1 %*********************************************************************************
3 %* John Hughes's and Simon Peyton Jones's Pretty Printer Combinators *
5 %* based on "The Design of a Pretty-printing Library" *
6 %* in Advanced Functional Programming, *
7 %* Johan Jeuring and Erik Meijer (eds), LNCS 925 *
8 %* http://www.cs.chalmers.se/~rjmh/Papers/pretty.ps *
10 %* Heavily modified by Simon Peyton Jones, Dec 96 *
12 %*********************************************************************************
14 Version 3.0 28 May 1997
15 * Cured massive performance bug. If you write
17 foldl <> empty (map (text.show) [1..10000])
19 you get quadratic behaviour with V2.0. Why? For just the same reason as you get
20 quadratic behaviour with left-associated (++) chains.
22 This is really bad news. One thing a pretty-printer abstraction should
23 certainly guarantee is insensivity to associativity. It matters: suddenly
24 GHC's compilation times went up by a factor of 100 when I switched to the
27 I fixed it with a bit of a hack (because I wanted to get GHC back on the
28 road). I added two new constructors to the Doc type, Above and Beside:
33 Then, where I need to get to a "TextBeside" or "NilAbove" form I "force"
34 the Doc to squeeze out these suspended calls to Beside and Above; but in so
35 doing I re-associate. It's quite simple, but I'm not satisfied that I've done
36 the best possible job. I'll send you the code if you are interested.
40 int, integer, float, double, rational,
41 lparen, rparen, lbrack, rbrack, lbrace, rbrace,
43 * fullRender's type signature has changed. Rather than producing a string it
44 now takes an extra couple of arguments that tells it how to glue fragments
49 -> Float -- Ribbons per line
50 -> (TextDetails -> a -> a) -- What to do with text
51 -> a -- What to do at the end
55 The "fragments" are encapsulated in the TextDetails data type:
56 data TextDetails = Chr Char
60 The Chr and Str constructors are obvious enough. The PStr constructor has a packed
61 string (FastString) inside it. It's generated by using the new "ptext" export.
63 An advantage of this new setup is that you can get the renderer to do output
64 directly (by passing in a function of type (TextDetails -> IO () -> IO ()),
65 rather than producing a string that you then print.
68 Version 2.0 24 April 1997
69 * Made empty into a left unit for <> as well as a right unit;
70 it is also now true that
72 which wasn't true before.
74 * Fixed an obscure bug in sep that occassionally gave very wierd behaviour
78 * Corrected and tidied up the laws and invariants
80 ======================================================================
81 Relative to John's original paper, there are the following new features:
83 1. There's an empty document, "empty". It's a left and right unit for
84 both <> and $$, and anywhere in the argument list for
85 sep, hcat, hsep, vcat, fcat etc.
87 It is Really Useful in practice.
89 2. There is a paragraph-fill combinator, fsep, that's much like sep,
90 only it keeps fitting things on one line until it can't fit any more.
92 3. Some random useful extra combinators are provided.
93 <+> puts its arguments beside each other with a space between them,
94 unless either argument is empty in which case it returns the other
97 hcat is a list version of <>
98 hsep is a list version of <+>
99 vcat is a list version of $$
101 sep (separate) is either like hsep or like vcat, depending on what fits
103 cat is behaves like sep, but it uses <> for horizontal conposition
104 fcat is behaves like fsep, but it uses <> for horizontal conposition
106 These new ones do the obvious things:
107 char, semi, comma, colon, space,
108 parens, brackets, braces,
111 4. The "above" combinator, $$, now overlaps its two arguments if the
112 last line of the top argument stops before the first line of the second begins.
113 For example: text "hi" $$ nest 5 "there"
120 There are two places this is really useful
122 a) When making labelled blocks, like this:
123 Left -> code for left
124 Right -> code for right
126 code for longlonglonglabel
127 The block is on the same line as the label if the label is
128 short, but on the next line otherwise.
130 b) When laying out lists like this:
135 which some people like. But if the list fits on one line
136 you want [first, second, third]. You can't do this with
137 John's original combinators, but it's quite easy with the
140 The combinator $+$ gives the original "never-overlap" behaviour.
142 5. Several different renderers are provided:
144 * one that uses cut-marks to avoid deeply-nested documents
145 simply piling up in the right-hand margin
146 * one that ignores indentation (fewer chars output; good for machines)
147 * one that ignores indentation and newlines (ditto, only more so)
149 6. Numerous implementation tidy-ups
150 Use of unboxed data types to speed up the implementation
155 {-# LANGUAGE BangPatterns #-}
156 {-# OPTIONS -fno-warn-unused-imports #-}
157 -- XXX GHC 6.9 seems to be confused by unpackCString# being used only in
162 Mode(..), TextDetails(..),
164 empty, isEmpty, nest,
166 char, text, ftext, ptext, zeroWidthText,
167 int, integer, float, double, rational,
168 parens, brackets, braces, quotes, doubleQuotes,
169 semi, comma, colon, space, equals,
170 lparen, rparen, lbrack, rbrack, lbrace, rbrace, cparen,
172 (<>), (<+>), hcat, hsep,
179 -- renderStyle, -- Haskell 1.3 only
180 render, fullRender, printDoc, showDocWith,
181 bufLeftRender -- performance hack
189 import Numeric (fromRat)
191 --import Foreign.Ptr (castPtr)
193 #if defined(__GLASGOW_HASKELL__)
195 import GHC.Base ( unpackCString# )
196 import GHC.Exts ( Int# )
197 import GHC.Ptr ( Ptr(..) )
200 -- Don't import Util( assertPanic ) because it makes a loop in the module structure
210 -- Disable ASSERT checks; they are expensive!
211 #define LOCAL_ASSERT(x)
216 %*********************************************************
218 \subsection{The interface}
220 %*********************************************************
222 The primitive @Doc@ values
226 isEmpty :: Doc -> Bool
227 -- | Some text, but without any width. Use for non-printing text
228 -- such as a HTML or Latex tags
229 zeroWidthText :: String -> Doc
231 text :: String -> Doc
234 semi, comma, colon, space, equals :: Doc
235 lparen, rparen, lbrack, rbrack, lbrace, rbrace :: Doc
237 parens, brackets, braces :: Doc -> Doc
238 quotes, doubleQuotes :: Doc -> Doc
241 integer :: Integer -> Doc
242 float :: Float -> Doc
243 double :: Double -> Doc
244 rational :: Rational -> Doc
247 Combining @Doc@ values
250 (<>) :: Doc -> Doc -> Doc -- Beside
251 hcat :: [Doc] -> Doc -- List version of <>
252 (<+>) :: Doc -> Doc -> Doc -- Beside, separated by space
253 hsep :: [Doc] -> Doc -- List version of <+>
255 ($$) :: Doc -> Doc -> Doc -- Above; if there is no
256 -- overlap it "dovetails" the two
257 vcat :: [Doc] -> Doc -- List version of $$
259 cat :: [Doc] -> Doc -- Either hcat or vcat
260 sep :: [Doc] -> Doc -- Either hsep or vcat
261 fcat :: [Doc] -> Doc -- ``Paragraph fill'' version of cat
262 fsep :: [Doc] -> Doc -- ``Paragraph fill'' version of sep
264 nest :: Int -> Doc -> Doc -- Nested
270 hang :: Doc -> Int -> Doc -> Doc
271 punctuate :: Doc -> [Doc] -> [Doc] -- punctuate p [d1, ... dn] = [d1 <> p, d2 <> p, ... dn-1 <> p, dn]
274 Displaying @Doc@ values.
277 instance Show Doc where
278 showsPrec _ doc cont = showDoc doc cont
280 render :: Doc -> String -- Uses default style
282 -> Int -- Line length
283 -> Float -- Ribbons per line
284 -> (TextDetails -> a -> a) -- What to do with text
285 -> a -- What to do at the end
289 {- When we start using 1.3
290 renderStyle :: Style -> Doc -> String
291 data Style = Style { lineLength :: Int, -- In chars
292 ribbonsPerLine :: Float, -- Ratio of ribbon length to line length
295 style :: Style -- The default style
296 style = Style { lineLength = 100, ribbonsPerLine = 2.5, mode = PageMode }
299 data Mode = PageMode -- Normal
300 | ZigZagMode -- With zig-zag cuts
301 | LeftMode -- No indentation, infinitely long lines
302 | OneLineMode -- All on one line
307 %*********************************************************
309 \subsection{The @Doc@ calculus}
311 %*********************************************************
313 The @Doc@ combinators satisfy the following laws:
317 <a1> (x $$ y) $$ z = x $$ (y $$ z)
325 <b1> (x <> y) <> z = x <> (y <> z)
326 <b2> empty <> x = empty
333 <t1> text s <> text t = text (s++t)
334 <t2> text "" <> x = x, if x non-empty
339 <n2> nest k (nest k' x) = nest (k+k') x
340 <n3> nest k (x <> y) = nest k z <> nest k y
341 <n4> nest k (x $$ y) = nest k x $$ nest k y
342 <n5> nest k empty = empty
343 <n6> x <> nest k y = x <> y, if x non-empty
345 - Note the side condition on <n6>! It is this that
346 makes it OK for empty to be a left unit for <>.
350 <m1> (text s <> x) $$ y = text s <> ((text "" <> x)) $$
353 <m2> (x $$ y) <> z = x $$ (y <> z)
357 Laws for list versions
358 ~~~~~~~~~~~~~~~~~~~~~~
359 <l1> sep (ps++[empty]++qs) = sep (ps ++ qs)
360 ...ditto hsep, hcat, vcat, fill...
362 <l2> nest k (sep ps) = sep (map (nest k) ps)
363 ...ditto hsep, hcat, vcat, fill...
367 <o1> oneLiner (nest k p) = nest k (oneLiner p)
368 <o2> oneLiner (x <> y) = oneLiner x <> oneLiner y
372 You might think that the following verion of <m1> would
375 <3 NO> (text s <> x) $$ y = text s <> ((empty <> x)) $$
378 But it doesn't work, for if x=empty, we would have
380 text s $$ y = text s <> (empty $$ nest (-length s) y)
381 = text s <> nest (-length s) y
386 %*********************************************************
388 \subsection{Simple derived definitions}
390 %*********************************************************
405 int n = text (show n)
406 integer n = text (show n)
407 float n = text (show n)
408 double n = text (show n)
409 rational n = text (show (fromRat n :: Double))
410 --rational n = text (show (fromRationalX n)) -- _showRational 30 n)
412 quotes p = char '`' <> p <> char '\''
413 doubleQuotes p = char '"' <> p <> char '"'
414 parens p = char '(' <> p <> char ')'
415 brackets p = char '[' <> p <> char ']'
416 braces p = char '{' <> p <> char '}'
418 cparen :: Bool -> Doc -> Doc
422 hcat = foldr (<>) empty
423 hsep = foldr (<+>) empty
424 vcat = foldr ($$) empty
426 hang d1 n d2 = sep [d1, nest n d2]
429 punctuate p (d:ds) = go d ds
432 go d (e:es) = (d <> p) : go e es
436 %*********************************************************
438 \subsection{The @Doc@ data type}
440 %*********************************************************
442 A @Doc@ represents a {\em set} of layouts. A @Doc@ with
443 no occurrences of @Union@ or @NoDoc@ represents just one layout.
447 | NilAbove Doc -- text "" $$ x
448 | TextBeside !TextDetails FastInt Doc -- text s <> x
449 | Nest FastInt Doc -- nest k x
450 | Union Doc Doc -- ul `union` ur
451 | NoDoc -- The empty set of documents
452 | Beside Doc Bool Doc -- True <=> space between
453 | Above Doc Bool Doc -- True <=> never overlap
455 type RDoc = Doc -- RDoc is a "reduced Doc", guaranteed not to have a top-level Above or Beside
458 reduceDoc :: Doc -> RDoc
459 reduceDoc (Beside p g q) = beside p g (reduceDoc q)
460 reduceDoc (Above p g q) = above p g (reduceDoc q)
464 data TextDetails = Chr {-#UNPACK#-}!Char
466 | PStr FastString -- a hashed string
467 | LStr {-#UNPACK#-}!LitString FastInt -- a '\0'-terminated
470 space_text :: TextDetails
472 nl_text :: TextDetails
476 Here are the invariants:
479 The argument of @NilAbove@ is never @Empty@. Therefore
480 a @NilAbove@ occupies at least two lines.
483 The arugment of @TextBeside@ is never @Nest@.
486 The layouts of the two arguments of @Union@ both flatten to the same string.
489 The arguments of @Union@ are either @TextBeside@, or @NilAbove@.
492 The right argument of a union cannot be equivalent to the empty set (@NoDoc@).
493 If the left argument of a union is equivalent to the empty set (@NoDoc@),
494 then the @NoDoc@ appears in the first line.
497 An empty document is always represented by @Empty@.
498 It can't be hidden inside a @Nest@, or a @Union@ of two @Empty@s.
501 The first line of every layout in the left argument of @Union@
502 is longer than the first line of any layout in the right argument.
503 (1) ensures that the left argument has a first line. In view of (3),
504 this invariant means that the right argument must have at least two
509 -- Arg of a NilAbove is always an RDoc
510 nilAbove_ :: Doc -> Doc
511 nilAbove_ p = LOCAL_ASSERT( _ok p ) NilAbove p
516 -- Arg of a TextBeside is always an RDoc
517 textBeside_ :: TextDetails -> FastInt -> Doc -> Doc
518 textBeside_ s sl p = TextBeside s sl (LOCAL_ASSERT( _ok p ) p)
520 _ok (Nest _ _) = False
523 -- Arg of Nest is always an RDoc
524 nest_ :: FastInt -> Doc -> Doc
525 nest_ k p = Nest k (LOCAL_ASSERT( _ok p ) p)
530 -- Args of union are always RDocs
531 union_ :: Doc -> Doc -> Doc
532 union_ p q = Union (LOCAL_ASSERT( _ok p ) p) (LOCAL_ASSERT( _ok q ) q)
534 _ok (TextBeside _ _ _) = True
535 _ok (NilAbove _) = True
536 _ok (Union _ _) = True
540 Notice the difference between
541 * NoDoc (no documents)
542 * Empty (one empty document; no height and no width)
543 * text "" (a document containing the empty string;
544 one line high, but has no width)
548 %*********************************************************
550 \subsection{@empty@, @text@, @nest@, @union@}
552 %*********************************************************
560 char c = textBeside_ (Chr c) (_ILIT(1)) Empty
561 text s = case iUnbox (length s) of {sl -> textBeside_ (Str s) sl Empty}
562 ftext :: FastString -> Doc
563 ftext s = case iUnbox (lengthFS s) of {sl -> textBeside_ (PStr s) sl Empty}
564 ptext :: LitString -> Doc
565 ptext s_= case iUnbox (lengthLS s) of {sl -> textBeside_ (LStr s sl) sl Empty}
566 where s = {-castPtr-} s_
567 zeroWidthText s = textBeside_ (Str s) (_ILIT(0)) Empty
569 #if defined(__GLASGOW_HASKELL__)
570 -- RULE that turns (text "abc") into (ptext (A# "abc"#)) to avoid the
571 -- intermediate packing/unpacking of the string.
573 "text/str" forall a. text (unpackCString# a) = ptext (Ptr a)
577 nest k p = mkNest (iUnbox k) (reduceDoc p) -- Externally callable version
579 -- mkNest checks for Nest's invariant that it doesn't have an Empty inside it
580 mkNest :: Int# -> Doc -> Doc
581 mkNest k (Nest k1 p) = mkNest (k +# k1) p
582 mkNest _ NoDoc = NoDoc
583 mkNest _ Empty = Empty
584 mkNest k p | k ==# _ILIT(0) = p -- Worth a try!
585 mkNest k p = nest_ k p
587 -- mkUnion checks for an empty document
588 mkUnion :: Doc -> Doc -> Doc
589 mkUnion Empty _ = Empty
590 mkUnion p q = p `union_` q
593 %*********************************************************
595 \subsection{Vertical composition @$$@}
597 %*********************************************************
601 p $$ q = Above p False q
602 ($+$) :: Doc -> Doc -> Doc
603 p $+$ q = Above p True q
605 above :: Doc -> Bool -> RDoc -> RDoc
606 above (Above p g1 q1) g2 q2 = above p g1 (above q1 g2 q2)
607 above p@(Beside _ _ _) g q = aboveNest (reduceDoc p) g (_ILIT(0)) (reduceDoc q)
608 above p g q = aboveNest p g (_ILIT(0)) (reduceDoc q)
610 aboveNest :: RDoc -> Bool -> FastInt -> RDoc -> RDoc
611 -- Specfication: aboveNest p g k q = p $g$ (nest k q)
613 aboveNest NoDoc _ _ _ = NoDoc
614 aboveNest (p1 `Union` p2) g k q = aboveNest p1 g k q `union_`
617 aboveNest Empty _ k q = mkNest k q
618 aboveNest (Nest k1 p) g k q = nest_ k1 (aboveNest p g (k -# k1) q)
619 -- p can't be Empty, so no need for mkNest
621 aboveNest (NilAbove p) g k q = nilAbove_ (aboveNest p g k q)
622 aboveNest (TextBeside s sl p) g k q = textBeside_ s sl rest
626 Empty -> nilAboveNest g k1 q
627 _ -> aboveNest p g k1 q
628 aboveNest _ _ _ _ = panic "aboveNest: Unhandled case"
632 nilAboveNest :: Bool -> FastInt -> RDoc -> RDoc
633 -- Specification: text s <> nilaboveNest g k q
634 -- = text s <> (text "" $g$ nest k q)
636 nilAboveNest _ _ Empty = Empty -- Here's why the "text s <>" is in the spec!
637 nilAboveNest g k (Nest k1 q) = nilAboveNest g (k +# k1) q
639 nilAboveNest g k q | (not g) && (k ># _ILIT(0)) -- No newline if no overlap
640 = textBeside_ (Str (spaces k)) k q
641 | otherwise -- Put them really above
642 = nilAbove_ (mkNest k q)
646 %*********************************************************
648 \subsection{Horizontal composition @<>@}
650 %*********************************************************
653 p <> q = Beside p False q
654 p <+> q = Beside p True q
656 beside :: Doc -> Bool -> RDoc -> RDoc
657 -- Specification: beside g p q = p <g> q
659 beside NoDoc _ _ = NoDoc
660 beside (p1 `Union` p2) g q = (beside p1 g q) `union_` (beside p2 g q)
662 beside (Nest k p) g q = nest_ k $! beside p g q -- p non-empty
663 beside p@(Beside p1 g1 q1) g2 q2
664 {- (A `op1` B) `op2` C == A `op1` (B `op2` C) iff op1 == op2
665 [ && (op1 == <> || op1 == <+>) ] -}
666 | g1 == g2 = beside p1 g1 $! beside q1 g2 q2
667 | otherwise = beside (reduceDoc p) g2 q2
668 beside p@(Above _ _ _) g q = let d = reduceDoc p in d `seq` beside d g q
669 beside (NilAbove p) g q = nilAbove_ $! beside p g q
670 beside (TextBeside s sl p) g q = textBeside_ s sl $! rest
673 Empty -> nilBeside g q
678 nilBeside :: Bool -> RDoc -> RDoc
679 -- Specification: text "" <> nilBeside g p
682 nilBeside _ Empty = Empty -- Hence the text "" in the spec
683 nilBeside g (Nest _ p) = nilBeside g p
684 nilBeside g p | g = textBeside_ space_text (_ILIT(1)) p
688 %*********************************************************
690 \subsection{Separate, @sep@, Hughes version}
692 %*********************************************************
695 -- Specification: sep ps = oneLiner (hsep ps)
699 sep = sepX True -- Separate with spaces
700 cat = sepX False -- Don't
702 sepX :: Bool -> [Doc] -> Doc
704 sepX x (p:ps) = sep1 x (reduceDoc p) (_ILIT(0)) ps
707 -- Specification: sep1 g k ys = sep (x : map (nest k) ys)
708 -- = oneLiner (x <g> nest k (hsep ys))
709 -- `union` x $$ nest k (vcat ys)
711 sep1 :: Bool -> RDoc -> FastInt -> [Doc] -> RDoc
712 sep1 _ NoDoc _ _ = NoDoc
713 sep1 g (p `Union` q) k ys = sep1 g p k ys
715 (aboveNest q False k (reduceDoc (vcat ys)))
717 sep1 g Empty k ys = mkNest k (sepX g ys)
718 sep1 g (Nest n p) k ys = nest_ n (sep1 g p (k -# n) ys)
720 sep1 _ (NilAbove p) k ys = nilAbove_ (aboveNest p False k (reduceDoc (vcat ys)))
721 sep1 g (TextBeside s sl p) k ys = textBeside_ s sl (sepNB g p (k -# sl) ys)
722 sep1 _ _ _ _ = panic "sep1: Unhandled case"
724 -- Specification: sepNB p k ys = sep1 (text "" <> p) k ys
725 -- Called when we have already found some text in the first item
726 -- We have to eat up nests
728 sepNB :: Bool -> Doc -> FastInt -> [Doc] -> Doc
729 sepNB g (Nest _ p) k ys = sepNB g p k ys
731 sepNB g Empty k ys = oneLiner (nilBeside g (reduceDoc rest))
733 nilAboveNest False k (reduceDoc (vcat ys))
736 | otherwise = hcat ys
738 sepNB g p k ys = sep1 g p k ys
741 %*********************************************************
745 %*********************************************************
754 -- fill (p1:p2:ps) = oneLiner p1 <#> nest (length p1)
755 -- (fill (oneLiner p2 : ps))
759 fill :: Bool -> [Doc] -> Doc
761 fill g (p:ps) = fill1 g (reduceDoc p) (_ILIT(0)) ps
764 fill1 :: Bool -> RDoc -> FastInt -> [Doc] -> Doc
765 fill1 _ NoDoc _ _ = NoDoc
766 fill1 g (p `Union` q) k ys = fill1 g p k ys
768 (aboveNest q False k (fill g ys))
770 fill1 g Empty k ys = mkNest k (fill g ys)
771 fill1 g (Nest n p) k ys = nest_ n (fill1 g p (k -# n) ys)
773 fill1 g (NilAbove p) k ys = nilAbove_ (aboveNest p False k (fill g ys))
774 fill1 g (TextBeside s sl p) k ys = textBeside_ s sl (fillNB g p (k -# sl) ys)
775 fill1 _ _ _ _ = panic "fill1: Unhandled case"
777 fillNB :: Bool -> Doc -> Int# -> [Doc] -> Doc
778 fillNB g (Nest _ p) k ys = fillNB g p k ys
779 fillNB _ Empty _ [] = Empty
780 fillNB g Empty k (y:ys) = nilBeside g (fill1 g (oneLiner (reduceDoc y)) k1 ys)
782 nilAboveNest False k (fill g (y:ys))
784 !k1 | g = k -# _ILIT(1)
787 fillNB g p k ys = fill1 g p k ys
791 %*********************************************************
793 \subsection{Selecting the best layout}
795 %*********************************************************
798 best :: Int -- Line length
799 -> Int -- Ribbon length
801 -> RDoc -- No unions in here!
807 get :: FastInt -- (Remaining) width of line
811 get w (NilAbove p) = nilAbove_ (get w p)
812 get w (TextBeside s sl p) = textBeside_ s sl (get1 w sl p)
813 get w (Nest k p) = nest_ k (get (w -# k) p)
814 get w (p `Union` q) = nicest w r (get w p) (get w q)
815 get _ _ = panic "best/get: Unhandled case"
817 get1 :: FastInt -- (Remaining) width of line
818 -> FastInt -- Amount of first line already eaten up
819 -> Doc -- This is an argument to TextBeside => eat Nests
820 -> Doc -- No unions in here!
822 get1 _ _ Empty = Empty
823 get1 _ _ NoDoc = NoDoc
824 get1 w sl (NilAbove p) = nilAbove_ (get (w -# sl) p)
825 get1 w sl (TextBeside t tl p) = textBeside_ t tl (get1 w (sl +# tl) p)
826 get1 w sl (Nest _ p) = get1 w sl p
827 get1 w sl (p `Union` q) = nicest1 w r sl (get1 w sl p)
829 get1 _ _ _ = panic "best/get1: Unhandled case"
831 nicest :: FastInt -> FastInt -> Doc -> Doc -> Doc
832 nicest w r p q = nicest1 w r (_ILIT(0)) p q
833 nicest1 :: FastInt -> FastInt -> Int# -> Doc -> Doc -> Doc
834 nicest1 w r sl p q | fits ((w `minFastInt` r) -# sl) p = p
837 fits :: FastInt -- Space available
839 -> Bool -- True if *first line* of Doc fits in space available
841 fits n _ | n <# _ILIT(0) = False
844 fits _ (NilAbove _) = True
845 fits n (TextBeside _ sl p) = fits (n -# sl) p
846 fits _ _ = panic "fits: Unhandled case"
849 @first@ and @nonEmptySet@ are similar to @nicest@ and @fits@, only simpler.
850 @first@ returns its first argument if it is non-empty, otherwise its second.
853 first :: Doc -> Doc -> Doc
854 first p q | nonEmptySet p = p
857 nonEmptySet :: Doc -> Bool
858 nonEmptySet NoDoc = False
859 nonEmptySet (_ `Union` _) = True
860 nonEmptySet Empty = True
861 nonEmptySet (NilAbove _) = True -- NoDoc always in first line
862 nonEmptySet (TextBeside _ _ p) = nonEmptySet p
863 nonEmptySet (Nest _ p) = nonEmptySet p
864 nonEmptySet _ = panic "nonEmptySet: Unhandled case"
867 @oneLiner@ returns the one-line members of the given set of @Doc@s.
870 oneLiner :: Doc -> Doc
871 oneLiner NoDoc = NoDoc
872 oneLiner Empty = Empty
873 oneLiner (NilAbove _) = NoDoc
874 oneLiner (TextBeside s sl p) = textBeside_ s sl (oneLiner p)
875 oneLiner (Nest k p) = nest_ k (oneLiner p)
876 oneLiner (p `Union` _) = oneLiner p
877 oneLiner _ = panic "oneLiner: Unhandled case"
882 %*********************************************************
884 \subsection{Displaying the best layout}
886 %*********************************************************
891 renderStyle Style{mode, lineLength, ribbonsPerLine} doc
892 = fullRender mode lineLength ribbonsPerLine doc ""
895 render doc = showDocWith PageMode doc
897 showDoc :: Doc -> String -> String
898 showDoc doc rest = showDocWithAppend PageMode doc rest
900 showDocWithAppend :: Mode -> Doc -> String -> String
901 showDocWithAppend mode doc rest = fullRender mode 100 1.5 string_txt rest doc
903 showDocWith :: Mode -> Doc -> String
904 showDocWith mode doc = showDocWithAppend mode doc ""
906 string_txt :: TextDetails -> String -> String
907 string_txt (Chr c) s = c:s
908 string_txt (Str s1) s2 = s1 ++ s2
909 string_txt (PStr s1) s2 = unpackFS s1 ++ s2
910 string_txt (LStr s1 _) s2 = unpackLitString s1 ++ s2
915 fullRender OneLineMode _ _ txt end doc
916 = lay (reduceDoc doc)
918 lay NoDoc = cant_fail
919 lay (Union _ q) = lay q -- Second arg can't be NoDoc
920 lay (Nest _ p) = lay p
922 lay (NilAbove p) = space_text `txt` lay p -- NoDoc always on
924 lay (TextBeside s _ p) = s `txt` lay p
925 lay _ = panic "fullRender/OneLineMode/lay: Unhandled case"
927 fullRender LeftMode _ _ txt end doc
928 = lay (reduceDoc doc)
930 lay NoDoc = cant_fail
931 lay (Union p q) = lay (first p q)
932 lay (Nest _ p) = lay p
934 lay (NilAbove p) = nl_text `txt` lay p -- NoDoc always on first line
935 lay (TextBeside s _ p) = s `txt` lay p
936 lay _ = panic "fullRender/LeftMode/lay: Unhandled case"
938 fullRender mode line_length ribbons_per_line txt end doc
939 = display mode line_length ribbon_length txt end best_doc
941 best_doc = best hacked_line_length ribbon_length (reduceDoc doc)
943 hacked_line_length, ribbon_length :: Int
944 ribbon_length = round (fromIntegral line_length / ribbons_per_line)
945 hacked_line_length = case mode of
946 ZigZagMode -> maxBound
949 display :: Mode -> Int -> Int -> (TextDetails -> t -> t) -> t -> Doc -> t
950 display mode page_width ribbon_width txt end doc
951 = case (iUnbox page_width) -# (iUnbox ribbon_width) of { gap_width ->
952 case gap_width `quotFastInt` _ILIT(2) of { shift ->
954 lay k (Nest k1 p) = lay (k +# k1) p
957 lay k (NilAbove p) = nl_text `txt` lay k p
959 lay k (TextBeside s sl p)
961 ZigZagMode | k >=# gap_width
963 Str (multi_ch shift '/') `txt` (
965 lay1 (k -# shift) s sl p)))
969 Str (multi_ch shift '\\') `txt` (
971 lay1 (k +# shift) s sl p )))
974 lay _ _ = panic "display/lay: Unhandled case"
976 lay1 k s sl p = indent k (s `txt` lay2 (k +# sl) p)
978 lay2 k (NilAbove p) = nl_text `txt` lay k p
979 lay2 k (TextBeside s sl p) = s `txt` (lay2 (k +# sl) p)
980 lay2 k (Nest _ p) = lay2 k p
982 lay2 _ _ = panic "display/lay2: Unhandled case"
984 -- optimise long indentations using LitString chunks of 8 spaces
985 indent n r | n >=# _ILIT(8) = LStr (sLit " ") (_ILIT(8)) `txt`
986 indent (n -# _ILIT(8)) r
987 | otherwise = Str (spaces n) `txt` r
993 cant_fail = error "easy_display: NoDoc"
995 multi_ch :: Int# -> Char -> String
996 multi_ch n ch | n <=# _ILIT(0) = ""
997 | otherwise = ch : multi_ch (n -# _ILIT(1)) ch
999 spaces :: Int# -> String
1000 spaces n | n <=# _ILIT(0) = ""
1001 | otherwise = ' ' : spaces (n -# _ILIT(1))
1007 pprCols = opt_PprCols
1009 printDoc :: Mode -> Handle -> Doc -> IO ()
1010 printDoc LeftMode hdl doc
1011 = do { printLeftRender hdl doc; hFlush hdl }
1012 printDoc mode hdl doc
1013 = do { fullRender mode pprCols 1.5 put done doc ;
1016 put (Chr c) next = hPutChar hdl c >> next
1017 put (Str s) next = hPutStr hdl s >> next
1018 put (PStr s) next = hPutStr hdl (unpackFS s) >> next
1019 -- NB. not hPutFS, we want this to go through
1020 -- the I/O library's encoding layer. (#3398)
1021 put (LStr s l) next = hPutLitString hdl s l >> next
1023 done = hPutChar hdl '\n'
1025 -- some versions of hPutBuf will barf if the length is zero
1026 hPutLitString :: Handle -> Ptr a -> Int# -> IO ()
1027 hPutLitString handle a l = if l ==# _ILIT(0)
1029 else hPutBuf handle a (iBox l)
1031 -- Printing output in LeftMode is performance critical: it's used when
1032 -- dumping C and assembly output, so we allow ourselves a few dirty
1035 -- (1) we specialise fullRender for LeftMode with IO output.
1037 -- (2) we add a layer of buffering on top of Handles. Handles
1038 -- don't perform well with lots of hPutChars, which is mostly
1039 -- what we're doing here, because Handles have to be thread-safe
1040 -- and async exception-safe. We only have a single thread and don't
1041 -- care about exceptions, so we add a layer of fast buffering
1042 -- over the Handle interface.
1044 -- (3) a few hacks in layLeft below to convince GHC to generate the right
1047 printLeftRender :: Handle -> Doc -> IO ()
1048 printLeftRender hdl doc = do
1049 b <- newBufHandle hdl
1053 bufLeftRender :: BufHandle -> Doc -> IO ()
1054 bufLeftRender b doc = layLeft b (reduceDoc doc)
1056 -- HACK ALERT! the "return () >>" below convinces GHC to eta-expand
1057 -- this function with the IO state lambda. Otherwise we end up with
1058 -- closures in all the case branches.
1059 layLeft :: BufHandle -> Doc -> IO ()
1060 layLeft b _ | b `seq` False = undefined -- make it strict in b
1061 layLeft _ NoDoc = cant_fail
1062 layLeft b (Union p q) = return () >> layLeft b (first p q)
1063 layLeft b (Nest _ p) = return () >> layLeft b p
1064 layLeft b Empty = bPutChar b '\n'
1065 layLeft b (NilAbove p) = bPutChar b '\n' >> layLeft b p
1066 layLeft b (TextBeside s _ p) = put b s >> layLeft b p
1068 put b _ | b `seq` False = undefined
1069 put b (Chr c) = bPutChar b c
1070 put b (Str s) = bPutStr b s
1071 put b (PStr s) = bPutFS b s
1072 put b (LStr s l) = bPutLitString b s l
1073 layLeft _ _ = panic "layLeft: Unhandled case"