Match.lhs 34.6 KB
Newer Older
1
%
Simon Marlow's avatar
Simon Marlow committed
2
% (c) The University of Glasgow 2006
3
% (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
4
%
Simon Marlow's avatar
Simon Marlow committed
5 6

The @match@ function
7 8

\begin{code}
9
{-# OPTIONS -fno-warn-incomplete-patterns #-}
10 11 12
-- The above warning supression flag is a temporary kludge.
-- While working on this module you are encouraged to remove it and fix
-- any warnings in the module. See
Ian Lynagh's avatar
Ian Lynagh committed
13
--     http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
14 15
-- for details

16
module Match ( match, matchEquations, matchWrapper, matchSimply, matchSinglePat ) where
17

18
#include "HsVersions.h"
19

20 21
import {-#SOURCE#-} DsExpr (dsLExpr)

Simon Marlow's avatar
Simon Marlow committed
22
import DynFlags
sof's avatar
sof committed
23
import HsSyn		
Simon Marlow's avatar
Simon Marlow committed
24 25
import TcHsSyn
import Check
26
import CoreSyn
Simon Marlow's avatar
Simon Marlow committed
27 28
import Literal
import CoreUtils
29
import MkCore
30
import DsMonad
Simon Marlow's avatar
Simon Marlow committed
31 32
import DsBinds
import DsGRHSs
33
import DsUtils
Simon Marlow's avatar
Simon Marlow committed
34 35 36 37 38 39 40 41 42 43 44 45
import Id
import DataCon
import MatchCon
import MatchLit
import PrelInfo
import Type
import TysWiredIn
import ListSetOps
import SrcLoc
import Maybes
import Util
import Name
46
import FiniteMap
47
import Outputable
48
import FastString
49 50
\end{code}

51 52 53 54 55 56 57 58
This function is a wrapper of @match@, it must be called from all the parts where 
it was called match, but only substitutes the firs call, ....
if the associated flags are declared, warnings will be issued.
It can not be called matchWrapper because this name already exists :-(

JJCQ 30-Nov-1997

\begin{code}
59 60 61
matchCheck ::  DsMatchContext
	    -> [Id]	        -- Vars rep'ing the exprs we're matching with
            -> Type             -- Type of the case expression
62 63 64
            -> [EquationInfo]   -- Info about patterns, etc. (type synonym below)
            -> DsM MatchResult  -- Desugared result!

65 66 67
matchCheck ctx vars ty qs = do
    dflags <- getDOptsDs
    matchCheck_really dflags ctx vars ty qs
68

69 70 71 72 73 74
matchCheck_really :: DynFlags
                  -> DsMatchContext
                  -> [Id]
                  -> Type
                  -> [EquationInfo]
                  -> DsM MatchResult
75
matchCheck_really dflags ctx vars ty qs
76 77 78
  | incomplete && shadow  = do
      dsShadowWarn ctx eqns_shadow
      dsIncompleteWarn ctx pats
79
      match vars ty qs
80 81
  | incomplete            = do
      dsIncompleteWarn ctx pats
82
      match vars ty qs
83 84
  | shadow                = do
      dsShadowWarn ctx eqns_shadow
85
      match vars ty qs
86
  | otherwise             =
87 88
      match vars ty qs
  where (pats, eqns_shadow) = check qs
89 90
        incomplete    = want_incomplete && (notNull pats)
        want_incomplete = case ctx of
91
                              DsMatchContext RecUpd _ ->
92 93 94
                                  dopt Opt_WarnIncompletePatternsRecUpd dflags
                              _ ->
                                  dopt Opt_WarnIncompletePatterns       dflags
95
        shadow        = dopt Opt_WarnOverlappingPatterns dflags
96
			&& not (null eqns_shadow)
97 98
\end{code}

99 100
This variable shows the maximum number of lines of output generated for warnings.
It will limit the number of patterns/equations displayed to@ maximum_output@.
101

sof's avatar
sof committed
102 103
(ToDo: add command-line option?)

104
\begin{code}
105
maximum_output :: Int
106 107 108
maximum_output = 4
\end{code}

109
The next two functions create the warning message.
110 111 112

\begin{code}
dsShadowWarn :: DsMatchContext -> [EquationInfo] -> DsM ()
113
dsShadowWarn ctx@(DsMatchContext kind loc) qs
114
  = putSrcSpanDs loc (warnDs warn)
115 116
  where
    warn | qs `lengthExceeds` maximum_output
Ian Lynagh's avatar
Ian Lynagh committed
117
         = pp_context ctx (ptext (sLit "are overlapped"))
118
		      (\ f -> vcat (map (ppr_eqn f kind) (take maximum_output qs)) $$
Ian Lynagh's avatar
Ian Lynagh committed
119
		      ptext (sLit "..."))
120
	 | otherwise
Ian Lynagh's avatar
Ian Lynagh committed
121
         = pp_context ctx (ptext (sLit "are overlapped"))
122
	              (\ f -> vcat $ map (ppr_eqn f kind) qs)
sof's avatar
sof committed
123

124 125

dsIncompleteWarn :: DsMatchContext -> [ExhaustivePat] -> DsM ()
126
dsIncompleteWarn ctx@(DsMatchContext kind loc) pats 
127
  = putSrcSpanDs loc (warnDs warn)
128
	where
Ian Lynagh's avatar
Ian Lynagh committed
129 130
	  warn = pp_context ctx (ptext (sLit "are non-exhaustive"))
                            (\_ -> hang (ptext (sLit "Patterns not matched:"))
131 132 133 134
		                   4 ((vcat $ map (ppr_incomplete_pats kind)
						  (take maximum_output pats))
		                      $$ dots))

Ian Lynagh's avatar
Ian Lynagh committed
135
	  dots | pats `lengthExceeds` maximum_output = ptext (sLit "...")
sof's avatar
sof committed
136
	       | otherwise                           = empty
137

138
pp_context :: DsMatchContext -> SDoc -> ((SDoc -> SDoc) -> SDoc) -> SDoc
139
pp_context (DsMatchContext kind _loc) msg rest_of_msg_fun
Ian Lynagh's avatar
Ian Lynagh committed
140 141
  = vcat [ptext (sLit "Pattern match(es)") <+> msg,
	  sep [ptext (sLit "In") <+> ppr_match <> char ':', nest 4 (rest_of_msg_fun pref)]]
142 143 144
  where
    (ppr_match, pref)
	= case kind of
145
	     FunRhs fun _ -> (pprMatchContext kind, \ pp -> ppr fun <+> pp)
146
             _            -> (pprMatchContext kind, \ pp -> pp)
147

148
ppr_pats :: Outputable a => [a] -> SDoc
149
ppr_pats pats = sep (map ppr pats)
150

151
ppr_shadow_pats :: HsMatchContext Name -> [Pat Id] -> SDoc
sof's avatar
sof committed
152
ppr_shadow_pats kind pats
Ian Lynagh's avatar
Ian Lynagh committed
153
  = sep [ppr_pats pats, matchSeparator kind, ptext (sLit "...")]
154 155 156 157

ppr_incomplete_pats :: HsMatchContext Name -> ExhaustivePat -> SDoc
ppr_incomplete_pats _ (pats,[]) = ppr_pats pats
ppr_incomplete_pats _ (pats,constraints) =
Ian Lynagh's avatar
Ian Lynagh committed
158
	                 sep [ppr_pats pats, ptext (sLit "with"), 
159
	                      sep (map ppr_constraint constraints)]
160

161
ppr_constraint :: (Name,[HsLit]) -> SDoc
Ian Lynagh's avatar
Ian Lynagh committed
162
ppr_constraint (var,pats) = sep [ppr var, ptext (sLit "`notElem`"), ppr pats]
163

164
ppr_eqn :: (SDoc -> SDoc) -> HsMatchContext Name -> EquationInfo -> SDoc
165
ppr_eqn prefixF kind eqn = prefixF (ppr_shadow_pats kind (eqn_pats eqn))
166 167 168
\end{code}


169 170 171 172 173 174
%************************************************************************
%*									*
		The main matching function
%*									*
%************************************************************************

175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
The function @match@ is basically the same as in the Wadler chapter,
except it is monadised, to carry around the name supply, info about
annotations, etc.

Notes on @match@'s arguments, assuming $m$ equations and $n$ patterns:
\begin{enumerate}
\item
A list of $n$ variable names, those variables presumably bound to the
$n$ expressions being matched against the $n$ patterns.  Using the
list of $n$ expressions as the first argument showed no benefit and
some inelegance.

\item
The second argument, a list giving the ``equation info'' for each of
the $m$ equations:
\begin{itemize}
\item
the $n$ patterns for that equation, and
\item
194
a list of Core bindings [@(Id, CoreExpr)@ pairs] to be ``stuck on
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
the front'' of the matching code, as in:
\begin{verbatim}
let <binds>
in  <matching-code>
\end{verbatim}
\item
and finally: (ToDo: fill in)

The right way to think about the ``after-match function'' is that it
is an embryonic @CoreExpr@ with a ``hole'' at the end for the
final ``else expression''.
\end{itemize}

There is a type synonym, @EquationInfo@, defined in module @DsUtils@.

An experiment with re-ordering this information about equations (in
particular, having the patterns available in column-major order)
showed no benefit.

\item
A default expression---what to evaluate if the overall pattern-match
fails.  This expression will (almost?) always be
217
a measly expression @Var@, unless we know it will only be used once
218 219 220
(as we do in @glue_success_exprs@).

Leaving out this third argument to @match@ (and slamming in lots of
221
@Var "fail"@s) is a positively {\em bad} idea, because it makes it
222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
impossible to share the default expressions.  (Also, it stands no
chance of working in our post-upheaval world of @Locals@.)
\end{enumerate}

Note: @match@ is often called via @matchWrapper@ (end of this module),
a function that does much of the house-keeping that goes with a call
to @match@.

It is also worth mentioning the {\em typical} way a block of equations
is desugared with @match@.  At each stage, it is the first column of
patterns that is examined.  The steps carried out are roughly:
\begin{enumerate}
\item
Tidy the patterns in column~1 with @tidyEqnInfo@ (this may add
bindings to the second component of the equation-info):
\begin{itemize}
\item
Remove the `as' patterns from column~1.
\item
Make all constructor patterns in column~1 into @ConPats@, notably
@ListPats@ and @TuplePats@.
\item
Handle any irrefutable (or ``twiddle'') @LazyPats@.
\end{itemize}
\item
Ian Lynagh's avatar
Ian Lynagh committed
247
Now {\em unmix} the equations into {\em blocks} [w\/ local function
248 249 250 251
@unmix_eqns@], in which the equations in a block all have variable
patterns in column~1, or they all have constructor patterns in ...
(see ``the mixture rule'' in SLPJ).
\item
252
Call @matchEqnBlock@ on each block of equations; it will do the
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
appropriate thing for each kind of column-1 pattern, usually ending up
in a recursive call to @match@.
\end{enumerate}

We are a little more paranoid about the ``empty rule'' (SLPJ, p.~87)
than the Wadler-chapter code for @match@ (p.~93, first @match@ clause).
And gluing the ``success expressions'' together isn't quite so pretty.

This (more interesting) clause of @match@ uses @tidy_and_unmix_eqns@
(a)~to get `as'- and `twiddle'-patterns out of the way (tidying), and
(b)~to do ``the mixture rule'' (SLPJ, p.~88) [which really {\em
un}mixes the equations], producing a list of equation-info
blocks, each block having as its first column of patterns either all
constructors, or all variables (or similar beasts), etc.

@match_unmixed_eqn_blks@ simply takes the place of the @foldr@ in the
Wadler-chapter @match@ (p.~93, last clause), and @match_unmixed_blk@
corresponds roughly to @matchVarCon@.

\begin{code}
Ian Lynagh's avatar
Ian Lynagh committed
273
match :: [Id]		  -- Variables rep\'ing the exprs we\'re matching with
274 275 276 277 278
      -> Type             -- Type of the case expression
      -> [EquationInfo]	  -- Info about patterns, etc. (type synonym below)
      -> DsM MatchResult  -- Desugared result!

match [] ty eqns
279
  = ASSERT2( not (null eqns), ppr ty )
280
    return (foldr1 combineMatchResults match_results)
281
  where
282 283 284 285 286 287 288 289 290 291 292
    match_results = [ ASSERT( null (eqn_pats eqn) ) 
		      eqn_rhs eqn
		    | eqn <- eqns ]

match vars@(v:_) ty eqns
  = ASSERT( not (null eqns ) )
    do	{ 	-- Tidy the first pattern, generating
		-- auxiliary bindings if necessary
	  (aux_binds, tidy_eqns) <- mapAndUnzipM (tidyEqnInfo v) eqns

		-- Group the equations and match each group in turn
293
       ; let grouped = groupEquations tidy_eqns
294 295

         -- print the view patterns that are commoned up to help debug
296
       ; ifOptM Opt_D_dump_view_pattern_commoning (debug grouped)
297 298

	; match_results <- mapM match_group grouped
299 300 301 302 303 304 305 306
	; return (adjustMatchResult (foldr1 (.) aux_binds) $
		  foldr1 combineMatchResults match_results) }
  where
    dropGroup :: [(PatGroup,EquationInfo)] -> [EquationInfo]
    dropGroup = map snd

    match_group :: [(PatGroup,EquationInfo)] -> DsM MatchResult
    match_group eqns@((group,_) : _)
307
        = case group of
308 309 310
            PgCon _    -> matchConFamily  vars ty (subGroup [(c,e) | (PgCon c, e) <- eqns])
            PgLit _    -> matchLiterals   vars ty (subGroup [(l,e) | (PgLit l, e) <- eqns])

311
            PgAny      -> matchVariables  vars ty (dropGroup eqns)
312
            PgN _      -> matchNPats      vars ty (dropGroup eqns)
313
            PgNpK _    -> matchNPlusKPats vars ty (dropGroup eqns)
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
            PgBang     -> matchBangs      vars ty (dropGroup eqns)
            PgCo _     -> matchCoercion   vars ty (dropGroup eqns)
            PgView _ _ -> matchView       vars ty (dropGroup eqns)

    -- FIXME: we should also warn about view patterns that should be
    -- commoned up but are not

    -- print some stuff to see what's getting grouped
    -- use -dppr-debug to see the resolution of overloaded lits
    debug eqns = 
        let gs = map (\group -> foldr (\ (p,_) -> \acc -> 
                                           case p of PgView e _ -> e:acc 
                                                     _ -> acc) [] group) eqns
            maybeWarn [] = return ()
            maybeWarn l = warnDs (vcat l)
        in 
          maybeWarn $ (map (\g -> text "Putting these view expressions into the same case:" <+> (ppr g))
                       (filter (not . null) gs))
332 333 334 335

matchVariables :: [Id] -> Type -> [EquationInfo] -> DsM MatchResult
-- Real true variables, just like in matchVar, SLPJ p 94
-- No binding to do: they'll all be wildcards by now (done in tidy)
336
matchVariables (_:vars) ty eqns = match vars ty (shiftEqns eqns)
337

338 339
matchBangs :: [Id] -> Type -> [EquationInfo] -> DsM MatchResult
matchBangs (var:vars) ty eqns
340
  = do	{ match_result <- match (var:vars) ty (map decomposeFirst_Bang eqns)
341 342 343 344
	; return (mkEvalMatchResult var ty match_result) }

matchCoercion :: [Id] -> Type -> [EquationInfo] -> DsM MatchResult
-- Apply the coercion to the match variable and then match that
345
matchCoercion (var:vars) ty (eqns@(eqn1:_))
346 347
  = do	{ let CoPat co pat _ = firstPat eqn1
	; var' <- newUniqueId (idName var) (hsPatType pat)
348
	; match_result <- match (var':vars) ty (map decomposeFirst_Coercion eqns)
349 350
	; rhs <- dsCoercion co (return (Var var))
	; return (mkCoLetMatchResult (NonRec var' rhs) match_result) }
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366

matchView :: [Id] -> Type -> [EquationInfo] -> DsM MatchResult
-- Apply the view function to the match variable and then match that
matchView (var:vars) ty (eqns@(eqn1:_))
  = do	{ -- we could pass in the expr from the PgView,
         -- but this needs to extract the pat anyway 
         -- to figure out the type of the fresh variable
         let ViewPat viewExpr (L _ pat) _ = firstPat eqn1
         -- do the rest of the compilation 
	; var' <- newUniqueId (idName var) (hsPatType pat)
	; match_result <- match (var':vars) ty (map decomposeFirst_View eqns)
         -- compile the view expressions
       ; viewExpr' <- dsLExpr viewExpr
	; return (mkViewMatchResult var' viewExpr' var match_result) }

-- decompose the first pattern and leave the rest alone
367
decomposeFirstPat :: (Pat Id -> Pat Id) -> EquationInfo -> EquationInfo
368 369 370
decomposeFirstPat extractpat (eqn@(EqnInfo { eqn_pats = pat : pats }))
	= eqn { eqn_pats = extractpat pat : pats}

371 372
decomposeFirst_Coercion, decomposeFirst_Bang, decomposeFirst_View :: EquationInfo -> EquationInfo

373 374 375 376
decomposeFirst_Coercion = decomposeFirstPat (\ (CoPat _ pat _) -> pat)
decomposeFirst_Bang     = decomposeFirstPat (\ (BangPat pat  ) -> unLoc pat)
decomposeFirst_View     = decomposeFirstPat (\ (ViewPat _ pat _) -> unLoc pat)

377 378
\end{code}

379 380 381 382 383 384
%************************************************************************
%*									*
		Tidying patterns
%*									*
%************************************************************************

385 386 387 388 389 390 391
Tidy up the leftmost pattern in an @EquationInfo@, given the variable @v@
which will be scrutinised.  This means:
\begin{itemize}
\item
Replace variable patterns @x@ (@x /= v@) with the pattern @_@,
together with the binding @x = v@.
\item
392
Replace the `as' pattern @x@@p@ with the pattern p and a binding @x = v@.
393 394 395
\item
Removing lazy (irrefutable) patterns (you don't want to know...).
\item
chak's avatar
chak committed
396 397
Converting explicit tuple-, list-, and parallel-array-pats into ordinary
@ConPats@. 
398 399
\item
Convert the literal pat "" to [].
400 401 402 403 404 405 406 407 408 409 410
\end{itemize}

The result of this tidying is that the column of patterns will include
{\em only}:
\begin{description}
\item[@WildPats@:]
The @VarPat@ information isn't needed any more after this.

\item[@ConPats@:]
@ListPats@, @TuplePats@, etc., are all converted into @ConPats@.

411 412
\item[@LitPats@ and @NPats@:]
@LitPats@/@NPats@ of ``known friendly types'' (Int, Char,
413
Float, 	Double, at least) are converted to unboxed form; e.g.,
414
\tr{(NPat (HsInt i) _ _)} is converted to:
415
\begin{verbatim}
416
(ConPat I# _ _ [LitPat (HsIntPrim i)])
417 418 419 420
\end{verbatim}
\end{description}

\begin{code}
421 422
tidyEqnInfo :: Id -> EquationInfo
	    -> DsM (DsWrapper, EquationInfo)
423
	-- DsM'd because of internal call to dsLHsBinds
424
	-- 	and mkSelectorBinds.
425 426
	-- "tidy1" does the interesting stuff, looking at
	-- one pattern and fiddling the list of bindings.
427 428 429 430 431 432 433 434 435
	--
	-- POST CONDITION: head pattern in the EqnInfo is
	--	WildPat
	--	ConPat
	--	NPat
	--	LitPat
	--	NPlusKPat
	-- but no other

436 437 438
tidyEqnInfo v eqn@(EqnInfo { eqn_pats = pat : pats }) = do
    (wrap, pat') <- tidy1 v pat
    return (wrap, eqn { eqn_pats = do pat' : pats })
439 440

tidy1 :: Id 			-- The Id being scrutinised
441
      -> Pat Id 		-- The pattern against which it is to be matched
442
      -> DsM (DsWrapper,	-- Extra bindings to do before the match
443
	      Pat Id) 		-- Equivalent pattern
444

445 446 447 448 449 450
-------------------------------------------------------
--	(pat', mr') = tidy1 v pat mr
-- tidies the *outer level only* of pat, giving pat'
-- It eliminates many pattern forms (as-patterns, variable patterns,
-- list patterns, etc) yielding one of:
--	WildPat
451
--	ConPatOut
452 453 454 455
--	LitPat
--	NPat
--	NPlusKPat

456 457
tidy1 v (ParPat pat)      = tidy1 v (unLoc pat) 
tidy1 v (SigPatOut pat _) = tidy1 v (unLoc pat) 
458
tidy1 _ (WildPat ty)      = return (idDsWrapper, WildPat ty)
459

460 461
	-- case v of { x -> mr[] }
	-- = case v of { _ -> let x=v in mr[] }
462
tidy1 v (VarPat var)
463
  = return (wrapBind var v, WildPat (idType var)) 
464

465
tidy1 v (VarPatOut var binds)
466
  = do	{ prs <- dsLHsBinds binds
467
	; return (wrapBind var v . mkCoreLet (Rec prs),
468
		  WildPat (idType var)) }
469

470 471
	-- case v of { x@p -> mr[] }
	-- = case v of { p -> let x=v in mr[] }
472 473 474
tidy1 v (AsPat (L _ var) pat)
  = do	{ (wrap, pat') <- tidy1 v (unLoc pat)
	; return (wrapBind var v . wrap, pat') }
475 476 477 478 479 480 481 482 483

{- now, here we handle lazy patterns:
    tidy1 v ~p bs = (v, v1 = case v of p -> v1 :
			v2 = case v of p -> v2 : ... : bs )

    where the v_i's are the binders in the pattern.

    ToDo: in "v_i = ... -> v_i", are the v_i's really the same thing?

484
    The case expr for v_i is just: match [v] [(p, [], \ x -> Var v_i)] any_expr
485 486
-}

487 488
tidy1 v (LazyPat pat)
  = do	{ sel_prs <- mkSelectorBinds pat (Var v)
489
	; let sel_binds =  [NonRec b rhs | (b,rhs) <- sel_prs]
490
	; return (mkCoreLets sel_binds, WildPat (idType v)) }
491

492
tidy1 _ (ListPat pats ty)
493
  = return (idDsWrapper, unLoc list_ConPat)
494
  where
495 496 497 498
    list_ty     = mkListTy ty
    list_ConPat = foldr (\ x y -> mkPrefixConPat consDataCon [x, y] list_ty)
	      	  	(mkNilPat list_ty)
	      	  	pats
499

500
-- Introduce fake parallel array constructors to be able to handle parallel
chak's avatar
chak committed
501
-- arrays with the existing machinery for constructor pattern
502
tidy1 _ (PArrPat pats ty)
503
  = return (idDsWrapper, unLoc parrConPat)
chak's avatar
chak committed
504 505
  where
    arity      = length pats
506
    parrConPat = mkPrefixConPat (parrFakeCon arity) pats (mkPArrTy ty)
chak's avatar
chak committed
507

508
tidy1 _ (TuplePat pats boxity ty)
509
  = return (idDsWrapper, unLoc tuple_ConPat)
510 511
  where
    arity = length pats
512
    tuple_ConPat = mkPrefixConPat (tupleCon boxity arity) pats ty
513

514
-- LitPats: we *might* be able to replace these w/ a simpler form
515
tidy1 _ (LitPat lit)
516
  = return (idDsWrapper, tidyLitPat lit)
517 518

-- NPats: we *might* be able to replace these w/ a simpler form
519
tidy1 _ (NPat lit mb_neg eq)
520
  = return (idDsWrapper, tidyNPat lit mb_neg eq)
521

522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
-- BangPatterns: Pattern matching is already strict in constructors,
-- tuples etc, so the last case strips off the bang for thoses patterns.
tidy1 v (BangPat (L _ (LazyPat p)))       = tidy1 v (BangPat p)
tidy1 v (BangPat (L _ (ParPat p)))        = tidy1 v (BangPat p)
tidy1 _ p@(BangPat (L _(VarPat _)))       = return (idDsWrapper, p)
tidy1 _ p@(BangPat (L _(VarPatOut _ _)))  = return (idDsWrapper, p)
tidy1 _ p@(BangPat (L _ (WildPat _)))     = return (idDsWrapper, p)
tidy1 _ p@(BangPat (L _ (CoPat _ _ _)))   = return (idDsWrapper, p)
tidy1 _ p@(BangPat (L _ (SigPatIn _ _)))  = return (idDsWrapper, p)
tidy1 _ p@(BangPat (L _ (SigPatOut _ _))) = return (idDsWrapper, p)
tidy1 v (BangPat (L _ (AsPat (L _ var) pat)))
  = do	{ (wrap, pat') <- tidy1 v (BangPat pat)
        ; return (wrapBind var v . wrap, pat') }
tidy1 v (BangPat (L _ p))                   = tidy1 v p

537
-- Everything else goes through unchanged...
538

539
tidy1 _ non_interesting_pat
540
  = return (idDsWrapper, non_interesting_pat)
541 542
\end{code}

543 544
\noindent
{\bf Previous @matchTwiddled@ stuff:}
545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632

Now we get to the only interesting part; note: there are choices for
translation [from Simon's notes]; translation~1:
\begin{verbatim}
deTwiddle [s,t] e
\end{verbatim}
returns
\begin{verbatim}
[ w = e,
  s = case w of [s,t] -> s
  t = case w of [s,t] -> t
]
\end{verbatim}

Here \tr{w} is a fresh variable, and the \tr{w}-binding prevents multiple
evaluation of \tr{e}.  An alternative translation (No.~2):
\begin{verbatim}
[ w = case e of [s,t] -> (s,t)
  s = case w of (s,t) -> s
  t = case w of (s,t) -> t
]
\end{verbatim}

%************************************************************************
%*									*
\subsubsection[improved-unmixing]{UNIMPLEMENTED idea for improved unmixing}
%*									*
%************************************************************************

We might be able to optimise unmixing when confronted by
only-one-constructor-possible, of which tuples are the most notable
examples.  Consider:
\begin{verbatim}
f (a,b,c) ... = ...
f d ... (e:f) = ...
f (g,h,i) ... = ...
f j ...       = ...
\end{verbatim}
This definition would normally be unmixed into four equation blocks,
one per equation.  But it could be unmixed into just one equation
block, because if the one equation matches (on the first column),
the others certainly will.

You have to be careful, though; the example
\begin{verbatim}
f j ...       = ...
-------------------
f (a,b,c) ... = ...
f d ... (e:f) = ...
f (g,h,i) ... = ...
\end{verbatim}
{\em must} be broken into two blocks at the line shown; otherwise, you
are forcing unnecessary evaluation.  In any case, the top-left pattern
always gives the cue.  You could then unmix blocks into groups of...
\begin{description}
\item[all variables:]
As it is now.
\item[constructors or variables (mixed):]
Need to make sure the right names get bound for the variable patterns.
\item[literals or variables (mixed):]
Presumably just a variant on the constructor case (as it is now).
\end{description}

%************************************************************************
%*									*
%*  matchWrapper: a convenient way to call @match@			*
%*									*
%************************************************************************
\subsection[matchWrapper]{@matchWrapper@: a convenient interface to @match@}

Calls to @match@ often involve similar (non-trivial) work; that work
is collected here, in @matchWrapper@.  This function takes as
arguments:
\begin{itemize}
\item
Typchecked @Matches@ (of a function definition, or a case or lambda
expression)---the main input;
\item
An error message to be inserted into any (runtime) pattern-matching
failure messages.
\end{itemize}

As results, @matchWrapper@ produces:
\begin{itemize}
\item
A list of variables (@Locals@) that the caller must ``promise'' to
bind to appropriate values; and
\item
633
a @CoreExpr@, the desugared output (main result).
634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
\end{itemize}

The main actions of @matchWrapper@ include:
\begin{enumerate}
\item
Flatten the @[TypecheckedMatch]@ into a suitable list of
@EquationInfo@s.
\item
Create as many new variables as there are patterns in a pattern-list
(in any one of the @EquationInfo@s).
\item
Create a suitable ``if it fails'' expression---a call to @error@ using
the error-string input; the {\em type} of this fail value can be found
by examining one of the RHS expressions in one of the @EquationInfo@s.
\item
Call @match@ with all of this information!
\end{enumerate}

\begin{code}
653
matchWrapper :: HsMatchContext Name	-- For shadowing warning messages
654
	     -> MatchGroup Id		-- Matches being desugared
655
	     -> DsM ([Id], CoreExpr) 	-- Results
656
\end{code}
657

658 659
 There is one small problem with the Lambda Patterns, when somebody
 writes something similar to:
660
\begin{verbatim}
661
    (\ (x:xs) -> ...)
662
\end{verbatim}
663
 he/she don't want a warning about incomplete patterns, that is done with 
664 665 666 667 668 669 670 671 672 673 674 675
 the flag @opt_WarnSimplePatterns@.
 This problem also appears in the:
\begin{itemize}
\item @do@ patterns, but if the @do@ can fail
      it creates another equation if the match can fail
      (see @DsExpr.doDo@ function)
\item @let@ patterns, are treated by @matchSimply@
   List Comprension Patterns, are treated by @matchSimply@ also
\end{itemize}

We can't call @matchSimply@ with Lambda patterns,
due to the fact that lambda patterns can have more than
676 677 678
one pattern, and match simply only accepts one pattern.

JJQC 30-Nov-1997
679

680
\begin{code}
681
matchWrapper ctxt (MatchGroup matches match_ty)
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
682 683
  = ASSERT( notNull matches )
    do	{ eqns_info   <- mapM mk_eqn_info matches
684
	; new_vars    <- selectMatchVars arg_pats
685
	; result_expr <- matchEquations ctxt new_vars eqns_info rhs_ty
686
	; return (new_vars, result_expr) }
687
  where
688 689 690
    arg_pats    = map unLoc (hsLMatchPats (head matches))
    n_pats	= length arg_pats
    (_, rhs_ty) = splitFunTysN n_pats match_ty
691 692 693 694

    mk_eqn_info (L _ (Match pats _ grhss))
      = do { let upats = map unLoc pats
	   ; match_result <- dsGRHSs ctxt upats grhss rhs_ty
695
	   ; return (EqnInfo { eqn_pats = upats, eqn_rhs  = match_result}) }
696

697 698 699 700 701 702 703 704

matchEquations  :: HsMatchContext Name
		-> [Id]	-> [EquationInfo] -> Type
		-> DsM CoreExpr
matchEquations ctxt vars eqns_info rhs_ty
  = do	{ dflags <- getDOptsDs
	; locn   <- getSrcSpanDs
	; let   ds_ctxt      = DsMatchContext ctxt locn
705
		error_doc = matchContextErrString ctxt
706 707 708

	; match_result <- match_fun dflags ds_ctxt vars rhs_ty eqns_info

709
	; fail_expr <- mkErrorAppDs pAT_ERROR_ID rhs_ty error_doc
710 711
	; extractMatchResult match_result fail_expr }
  where 
712 713 714 715 716
    match_fun dflags ds_ctxt
       = case ctxt of 
           LambdaExpr | dopt Opt_WarnSimplePatterns dflags -> matchCheck ds_ctxt
                      | otherwise                          -> match
           _                                               -> matchCheck ds_ctxt
717 718 719 720 721 722 723 724 725 726 727 728 729
\end{code}

%************************************************************************
%*									*
\subsection[matchSimply]{@matchSimply@: match a single expression against a single pattern}
%*									*
%************************************************************************

@mkSimpleMatch@ is a wrapper for @match@ which deals with the
situation where we want to match a single expression against a single
pattern. It returns an expression.

\begin{code}
730
matchSimply :: CoreExpr			-- Scrutinee
731 732
	    -> HsMatchContext Name	-- Match kind
	    -> LPat Id			-- Pattern it should match
733 734
	    -> CoreExpr			-- Return this if it matches
	    -> CoreExpr			-- Return this if it doesn't
735
	    -> DsM CoreExpr
736

737 738
matchSimply scrut hs_ctx pat result_expr fail_expr = do
    let
739
      match_result = cantFailMatchResult result_expr
740 741 742
      rhs_ty       = exprType fail_expr
        -- Use exprType of fail_expr, because won't refine in the case of failure!
    match_result' <- matchSinglePat scrut hs_ctx pat rhs_ty match_result
743
    extractMatchResult match_result' fail_expr
744

745

746
matchSinglePat :: CoreExpr -> HsMatchContext Name -> LPat Id
747
	       -> Type -> MatchResult -> DsM MatchResult
748 749 750
matchSinglePat (Var var) hs_ctx (L _ pat) ty match_result = do
    dflags <- getDOptsDs
    locn <- getSrcSpanDs
751
    let
752
        match_fn dflags
753
           | dopt Opt_WarnSimplePatterns dflags = matchCheck ds_ctx
754 755 756
           | otherwise                          = match
           where
             ds_ctx = DsMatchContext hs_ctx locn
757
    match_fn dflags [var] ty [EqnInfo { eqn_pats = [pat], eqn_rhs  = match_result }]
758

759 760 761 762
matchSinglePat scrut hs_ctx pat ty match_result = do
    var <- selectSimpleMatchVarL pat
    match_result' <- matchSinglePat (Var var) hs_ctx pat ty match_result
    return (adjustMatchResult (bindNonRec var scrut) match_result')
763 764
\end{code}

765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782

%************************************************************************
%*									*
		Pattern classification
%*									*
%************************************************************************

\begin{code}
data PatGroup
  = PgAny		-- Immediate match: variables, wildcards, 
			--		    lazy patterns
  | PgCon DataCon	-- Constructor patterns (incl list, tuple)
  | PgLit Literal	-- Literal patterns
  | PgN   Literal	-- Overloaded literals
  | PgNpK Literal	-- n+k patterns
  | PgBang		-- Bang patterns
  | PgCo Type		-- Coercion patterns; the type is the type
			--	of the pattern *inside*
783 784 785
  | PgView (LHsExpr Id) -- view pattern (e -> p):
                        -- the LHsExpr is the expression e
           Type         -- the Type is the type of p (equivalently, the result type of e)
786 787

groupEquations :: [EquationInfo] -> [[(PatGroup, EquationInfo)]]
788 789 790
-- If the result is of form [g1, g2, g3], 
-- (a) all the (pg,eq) pairs in g1 have the same pg
-- (b) none of the gi are empty
791
-- The ordering of equations is unchanged
792 793 794 795 796 797
groupEquations eqns
  = runs same_gp [(patGroup (firstPat eqn), eqn) | eqn <- eqns]
  where
    same_gp :: (PatGroup,EquationInfo) -> (PatGroup,EquationInfo) -> Bool
    (pg1,_) `same_gp` (pg2,_) = pg1 `sameGroup` pg2

798
subGroup :: Ord a => [(a, EquationInfo)] -> [[EquationInfo]]
799 800
-- Input is a particular group.  The result sub-groups the 
-- equations by with particular constructor, literal etc they match.
801 802 803 804
-- Each sub-list in the result has the same PatGroup
-- See Note [Take care with pattern order]
subGroup group 
    = map reverse $ eltsFM $ foldl accumulate emptyFM group
805
  where
806 807 808 809 810 811 812 813
    accumulate pg_map (pg, eqn)
      = case lookupFM pg_map pg of
          Just eqns -> addToFM pg_map pg (eqn:eqns)
          Nothing   -> addToFM pg_map pg [eqn]

    -- pg_map :: FiniteMap a [EquationInfo]
    -- Equations seen so far in reverse order of appearance
\end{code}
814

815 816 817 818 819 820 821 822 823
Note [Take care with pattern order]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In the subGroup function we must be very careful about pattern re-ordering,
Consider the patterns [ (True, Nothing), (False, x), (True, y) ]
Then in bringing together the patterns for True, we must not 
swap the Nothing and y!


\begin{code}
824 825 826 827 828 829 830 831
sameGroup :: PatGroup -> PatGroup -> Bool
-- Same group means that a single case expression 
-- or test will suffice to match both, *and* the order
-- of testing within the group is insignificant.
sameGroup PgAny      PgAny      = True
sameGroup PgBang     PgBang     = True
sameGroup (PgCon _)  (PgCon _)  = True		-- One case expression
sameGroup (PgLit _)  (PgLit _)  = True		-- One case expression
832 833
sameGroup (PgN l1)   (PgN l2)   = l1==l2	-- Order is significant
sameGroup (PgNpK l1) (PgNpK l2) = l1==l2	-- See Note [Grouping overloaded literal patterns]
834
sameGroup (PgCo	t1)  (PgCo t2)  = t1 `coreEqType` t2
835 836 837 838
	-- CoPats are in the same goup only if the type of the
	-- enclosed pattern is the same. The patterns outside the CoPat
	-- always have the same type, so this boils down to saying that
	-- the two coercions are identical.
839 840 841
sameGroup (PgView e1 t1) (PgView e2 t2) = viewLExprEq (e1,t1) (e2,t2) 
       -- ViewPats are in the same gorup iff the expressions
       -- are "equal"---conservatively, we use syntactic equality
842
sameGroup _          _          = False
843

844
-- An approximation of syntactic equality used for determining when view
845
-- exprs are in the same group.
846
-- This function can always safely return false;
847 848
-- but doing so will result in the application of the view function being repeated.
--
849
-- Currently: compare applications of literals and variables
850 851 852 853 854 855 856
--            and anything else that we can do without involving other
--            HsSyn types in the recursion
--
-- NB we can't assume that the two view expressions have the same type.  Consider
--   f (e1 -> True) = ...
--   f (e2 -> "hi") = ...
viewLExprEq :: (LHsExpr Id,Type) -> (LHsExpr Id,Type) -> Bool
857
viewLExprEq (e1,_) (e2,_) =
858 859 860 861
    let 
        -- short name for recursive call on unLoc
        lexp e e' = exp (unLoc e) (unLoc e')

862 863 864 865 866
	eq_list :: (a->a->Bool) -> [a] -> [a] -> Bool
        eq_list _  []     []     = True
        eq_list _  []     (_:_)  = False
        eq_list _  (_:_)  []     = False
        eq_list eq (x:xs) (y:ys) = eq x y && eq_list eq xs ys
867 868 869 870 871 872 873 874 875

        -- conservative, in that it demands that wrappers be
        -- syntactically identical and doesn't look under binders
        --
        -- coarser notions of equality are possible
        -- (e.g., reassociating compositions,
        --        equating different ways of writing a coercion)
        wrap WpHole WpHole = True
        wrap (WpCompose w1 w2) (WpCompose w1' w2') = wrap w1 w1' && wrap w2 w2'
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
876 877
        wrap (WpCast c)  (WpCast c')  = tcEqType c c'
        wrap (WpApp d)   (WpApp d')   = d == d'
878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894
        wrap (WpTyApp t) (WpTyApp t') = tcEqType t t'
        -- Enhancement: could implement equality for more wrappers
        --   if it seems useful (lams and lets)
        wrap _ _ = False

        -- real comparison is on HsExpr's
        -- strip parens 
        exp (HsPar (L _ e)) e'   = exp e e'
        exp e (HsPar (L _ e'))   = exp e e'
        -- because the expressions do not necessarily have the same type,
        -- we have to compare the wrappers
        exp (HsWrap h e) (HsWrap h' e') = wrap h h' && exp e e'
        exp (HsVar i) (HsVar i') =  i == i' 
        -- the instance for IPName derives using the id, so this works if the
        -- above does
        exp (HsIPVar i) (HsIPVar i') = i == i' 
        exp (HsOverLit l) (HsOverLit l') = 
895
            -- Overloaded lits are equal if they have the same type
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911
            -- and the data is the same.
            -- this is coarser than comparing the SyntaxExpr's in l and l',
            -- which resolve the overloading (e.g., fromInteger 1),
            -- because these expressions get written as a bunch of different variables
            -- (presumably to improve sharing)
            tcEqType (overLitType l) (overLitType l') && l == l'
        exp (HsApp e1 e2) (HsApp e1' e2') = lexp e1 e1' && lexp e2 e2'
        -- the fixities have been straightened out by now, so it's safe
        -- to ignore them?
        exp (OpApp l o _ ri) (OpApp l' o' _ ri') = 
            lexp l l' && lexp o o' && lexp ri ri'
        exp (NegApp e n) (NegApp e' n') = lexp e e' && exp n n'
        exp (SectionL e1 e2) (SectionL e1' e2') = 
            lexp e1 e1' && lexp e2 e2'
        exp (SectionR e1 e2) (SectionR e1' e2') = 
            lexp e1 e1' && lexp e2 e2'
912 913
        exp (ExplicitTuple es1 _) (ExplicitTuple es2 _) =
            eq_list tup_arg es1 es2
914 915
        exp (HsIf e e1 e2) (HsIf e' e1' e2') =
            lexp e e' && lexp e1 e1' && lexp e2 e2'
916

917 918
        -- Enhancement: could implement equality for more expressions
        --   if it seems useful
919 920
	-- But no need for HsLit, ExplicitList, ExplicitTuple, 
	-- because they cannot be functions
921
        exp _ _  = False
922 923 924 925

        tup_arg (Present e1) (Present e2) = lexp e1 e2
        tup_arg (Missing t1) (Missing t2) = tcEqType t1 t2
        tup_arg _ _ = False
926 927 928
    in
      lexp e1 e2

929 930 931 932 933
patGroup :: Pat Id -> PatGroup
patGroup (WildPat {})       	      = PgAny
patGroup (BangPat {})       	      = PgBang  
patGroup (ConPatOut { pat_con = dc }) = PgCon (unLoc dc)
patGroup (LitPat lit)		      = PgLit (hsLitKey lit)
934
patGroup (NPat olit mb_neg _)	      = PgN   (hsOverLitKey olit (isJust mb_neg))
935
patGroup (NPlusKPat _ olit _ _)	      = PgNpK (hsOverLitKey olit False)
936 937
patGroup (CoPat _ p _)		      = PgCo  (hsPatType p)	-- Type of innelexp pattern
patGroup (ViewPat expr p _)               = PgView expr (hsPatType (unLoc p))
938 939 940
patGroup pat = pprPanic "patGroup" (ppr pat)
\end{code}

941 942
Note [Grouping overloaded literal patterns]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
943 944 945 946 947 948 949
WATCH OUT!  Consider

	f (n+1) = ...
	f (n+2) = ...
	f (n+1) = ...

We can't group the first and third together, because the second may match 
950 951 952 953 954 955 956 957
the same thing as the first.  Same goes for *overloaded* literal patterns
	f 1 True = ...
	f 2 False = ...
	f 1 False = ...
If the first arg matches '1' but the second does not match 'True', we
cannot jump to the third equation!  Because the same argument might
match '2'!
Hence we don't regard 1 and 2, or (n+1) and (n+2), as part of the same group.