File: TestTemplate.hs

package info (click to toggle)
haskell-what4 1.5.1-4
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 2,240 kB
  • sloc: haskell: 34,630; makefile: 5
file content (815 lines) | stat: -rw-r--r-- 29,501 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
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
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
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
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
{-# LANGUAGE BlockArguments #-}
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE GADTs #-}
{-# LANGUAGE LambdaCase #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TypeApplications #-}
{-# LANGUAGE TypeOperators #-}
--module TestTemplate where

module Main where

import Control.Exception
import Control.Monad ((<=<), unless)
import Control.Monad.IO.Class (liftIO)
import           Control.Monad.Trans.Maybe
import Data.Bits
import Data.Parameterized.Map (MapF)
import qualified Data.Parameterized.Map as MapF
import Data.Parameterized.Nonce
import Data.Parameterized.Pair
import Data.Parameterized.Some
import Data.String
import Numeric (showHex)
import System.Exit (exitFailure)
-- import System.IO

import LibBF
import qualified Data.BitVector.Sized as BV

import What4.BaseTypes
import What4.Config
import What4.Interface

import           What4.Protocol.SMTWriter ((.==), mkSMTTerm)
import qualified What4.Protocol.SMTWriter as SMT
import qualified What4.Protocol.SMTLib2 as SMT2
import qualified What4.Protocol.Online as Online
import           What4.Protocol.Online (SolverProcess(..), OnlineSolver(..))
import qualified What4.Solver.CVC4 as CVC4

import What4.Expr
import What4.Expr.App (reduceApp)
import What4.Expr.Builder
import What4.Expr.GroundEval
import What4.SatResult

import What4.Utils.Arithmetic
import What4.Utils.FloatHelpers

import           Hedgehog
import qualified Hedgehog.Gen as Gen
import qualified Hedgehog.Range as Gen
import GHC.Stack

--import Debug.Trace (trace)



main :: IO ()
main =
  do let fpp = knownRepr :: FloatPrecisionRepr Prec32

     let xs = castTemplates RNE <>
              (Some <$> floatTestTemplates [] 0 fpp) <>
              (do r <- roundingModes
                  (Some <$> floatTemplates [r] 1 fpp))

     sym <- newExprBuilder FloatIEEERepr EmptyExprBuilderState globalNonceGenerator

     extendConfig CVC4.cvc4Options (getConfiguration sym)
     proc <- Online.startSolverProcess @(SMT2.Writer CVC4.CVC4) CVC4.cvc4Features Nothing sym

     let testnum = 500

     tests <- sequence [ do p <- templateGroundEvalTestAlt sym proc t testnum
                            --p <- templateGroundEvalTest sym proc t testnum
                            --p <- templateConstantFoldTest sym t testnum
                            pure (fromString (show t), p)
                       | Some t <- xs
                       ]
     testsPassed <- checkSequential $ Group "Float tests" tests
     unless testsPassed exitFailure


data FUnOp
  = FNeg
  | FAbs
  | FSqrt RoundingMode
  | FRound RoundingMode
 deriving (Show)

data FBinOp
  = FAdd RoundingMode
  | FSub RoundingMode
  | FMul RoundingMode
  | FDiv RoundingMode
  | FRem
  | FMin
  | FMax
 deriving (Show)

data FTestOp
  = FIsNaN
  | FIsInf
  | FIsZero
  | FIsPos
  | FIsNeg
  | FIsSubnorm
  | FIsNorm
 deriving Show

data FRelOp
  = FLogicEq
  | FLogicNeq
  | FEq
  | FApart
  | FUnordered
  | FLe
  | FLt
  | FGe
  | FGt
 deriving Show

-- | This datatype essentially mirrors the public API of the
--   What4 interface.  There should (eventually) be one element
--   of this datatype per syntax former method in "What4.Interface".
--   Each template represents a fragment of syntax that could be
--   generated.  We use these templates to test constant folding,
--   ground evaluation, term simplifications, fidelity
--   WRT solver term semantics and soundness of abstract domains.
--
--   The overall idea is that we want to enumerate all small
--   templates and use each template to generate a collection of test
--   cases.  Hopefully we can achieve high code path coverages with
--   a relatively small depth of templates, which should keep this process
--   managable.  We also have the option to manually generate templates
--   if necessary to increase test coverage.
data TestTemplate tp where
  TVar :: BaseTypeRepr tp -> TestTemplate tp

  TFloatPZero :: FloatPrecisionRepr fpp -> TestTemplate (BaseFloatType fpp)
  TFloatNZero :: FloatPrecisionRepr fpp -> TestTemplate (BaseFloatType fpp)
  TFloatNaN   :: FloatPrecisionRepr fpp -> TestTemplate (BaseFloatType fpp)

  TFloatUnOp ::
    FUnOp ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp)

  TFloatBinOp ::
    FBinOp ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp)

  TFloatTest ::
    FTestOp ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate BaseBoolType

  TFloatRel ::
    FRelOp ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate BaseBoolType

  TFloatFromBits :: (2 <= eb, 2 <= sb) =>
    FloatPrecisionRepr (FloatingPointPrecision eb sb) ->
    TestTemplate (BaseBVType (eb + sb)) ->
    TestTemplate (BaseFloatType (FloatingPointPrecision eb sb))

  TFloatToBits :: (2 <= eb, 2 <= sb) =>
    TestTemplate (BaseFloatType (FloatingPointPrecision eb sb)) ->
    TestTemplate (BaseBVType (eb + sb))

  TBVToFloat :: (1 <= w) =>
    FloatPrecisionRepr fpp ->
    RoundingMode ->
    Bool {- False = unsigned -} ->
    TestTemplate (BaseBVType w) ->
    TestTemplate (BaseFloatType fpp)

  TFloatToBV :: (1 <= w) =>
    NatRepr w ->
    RoundingMode ->
    Bool {- False = unsigned -} ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseBVType w)

  TFloatCast ::
    FloatPrecisionRepr fpp ->
    RoundingMode ->
    TestTemplate (BaseFloatType fpp') ->
    TestTemplate (BaseFloatType fpp)

  TFloatToReal ::
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate BaseRealType

  TRealToFloat ::
    FloatPrecisionRepr fpp ->
    RoundingMode ->
    TestTemplate BaseRealType ->
    TestTemplate (BaseFloatType fpp)

  TFloatFMA ::
    RoundingMode ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp) ->
    TestTemplate (BaseFloatType fpp)



instance Show (TestTemplate tp) where
  showsPrec d tt = case tt of
    TVar{}        -> showString "<var>"
    TFloatPZero{} -> showString "+0.0"
    TFloatNZero{} -> showString "-0.0"
    TFloatNaN{}   -> showString "NaN"
    TFloatUnOp op x -> showParen (d > 0) $
       shows op . showString " " . showsPrec 10 x
    TFloatBinOp op x y -> showParen (d > 0) $
       shows op . showString " " . showsPrec 10 x . showString " " . showsPrec 10 y
    TFloatTest op x -> showParen (d > 0) $
       shows op . showString " " . showsPrec 10 x
    TFloatRel op x y ->  showParen (d > 0) $
       shows op . showString " " . showsPrec 10 x . showString " " . showsPrec 10 y
    TFloatFMA r x y z -> showParen (d > 0) $
       showString "FMA" . showString " " . shows r . showString " " .
       showsPrec 10 x . showString " " . showsPrec 10 y . showString " " . showsPrec 10 z
    TFloatFromBits _fpp x -> showParen (d > 0) $
       showString "FloatFromBits " . showsPrec 10 x
    TFloatToBits x -> showParen (d > 0) $
       showString "FloatToBits " . showsPrec 10 x
    TFloatCast fpp r x -> showParen (d > 0) $
       showString "FloatCast " . shows fpp . showString " " . shows r . showString " " . showsPrec 10 x
    TFloatToReal x -> showParen (d > 0) $
       showString "FloatToReal " . showsPrec 10 x
    TRealToFloat fpp r x -> showParen (d > 0) $
       showString "RealToFloat " . shows fpp . showString " " . shows r . showString " " . showsPrec 10 x
    TBVToFloat fpp r sgn x ->
       showString (if sgn then "SBVToFloat " else "BVToFloat ") .
       shows fpp . showString " " .
       shows r . showString " " .
       showsPrec 10 x
    TFloatToBV w r sgn x ->
       showString (if sgn then "FloatToSBV" else "FloatToBV ") .
       shows w . showString " " .
       shows r . showString " " .
       showsPrec 10 x

-- | Compute the maximum depth of the given test template
templateDepth :: TestTemplate tp -> Integer
templateDepth = f
  where
  f :: TestTemplate tp -> Integer
  f t = case t of
          TVar{} -> 0

          TFloatPZero{} -> 0
          TFloatNZero{} -> 0
          TFloatNaN{}  -> 0

          TFloatUnOp _op x    -> 1 + (f x)
          TFloatBinOp _op x y -> 1 + max (f x) (f y)
          TFloatTest _op x    -> 1 + (f x)
          TFloatRel _op x y   -> 1 + max (f x) (f y)
          TFloatFMA _ x y z   -> 1 + max (f x) (max (f y) (f z))
          TFloatFromBits _ x  -> 1 + f x
          TFloatToBits x      -> 1 + f x
          TFloatCast _ _ x    -> 1 + f x
          TFloatToReal x      -> 1 + f x
          TRealToFloat _ _ x  -> 1 + f x
          TBVToFloat _ _ _ x  -> 1 + f x
          TFloatToBV _ _ _ x  -> 1 + f x


-- | A manually provided collection test templates that test coercions between types.
castTemplates :: RoundingMode -> [Some TestTemplate]
castTemplates r =
  [ Some (TFloatFromBits (knownRepr :: FloatPrecisionRepr Prec32) (TVar (BaseBVRepr (knownNat @32))))
  , Some (TFloatToBits (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec32))))
  , Some (TFloatCast (knownRepr :: FloatPrecisionRepr Prec32) r
              (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec64))))
  , Some (TFloatCast (knownRepr :: FloatPrecisionRepr Prec64) r
              (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec32))))
  , Some (TFloatToReal (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec32))))
  , Some (TRealToFloat (knownRepr :: FloatPrecisionRepr Prec32) r (TVar knownRepr))

  , Some (TBVToFloat (knownRepr :: FloatPrecisionRepr Prec32) r False
              (TVar (knownRepr :: BaseTypeRepr (BaseBVType 32))))
  , Some (TBVToFloat (knownRepr :: FloatPrecisionRepr Prec32) r True
              (TVar (knownRepr :: BaseTypeRepr (BaseBVType 32))))

  , Some (TFloatToBV (knownNat :: NatRepr 32) r False
              (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec32))))
  , Some (TFloatToBV (knownNat :: NatRepr 32) r True
              (TVar (knownRepr :: BaseTypeRepr (BaseFloatType Prec32))))
  ]


-- | Generate test templates for all predicates and relations
--   on folating point values, whose subterms are generated
--   by calling @floatTemplates.  With the given inputs.
--
--   CAUTION! This function blows up very quickly!
floatTestTemplates ::
  [RoundingMode] ->
  Integer ->
  FloatPrecisionRepr fpp ->
  [TestTemplate BaseBoolType]
floatTestTemplates rs n fpp = tops <> relops
 where
   subterms = floatTemplates rs n fpp
   tops   = [ TFloatTest op x | op <- fTestOps, x <- subterms ]
   relops = [ TFloatRel op x y | op <- fRelOps, x <- subterms, y <- subterms ]

-- | Generate floating-point test templates of the given
--   depth, iterating through each of the given rounding
--   modes for operations that require rounding.
--
--   CAUTION! This function blows up very quickly!
floatTemplates ::
  [RoundingMode] ->
  Integer ->
  FloatPrecisionRepr fpp ->
  [TestTemplate (BaseFloatType fpp)]
floatTemplates rs n fpp
  | n <= 0 = base
  | n == 1 = base <> floatOps fpp rs base
  | otherwise = f n

  where
   base = [ TVar (BaseFloatRepr fpp), TFloatPZero fpp, TFloatNZero fpp, TFloatNaN fpp ]

   f d | d < 1 = [ TVar (BaseFloatRepr fpp) ]
   f d = [ TVar (BaseFloatRepr fpp) ] <> floatOps fpp rs (f (d-1))

floatOps ::
  FloatPrecisionRepr fpp ->
  [RoundingMode] ->
  [TestTemplate (BaseFloatType fpp)] ->
  [TestTemplate (BaseFloatType fpp)]
floatOps fpp@(FloatingPointPrecisionRepr eb sb) rs subterms = casts <> uops <> bops <> fma
  where
    uops  = [ TFloatUnOp op x | op <- fUnOps rs, x <- subterms ]
    bops  = [ TFloatBinOp op x y | op <- fBinOps rs, x <- subterms, y <- subterms ]
    fma   = [ TFloatFMA r x y z | r <- rs, x <- subterms, y <- subterms, z <- subterms ]
    casts = [ case isPosNat (addNat eb sb) of
                Just LeqProof -> TFloatFromBits fpp (TVar (BaseBVRepr (addNat eb sb)))
                Nothing -> error $ unwords ["floatOps", "bad fpp", show fpp]
            ]

roundingModes :: [RoundingMode]
roundingModes = [ RNE, RNA, RTP, RTN, RTZ ]

fBinOps :: [RoundingMode] -> [FBinOp]
fBinOps rs =
  (FAdd <$> rs) <>
  (FSub <$> rs) <>
  (FMul <$> rs) <>
  (FDiv <$> rs) <>
  [ FRem, FMin, FMax ]

fUnOps :: [RoundingMode] -> [FUnOp]
fUnOps rs =
  [ FNeg, FAbs ] <>
  (FSqrt <$> rs) <>
  (FRound <$> rs)

fTestOps :: [FTestOp]
fTestOps = [ FIsNaN, FIsInf, FIsZero, FIsPos, FIsNeg, FIsSubnorm, FIsNorm ]

fRelOps :: [FRelOp]
fRelOps = [ FLogicEq, FLogicNeq, FEq, FApart, FUnordered, FLe, FLt, FGe, FGt ]



generateByType :: BaseTypeRepr tp -> Gen (GroundValue tp)
generateByType BaseBoolRepr = Gen.bool
generateByType (BaseFloatRepr fpp) = genFloat fpp
generateByType (BaseBVRepr w) = genBV w
generateByType BaseRealRepr  = genReal
generateByType tp = error ("generateByType! TODO " ++ show tp)

genReal :: Gen Rational
genReal = Gen.realFrac_ (Gen.linearFracFrom 0 (negate mx) mx)
  where mx = 1 ^^ (200::Integer)

genBV :: (1 <= w) => NatRepr w -> Gen (BV.BV w)
genBV w =
  do val <- Gen.integral (Gen.linearFrom 0 (minSigned w) (maxSigned w))
     pure (BV.mkBV w val)

-- | A random generator for floating-point values that tries to
--   get good coverage for all the various special and normal values.
genFloat :: FloatPrecisionRepr fpp -> Gen BigFloat
genFloat (FloatingPointPrecisionRepr eb sb) =
    Gen.frequency
        [ ( 1, pure bfPosZero)
        , ( 1, pure bfNegZero)
        , ( 1, pure bfPosInf)
        , ( 1, pure bfNegInf)
        , ( 1, pure bfNaN)
        , (50, genNormal)
        , ( 5, genSubnormal)
        , (45, genBinary)
        ]
 where
  emax = bit (fromInteger (intValue eb - 1)) - 1
  smax = bit (fromInteger (intValue sb)) - 1
  opts = fpOpts (intValue eb) (intValue sb) Away
  numBits = intValue eb + intValue sb

  -- generates non-shrinkable floats uniformly chosen from among all bitpatterns
  genBinary =
    do bits <- Gen.integral_ (Gen.linear 0 (bit (fromInteger numBits) - 1))
       pure (bfFromBits opts bits)

  -- generates non-shrinkable floats corresponding to subnormal values.  These are
  -- values with 0 biased exponent and nonzero mantissa.
  genSubnormal =
    do sgn  <- Gen.bool
       bits <- Gen.integral_ (Gen.linear 1 (bit (fromInteger (intValue sb)) - 1))
       let x0 = bfFromBits opts bits
       let x  = if sgn then bfNeg x0 else x0
       pure $! x

  -- tries to generate shrinkable floats, prefering "smaller" values
  genNormal =
    do sgn <- Gen.bool
       ex  <- Gen.integral (Gen.linearFrom 0 (1-emax) emax)
       mag <- Gen.integral (Gen.linear 1 smax)
       let x0 = bfStatus (bfMul2Exp opts (bfFromInteger mag) (ex - fromIntegral (lgCeil mag)))
       let x  = if sgn then bfNeg x0 else x0
       pure $! x

-- | Use the given map to bind the values in an expression and compute the value
--   of the expression, if it can be computed.
mapGroundEval :: MapF (Expr t) GroundValueWrapper -> Expr t tp -> MaybeT IO (GroundValue tp)
mapGroundEval m x =
  case MapF.lookup x m of
    Just v -> pure (unGVW v)
    Nothing -> tryEvalGroundExpr (mapGroundEval m) x

-- | Inject ground values into expressions based on their type.
groundLit ::  ExprBuilder t st fs -> BaseTypeRepr tp -> GroundValue tp -> IO (Expr t tp)
groundLit sym tp v =
  case tp of
    BaseFloatRepr fpp -> floatLit sym fpp v
    BaseBVRepr w      -> bvLit sym w v
    BaseBoolRepr      -> pure (backendPred sym v)
    BaseRealRepr      -> realLit sym v
    _ -> error $ unwords ["groundLit TODO", show tp]

-- | Given a map binding varables to expressions, rebuild the given expression by reapplying
--   the expression formers appearing in it.  This is used to test the constant-folding
--   rules of the expression builder.
reduceEval :: ExprBuilder t st fs -> MapF (Expr t) GroundValueWrapper -> Expr t tp -> IO (Expr t tp)
reduceEval sym m e
  | Just v <- MapF.lookup e m = groundLit sym (exprType e) (unGVW v)
  | Just a <- asApp e = reduceApp sym bvUnary =<< traverseApp (reduceEval sym m) a
  | otherwise = pure e

-- | Use the given solver process as an evaluation oracle to
--   verify that a given expression must have the given
--   value when the variables in it are bound via the
--   given map.
--
--   A value as computed via @solverEval@ may nonetheless fail
--   this test if functions appearing in it are underconstrained.
verifySolverEval :: forall t st fs solver tp.
  OnlineSolver solver =>
  ExprBuilder t st fs ->
  SolverProcess t solver ->
  MapF (Expr t) GroundValueWrapper ->
  Expr t tp ->
  GroundValue tp ->
  IO Bool
verifySolverEval _sym proc gmap expr val =
  do let c = Online.solverConn proc
     let f :: Pair (Expr t) GroundValueWrapper -> IO (SMT.Term solver)
         f (Pair e (GVW v)) =
            case exprType e of
              BaseFloatRepr fpp ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.floatTerm fpp v)
              BaseBVRepr w ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.bvTerm w v)
              BaseBoolRepr ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.boolExpr v)
              BaseRealRepr ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.rationalTerm v)

              tp -> fail ("verifySolverEval: TODO " ++ show tp)

     Online.inNewFrame proc do
       mapM_ (SMT.assumeFormula c <=< f) (MapF.toList gmap)

       gl <- f (Pair expr (GVW val))
       SMT.assumeFormula c (SMT.notExpr gl)

       res <- Online.check proc "eval"
       case res of
         Unknown -> fail "Expected UNSAT, but got UNKNOWN"
         Unsat _ -> pure True
         Sat _   -> pure False

-- | Use the given solver process as an evaluation oracle to
--   compute the a value of the given expression when given
--   a binding of variables that appear in the expression.
--   Return the value computed by the solver.
--
--   In principle, the solver might return one of several
--   different values for the expression if any of the
--   functions appearing in it are partial or underspecified.
solverEval :: forall t st fs solver tp.
  OnlineSolver solver =>
  ExprBuilder t st fs ->
  SolverProcess t solver ->
  MapF (Expr t) GroundValueWrapper ->
  Expr t tp ->
  IO (GroundValue tp)
solverEval _sym proc gmap expr =
  do let c = Online.solverConn proc
     let f :: Pair (Expr t) GroundValueWrapper -> IO (SMT.Term solver)
         f (Pair e (GVW v)) =
            case exprType e of
              BaseFloatRepr fpp ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.floatTerm fpp v)
              BaseBVRepr w ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.bvTerm w v)
              BaseBoolRepr ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.boolExpr v)
              BaseRealRepr ->
                do e' <- mkSMTTerm c e
                   return (e' .== SMT.rationalTerm v)

              tp -> fail ("solverEval: TODO " ++ show tp)

     Online.inNewFrame proc do
       mapM_ (SMT.assumeFormula c <=< f) (MapF.toList gmap)
       e' <- mkSMTTerm c expr
       res <- Online.check proc "eval"
       case res of
         Unsat _ -> fail "Expected SAT, but got UNSAT"
         Unknown -> fail "Expected SAT, but got UNKNOWN"
         Sat _ ->
           case exprType expr of
             BaseFloatRepr fpp ->
               do bv <- SMT.smtEvalFloat (Online.solverEvalFuns proc) fpp e'
                  return (bfFromBits (fppOpts fpp RNE) (BV.asUnsigned bv))
             BaseBVRepr w ->
               SMT.smtEvalBV (Online.solverEvalFuns proc) w e'
             BaseRealRepr ->
               SMT.smtEvalReal (Online.solverEvalFuns proc) e'
             BaseBoolRepr ->
               SMT.smtEvalBool (Online.solverEvalFuns proc) e'

             tp -> fail ("solverEval: TODO2 " ++ show tp)

showMap :: MapF (Expr t) GroundValueWrapper -> String
showMap gmap = unlines (map f (MapF.toList gmap))
  where
    f :: Pair (Expr t) (GroundValueWrapper) -> String
    f (Pair e (GVW v)) = show (printSymExpr e) <> " |-> " <> showGroundVal (exprType e) v

showGroundVal :: HasCallStack => BaseTypeRepr tp -> GroundValue tp -> String
showGroundVal tp v =
  case tp of
    BaseFloatRepr fpp ->
      let i = bfToBits (fppOpts fpp RNE) v in
        show v <> " 0x" <> showHex i ""
    BaseBVRepr w -> BV.ppHex w v
    BaseRealRepr -> show v
    BaseBoolRepr -> show v
    _ -> "showGroundVal: TODO " <> show tp

-- | This property generator takes a template and uses it to
--   compare our Haskell-side ground evaulation code against
--   the computations performed by an online solver, which
--   is used as a computational oracle.  Random values are
--   chosen for the free variables, and the expression is evaluated
--   by the ground evaluator using the generated values for
--   the variables.  Next, in the solver, we assert the equality of
--   the same variables to their concrete values and ask for a satisfying
--   model; then we ask the solver for the value of the expression in
--   that model and check that the two computations agree.
--
--   Some expressions are underspecified, which means their output is
--   unconstrained for some inputs (e.g, division by 0). In these cases
--   the ground evaluator may compute no value at all; these cases
--   are considered successful tests.
templateGroundEvalTest ::
  OnlineSolver solver =>
  ExprBuilder t st fs ->
  SolverProcess t solver ->
  TestTemplate tp ->
  Int ->
  IO Property
templateGroundEvalTest sym proc t numTests =
  do (sz, gmapGen, expr) <- templateGen sym t
     pure $ withTests (fromIntegral (max 1 (numTests * sz))) $ property $
       do annotateShow (printSymExpr expr)
          gmap <- forAllWith showMap gmapGen
          v  <- liftIO (runMaybeT (mapGroundEval gmap expr))
          annotate (maybe "Nothing" (showGroundVal (exprType expr)) v)
          res <- liftIO (try (solverEval sym proc gmap expr))
          case res of
            Left (ex :: IOError) -> footnote (show ex) >> failure
            Right v' ->
              do annotate (showGroundVal (exprType expr) v')
                 case v of
                   Just v_ -> Just True === groundEq (exprType expr) v_ v'
                   Nothing -> success

-- | This property generator takes a template and uses it to
--   compare our Haskell-side ground evaulation code against
--   the computations performed by an online solver, which
--   is used as a computational oracle.  Random values are
--   chosen for the free variables, and the expression is evaluated
--   by the ground evaluator using the generated values for
--   the variables.  Next, in the solver, we assert the equality of
--   the same variables to their concrete values, and ask the solver
--   to prove that it has the same value as we computed in the
--   ground evaluator.  This complements the above test; together
--   they demonstrate that the ground evaluator, when it computes a
--   value at all, computes the unique value that a solver may assign to it.
templateGroundEvalTestAlt ::
  OnlineSolver solver =>
  ExprBuilder t st fs ->
  SolverProcess t solver ->
  TestTemplate tp ->
  Int ->
  IO Property
templateGroundEvalTestAlt sym proc t numTests =
  do (sz, gmapGen, expr) <- templateGen sym t
     pure $ withTests (fromIntegral (max 1 (numTests * sz))) $ property $
       do annotateShow (printSymExpr expr)
          gmap <- forAllWith showMap gmapGen
          v  <- liftIO (runMaybeT (mapGroundEval gmap expr))
          case v of
            Nothing -> success
            Just v_ ->
              do annotate (showGroundVal (exprType expr) v_)
                 res <- liftIO (try (verifySolverEval sym proc gmap expr v_))
                 case res of
                   Left (ex :: IOError) -> footnote (show ex) >> failure
                   Right b -> if b then success else failure


-- | This property generator takes a template and uses it to
--   compare the ground evaluation code against the constant-folding
--   rules used when constructing terms.  Similar to the test above,
--   we compute an expression and then use ground evaluation to
--   compute a value for randomly-chosen values of the variables.
--   Next, we \"reduce\" the expression by reapplying the syntactic
--   constructors, replacing the variables with literal expressions.
--   Finally, we check that the expression has constant-folded to
--   a literal expression that agrees with the ground value computed
--   before. Moreover, ground evaluation should fail to compute a value
--   iff the reduced expression does not constant-fold to a literal.
templateConstantFoldTest ::
  ExprBuilder t st fs ->
  TestTemplate tp ->
  Int ->
  IO Property
templateConstantFoldTest sym t numTests =
  do (sz, gmapGen, expr) <- templateGen sym t
     pure $ withTests (fromIntegral (max 1 (numTests * sz))) $ property $
       do annotateShow (printSymExpr expr)
          gmap <- forAllWith showMap gmapGen

          v  <- liftIO (runMaybeT (mapGroundEval gmap expr))
          annotate (maybe "Nothing" (showGroundVal (exprType expr)) v)

          v' <- liftIO (reduceEval sym gmap expr)
          annotateShow (printSymExpr v')

          case v of
            Just v_ ->
              do p <- liftIO (isEq sym v' =<< groundLit sym (exprType expr) v_)
                 Just True === asConstantPred p
            Nothing -> False === baseIsConcrete v'

-- | Given a test template, compute data that can be used to drive one of the
--   test predicates above.  We return an @Int@ that counts how many variables
--   appear in the template, a generator action that computes ground values
--   for the variables appearing in the template, and an expression over
--   those variables according to the template.
templateGen :: forall t st fs tp.
  ExprBuilder t st fs -> TestTemplate tp -> IO (Int, Gen (MapF (Expr t) GroundValueWrapper), Expr t tp)
templateGen sym = f
  where
    f :: forall tp'. TestTemplate tp' -> IO (Int, Gen (MapF (Expr t) GroundValueWrapper), Expr t tp')
    f (TVar bt) =
       do v <- freshConstant sym emptySymbol bt
          return (1, MapF.singleton v . GVW <$> generateByType bt, v)

    f (TFloatPZero fpp) =
       do e <- floatPZero sym fpp
          return (0, pure MapF.empty, e)

    f (TFloatNZero fpp) =
       do e <- floatNZero sym fpp
          return (0, pure MapF.empty, e)

    f (TFloatNaN fpp) =
        do e <- floatNaN sym fpp
           return (0, pure MapF.empty, e)

    f (TFloatUnOp op x) =
        do (xn,xg,xe) <- f x
           e <- case op of
                  FNeg     -> floatNeg sym xe
                  FAbs     -> floatAbs sym xe
                  FSqrt  r -> floatSqrt sym r xe
                  FRound r -> floatRound sym r xe
           return (xn,xg, e)

    f (TFloatBinOp op x y) =
        do (xn,xg,xe) <- f x
           (yn,yg,ye) <- f y
           e <- case op of
                  FAdd r -> floatAdd sym r xe ye
                  FSub r -> floatSub sym r xe ye
                  FMul r -> floatMul sym r xe ye
                  FDiv r -> floatDiv sym r xe ye
                  FRem   -> floatRem sym xe ye
                  FMin   -> floatMin sym xe ye
                  FMax   -> floatMax sym xe ye

           return (xn+yn, MapF.union <$> xg <*> yg, e)

    f (TFloatTest op x) =
        do (xn,xg,xe) <- f x
           e <- case op of
                  FIsNaN  -> floatIsNaN sym xe
                  FIsInf  -> floatIsInf sym xe
                  FIsZero -> floatIsZero sym xe
                  FIsPos  -> floatIsPos sym xe
                  FIsNeg  -> floatIsNeg sym xe
                  FIsSubnorm -> floatIsSubnorm sym xe
                  FIsNorm -> floatIsNorm sym xe
           return (xn, xg, e)

    f (TFloatRel op x y) =
        do (xn,xg,xe) <- f x
           (yn,yg,ye) <- f y
           e <- case op of
                  FLogicEq   -> floatEq sym xe ye
                  FLogicNeq  -> floatNe sym xe ye
                  FEq        -> floatFpEq sym xe ye
                  FApart     -> floatFpApart sym xe ye
                  FUnordered -> floatFpUnordered sym xe ye
                  FLe        -> floatLe sym xe ye
                  FLt        -> floatLt sym xe ye
                  FGe        -> floatGe sym xe ye
                  FGt        -> floatGt sym xe ye

           return (xn+yn, MapF.union <$> xg <*> yg, e)

    f (TFloatFMA r x y z) =
        do (xn,xg,xe) <- f x
           (yn,yg,ye) <- f y
           (zn,zg,ze) <- f z
           e <- floatFMA sym r xe ye ze
           return (xn+yn+zn, foldr MapF.union MapF.empty <$> sequence [xg,yg,zg], e)

    f (TFloatFromBits fpp x) =
        do (xn,xg,xe) <- f x
           e <- floatFromBinary sym fpp xe
           return (xn,xg,e)

    f (TFloatToBits x) =
        do (xn,xg,xe) <- f x
           e <- floatToBinary sym xe
           return (xn,xg,e)

    f (TFloatCast fpp r x) =
        do (xn,xg,xe) <- f x
           e <- floatCast sym fpp r xe
           return (xn,xg,e)

    f (TFloatToReal x) =
        do (xn,xg,xe) <- f x
           e <- floatToReal sym xe
           return (xn,xg,e)

    f (TRealToFloat fpp r x) =
        do (xn,xg,xe) <- f x
           e <- realToFloat sym fpp r xe
           return (xn,xg,e)

    f (TBVToFloat fpp r sgn x) =
        do (xn,xg,xe) <- f x
           e <- case sgn of
                  False -> bvToFloat sym fpp r xe
                  True  -> sbvToFloat sym fpp r xe
           return (xn,xg,e)

    f (TFloatToBV w r sgn x) =
        do (xn,xg,xe) <- f x
           e <- case sgn of
                  False -> floatToBV sym w r xe
                  True  -> floatToSBV sym w r xe
           return (xn,xg,e)