2 % (c) The University of Glasgow 1992-2002
4 \section[Util]{Highly random utility functions}
9 -- general list processing
10 zipEqual, zipWithEqual, zipWith3Equal, zipWith4Equal,
11 zipLazy, stretchZipWith,
12 mapAndUnzip, mapAndUnzip3,
14 lengthExceeds, lengthIs, lengthAtLeast, listLengthCmp, atLength,
26 -- transitive closures
30 mapAccumL, mapAccumR, mapAccumB,
33 takeList, dropList, splitAtList,
36 eqListBy, equalLength, compareLength,
37 thenCmp, cmpList, prefixMatch, suffixMatch, maybePrefixMatch,
52 -- Floating point stuff
56 #include "../includes/ghcconfig.h"
57 #include "HsVersions.h"
59 import Panic ( panic, trace )
62 #if __GLASGOW_HASKELL__ <= 408
63 import EXCEPTION ( catchIO, justIoErrors, raiseInThread )
65 import DATA_IOREF ( IORef, newIORef )
66 import UNSAFE_IO ( unsafePerformIO )
68 import qualified List ( elem, notElem )
71 import List ( zipWith4 )
74 import Char ( isUpper, isAlphaNum, isSpace, ord, isDigit )
80 %************************************************************************
82 \subsection{The Eager monad}
84 %************************************************************************
86 The @Eager@ monad is just an encoding of continuation-passing style,
87 used to allow you to express "do this and then that", mainly to avoid
88 space leaks. It's done with a type synonym to save bureaucracy.
93 type Eager ans a = (a -> ans) -> ans
95 runEager :: Eager a a -> a
96 runEager m = m (\x -> x)
98 appEager :: Eager ans a -> (a -> ans) -> ans
99 appEager m cont = m cont
101 thenEager :: Eager ans a -> (a -> Eager ans b) -> Eager ans b
102 thenEager m k cont = m (\r -> k r cont)
104 returnEager :: a -> Eager ans a
105 returnEager v cont = cont v
107 mapEager :: (a -> Eager ans b) -> [a] -> Eager ans [b]
108 mapEager f [] = returnEager []
109 mapEager f (x:xs) = f x `thenEager` \ y ->
110 mapEager f xs `thenEager` \ ys ->
115 %************************************************************************
117 \subsection{A for loop}
119 %************************************************************************
122 -- Compose a function with itself n times. (nth rather than twice)
123 nTimes :: Int -> (a -> a) -> (a -> a)
126 nTimes n f = f . nTimes (n-1) f
129 %************************************************************************
131 \subsection[Utils-lists]{General list processing}
133 %************************************************************************
136 filterOut :: (a->Bool) -> [a] -> [a]
137 -- Like filter, only reverses the sense of the test
139 filterOut p (x:xs) | p x = filterOut p xs
140 | otherwise = x : filterOut p xs
143 A paranoid @zip@ (and some @zipWith@ friends) that checks the lists
144 are of equal length. Alastair Reid thinks this should only happen if
145 DEBUGging on; hey, why not?
148 zipEqual :: String -> [a] -> [b] -> [(a,b)]
149 zipWithEqual :: String -> (a->b->c) -> [a]->[b]->[c]
150 zipWith3Equal :: String -> (a->b->c->d) -> [a]->[b]->[c]->[d]
151 zipWith4Equal :: String -> (a->b->c->d->e) -> [a]->[b]->[c]->[d]->[e]
155 zipWithEqual _ = zipWith
156 zipWith3Equal _ = zipWith3
157 zipWith4Equal _ = zipWith4
159 zipEqual msg [] [] = []
160 zipEqual msg (a:as) (b:bs) = (a,b) : zipEqual msg as bs
161 zipEqual msg as bs = panic ("zipEqual: unequal lists:"++msg)
163 zipWithEqual msg z (a:as) (b:bs)= z a b : zipWithEqual msg z as bs
164 zipWithEqual msg _ [] [] = []
165 zipWithEqual msg _ _ _ = panic ("zipWithEqual: unequal lists:"++msg)
167 zipWith3Equal msg z (a:as) (b:bs) (c:cs)
168 = z a b c : zipWith3Equal msg z as bs cs
169 zipWith3Equal msg _ [] [] [] = []
170 zipWith3Equal msg _ _ _ _ = panic ("zipWith3Equal: unequal lists:"++msg)
172 zipWith4Equal msg z (a:as) (b:bs) (c:cs) (d:ds)
173 = z a b c d : zipWith4Equal msg z as bs cs ds
174 zipWith4Equal msg _ [] [] [] [] = []
175 zipWith4Equal msg _ _ _ _ _ = panic ("zipWith4Equal: unequal lists:"++msg)
180 -- zipLazy is lazy in the second list (observe the ~)
182 zipLazy :: [a] -> [b] -> [(a,b)]
184 zipLazy (x:xs) ~(y:ys) = (x,y) : zipLazy xs ys
189 stretchZipWith :: (a -> Bool) -> b -> (a->b->c) -> [a] -> [b] -> [c]
190 -- (stretchZipWith p z f xs ys) stretches ys by inserting z in
191 -- the places where p returns *True*
193 stretchZipWith p z f [] ys = []
194 stretchZipWith p z f (x:xs) ys
195 | p x = f x z : stretchZipWith p z f xs ys
196 | otherwise = case ys of
198 (y:ys) -> f x y : stretchZipWith p z f xs ys
203 mapAndUnzip :: (a -> (b, c)) -> [a] -> ([b], [c])
205 mapAndUnzip f [] = ([],[])
209 (rs1, rs2) = mapAndUnzip f xs
213 mapAndUnzip3 :: (a -> (b, c, d)) -> [a] -> ([b], [c], [d])
215 mapAndUnzip3 f [] = ([],[],[])
216 mapAndUnzip3 f (x:xs)
219 (rs1, rs2, rs3) = mapAndUnzip3 f xs
221 (r1:rs1, r2:rs2, r3:rs3)
225 nOfThem :: Int -> a -> [a]
226 nOfThem n thing = replicate n thing
228 -- 'atLength atLen atEnd ls n' unravels list 'ls' to position 'n';
231 -- atLength atLenPred atEndPred ls n
232 -- | n < 0 = atLenPred n
233 -- | length ls < n = atEndPred (n - length ls)
234 -- | otherwise = atLenPred (drop n ls)
236 atLength :: ([a] -> b)
241 atLength atLenPred atEndPred ls n
242 | n < 0 = atEndPred n
243 | otherwise = go n ls
245 go n [] = atEndPred n
246 go 0 ls = atLenPred ls
247 go n (_:xs) = go (n-1) xs
250 lengthExceeds :: [a] -> Int -> Bool
251 -- (lengthExceeds xs n) = (length xs > n)
252 lengthExceeds = atLength notNull (const False)
254 lengthAtLeast :: [a] -> Int -> Bool
255 lengthAtLeast = atLength notNull (== 0)
257 lengthIs :: [a] -> Int -> Bool
258 lengthIs = atLength null (==0)
260 listLengthCmp :: [a] -> Int -> Ordering
261 listLengthCmp = atLength atLen atEnd
265 | x > 0 = LT -- not yet seen 'n' elts, so list length is < n.
271 isSingleton :: [a] -> Bool
272 isSingleton [x] = True
273 isSingleton _ = False
275 notNull :: [a] -> Bool
279 snocView :: [a] -> Maybe ([a],a)
280 -- Split off the last element
281 snocView [] = Nothing
282 snocView xs = go [] xs
284 -- Invariant: second arg is non-empty
285 go acc [x] = Just (reverse acc, x)
286 go acc (x:xs) = go (x:acc) xs
296 Debugging/specialising versions of \tr{elem} and \tr{notElem}
299 isIn, isn'tIn :: (Eq a) => String -> a -> [a] -> Bool
302 isIn msg x ys = elem__ x ys
303 isn'tIn msg x ys = notElem__ x ys
305 --these are here to be SPECIALIZEd (automagically)
307 elem__ x (y:ys) = x==y || elem__ x ys
309 notElem__ x [] = True
310 notElem__ x (y:ys) = x /= y && notElem__ x ys
314 = elem (_ILIT 0) x ys
318 | i ># _ILIT 100 = trace ("Over-long elem in " ++ msg) $
320 | otherwise = x == y || elem (i +# _ILIT(1)) x ys
323 = notElem (_ILIT 0) x ys
325 notElem i x [] = True
327 | i ># _ILIT 100 = trace ("Over-long notElem in " ++ msg) $
328 x `List.notElem` (y:ys)
329 | otherwise = x /= y && notElem (i +# _ILIT(1)) x ys
333 %************************************************************************
335 \subsubsection[Utils-Carsten-mergesort]{A mergesort from Carsten}
337 %************************************************************************
340 Date: Mon, 3 May 93 20:45:23 +0200
341 From: Carsten Kehler Holst <kehler@cs.chalmers.se>
342 To: partain@dcs.gla.ac.uk
343 Subject: natural merge sort beats quick sort [ and it is prettier ]
345 Here is a piece of Haskell code that I'm rather fond of. See it as an
346 attempt to get rid of the ridiculous quick-sort routine. group is
347 quite useful by itself I think it was John's idea originally though I
348 believe the lazy version is due to me [surprisingly complicated].
349 gamma [used to be called] is called gamma because I got inspired by
350 the Gamma calculus. It is not very close to the calculus but does
351 behave less sequentially than both foldr and foldl. One could imagine
352 a version of gamma that took a unit element as well thereby avoiding
353 the problem with empty lists.
355 I've tried this code against
357 1) insertion sort - as provided by haskell
358 2) the normal implementation of quick sort
359 3) a deforested version of quick sort due to Jan Sparud
360 4) a super-optimized-quick-sort of Lennart's
362 If the list is partially sorted both merge sort and in particular
363 natural merge sort wins. If the list is random [ average length of
364 rising subsequences = approx 2 ] mergesort still wins and natural
365 merge sort is marginally beaten by Lennart's soqs. The space
366 consumption of merge sort is a bit worse than Lennart's quick sort
367 approx a factor of 2. And a lot worse if Sparud's bug-fix [see his
368 fpca article ] isn't used because of group.
375 group :: (a -> a -> Bool) -> [a] -> [[a]]
376 -- Given a <= function, group finds maximal contiguous up-runs
377 -- or down-runs in the input list.
378 -- It's stable, in the sense that it never re-orders equal elements
380 -- Date: Mon, 12 Feb 1996 15:09:41 +0000
381 -- From: Andy Gill <andy@dcs.gla.ac.uk>
382 -- Here is a `better' definition of group.
385 group p (x:xs) = group' xs x x (x :)
387 group' [] _ _ s = [s []]
388 group' (x:xs) x_min x_max s
389 | x_max `p` x = group' xs x_min x (s . (x :))
390 | not (x_min `p` x) = group' xs x x_max ((x :) . s)
391 | otherwise = s [] : group' xs x x (x :)
392 -- NB: the 'not' is essential for stablity
393 -- x `p` x_min would reverse equal elements
395 generalMerge :: (a -> a -> Bool) -> [a] -> [a] -> [a]
396 generalMerge p xs [] = xs
397 generalMerge p [] ys = ys
398 generalMerge p (x:xs) (y:ys) | x `p` y = x : generalMerge p xs (y:ys)
399 | otherwise = y : generalMerge p (x:xs) ys
401 -- gamma is now called balancedFold
403 balancedFold :: (a -> a -> a) -> [a] -> a
404 balancedFold f [] = error "can't reduce an empty list using balancedFold"
405 balancedFold f [x] = x
406 balancedFold f l = balancedFold f (balancedFold' f l)
408 balancedFold' :: (a -> a -> a) -> [a] -> [a]
409 balancedFold' f (x:y:xs) = f x y : balancedFold' f xs
410 balancedFold' f xs = xs
412 generalNaturalMergeSort p [] = []
413 generalNaturalMergeSort p xs = (balancedFold (generalMerge p) . group p) xs
416 generalMergeSort p [] = []
417 generalMergeSort p xs = (balancedFold (generalMerge p) . map (: [])) xs
419 mergeSort, naturalMergeSort :: Ord a => [a] -> [a]
421 mergeSort = generalMergeSort (<=)
422 naturalMergeSort = generalNaturalMergeSort (<=)
424 mergeSortLe le = generalMergeSort le
427 sortLe :: (a->a->Bool) -> [a] -> [a]
428 sortLe le = generalNaturalMergeSort le
431 %************************************************************************
433 \subsection[Utils-transitive-closure]{Transitive closure}
435 %************************************************************************
437 This algorithm for transitive closure is straightforward, albeit quadratic.
440 transitiveClosure :: (a -> [a]) -- Successor function
441 -> (a -> a -> Bool) -- Equality predicate
443 -> [a] -- The transitive closure
445 transitiveClosure succ eq xs
449 go done (x:xs) | x `is_in` done = go done xs
450 | otherwise = go (x:done) (succ x ++ xs)
453 x `is_in` (y:ys) | eq x y = True
454 | otherwise = x `is_in` ys
457 %************************************************************************
459 \subsection[Utils-accum]{Accumulating}
461 %************************************************************************
463 @mapAccumL@ behaves like a combination
464 of @map@ and @foldl@;
465 it applies a function to each element of a list, passing an accumulating
466 parameter from left to right, and returning a final value of this
467 accumulator together with the new list.
470 mapAccumL :: (acc -> x -> (acc, y)) -- Function of elt of input list
471 -- and accumulator, returning new
472 -- accumulator and elt of result list
473 -> acc -- Initial accumulator
475 -> (acc, [y]) -- Final accumulator and result list
477 mapAccumL f b [] = (b, [])
478 mapAccumL f b (x:xs) = (b'', x':xs') where
480 (b'', xs') = mapAccumL f b' xs
483 @mapAccumR@ does the same, but working from right to left instead. Its type is
484 the same as @mapAccumL@, though.
487 mapAccumR :: (acc -> x -> (acc, y)) -- Function of elt of input list
488 -- and accumulator, returning new
489 -- accumulator and elt of result list
490 -> acc -- Initial accumulator
492 -> (acc, [y]) -- Final accumulator and result list
494 mapAccumR f b [] = (b, [])
495 mapAccumR f b (x:xs) = (b'', x':xs') where
497 (b', xs') = mapAccumR f b xs
500 Here is the bi-directional version, that works from both left and right.
503 mapAccumB :: (accl -> accr -> x -> (accl, accr,y))
504 -- Function of elt of input list
505 -- and accumulator, returning new
506 -- accumulator and elt of result list
507 -> accl -- Initial accumulator from left
508 -> accr -- Initial accumulator from right
510 -> (accl, accr, [y]) -- Final accumulators and result list
512 mapAccumB f a b [] = (a,b,[])
513 mapAccumB f a b (x:xs) = (a'',b'',y:ys)
515 (a',b'',y) = f a b' x
516 (a'',b',ys) = mapAccumB f a' b xs
519 A strict version of foldl.
522 foldl' :: (a -> b -> a) -> a -> [b] -> a
523 foldl' f z xs = lgo z xs
526 lgo z (x:xs) = (lgo $! (f z x)) xs
529 A combination of foldl with zip. It works with equal length lists.
532 foldl2 :: (acc -> a -> b -> acc) -> acc -> [a] -> [b] -> acc
534 foldl2 k z (a:as) (b:bs) = foldl2 k (k z a b) as bs
537 Count the number of times a predicate is true
540 count :: (a -> Bool) -> [a] -> Int
542 count p (x:xs) | p x = 1 + count p xs
543 | otherwise = count p xs
546 @splitAt@, @take@, and @drop@ but with length of another
547 list giving the break-off point:
550 takeList :: [b] -> [a] -> [a]
555 (y:ys) -> y : takeList xs ys
557 dropList :: [b] -> [a] -> [a]
559 dropList _ xs@[] = xs
560 dropList (_:xs) (_:ys) = dropList xs ys
563 splitAtList :: [b] -> [a] -> ([a], [a])
564 splitAtList [] xs = ([], xs)
565 splitAtList _ xs@[] = (xs, xs)
566 splitAtList (_:xs) (y:ys) = (y:ys', ys'')
568 (ys', ys'') = splitAtList xs ys
573 %************************************************************************
575 \subsection[Utils-comparison]{Comparisons}
577 %************************************************************************
580 eqListBy :: (a->a->Bool) -> [a] -> [a] -> Bool
581 eqListBy eq [] [] = True
582 eqListBy eq (x:xs) (y:ys) = eq x y && eqListBy eq xs ys
583 eqListBy eq xs ys = False
585 equalLength :: [a] -> [b] -> Bool
586 equalLength [] [] = True
587 equalLength (_:xs) (_:ys) = equalLength xs ys
588 equalLength xs ys = False
590 compareLength :: [a] -> [b] -> Ordering
591 compareLength [] [] = EQ
592 compareLength (_:xs) (_:ys) = compareLength xs ys
593 compareLength [] _ys = LT
594 compareLength _xs [] = GT
596 thenCmp :: Ordering -> Ordering -> Ordering
597 {-# INLINE thenCmp #-}
599 thenCmp other any = other
601 cmpList :: (a -> a -> Ordering) -> [a] -> [a] -> Ordering
602 -- `cmpList' uses a user-specified comparer
604 cmpList cmp [] [] = EQ
605 cmpList cmp [] _ = LT
606 cmpList cmp _ [] = GT
607 cmpList cmp (a:as) (b:bs)
608 = case cmp a b of { EQ -> cmpList cmp as bs; xxx -> xxx }
612 prefixMatch :: Eq a => [a] -> [a] -> Bool
613 prefixMatch [] _str = True
614 prefixMatch _pat [] = False
615 prefixMatch (p:ps) (s:ss) | p == s = prefixMatch ps ss
618 maybePrefixMatch :: String -> String -> Maybe String
619 maybePrefixMatch [] rest = Just rest
620 maybePrefixMatch (_:_) [] = Nothing
621 maybePrefixMatch (p:pat) (r:rest)
622 | p == r = maybePrefixMatch pat rest
623 | otherwise = Nothing
625 suffixMatch :: Eq a => [a] -> [a] -> Bool
626 suffixMatch pat str = prefixMatch (reverse pat) (reverse str)
629 %************************************************************************
631 \subsection[Utils-pairs]{Pairs}
633 %************************************************************************
635 The following are curried versions of @fst@ and @snd@.
639 cfst :: a -> b -> a -- stranal-sem only (Note)
644 The following provide us higher order functions that, when applied
645 to a function, operate on pairs.
649 applyToPair :: ((a -> c),(b -> d)) -> (a,b) -> (c,d)
650 applyToPair (f,g) (x,y) = (f x, g y)
652 applyToFst :: (a -> c) -> (a,b)-> (c,b)
653 applyToFst f (x,y) = (f x,y)
655 applyToSnd :: (b -> d) -> (a,b) -> (a,d)
656 applyToSnd f (x,y) = (x,f y)
661 unzipWith :: (a -> b -> c) -> [(a, b)] -> [c]
662 unzipWith f pairs = map ( \ (a, b) -> f a b ) pairs
666 seqList :: [a] -> b -> b
668 seqList (x:xs) b = x `seq` seqList xs b
674 global :: a -> IORef a
675 global a = unsafePerformIO (newIORef a)
681 looksLikeModuleName [] = False
682 looksLikeModuleName (c:cs) = isUpper c && go cs
684 go ('.':cs) = looksLikeModuleName cs
685 go (c:cs) = (isAlphaNum c || c == '_') && go cs
688 Akin to @Prelude.words@, but sensitive to dquoted entities treating
689 them as single words.
692 toArgs :: String -> [String]
695 case break (\ ch -> isSpace ch || ch == '"') (dropWhile isSpace s) of -- "
697 (\ ws -> if null w then ws else w : ws) $
701 | x /= '"' -> toArgs xs
704 ((str,rs):_) -> stripQuotes str : toArgs rs
707 -- strip away dquotes; assume first and last chars contain quotes.
708 stripQuotes :: String -> String
709 stripQuotes ('"':xs) = init xs
713 -- -----------------------------------------------------------------------------
717 readRational__ :: ReadS Rational -- NB: doesn't handle leading "-"
718 readRational__ r = do
721 return ((n%1)*10^^(k-d), t)
724 (ds,s) <- lexDecDigits r
725 (ds',t) <- lexDotDigits s
726 return (read (ds++ds'), length ds', t)
728 readExp (e:s) | e `elem` "eE" = readExp' s
729 readExp s = return (0,s)
731 readExp' ('+':s) = readDec s
732 readExp' ('-':s) = do
735 readExp' s = readDec s
738 (ds,r) <- nonnull isDigit s
739 return (foldl1 (\n d -> n * 10 + d) [ ord d - ord '0' | d <- ds ],
742 lexDecDigits = nonnull isDigit
744 lexDotDigits ('.':s) = return (span isDigit s)
745 lexDotDigits s = return ("",s)
747 nonnull p s = do (cs@(_:_),t) <- return (span p s)
750 readRational :: String -> Rational -- NB: *does* handle a leading "-"
753 '-' : xs -> - (read_me xs)
757 = case (do { (x,"") <- readRational__ s ; return x }) of
759 [] -> error ("readRational: no parse:" ++ top_s)
760 _ -> error ("readRational: ambiguous parse:" ++ top_s)