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
156 -- The above warning supression flag is a temporary kludge.
157 -- While working on this module you are encouraged to remove it and fix
158 -- any warnings in the module. See
159 -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
164 Mode(..), TextDetails(..),
166 empty, isEmpty, nest,
168 text, char, ftext, ptext,
169 int, integer, float, double, rational,
170 parens, brackets, braces, quotes, doubleQuotes,
171 semi, comma, colon, space, equals,
172 lparen, rparen, lbrack, rbrack, lbrace, rbrace, cparen,
174 (<>), (<+>), hcat, hsep,
181 -- renderStyle, -- Haskell 1.3 only
182 render, fullRender, printDoc, showDocWith
185 #include "HsVersions.h"
191 import Numeric (fromRat)
193 --import Foreign.Ptr (castPtr)
195 #if defined(__GLASGOW_HASKELL__)
197 import GHC.Base ( unpackCString# )
198 import GHC.Ptr ( Ptr(..) )
201 -- Don't import Util( assertPanic ) because it makes a loop in the module structure
211 -- Disable ASSERT checks; they are expensive!
212 #define LOCAL_ASSERT(x)
217 *********************************************************
219 \subsection{The interface}
221 *********************************************************
223 The primitive @Doc@ values
227 isEmpty :: Doc -> Bool
228 text :: String -> Doc
231 semi, comma, colon, space, equals :: Doc
232 lparen, rparen, lbrack, rbrack, lbrace, rbrace :: Doc
234 parens, brackets, braces :: Doc -> Doc
235 quotes, doubleQuotes :: Doc -> Doc
238 integer :: Integer -> Doc
239 float :: Float -> Doc
240 double :: Double -> Doc
241 rational :: Rational -> Doc
244 Combining @Doc@ values
247 (<>) :: Doc -> Doc -> Doc -- Beside
248 hcat :: [Doc] -> Doc -- List version of <>
249 (<+>) :: Doc -> Doc -> Doc -- Beside, separated by space
250 hsep :: [Doc] -> Doc -- List version of <+>
252 ($$) :: Doc -> Doc -> Doc -- Above; if there is no
253 -- overlap it "dovetails" the two
254 vcat :: [Doc] -> Doc -- List version of $$
256 cat :: [Doc] -> Doc -- Either hcat or vcat
257 sep :: [Doc] -> Doc -- Either hsep or vcat
258 fcat :: [Doc] -> Doc -- ``Paragraph fill'' version of cat
259 fsep :: [Doc] -> Doc -- ``Paragraph fill'' version of sep
261 nest :: Int -> Doc -> Doc -- Nested
267 hang :: Doc -> Int -> Doc -> Doc
268 punctuate :: Doc -> [Doc] -> [Doc] -- punctuate p [d1, ... dn] = [d1 <> p, d2 <> p, ... dn-1 <> p, dn]
271 Displaying @Doc@ values.
274 instance Show Doc where
275 showsPrec prec doc cont = showDoc doc cont
277 render :: Doc -> String -- Uses default style
279 -> Int -- Line length
280 -> Float -- Ribbons per line
281 -> (TextDetails -> a -> a) -- What to do with text
282 -> a -- What to do at the end
286 {- When we start using 1.3
287 renderStyle :: Style -> Doc -> String
288 data Style = Style { lineLength :: Int, -- In chars
289 ribbonsPerLine :: Float, -- Ratio of ribbon length to line length
292 style :: Style -- The default style
293 style = Style { lineLength = 100, ribbonsPerLine = 2.5, mode = PageMode }
296 data Mode = PageMode -- Normal
297 | ZigZagMode -- With zig-zag cuts
298 | LeftMode -- No indentation, infinitely long lines
299 | OneLineMode -- All on one line
304 *********************************************************
306 \subsection{The @Doc@ calculus}
308 *********************************************************
310 The @Doc@ combinators satisfy the following laws:
314 <a1> (x $$ y) $$ z = x $$ (y $$ z)
322 <b1> (x <> y) <> z = x <> (y <> z)
323 <b2> empty <> x = empty
330 <t1> text s <> text t = text (s++t)
331 <t2> text "" <> x = x, if x non-empty
336 <n2> nest k (nest k' x) = nest (k+k') x
337 <n3> nest k (x <> y) = nest k z <> nest k y
338 <n4> nest k (x $$ y) = nest k x $$ nest k y
339 <n5> nest k empty = empty
340 <n6> x <> nest k y = x <> y, if x non-empty
342 ** Note the side condition on <n6>! It is this that
343 ** makes it OK for empty to be a left unit for <>.
347 <m1> (text s <> x) $$ y = text s <> ((text "" <> x)) $$
350 <m2> (x $$ y) <> z = x $$ (y <> z)
354 Laws for list versions
355 ~~~~~~~~~~~~~~~~~~~~~~
356 <l1> sep (ps++[empty]++qs) = sep (ps ++ qs)
357 ...ditto hsep, hcat, vcat, fill...
359 <l2> nest k (sep ps) = sep (map (nest k) ps)
360 ...ditto hsep, hcat, vcat, fill...
364 <o1> oneLiner (nest k p) = nest k (oneLiner p)
365 <o2> oneLiner (x <> y) = oneLiner x <> oneLiner y
369 You might think that the following verion of <m1> would
372 <3 NO> (text s <> x) $$ y = text s <> ((empty <> x)) $$
375 But it doesn't work, for if x=empty, we would have
377 text s $$ y = text s <> (empty $$ nest (-length s) y)
378 = text s <> nest (-length s) y
383 *********************************************************
385 \subsection{Simple derived definitions}
387 *********************************************************
402 int n = text (show n)
403 integer n = text (show n)
404 float n = text (show n)
405 double n = text (show n)
406 rational n = text (show (fromRat n :: Double))
407 --rational n = text (show (fromRationalX n)) -- _showRational 30 n)
409 quotes p = char '`' <> p <> char '\''
410 doubleQuotes p = char '"' <> p <> char '"'
411 parens p = char '(' <> p <> char ')'
412 brackets p = char '[' <> p <> char ']'
413 braces p = char '{' <> p <> char '}'
418 hcat = foldr (<>) empty
419 hsep = foldr (<+>) empty
420 vcat = foldr ($$) empty
422 hang d1 n d2 = sep [d1, nest n d2]
425 punctuate p (d:ds) = go d ds
428 go d (e:es) = (d <> p) : go e es
432 *********************************************************
434 \subsection{The @Doc@ data type}
436 *********************************************************
438 A @Doc@ represents a {\em set} of layouts. A @Doc@ with
439 no occurrences of @Union@ or @NoDoc@ represents just one layout.
443 | NilAbove Doc -- text "" $$ x
444 | TextBeside !TextDetails FastInt Doc -- text s <> x
445 | Nest FastInt Doc -- nest k x
446 | Union Doc Doc -- ul `union` ur
447 | NoDoc -- The empty set of documents
448 | Beside Doc Bool Doc -- True <=> space between
449 | Above Doc Bool Doc -- True <=> never overlap
451 type RDoc = Doc -- RDoc is a "reduced Doc", guaranteed not to have a top-level Above or Beside
454 reduceDoc :: Doc -> RDoc
455 reduceDoc (Beside p g q) = beside p g (reduceDoc q)
456 reduceDoc (Above p g q) = above p g (reduceDoc q)
460 data TextDetails = Chr {-#UNPACK#-}!Char
462 | PStr FastString -- a hashed string
463 | LStr {-#UNPACK#-}!LitString FastInt -- a '\0'-terminated array of bytes
469 Here are the invariants:
472 The argument of @NilAbove@ is never @Empty@. Therefore
473 a @NilAbove@ occupies at least two lines.
476 The arugment of @TextBeside@ is never @Nest@.
479 The layouts of the two arguments of @Union@ both flatten to the same string.
482 The arguments of @Union@ are either @TextBeside@, or @NilAbove@.
485 The right argument of a union cannot be equivalent to the empty set (@NoDoc@).
486 If the left argument of a union is equivalent to the empty set (@NoDoc@),
487 then the @NoDoc@ appears in the first line.
490 An empty document is always represented by @Empty@.
491 It can't be hidden inside a @Nest@, or a @Union@ of two @Empty@s.
494 The first line of every layout in the left argument of @Union@
495 is longer than the first line of any layout in the right argument.
496 (1) ensures that the left argument has a first line. In view of (3),
497 this invariant means that the right argument must have at least two
502 -- Arg of a NilAbove is always an RDoc
503 nilAbove_ p = LOCAL_ASSERT( ok p ) NilAbove p
508 -- Arg of a TextBeside is always an RDoc
509 textBeside_ s sl p = TextBeside s sl (LOCAL_ASSERT( ok p ) p)
511 ok (Nest _ _) = False
514 -- Arg of Nest is always an RDoc
515 nest_ k p = Nest k (LOCAL_ASSERT( ok p ) p)
520 -- Args of union are always RDocs
521 union_ p q = Union (LOCAL_ASSERT( ok p ) p) (LOCAL_ASSERT( ok q ) q)
523 ok (TextBeside _ _ _) = True
524 ok (NilAbove _) = True
525 ok (Union _ _) = True
530 Notice the difference between
531 * NoDoc (no documents)
532 * Empty (one empty document; no height and no width)
533 * text "" (a document containing the empty string;
534 one line high, but has no width)
538 *********************************************************
540 \subsection{@empty@, @text@, @nest@, @union@}
542 *********************************************************
550 char c = textBeside_ (Chr c) (_ILIT(1)) Empty
551 text s = case iUnbox (length s) of {sl -> textBeside_ (Str s) sl Empty}
552 ftext s = case iUnbox (lengthFS s) of {sl -> textBeside_ (PStr s) sl Empty}
553 ptext s_= case iUnbox (strLength s) of {sl -> textBeside_ (LStr s sl) sl Empty}
554 where s = {-castPtr-} s_
556 #if defined(__GLASGOW_HASKELL__)
557 -- RULE that turns (text "abc") into (ptext (A# "abc"#)) to avoid the
558 -- intermediate packing/unpacking of the string.
560 "text/str" forall a. text (unpackCString# a) = ptext (Ptr a)
564 nest k p = mkNest (iUnbox k) (reduceDoc p) -- Externally callable version
566 -- mkNest checks for Nest's invariant that it doesn't have an Empty inside it
567 mkNest k (Nest k1 p) = mkNest (k +# k1) p
568 mkNest k NoDoc = NoDoc
569 mkNest k Empty = Empty
570 mkNest k p | k ==# _ILIT(0) = p -- Worth a try!
571 mkNest k p = nest_ k p
573 -- mkUnion checks for an empty document
574 mkUnion Empty q = Empty
575 mkUnion p q = p `union_` q
578 *********************************************************
580 \subsection{Vertical composition @$$@}
582 *********************************************************
586 p $$ q = Above p False q
587 p $+$ q = Above p True q
589 above :: Doc -> Bool -> RDoc -> RDoc
590 above (Above p g1 q1) g2 q2 = above p g1 (above q1 g2 q2)
591 above p@(Beside _ _ _) g q = aboveNest (reduceDoc p) g (_ILIT(0)) (reduceDoc q)
592 above p g q = aboveNest p g (_ILIT(0)) (reduceDoc q)
594 aboveNest :: RDoc -> Bool -> FastInt -> RDoc -> RDoc
595 -- Specfication: aboveNest p g k q = p $g$ (nest k q)
597 aboveNest NoDoc g k q = NoDoc
598 aboveNest (p1 `Union` p2) g k q = aboveNest p1 g k q `union_`
601 aboveNest Empty g k q = mkNest k q
602 aboveNest (Nest k1 p) g k q = nest_ k1 (aboveNest p g (k -# k1) q)
603 -- p can't be Empty, so no need for mkNest
605 aboveNest (NilAbove p) g k q = nilAbove_ (aboveNest p g k q)
606 aboveNest (TextBeside s sl p) g k q = textBeside_ s sl rest
610 Empty -> nilAboveNest g k1 q
611 other -> aboveNest p g k1 q
615 nilAboveNest :: Bool -> FastInt -> RDoc -> RDoc
616 -- Specification: text s <> nilaboveNest g k q
617 -- = text s <> (text "" $g$ nest k q)
619 nilAboveNest g k Empty = Empty -- Here's why the "text s <>" is in the spec!
620 nilAboveNest g k (Nest k1 q) = nilAboveNest g (k +# k1) q
622 nilAboveNest g k q | (not g) && (k ># _ILIT(0)) -- No newline if no overlap
623 = textBeside_ (Str (spaces k)) k q
624 | otherwise -- Put them really above
625 = nilAbove_ (mkNest k q)
629 *********************************************************
631 \subsection{Horizontal composition @<>@}
633 *********************************************************
636 p <> q = Beside p False q
637 p <+> q = Beside p True q
639 beside :: Doc -> Bool -> RDoc -> RDoc
640 -- Specification: beside g p q = p <g> q
642 beside NoDoc g q = NoDoc
643 beside (p1 `Union` p2) g q = (beside p1 g q) `union_` (beside p2 g q)
645 beside (Nest k p) g q = nest_ k $! beside p g q -- p non-empty
646 beside p@(Beside p1 g1 q1) g2 q2
647 {- (A `op1` B) `op2` C == A `op1` (B `op2` C) iff op1 == op2
648 [ && (op1 == <> || op1 == <+>) ] -}
649 | g1 == g2 = beside p1 g1 $! beside q1 g2 q2
650 | otherwise = beside (reduceDoc p) g2 q2
651 beside p@(Above _ _ _) g q = let d = reduceDoc p in d `seq` beside d g q
652 beside (NilAbove p) g q = nilAbove_ $! beside p g q
653 beside (TextBeside s sl p) g q = textBeside_ s sl $! rest
656 Empty -> nilBeside g q
657 other -> beside p g q
661 nilBeside :: Bool -> RDoc -> RDoc
662 -- Specification: text "" <> nilBeside g p
665 nilBeside g Empty = Empty -- Hence the text "" in the spec
666 nilBeside g (Nest _ p) = nilBeside g p
667 nilBeside g p | g = textBeside_ space_text (_ILIT(1)) p
671 *********************************************************
673 \subsection{Separate, @sep@, Hughes version}
675 *********************************************************
678 -- Specification: sep ps = oneLiner (hsep ps)
682 sep = sepX True -- Separate with spaces
683 cat = sepX False -- Don't
686 sepX x (p:ps) = sep1 x (reduceDoc p) (_ILIT(0)) ps
689 -- Specification: sep1 g k ys = sep (x : map (nest k) ys)
690 -- = oneLiner (x <g> nest k (hsep ys))
691 -- `union` x $$ nest k (vcat ys)
693 sep1 :: Bool -> RDoc -> FastInt -> [Doc] -> RDoc
694 sep1 g NoDoc k ys = NoDoc
695 sep1 g (p `Union` q) k ys = sep1 g p k ys
697 (aboveNest q False k (reduceDoc (vcat ys)))
699 sep1 g Empty k ys = mkNest k (sepX g ys)
700 sep1 g (Nest n p) k ys = nest_ n (sep1 g p (k -# n) ys)
702 sep1 g (NilAbove p) k ys = nilAbove_ (aboveNest p False k (reduceDoc (vcat ys)))
703 sep1 g (TextBeside s sl p) k ys = textBeside_ s sl (sepNB g p (k -# sl) ys)
705 -- Specification: sepNB p k ys = sep1 (text "" <> p) k ys
706 -- Called when we have already found some text in the first item
707 -- We have to eat up nests
709 sepNB g (Nest _ p) k ys = sepNB g p k ys
711 sepNB g Empty k ys = oneLiner (nilBeside g (reduceDoc rest))
713 nilAboveNest False k (reduceDoc (vcat ys))
716 | otherwise = hcat ys
718 sepNB g p k ys = sep1 g p k ys
721 *********************************************************
725 *********************************************************
734 -- fill (p1:p2:ps) = oneLiner p1 <#> nest (length p1)
735 -- (fill (oneLiner p2 : ps))
740 fill g (p:ps) = fill1 g (reduceDoc p) (_ILIT(0)) ps
743 fill1 :: Bool -> RDoc -> FastInt -> [Doc] -> Doc
744 fill1 g NoDoc k ys = NoDoc
745 fill1 g (p `Union` q) k ys = fill1 g p k ys
747 (aboveNest q False k (fill g ys))
749 fill1 g Empty k ys = mkNest k (fill g ys)
750 fill1 g (Nest n p) k ys = nest_ n (fill1 g p (k -# n) ys)
752 fill1 g (NilAbove p) k ys = nilAbove_ (aboveNest p False k (fill g ys))
753 fill1 g (TextBeside s sl p) k ys = textBeside_ s sl (fillNB g p (k -# sl) ys)
755 fillNB g (Nest _ p) k ys = fillNB g p k ys
756 fillNB g Empty k [] = Empty
757 fillNB g Empty k (y:ys) = nilBeside g (fill1 g (oneLiner (reduceDoc y)) k1 ys)
759 nilAboveNest False k (fill g (y:ys))
761 k1 | g = k -# _ILIT(1)
764 fillNB g p k ys = fill1 g p k ys
768 *********************************************************
770 \subsection{Selecting the best layout}
772 *********************************************************
775 best :: Int -- Line length
776 -> Int -- Ribbon length
778 -> RDoc -- No unions in here!
784 get :: FastInt -- (Remaining) width of line
788 get w (NilAbove p) = nilAbove_ (get w p)
789 get w (TextBeside s sl p) = textBeside_ s sl (get1 w sl p)
790 get w (Nest k p) = nest_ k (get (w -# k) p)
791 get w (p `Union` q) = nicest w r (get w p) (get w q)
793 get1 :: FastInt -- (Remaining) width of line
794 -> FastInt -- Amount of first line already eaten up
795 -> Doc -- This is an argument to TextBeside => eat Nests
796 -> Doc -- No unions in here!
798 get1 w sl Empty = Empty
799 get1 w sl NoDoc = NoDoc
800 get1 w sl (NilAbove p) = nilAbove_ (get (w -# sl) p)
801 get1 w sl (TextBeside t tl p) = textBeside_ t tl (get1 w (sl +# tl) p)
802 get1 w sl (Nest k p) = get1 w sl p
803 get1 w sl (p `Union` q) = nicest1 w r sl (get1 w sl p)
806 nicest w r p q = nicest1 w r (_ILIT(0)) p q
807 nicest1 w r sl p q | fits ((w `minFastInt` r) -# sl) p = p
810 fits :: FastInt -- Space available
812 -> Bool -- True if *first line* of Doc fits in space available
814 fits n p | n <# _ILIT(0) = False
817 fits n (NilAbove _) = True
818 fits n (TextBeside _ sl p) = fits (n -# sl) p
821 @first@ and @nonEmptySet@ are similar to @nicest@ and @fits@, only simpler.
822 @first@ returns its first argument if it is non-empty, otherwise its second.
825 first p q | nonEmptySet p = p
828 nonEmptySet NoDoc = False
829 nonEmptySet (p `Union` q) = True
830 nonEmptySet Empty = True
831 nonEmptySet (NilAbove p) = True -- NoDoc always in first line
832 nonEmptySet (TextBeside _ _ p) = nonEmptySet p
833 nonEmptySet (Nest _ p) = nonEmptySet p
836 @oneLiner@ returns the one-line members of the given set of @Doc@s.
839 oneLiner :: Doc -> Doc
840 oneLiner NoDoc = NoDoc
841 oneLiner Empty = Empty
842 oneLiner (NilAbove p) = NoDoc
843 oneLiner (TextBeside s sl p) = textBeside_ s sl (oneLiner p)
844 oneLiner (Nest k p) = nest_ k (oneLiner p)
845 oneLiner (p `Union` q) = oneLiner p
850 *********************************************************
852 \subsection{Displaying the best layout}
854 *********************************************************
859 renderStyle Style{mode, lineLength, ribbonsPerLine} doc
860 = fullRender mode lineLength ribbonsPerLine doc ""
863 render doc = showDocWith PageMode doc
864 showDoc doc rest = showDocWithAppend PageMode doc rest
866 showDocWithAppend :: Mode -> Doc -> String -> String
867 showDocWithAppend mode doc rest = fullRender mode 100 1.5 string_txt rest doc
869 showDocWith :: Mode -> Doc -> String
870 showDocWith mode doc = showDocWithAppend mode doc ""
872 string_txt (Chr c) s = c:s
873 string_txt (Str s1) s2 = s1 ++ s2
874 string_txt (PStr s1) s2 = unpackFS s1 ++ s2
875 string_txt (LStr s1 _) s2 = unpackLitString s1 ++ s2
880 fullRender OneLineMode _ _ txt end doc
881 = lay (reduceDoc doc)
883 lay NoDoc = cant_fail
884 lay (Union p q) = (lay q) -- Second arg can't be NoDoc
885 lay (Nest k p) = lay p
887 lay (NilAbove p) = space_text `txt` lay p -- NoDoc always on first line
888 lay (TextBeside s sl p) = s `txt` lay p
890 fullRender LeftMode _ _ txt end doc
891 = lay (reduceDoc doc)
893 lay NoDoc = cant_fail
894 lay (Union p q) = lay (first p q)
895 lay (Nest k p) = lay p
897 lay (NilAbove p) = nl_text `txt` lay p -- NoDoc always on first line
898 lay (TextBeside s sl p) = s `txt` lay p
900 fullRender mode line_length ribbons_per_line txt end doc
901 = display mode line_length ribbon_length txt end best_doc
903 best_doc = best hacked_line_length ribbon_length (reduceDoc doc)
905 hacked_line_length, ribbon_length :: Int
906 ribbon_length = round (fromIntegral line_length / ribbons_per_line)
907 hacked_line_length = case mode of { ZigZagMode -> maxBound; other -> line_length }
909 display mode page_width ribbon_width txt end doc
910 = case (iUnbox page_width) -# (iUnbox ribbon_width) of { gap_width ->
911 case gap_width `quotFastInt` _ILIT(2) of { shift ->
913 lay k (Nest k1 p) = lay (k +# k1) p
916 lay k (NilAbove p) = nl_text `txt` lay k p
918 lay k (TextBeside s sl p)
920 ZigZagMode | k >=# gap_width
922 Str (multi_ch shift '/') `txt` (
924 lay1 (k -# shift) s sl p)))
928 Str (multi_ch shift '\\') `txt` (
930 lay1 (k +# shift) s sl p )))
932 other -> lay1 k s sl p
934 lay1 k s sl p = indent k (s `txt` lay2 (k +# sl) p)
936 lay2 k (NilAbove p) = nl_text `txt` lay k p
937 lay2 k (TextBeside s sl p) = s `txt` (lay2 (k +# sl) p)
938 lay2 k (Nest _ p) = lay2 k p
941 -- optimise long indentations using LitString chunks of 8 spaces
942 indent n r | n >=# _ILIT(8) = LStr SLIT(" ") (_ILIT(8)) `txt`
943 indent (n -# _ILIT(8)) r
944 | otherwise = Str (spaces n) `txt` r
949 cant_fail = error "easy_display: NoDoc"
951 multi_ch n ch | n <=# _ILIT(0) = ""
952 | otherwise = ch : multi_ch (n -# _ILIT(1)) ch
954 spaces n | n <=# _ILIT(0) = ""
955 | otherwise = ' ' : spaces (n -# _ILIT(1))
960 pprCols = (120 :: Int) -- could make configurable
962 printDoc :: Mode -> Handle -> Doc -> IO ()
963 printDoc LeftMode hdl doc
964 = do { printLeftRender hdl doc; hFlush hdl }
965 printDoc mode hdl doc
966 = do { fullRender mode pprCols 1.5 put done doc ;
969 put (Chr c) next = hPutChar hdl c >> next
970 put (Str s) next = hPutStr hdl s >> next
971 put (PStr s) next = hPutFS hdl s >> next
972 put (LStr s l) next = hPutLitString hdl s l >> next
974 done = hPutChar hdl '\n'
976 -- some versions of hPutBuf will barf if the length is zero
977 hPutLitString handle a l = if l ==# _ILIT(0)
979 else hPutBuf handle a (iBox l)
981 -- Printing output in LeftMode is performance critical: it's used when
982 -- dumping C and assembly output, so we allow ourselves a few dirty
985 -- (1) we specialise fullRender for LeftMode with IO output.
987 -- (2) we add a layer of buffering on top of Handles. Handles
988 -- don't perform well with lots of hPutChars, which is mostly
989 -- what we're doing here, because Handles have to be thread-safe
990 -- and async exception-safe. We only have a single thread and don't
991 -- care about exceptions, so we add a layer of fast buffering
992 -- over the Handle interface.
994 -- (3) a few hacks in layLeft below to convince GHC to generate the right
997 printLeftRender :: Handle -> Doc -> IO ()
998 printLeftRender hdl doc = do
999 b <- newBufHandle hdl
1000 layLeft b (reduceDoc doc)
1003 -- HACK ALERT! the "return () >>" below convinces GHC to eta-expand
1004 -- this function with the IO state lambda. Otherwise we end up with
1005 -- closures in all the case branches.
1006 layLeft b _ | b `seq` False = undefined -- make it strict in b
1007 layLeft b NoDoc = cant_fail
1008 layLeft b (Union p q) = return () >> layLeft b (first p q)
1009 layLeft b (Nest k p) = return () >> layLeft b p
1010 layLeft b Empty = bPutChar b '\n'
1011 layLeft b (NilAbove p) = bPutChar b '\n' >> layLeft b p
1012 layLeft b (TextBeside s sl p) = put b s >> layLeft b p
1014 put b _ | b `seq` False = undefined
1015 put b (Chr c) = bPutChar b c
1016 put b (Str s) = bPutStr b s
1017 put b (PStr s) = bPutFS b s
1018 put b (LStr s l) = bPutLitString b s l