox b)
-         => (a -> Int -> b -> m a) -> a -> Vector b -> m ()
-{-# INLINE ifoldM'_ #-}
-ifoldM'_ = G.ifoldM'_
-
--- | /O(n)/ Monadic fold over non-empty vectors with strict accumulator
--- that discards the result
-fold1M'_ :: (Monad m, Unbox a) => (a -> a -> m a) -> Vector a -> m ()
-{-# INLINE fold1M'_ #-}
-fold1M'_ = G.fold1M'_
-
--- Prefix sums (scans)
--- -------------------
-
--- | /O(n)/ Prescan
---
--- @
--- prescanl f z = 'init' . 'scanl' f z
--- @
---
--- Example: @prescanl (+) 0 \<1,2,3,4\> = \<0,1,3,6\>@
---
-prescanl :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl #-}
-prescanl = G.prescanl
-
--- | /O(n)/ Prescan with strict accumulator
-prescanl' :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE prescanl' #-}
-prescanl' = G.prescanl'
-
--- | /O(n)/ Scan
---
--- @
--- postscanl f z = 'tail' . 'scanl' f z
--- @
---
--- Example: @postscanl (+) 0 \<1,2,3,4\> = \<1,3,6,10\>@
---
-postscanl :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl #-}
-postscanl = G.postscanl
-
--- | /O(n)/ Scan with strict accumulator
-postscanl' :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE postscanl' #-}
-postscanl' = G.postscanl'
-
--- | /O(n)/ Haskell-style scan
---
--- > scanl f z <x1,...,xn> = <y1,...,y(n+1)>
--- >   where y1 = z
--- >         yi = f y(i-1) x(i-1)
---
--- Example: @scanl (+) 0 \<1,2,3,4\> = \<0,1,3,6,10\>@
---
-scanl :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl #-}
-scanl = G.scanl
-
--- | /O(n)/ Haskell-style scan with strict accumulator
-scanl' :: (Unbox a, Unbox b) => (a -> b -> a) -> a -> Vector b -> Vector a
-{-# INLINE scanl' #-}
-scanl' = G.scanl'
-
--- | /O(n)/ Scan over a non-empty vector
---
--- > scanl f <x1,...,xn> = <y1,...,yn>
--- >   where y1 = x1
--- >         yi = f y(i-1) xi
---
-scanl1 :: Unbox a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1 #-}
-scanl1 = G.scanl1
-
--- | /O(n)/ Scan over a non-empty vector with a strict accumulator
-scanl1' :: Unbox a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanl1' #-}
-scanl1' = G.scanl1'
-
--- | /O(n)/ Right-to-left prescan
---
--- @
--- prescanr f z = 'reverse' . 'prescanl' (flip f) z . 'reverse'
--- @
---
-prescanr :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr #-}
-prescanr = G.prescanr
-
--- | /O(n)/ Right-to-left prescan with strict accumulator
-prescanr' :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE prescanr' #-}
-prescanr' = G.prescanr'
-
--- | /O(n)/ Right-to-left scan
-postscanr :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr #-}
-postscanr = G.postscanr
-
--- | /O(n)/ Right-to-left scan with strict accumulator
-postscanr' :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE postscanr' #-}
-postscanr' = G.postscanr'
-
--- | /O(n)/ Right-to-left Haskell-style scan
-scanr :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr #-}
-scanr = G.scanr
-
--- | /O(n)/ Right-to-left Haskell-style scan with strict accumulator
-scanr' :: (Unbox a, Unbox b) => (a -> b -> b) -> b -> Vector a -> Vector b
-{-# INLINE scanr' #-}
-scanr' = G.scanr'
-
--- | /O(n)/ Right-to-left scan over a non-empty vector
-scanr1 :: Unbox a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1 #-}
-scanr1 = G.scanr1
-
--- | /O(n)/ Right-to-left scan over a non-empty vector with a strict
--- accumulator
-scanr1' :: Unbox a => (a -> a -> a) -> Vector a -> Vector a
-{-# INLINE scanr1' #-}
-scanr1' = G.scanr1'
-
--- Conversions - Lists
--- ------------------------
-
--- | /O(n)/ Convert a vector to a list
-toList :: Unbox a => Vector a -> [a]
-{-# INLINE toList #-}
-toList = G.toList
-
--- | /O(n)/ Convert a list to a vector
-fromList :: Unbox a => [a] -> Vector a
-{-# INLINE fromList #-}
-fromList = G.fromList
-
--- | /O(n)/ Convert the first @n@ elements of a list to a vector
---
--- @
--- fromListN n xs = 'fromList' ('take' n xs)
--- @
-fromListN :: Unbox a => Int -> [a] -> Vector a
-{-# INLINE fromListN #-}
-fromListN = G.fromListN
-
--- Conversions - Mutable vectors
--- -----------------------------
-
--- | /O(1)/ Unsafe convert a mutable vector to an immutable one without
--- copying. The mutable vector may not be used after this operation.
-unsafeFreeze :: (Unbox a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE unsafeFreeze #-}
-unsafeFreeze = G.unsafeFreeze
-
--- | /O(1)/ Unsafely convert an immutable vector to a mutable one without
--- copying. The immutable vector may not be used after this operation.
-unsafeThaw :: (Unbox a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE unsafeThaw #-}
-unsafeThaw = G.unsafeThaw
-
--- | /O(n)/ Yield a mutable copy of the immutable vector.
-thaw :: (Unbox a, PrimMonad m) => Vector a -> m (MVector (PrimState m) a)
-{-# INLINE thaw #-}
-thaw = G.thaw
-
--- | /O(n)/ Yield an immutable copy of the mutable vector.
-freeze :: (Unbox a, PrimMonad m) => MVector (PrimState m) a -> m (Vector a)
-{-# INLINE freeze #-}
-freeze = G.freeze
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length. This is not checked.
-unsafeCopy
-  :: (Unbox a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | /O(n)/ Copy an immutable vector into a mutable one. The two vectors must
--- have the same length.
-copy :: (Unbox a, PrimMonad m) => MVector (PrimState m) a -> Vector a -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
-
-#define DEFINE_IMMUTABLE
-#include "unbox-tuple-instances"
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Base.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Base.hs
deleted file mode 100644
index a88795c5b4bc..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Base.hs
+++ /dev/null
@@ -1,408 +0,0 @@
-{-# LANGUAGE BangPatterns, CPP, MultiParamTypeClasses, TypeFamilies, FlexibleContexts #-}
-#if __GLASGOW_HASKELL__ >= 707
-{-# LANGUAGE DeriveDataTypeable, StandaloneDeriving #-}
-#endif
-{-# OPTIONS_HADDOCK hide #-}
-
--- |
--- Module      : Data.Vector.Unboxed.Base
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Adaptive unboxed vectors: basic implementation
---
-
-module Data.Vector.Unboxed.Base (
-  MVector(..), IOVector, STVector, Vector(..), Unbox
-) where
-
-import qualified Data.Vector.Generic         as G
-import qualified Data.Vector.Generic.Mutable as M
-
-import qualified Data.Vector.Primitive as P
-
-import Control.DeepSeq ( NFData(rnf) )
-
-import Control.Monad.Primitive
-import Control.Monad ( liftM )
-
-import Data.Word ( Word8, Word16, Word32, Word64 )
-import Data.Int  ( Int8, Int16, Int32, Int64 )
-import Data.Complex
-
-#if !MIN_VERSION_base(4,8,0)
-import Data.Word ( Word )
-#endif
-
-#if __GLASGOW_HASKELL__ >= 707
-import Data.Typeable ( Typeable )
-#else
-import Data.Typeable ( Typeable1(..), Typeable2(..), mkTyConApp,
-                       mkTyCon3
-                     )
-#endif
-
-import Data.Data     ( Data(..) )
-
--- Data.Vector.Internal.Check is unused
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
-data family MVector s a
-data family Vector    a
-
-type IOVector = MVector RealWorld
-type STVector s = MVector s
-
-type instance G.Mutable Vector = MVector
-
-class (G.Vector Vector a, M.MVector MVector a) => Unbox a
-
-instance NFData (Vector a) where rnf !_ = ()
-instance NFData (MVector s a) where rnf !_ = ()
-
--- -----------------
--- Data and Typeable
--- -----------------
-#if __GLASGOW_HASKELL__ >= 707
-deriving instance Typeable Vector
-deriving instance Typeable MVector
-#else
-vectorTyCon = mkTyCon3 "vector"
-
-instance Typeable1 Vector where
-  typeOf1 _ = mkTyConApp (vectorTyCon "Data.Vector.Unboxed" "Vector") []
-
-instance Typeable2 MVector where
-  typeOf2 _ = mkTyConApp (vectorTyCon "Data.Vector.Unboxed.Mutable" "MVector") []
-#endif
-
-instance (Data a, Unbox a) => Data (Vector a) where
-  gfoldl       = G.gfoldl
-  toConstr _   = error "toConstr"
-  gunfold _ _  = error "gunfold"
-  dataTypeOf _ = G.mkType "Data.Vector.Unboxed.Vector"
-  dataCast1    = G.dataCast
-
--- ----
--- Unit
--- ----
-
-newtype instance MVector s () = MV_Unit Int
-newtype instance Vector    () = V_Unit Int
-
-instance Unbox ()
-
-instance M.MVector MVector () where
-  {-# INLINE basicLength #-}
-  {-# INLINE basicUnsafeSlice #-}
-  {-# INLINE basicOverlaps #-}
-  {-# INLINE basicUnsafeNew #-}
-  {-# INLINE basicInitialize #-}
-  {-# INLINE basicUnsafeRead #-}
-  {-# INLINE basicUnsafeWrite #-}
-  {-# INLINE basicClear #-}
-  {-# INLINE basicSet #-}
-  {-# INLINE basicUnsafeCopy #-}
-  {-# INLINE basicUnsafeGrow #-}
-
-  basicLength (MV_Unit n) = n
-
-  basicUnsafeSlice _ m (MV_Unit _) = MV_Unit m
-
-  basicOverlaps _ _ = False
-
-  basicUnsafeNew n = return (MV_Unit n)
-
-  -- Nothing to initialize
-  basicInitialize _ = return ()
-
-  basicUnsafeRead (MV_Unit _) _ = return ()
-
-  basicUnsafeWrite (MV_Unit _) _ () = return ()
-
-  basicClear _ = return ()
-
-  basicSet (MV_Unit _) () = return ()
-
-  basicUnsafeCopy (MV_Unit _) (MV_Unit _) = return ()
-
-  basicUnsafeGrow (MV_Unit n) m = return $ MV_Unit (n+m)
-
-instance G.Vector Vector () where
-  {-# INLINE basicUnsafeFreeze #-}
-  basicUnsafeFreeze (MV_Unit n) = return $ V_Unit n
-
-  {-# INLINE basicUnsafeThaw #-}
-  basicUnsafeThaw (V_Unit n) = return $ MV_Unit n
-
-  {-# INLINE basicLength #-}
-  basicLength (V_Unit n) = n
-
-  {-# INLINE basicUnsafeSlice #-}
-  basicUnsafeSlice _ m (V_Unit _) = V_Unit m
-
-  {-# INLINE basicUnsafeIndexM #-}
-  basicUnsafeIndexM (V_Unit _) _ = return ()
-
-  {-# INLINE basicUnsafeCopy #-}
-  basicUnsafeCopy (MV_Unit _) (V_Unit _) = return ()
-
-  {-# INLINE elemseq #-}
-  elemseq _ = seq
-
-
--- ---------------
--- Primitive types
--- ---------------
-
-#define primMVector(ty,con)                                             \
-instance M.MVector MVector ty where {                                   \
-  {-# INLINE basicLength #-}                                            \
-; {-# INLINE basicUnsafeSlice #-}                                       \
-; {-# INLINE basicOverlaps #-}                                          \
-; {-# INLINE basicUnsafeNew #-}                                         \
-; {-# INLINE basicInitialize #-}                                        \
-; {-# INLINE basicUnsafeReplicate #-}                                   \
-; {-# INLINE basicUnsafeRead #-}                                        \
-; {-# INLINE basicUnsafeWrite #-}                                       \
-; {-# INLINE basicClear #-}                                             \
-; {-# INLINE basicSet #-}                                               \
-; {-# INLINE basicUnsafeCopy #-}                                        \
-; {-# INLINE basicUnsafeGrow #-}                                        \
-; basicLength (con v) = M.basicLength v                                 \
-; basicUnsafeSlice i n (con v) = con $ M.basicUnsafeSlice i n v         \
-; basicOverlaps (con v1) (con v2) = M.basicOverlaps v1 v2               \
-; basicUnsafeNew n = con `liftM` M.basicUnsafeNew n                     \
-; basicInitialize (con v) = M.basicInitialize v                         \
-; basicUnsafeReplicate n x = con `liftM` M.basicUnsafeReplicate n x     \
-; basicUnsafeRead (con v) i = M.basicUnsafeRead v i                     \
-; basicUnsafeWrite (con v) i x = M.basicUnsafeWrite v i x               \
-; basicClear (con v) = M.basicClear v                                   \
-; basicSet (con v) x = M.basicSet v x                                   \
-; basicUnsafeCopy (con v1) (con v2) = M.basicUnsafeCopy v1 v2           \
-; basicUnsafeMove (con v1) (con v2) = M.basicUnsafeMove v1 v2           \
-; basicUnsafeGrow (con v) n = con `liftM` M.basicUnsafeGrow v n }
-
-#define primVector(ty,con,mcon)                                         \
-instance G.Vector Vector ty where {                                     \
-  {-# INLINE basicUnsafeFreeze #-}                                      \
-; {-# INLINE basicUnsafeThaw #-}                                        \
-; {-# INLINE basicLength #-}                                            \
-; {-# INLINE basicUnsafeSlice #-}                                       \
-; {-# INLINE basicUnsafeIndexM #-}                                      \
-; {-# INLINE elemseq #-}                                                \
-; basicUnsafeFreeze (mcon v) = con `liftM` G.basicUnsafeFreeze v        \
-; basicUnsafeThaw (con v) = mcon `liftM` G.basicUnsafeThaw v            \
-; basicLength (con v) = G.basicLength v                                 \
-; basicUnsafeSlice i n (con v) = con $ G.basicUnsafeSlice i n v         \
-; basicUnsafeIndexM (con v) i = G.basicUnsafeIndexM v i                 \
-; basicUnsafeCopy (mcon mv) (con v) = G.basicUnsafeCopy mv v            \
-; elemseq _ = seq }
-
-newtype instance MVector s Int = MV_Int (P.MVector s Int)
-newtype instance Vector    Int = V_Int  (P.Vector    Int)
-instance Unbox Int
-primMVector(Int, MV_Int)
-primVector(Int, V_Int, MV_Int)
-
-newtype instance MVector s Int8 = MV_Int8 (P.MVector s Int8)
-newtype instance Vector    Int8 = V_Int8  (P.Vector    Int8)
-instance Unbox Int8
-primMVector(Int8, MV_Int8)
-primVector(Int8, V_Int8, MV_Int8)
-
-newtype instance MVector s Int16 = MV_Int16 (P.MVector s Int16)
-newtype instance Vector    Int16 = V_Int16  (P.Vector    Int16)
-instance Unbox Int16
-primMVector(Int16, MV_Int16)
-primVector(Int16, V_Int16, MV_Int16)
-
-newtype instance MVector s Int32 = MV_Int32 (P.MVector s Int32)
-newtype instance Vector    Int32 = V_Int32  (P.Vector    Int32)
-instance Unbox Int32
-primMVector(Int32, MV_Int32)
-primVector(Int32, V_Int32, MV_Int32)
-
-newtype instance MVector s Int64 = MV_Int64 (P.MVector s Int64)
-newtype instance Vector    Int64 = V_Int64  (P.Vector    Int64)
-instance Unbox Int64
-primMVector(Int64, MV_Int64)
-primVector(Int64, V_Int64, MV_Int64)
-
-
-newtype instance MVector s Word = MV_Word (P.MVector s Word)
-newtype instance Vector    Word = V_Word  (P.Vector    Word)
-instance Unbox Word
-primMVector(Word, MV_Word)
-primVector(Word, V_Word, MV_Word)
-
-newtype instance MVector s Word8 = MV_Word8 (P.MVector s Word8)
-newtype instance Vector    Word8 = V_Word8  (P.Vector    Word8)
-instance Unbox Word8
-primMVector(Word8, MV_Word8)
-primVector(Word8, V_Word8, MV_Word8)
-
-newtype instance MVector s Word16 = MV_Word16 (P.MVector s Word16)
-newtype instance Vector    Word16 = V_Word16  (P.Vector    Word16)
-instance Unbox Word16
-primMVector(Word16, MV_Word16)
-primVector(Word16, V_Word16, MV_Word16)
-
-newtype instance MVector s Word32 = MV_Word32 (P.MVector s Word32)
-newtype instance Vector    Word32 = V_Word32  (P.Vector    Word32)
-instance Unbox Word32
-primMVector(Word32, MV_Word32)
-primVector(Word32, V_Word32, MV_Word32)
-
-newtype instance MVector s Word64 = MV_Word64 (P.MVector s Word64)
-newtype instance Vector    Word64 = V_Word64  (P.Vector    Word64)
-instance Unbox Word64
-primMVector(Word64, MV_Word64)
-primVector(Word64, V_Word64, MV_Word64)
-
-
-newtype instance MVector s Float = MV_Float (P.MVector s Float)
-newtype instance Vector    Float = V_Float  (P.Vector    Float)
-instance Unbox Float
-primMVector(Float, MV_Float)
-primVector(Float, V_Float, MV_Float)
-
-newtype instance MVector s Double = MV_Double (P.MVector s Double)
-newtype instance Vector    Double = V_Double  (P.Vector    Double)
-instance Unbox Double
-primMVector(Double, MV_Double)
-primVector(Double, V_Double, MV_Double)
-
-
-newtype instance MVector s Char = MV_Char (P.MVector s Char)
-newtype instance Vector    Char = V_Char  (P.Vector    Char)
-instance Unbox Char
-primMVector(Char, MV_Char)
-primVector(Char, V_Char, MV_Char)
-
--- ----
--- Bool
--- ----
-
-fromBool :: Bool -> Word8
-{-# INLINE fromBool #-}
-fromBool True = 1
-fromBool False = 0
-
-toBool :: Word8 -> Bool
-{-# INLINE toBool #-}
-toBool 0 = False
-toBool _ = True
-
-newtype instance MVector s Bool = MV_Bool (P.MVector s Word8)
-newtype instance Vector    Bool = V_Bool  (P.Vector    Word8)
-
-instance Unbox Bool
-
-instance M.MVector MVector Bool where
-  {-# INLINE basicLength #-}
-  {-# INLINE basicUnsafeSlice #-}
-  {-# INLINE basicOverlaps #-}
-  {-# INLINE basicUnsafeNew #-}
-  {-# INLINE basicInitialize #-}
-  {-# INLINE basicUnsafeReplicate #-}
-  {-# INLINE basicUnsafeRead #-}
-  {-# INLINE basicUnsafeWrite #-}
-  {-# INLINE basicClear #-}
-  {-# INLINE basicSet #-}
-  {-# INLINE basicUnsafeCopy #-}
-  {-# INLINE basicUnsafeGrow #-}
-  basicLength (MV_Bool v) = M.basicLength v
-  basicUnsafeSlice i n (MV_Bool v) = MV_Bool $ M.basicUnsafeSlice i n v
-  basicOverlaps (MV_Bool v1) (MV_Bool v2) = M.basicOverlaps v1 v2
-  basicUnsafeNew n = MV_Bool `liftM` M.basicUnsafeNew n
-  basicInitialize (MV_Bool v) = M.basicInitialize v
-  basicUnsafeReplicate n x = MV_Bool `liftM` M.basicUnsafeReplicate n (fromBool x)
-  basicUnsafeRead (MV_Bool v) i = toBool `liftM` M.basicUnsafeRead v i
-  basicUnsafeWrite (MV_Bool v) i x = M.basicUnsafeWrite v i (fromBool x)
-  basicClear (MV_Bool v) = M.basicClear v
-  basicSet (MV_Bool v) x = M.basicSet v (fromBool x)
-  basicUnsafeCopy (MV_Bool v1) (MV_Bool v2) = M.basicUnsafeCopy v1 v2
-  basicUnsafeMove (MV_Bool v1) (MV_Bool v2) = M.basicUnsafeMove v1 v2
-  basicUnsafeGrow (MV_Bool v) n = MV_Bool `liftM` M.basicUnsafeGrow v n
-
-instance G.Vector Vector Bool where
-  {-# INLINE basicUnsafeFreeze #-}
-  {-# INLINE basicUnsafeThaw #-}
-  {-# INLINE basicLength #-}
-  {-# INLINE basicUnsafeSlice #-}
-  {-# INLINE basicUnsafeIndexM #-}
-  {-# INLINE elemseq #-}
-  basicUnsafeFreeze (MV_Bool v) = V_Bool `liftM` G.basicUnsafeFreeze v
-  basicUnsafeThaw (V_Bool v) = MV_Bool `liftM` G.basicUnsafeThaw v
-  basicLength (V_Bool v) = G.basicLength v
-  basicUnsafeSlice i n (V_Bool v) = V_Bool $ G.basicUnsafeSlice i n v
-  basicUnsafeIndexM (V_Bool v) i = toBool `liftM` G.basicUnsafeIndexM v i
-  basicUnsafeCopy (MV_Bool mv) (V_Bool v) = G.basicUnsafeCopy mv v
-  elemseq _ = seq
-
--- -------
--- Complex
--- -------
-
-newtype instance MVector s (Complex a) = MV_Complex (MVector s (a,a))
-newtype instance Vector    (Complex a) = V_Complex  (Vector    (a,a))
-
-instance (Unbox a) => Unbox (Complex a)
-
-instance (Unbox a) => M.MVector MVector (Complex a) where
-  {-# INLINE basicLength #-}
-  {-# INLINE basicUnsafeSlice #-}
-  {-# INLINE basicOverlaps #-}
-  {-# INLINE basicUnsafeNew #-}
-  {-# INLINE basicInitialize #-}
-  {-# INLINE basicUnsafeReplicate #-}
-  {-# INLINE basicUnsafeRead #-}
-  {-# INLINE basicUnsafeWrite #-}
-  {-# INLINE basicClear #-}
-  {-# INLINE basicSet #-}
-  {-# INLINE basicUnsafeCopy #-}
-  {-# INLINE basicUnsafeGrow #-}
-  basicLength (MV_Complex v) = M.basicLength v
-  basicUnsafeSlice i n (MV_Complex v) = MV_Complex $ M.basicUnsafeSlice i n v
-  basicOverlaps (MV_Complex v1) (MV_Complex v2) = M.basicOverlaps v1 v2
-  basicUnsafeNew n = MV_Complex `liftM` M.basicUnsafeNew n
-  basicInitialize (MV_Complex v) = M.basicInitialize v
-  basicUnsafeReplicate n (x :+ y) = MV_Complex `liftM` M.basicUnsafeReplicate n (x,y)
-  basicUnsafeRead (MV_Complex v) i = uncurry (:+) `liftM` M.basicUnsafeRead v i
-  basicUnsafeWrite (MV_Complex v) i (x :+ y) = M.basicUnsafeWrite v i (x,y)
-  basicClear (MV_Complex v) = M.basicClear v
-  basicSet (MV_Complex v) (x :+ y) = M.basicSet v (x,y)
-  basicUnsafeCopy (MV_Complex v1) (MV_Complex v2) = M.basicUnsafeCopy v1 v2
-  basicUnsafeMove (MV_Complex v1) (MV_Complex v2) = M.basicUnsafeMove v1 v2
-  basicUnsafeGrow (MV_Complex v) n = MV_Complex `liftM` M.basicUnsafeGrow v n
-
-instance (Unbox a) => G.Vector Vector (Complex a) where
-  {-# INLINE basicUnsafeFreeze #-}
-  {-# INLINE basicUnsafeThaw #-}
-  {-# INLINE basicLength #-}
-  {-# INLINE basicUnsafeSlice #-}
-  {-# INLINE basicUnsafeIndexM #-}
-  {-# INLINE elemseq #-}
-  basicUnsafeFreeze (MV_Complex v) = V_Complex `liftM` G.basicUnsafeFreeze v
-  basicUnsafeThaw (V_Complex v) = MV_Complex `liftM` G.basicUnsafeThaw v
-  basicLength (V_Complex v) = G.basicLength v
-  basicUnsafeSlice i n (V_Complex v) = V_Complex $ G.basicUnsafeSlice i n v
-  basicUnsafeIndexM (V_Complex v) i
-                = uncurry (:+) `liftM` G.basicUnsafeIndexM v i
-  basicUnsafeCopy (MV_Complex mv) (V_Complex v)
-                = G.basicUnsafeCopy mv v
-  elemseq _ (x :+ y) z = G.elemseq (undefined :: Vector a) x
-                       $ G.elemseq (undefined :: Vector a) y z
-
--- ------
--- Tuples
--- ------
-
-#define DEFINE_INSTANCES
-#include "unbox-tuple-instances"
diff --git a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Mutable.hs b/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Mutable.hs
deleted file mode 100644
index cb82acea8f87..000000000000
--- a/third_party/bazel/rules_haskell/examples/vector/Data/Vector/Unboxed/Mutable.hs
+++ /dev/null
@@ -1,307 +0,0 @@
-{-# LANGUAGE CPP #-}
-
--- |
--- Module      : Data.Vector.Unboxed.Mutable
--- Copyright   : (c) Roman Leshchinskiy 2009-2010
--- License     : BSD-style
---
--- Maintainer  : Roman Leshchinskiy <rl@cse.unsw.edu.au>
--- Stability   : experimental
--- Portability : non-portable
---
--- Mutable adaptive unboxed vectors
---
-
-module Data.Vector.Unboxed.Mutable (
-  -- * Mutable vectors of primitive types
-  MVector(..), IOVector, STVector, Unbox,
-
-  -- * Accessors
-
-  -- ** Length information
-  length, null,
-
-  -- ** Extracting subvectors
-  slice, init, tail, take, drop, splitAt,
-  unsafeSlice, unsafeInit, unsafeTail, unsafeTake, unsafeDrop,
-
-  -- ** Overlapping
-  overlaps,
-
-  -- * Construction
-
-  -- ** Initialisation
-  new, unsafeNew, replicate, replicateM, clone,
-
-  -- ** Growing
-  grow, unsafeGrow,
-
-  -- ** Restricting memory usage
-  clear,
-
-  -- * Zipping and unzipping
-  zip, zip3, zip4, zip5, zip6,
-  unzip, unzip3, unzip4, unzip5, unzip6,
-
-  -- * Accessing individual elements
-  read, write, modify, swap,
-  unsafeRead, unsafeWrite, unsafeModify, unsafeSwap,
-
-  -- * Modifying vectors
-  nextPermutation,
-
-  -- ** Filling and copying
-  set, copy, move, unsafeCopy, unsafeMove
-) where
-
-import Data.Vector.Unboxed.Base
-import qualified Data.Vector.Generic.Mutable as G
-import Data.Vector.Fusion.Util ( delayed_min )
-import Control.Monad.Primitive
-
-import Prelude hiding ( length, null, replicate, reverse, map, read,
-                        take, drop, splitAt, init, tail,
-                        zip, zip3, unzip, unzip3 )
-
--- don't import an unused Data.Vector.Internal.Check
-#define NOT_VECTOR_MODULE
-#include "vector.h"
-
--- Length information
--- ------------------
-
--- | Length of the mutable vector.
-length :: Unbox a => MVector s a -> Int
-{-# INLINE length #-}
-length = G.length
-
--- | Check whether the vector is empty
-null :: Unbox a => MVector s a -> Bool
-{-# INLINE null #-}
-null = G.null
-
--- Extracting subvectors
--- ---------------------
-
--- | Yield a part of the mutable vector without copying it.
-slice :: Unbox a => Int -> Int -> MVector s a -> MVector s a
-{-# INLINE slice #-}
-slice = G.slice
-
-take :: Unbox a => Int -> MVector s a -> MVector s a
-{-# INLINE take #-}
-take = G.take
-
-drop :: Unbox a => Int -> MVector s a -> MVector s a
-{-# INLINE drop #-}
-drop = G.drop
-
-splitAt :: Unbox a => Int -> MVector s a -> (MVector s a, MVector s a)
-{-# INLINE splitAt #-}
-splitAt = G.splitAt
-
-init :: Unbox a => MVector s a -> MVector s a
-{-# INLINE init #-}
-init = G.init
-
-tail :: Unbox a => MVector s a -> MVector s a
-{-# INLINE tail #-}
-tail = G.tail
-
--- | Yield a part of the mutable vector without copying it. No bounds checks
--- are performed.
-unsafeSlice :: Unbox a
-            => Int  -- ^ starting index
-            -> Int  -- ^ length of the slice
-            -> MVector s a
-            -> MVector s a
-{-# INLINE unsafeSlice #-}
-unsafeSlice = G.unsafeSlice
-
-unsafeTake :: Unbox a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeTake #-}
-unsafeTake = G.unsafeTake
-
-unsafeDrop :: Unbox a => Int -> MVector s a -> MVector s a
-{-# INLINE unsafeDrop #-}
-unsafeDrop = G.unsafeDrop
-
-unsafeInit :: Unbox a => MVector s a -> MVector s a
-{-# INLINE unsafeInit #-}
-unsafeInit = G.unsafeInit
-
-unsafeTail :: Unbox a => MVector s a -> MVector s a
-{-# INLINE unsafeTail #-}
-unsafeTail = G.unsafeTail
-
--- Overlapping
--- -----------
-
--- | Check whether two vectors overlap.
-overlaps :: Unbox a => MVector s a -> MVector s a -> Bool
-{-# INLINE overlaps #-}
-overlaps = G.overlaps
-
--- Initialisation
--- --------------
-
--- | Create a mutable vector of the given length.
-new :: (PrimMonad m, Unbox a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE new #-}
-new = G.new
-
--- | Create a mutable vector of the given length. The memory is not initialized.
-unsafeNew :: (PrimMonad m, Unbox a) => Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeNew #-}
-unsafeNew = G.unsafeNew
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with an initial value.
-replicate :: (PrimMonad m, Unbox a) => Int -> a -> m (MVector (PrimState m) a)
-{-# INLINE replicate #-}
-replicate = G.replicate
-
--- | Create a mutable vector of the given length (0 if the length is negative)
--- and fill it with values produced by repeatedly executing the monadic action.
-replicateM :: (PrimMonad m, Unbox a) => Int -> m a -> m (MVector (PrimState m) a)
-{-# INLINE replicateM #-}
-replicateM = G.replicateM
-
--- | Create a copy of a mutable vector.
-clone :: (PrimMonad m, Unbox a)
-      => MVector (PrimState m) a -> m (MVector (PrimState m) a)
-{-# INLINE clone #-}
-clone = G.clone
-
--- Growing
--- -------
-
--- | Grow a vector by the given number of elements. The number must be
--- positive.
-grow :: (PrimMonad m, Unbox a)
-              => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE grow #-}
-grow = G.grow
-
--- | Grow a vector by the given number of elements. The number must be
--- positive but this is not checked.
-unsafeGrow :: (PrimMonad m, Unbox a)
-               => MVector (PrimState m) a -> Int -> m (MVector (PrimState m) a)
-{-# INLINE unsafeGrow #-}
-unsafeGrow = G.unsafeGrow
-
--- Restricting memory usage
--- ------------------------
-
--- | Reset all elements of the vector to some undefined value, clearing all
--- references to external objects. This is usually a noop for unboxed vectors.
-clear :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> m ()
-{-# INLINE clear #-}
-clear = G.clear
-
--- Accessing individual elements
--- -----------------------------
-
--- | Yield the element at the given position.
-read :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE read #-}
-read = G.read
-
--- | Replace the element at the given position.
-write :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE write #-}
-write = G.write
-
--- | Modify the element at the given position.
-modify :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE modify #-}
-modify = G.modify
-
--- | Swap the elements at the given positions.
-swap :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE swap #-}
-swap = G.swap
-
-
--- | Yield the element at the given position. No bounds checks are performed.
-unsafeRead :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> Int -> m a
-{-# INLINE unsafeRead #-}
-unsafeRead = G.unsafeRead
-
--- | Replace the element at the given position. No bounds checks are performed.
-unsafeWrite
-    :: (PrimMonad m, Unbox a) =>  MVector (PrimState m) a -> Int -> a -> m ()
-{-# INLINE unsafeWrite #-}
-unsafeWrite = G.unsafeWrite
-
--- | Modify the element at the given position. No bounds checks are performed.
-unsafeModify :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> (a -> a) -> Int -> m ()
-{-# INLINE unsafeModify #-}
-unsafeModify = G.unsafeModify
-
--- | Swap the elements at the given positions. No bounds checks are performed.
-unsafeSwap
-    :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> Int -> Int -> m ()
-{-# INLINE unsafeSwap #-}
-unsafeSwap = G.unsafeSwap
-
--- Filling and copying
--- -------------------
-
--- | Set all elements of the vector to the given value.
-set :: (PrimMonad m, Unbox a) => MVector (PrimState m) a -> a -> m ()
-{-# INLINE set #-}
-set = G.set
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap.
-copy :: (PrimMonad m, Unbox a)
-     => MVector (PrimState m) a   -- ^ target
-     -> MVector (PrimState m) a   -- ^ source
-     -> m ()
-{-# INLINE copy #-}
-copy = G.copy
-
--- | Copy a vector. The two vectors must have the same length and may not
--- overlap. This is not checked.
-unsafeCopy :: (PrimMonad m, Unbox a)
-           => MVector (PrimState m) a   -- ^ target
-           -> MVector (PrimState m) a   -- ^ source
-           -> m ()
-{-# INLINE unsafeCopy #-}
-unsafeCopy = G.unsafeCopy
-
--- | Move the contents of a vector. The two vectors must have the same
--- length.
---
--- If the vectors do not overlap, then this is equivalent to 'copy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-move :: (PrimMonad m, Unbox a)
-                 => MVector (PrimState m) a -> MVector (PrimState m) a -> m ()
-{-# INLINE move #-}
-move = G.move
-
--- | Move the contents of a vector. The two vectors must have the same
--- length, but this is not checked.
---
--- If the vectors do not overlap, then this is equivalent to 'unsafeCopy'.
--- Otherwise, the copying is performed as if the source vector were
--- copied to a temporary vector and then the temporary vector was copied
--- to the target vector.
-unsafeMove :: (PrimMonad m, Unbox a)
-                          => MVector (PrimState m) a   -- ^ target
-                          -> MVector (PrimState m) a   -- ^ source
-                          -> m ()
-{-# INLINE unsafeMove #-}
-unsafeMove = G.unsafeMove
-
--- | Compute the next (lexicographically) permutation of given vector in-place.
---   Returns False when input is the last permtuation
-nextPermutation :: (PrimMonad m,Ord e,Unbox e) => MVector (PrimState m) e -> m Bool
-{-# INLINE nextPermutation #-}
-nextPermutation = G.nextPermutation
-
-#define DEFINE_MUTABLE
-#include "unbox-tuple-instances"