Type-indexed Typeable

This at long last realizes the ideas for type-indexed Typeable discussed in A
Reflection on Types (#11011). The general sketch of the project is described on
the Wiki (Typeable/BenGamari). The general idea is that we are adding a type
index to `TypeRep`,

    data TypeRep (a :: k)

This index allows the typechecker to reason about the type represented by the `TypeRep`.
This index representation mechanism is exposed as `Type.Reflection`, which also provides
a number of patterns for inspecting `TypeRep`s,

```lang=haskell
pattern TRFun :: forall k (fun :: k). ()
              => forall (r1 :: RuntimeRep) (r2 :: RuntimeRep)
                        (arg :: TYPE r1) (res :: TYPE r2).
                 (k ~ Type, fun ~~ (arg -> res))
              => TypeRep arg
              -> TypeRep res
              -> TypeRep fun

pattern TRApp :: forall k2 (t :: k2). ()
              => forall k1 (a :: k1 -> k2) (b :: k1). (t ~ a b)
              => TypeRep a -> TypeRep b -> TypeRep t

-- | Pattern match on a type constructor.
pattern TRCon :: forall k (a :: k). TyCon -> TypeRep a

-- | Pattern match on a type constructor including its instantiated kind
-- variables.
pattern TRCon' :: forall k (a :: k). TyCon -> [SomeTypeRep] -> TypeRep a
```

In addition, we give the user access to the kind of a `TypeRep` (#10343),

    typeRepKind :: TypeRep (a :: k) -> TypeRep k

Moreover, all of this plays nicely with 8.2's levity polymorphism, including the
newly levity polymorphic (->) type constructor.

Library changes
---------------

The primary change here is the introduction of a Type.Reflection module to base.
This module provides access to the new type-indexed TypeRep introduced in this
patch. We also continue to provide the unindexed Data.Typeable interface, which
is simply a type synonym for the existentially quantified SomeTypeRep,

    data SomeTypeRep where SomeTypeRep :: TypeRep a -> SomeTypeRep

Naturally, this change also touched Data.Dynamic, which can now export the
Dynamic data constructor. Moreover, I removed a blanket reexport of
Data.Typeable from Data.Dynamic (which itself doesn't even import Data.Typeable
now).

We also add a kind heterogeneous type equality type, (:~~:), to
Data.Type.Equality.

Implementation
--------------

The implementation strategy is described in Note [Grand plan for Typeable] in
TcTypeable. None of it was difficult, but it did exercise a number of parts of
the new levity polymorphism story which had not yet been exercised, which took
some sorting out.

The rough idea is that we augment the TyCon produced for each type constructor
with information about the constructor's kind (which we call a KindRep). This
allows us to reconstruct the monomorphic result kind of an particular
instantiation of a type constructor given its kind arguments.

Unfortunately all of this takes a fair amount of work to generate and send
through the compilation pipeline. In particular, the KindReps can unfortunately
get quite large. Moreover, the simplifier will float out various pieces of them,
resulting in numerous top-level bindings. Consequently we mark the KindRep
bindings as noinline, ensuring that the float-outs don't make it into the
interface file. This is important since there is generally little benefit to
inlining KindReps and they would otherwise strongly affect compiler performance.

Performance
-----------

Initially I was hoping to also clear up the remaining holes in Typeable's
coverage by adding support for both unboxed tuples (#12409) and unboxed sums
(#13276). While the former was fairly straightforward, the latter ended up being
quite difficult: while the implementation can support them easily, enabling this
support causes thousands of Typeable bindings to be emitted to the GHC.Types as
each arity-N sum tycon brings with it N promoted datacons, each of which has a
KindRep whose size which itself scales with N. Doing this was simply too
expensive to be practical; consequently I've disabled support for the time
being.

Even after disabling sums this change regresses compiler performance far more
than I would like. In particular there are several testcases in the testsuite
which consist mostly of types which regress by over 30% in compiler allocations.
These include (considering the "bytes allocated" metric),

 * T1969:  +10%
 * T10858: +23%
 * T3294:  +19%
 * T5631:  +41%
 * T6048:  +23%
 * T9675:  +20%
 * T9872a: +5.2%
 * T9872d: +12%
 * T9233:  +10%
 * T10370: +34%
 * T12425: +30%
 * T12234: +16%
 * 13035:  +17%
 * T4029:  +6.1%

I've spent quite some time chasing down the source of this regression and while
I was able to make som improvements, I think this approach of generating
Typeable bindings at time of type definition is doomed to give us unnecessarily
large compile-time overhead.

In the future I think we should consider moving some of all of the Typeable
binding generation logic back to the solver (where it was prior to
91c6b1f54aea658b0056caec45655475897f1972). I've opened #13261 documenting this
proposal.

4 files changed, 74 insertions, 56 deletions

diff --git a/testsuite/tests/perf/compiler/all.T b/testsuite/tests/perf/compiler/all.T
index 5f898fbbee..24b03d0326 100644
--- a/testsuite/tests/perf/compiler/all.T
+++ b/testsuite/tests/perf/compiler/all.T
@@ -39,7 +39,7 @@ test('T1969',
              # 2013-11-13 17 (x86/Windows, 64bit machine)
              # 2015-07-11 21 (x86/Linux, 64bit machine) use +RTS -G1
              # 2016-04-06 30 (x86/Linux, 64bit machine)
-           (wordsize(64), 68, 20)]),
+           (wordsize(64), 83, 20)]),
              #            28 (amd64/Linux)
              #            34 (amd64/Linux)
              # 2012-09-20 23 (amd64/Linux)
@@ -53,6 +53,7 @@ test('T1969',
              # 2015-10-28 55, (amd64/Linux) emit Typeable at definition site
              # 2016-10-20 68, (amd64/Linux) allow top-level string literals
              #                See the comment 16 on #8472.
+             # 2017-02-17 83  (amd64/Linux) Type-indexed Typeable
       compiler_stats_num_field('max_bytes_used',
           [(platform('i386-unknown-mingw32'), 5719436, 20),
                                  # 2010-05-17 5717704 (x86/Windows)
@@ -96,27 +97,28 @@ test('T1969',
              # 2014-06-29 303300692 (x86/Linux)
              # 2015-07-11 288699104 (x86/Linux, 64-bit machine) use +RTS -G1
              # 2016-04-06 344730660 (x86/Linux, 64-bit machine)
-           (wordsize(64), 756138176, 5)]),
-             # 17/11/2009 434845560 (amd64/Linux)
-             # 08/12/2009 459776680 (amd64/Linux)
-             # 17/05/2010 519377728 (amd64/Linux)
-             # 05/08/2011 561382568 (amd64/OS X)
-             # 16/07/2012 589168872 (amd64/Linux)
-             # 20/07/2012 595936240 (amd64/Linux)
-             # 23/08/2012 606230880 (amd64/Linux)
-             # 29/08/2012 633334184 (amd64/Linux) new codegen
-             # 18/09/2012 641959976 (amd64/Linux)
-             # 19/10/2012 661832592 (amd64/Linux) -fPIC turned on
-             # 23/10/2012 642594312 (amd64/Linux) -fPIC turned off again
-             # 12/11/2012 658786936 (amd64/Linux) UNKNOWN REASON
-             # 17/1/13:   667160192 (x86_64/Linux) new demand analyser
-             # 18/10/2013 698612512 (x86_64/Linux) fix for #8456
-             # 10/02/2014 660922376 (x86_64/Linux) call arity analysis
-             # 17/07/2014 651626680 (x86_64/Linux) roundabout update
-             # 10/09/2014 630299456 (x86_64/Linux) post-AMP-cleanup
-             # 03/06/2015 581460896 (x86_64/Linux) use +RTS -G1
-             # 28/10/2015 695430728 (x86_64/Linux) emit Typeable at definition site
-             # 28/10/2015 756138176 (x86_64/Linux) inst-decl defaults go via typechecker (#12220)
+           (wordsize(64), 831733376, 5)]),
+             # 2009-11-17 434845560 (amd64/Linux)
+             # 2009-12-08 459776680 (amd64/Linux)
+             # 2010-05-17 519377728 (amd64/Linux)
+             # 2011-08-05 561382568 (amd64/OS X)
+             # 2012-07-16 589168872 (amd64/Linux)
+             # 2012-07-20 595936240 (amd64/Linux)
+             # 2012-08-23 606230880 (amd64/Linux)
+             # 2012-08-29 633334184 (amd64/Linux) new codegen
+             # 2012-09-18 641959976 (amd64/Linux)
+             # 2012-10-19 661832592 (amd64/Linux) -fPIC turned on
+             # 2012-10-23 642594312 (amd64/Linux) -fPIC turned off again
+             # 2012-11-12 658786936 (amd64/Linux) UNKNOWN REASON
+             # 2013-91-17 667160192 (x86_64/Linux) new demand analyser
+             # 2013-10-18 698612512 (x86_64/Linux) fix for #8456
+             # 2014-02-10 660922376 (x86_64/Linux) call arity analysis
+             # 2014-07-17 651626680 (x86_64/Linux) roundabout update
+             # 2014-09-10 630299456 (x86_64/Linux) post-AMP-cleanup
+             # 2015-06-03 581460896 (x86_64/Linux) use +RTS -G1
+             # 2015-10-28 695430728 (x86_64/Linux) emit Typeable at definition site
+             # 2015-10-28 756138176 (x86_64/Linux) inst-decl defaults go via typechecker (#12220)
+             # 2017-02-17 831733376 (x86_64/Linux) Type-indexed Typeable
       only_ways(['normal']),
 
       extra_hc_opts('-dcore-lint -static'),
@@ -155,7 +157,7 @@ test('T3294',
              # 2015-07-11 43196344 (x86/Linux, 64-bit machine) use +RTS -G1
              # 2016-04-06 28686588 (x86/Linux, 64-bit machine)
 
-           (wordsize(64), 52992688, 20)]),
+           (wordsize(64), 63131248, 20)]),
              # prev:           25753192 (amd64/Linux)
              # 29/08/2012:     37724352 (amd64/Linux)
              #  (increase due to new codegen, see #7198)
@@ -173,6 +175,7 @@ test('T3294',
              #  D757: emit Typeable instances at site of type definition
              # 2016-07-11:     54609256  (Windows) before fix for #12227
              # 2016-07-11:     52992688  (Windows) after fix for #12227
+             # 2017-02-17:     63131248  (amd64/Linux) Type indexed Typeable
 
       compiler_stats_num_field('bytes allocated',
           [(wordsize(32), 1377050640, 5),
@@ -182,7 +185,7 @@ test('T3294',
            # 2013-11-13: 1478325844  (x86/Windows, 64bit machine)
            # 2014-01-12: 1565185140  (x86/Linux)
            # 2013-04-04: 1377050640  (x86/Windows, 64bit machine)
-           (wordsize(64), 2739731144, 5)]),
+           (wordsize(64), 2758641264, 5)]),
             # old:        1357587088 (amd64/Linux)
             # 29/08/2012: 2961778696 (amd64/Linux)
             # (^ increase due to new codegen, see #7198)
@@ -195,6 +198,7 @@ test('T3294',
             # 2014-09-10: 2709595808 (amd64/Linux) post-AMP cleanup
             # 2016-07-11: 2664479936 (Windows) before fix for #12227
             # 2016-07-11: 2739731144 (Windows) after fix for #12227 (ignoring)
+            # 2016-02-17: 2758641264 (amd64/Linux) (Type indexed Typeable)
       conf_3294,
 
       # Use `+RTS -G1` for more stable residency measurements. Note [residency].
@@ -419,7 +423,7 @@ test('T5631',
         # 2014-04-04:     346389856 (x86 Windows, 64 bit machine)
         # 2014-12-01:     390199244 (Windows laptop)
         # 2016-04-06:     570137436 (amd64/Linux) many reasons
-           (wordsize(64), 1077429456, 5)]),
+           (wordsize(64), 1517484488, 5)]),
         # expected value: 774595008 (amd64/Linux):
         # expected value: 735486328 (amd64/Linux) 2012/12/12:
         # expected value: 690742040 (amd64/Linux) Call Arity improvements
@@ -431,7 +435,7 @@ test('T5631',
         # 2015-03-18:     1124068664 (Mac) optimize Unify & zonking
         # 2016-10-19:     1024926024 (amd64/Linux) Refactor traceRn interface (#12617)
         # 2016-11-10:     1077429456 (amd64/Linux) Stop -dno-debug-output suppressing -ddump-tc-trace
-
+        # 2017-02-17:     1517484488 (amd64/Linux) Type-indexed Typeable
        only_ways(['normal'])
       ],
      compile,
@@ -655,18 +659,19 @@ test('T6048',
             # 2014-12-01: 49987836 (x86 Windows)
             # 2016-04-06: 55701280 (x86/Linux, 64-bit machine)
 
-           (wordsize(64),  94327392, 10)])
-             # 18/09/2012  97247032 amd64/Linux
-             # 16/01/2014 108578664 amd64/Linux (unknown, likely foldl-via-foldr)
-             # 18/01/2014  95960720 amd64/Linux Call Arity improvements
-             # 28/02/2014 105556793 amd64/Linux (unknown, tweak in base/4d9e7c9e3 resulted in change)
-             # 05/03/2014 110646312 amd64/Linux Call Arity became more elaborate
-             # 14/07/2014 125431448 amd64/Linux unknown reason. Even worse in GHC-7.8.3. *shurg*
-             # 29/08/2014 108354472 amd64/Linux w/w for INLINABLE things
-             # 14/09/2014  88186056 amd64/Linux BPP part1 change (more NoImplicitPreludes in base)
-             # 08/01/2014  95946688 amd64/Linux Mostly 4c834fd. Occasional spikes to 103822120!
-             # 11/03/2016 108225624 amd64/Linux unknown reason sadly; likely gradual creep.
-             # 25/11/2016  94327392 amd64/Linux Back down again hooray; still not sure why
+           (wordsize(64), 115714216, 10)])
+             # 2012-09-18  97247032 amd64/Linux
+             # 2014-01-16 108578664 amd64/Linux (unknown, likely foldl-via-foldr)
+             # 2014-01-18  95960720 amd64/Linux Call Arity improvements
+             # 2014-02-28 105556793 amd64/Linux (unknown, tweak in base/4d9e7c9e3 resulted in change)
+             # 2014-03-05 110646312 amd64/Linux Call Arity became more elaborate
+             # 2014-07-14 125431448 amd64/Linux unknown reason. Even worse in GHC-7.8.3. *shurg*
+             # 2014-08-29 108354472 amd64/Linux w/w for INLINABLE things
+             # 2014-09-14  88186056 amd64/Linux BPP part1 change (more NoImplicitPreludes in base)
+             # 2014-01-08  95946688 amd64/Linux Mostly 4c834fd. Occasional spikes to 103822120!
+             # 2016-03-11 108225624 amd64/Linux unknown reason sadly; likely gradual creep.
+             # 2016-11-25  94327392 amd64/Linux Back down again hooray; still not sure why
+             # 2017-02-17 115715592 amd64/Linux Type-indexed Typeable
       ],
       compile,[''])
 
@@ -721,9 +726,10 @@ test('T9675',
           # 2015-07-11    56         (x86/Linux, 64-bit machine) use +RTS -G1
           ]),
        compiler_stats_num_field('bytes allocated',
-          [(wordsize(64), 608284152, 10)
+          [(wordsize(64), 731171072, 10)
           # 2014-10-13    544489040
           # 2015-10-28    608284152  emit Typeable at definition site
+          # 2017-02-17    731171072  Type-indexed Typeable
           ,(wordsize(32), 279480696, 10)
           # 2015-07-11    279480696  (x86/Linux, 64-bit machine) use +RTS -G1
           ]),
@@ -737,14 +743,14 @@ test('T9675',
 test('T9872a',
      [ only_ways(['normal']),
        compiler_stats_num_field('bytes allocated',
-          [(wordsize(64), 3304620816, 5),
+          [(wordsize(64), 3298422648, 5),
           # 2014-12-10    5521332656    Initally created
           # 2014-12-16    5848657456    Flattener parameterized over roles
           # 2014-12-18    2680733672    Reduce type families even more eagerly
           # 2015-12-11    3581500440    TypeInType (see #11196)
           # 2016-04-07    3352882080    CSE improvements
           # 2016-10-19    3134866040    Refactor traceRn interface (#12617)
-          # 2017-02-01    3304620816
+          # 2017-02-17    3298422648    Type-indexed Typeable
            (wordsize(32), 1740903516, 5)
           # was           1325592896
           # 2016-04-06    1740903516    x86/Linux
@@ -792,7 +798,7 @@ test('T9872c',
 test('T9872d',
      [ only_ways(['normal']),
        compiler_stats_num_field('bytes allocated',
-          [(wordsize(64), 478169352, 5),
+          [(wordsize(64), 535565128, 5),
           # 2014-12-18    796071864   Initally created
           # 2014-12-18    739189056   Reduce type families even more eagerly
           # 2015-01-07    687562440   TrieMap leaf compression
@@ -802,6 +808,7 @@ test('T9872d',
           # 2016-03-18    506691240   optimize Unify & zonking
           # 2016-12-05    478169352   using tyConIsTyFamFree, I think, but only
           #                           a 1% improvement 482 -> 478
+          # 2017-02-17    535565128   Type-indexed Typeable
            (wordsize(32), 264566040, 5)
           # some date     328810212
           # 2015-07-11    350369584
@@ -835,7 +842,7 @@ test('T9961',
 test('T9233',
     [ only_ways(['normal']),
       compiler_stats_num_field('bytes allocated',
-        [(wordsize(64), 884436192, 5),
+        [(wordsize(64),  974530192, 5),
          # 2015-08-04    999826288     initial value
          # 2016-04-14   1066246248     Final demand analyzer run
          # 2016-06-18    984268712     shuffling around of Data.Functor.Identity
@@ -845,6 +852,7 @@ test('T9233',
          # 2017-01-23    861862608     worker/wrapper evald-ness flags; another 5% improvement!
          # 2017-02-01    894486272     Join points
          # 2017-02-07    884436192     Another improvement to SetLevels
+         # 2017-02-17    974530192     Type-indexed Typeable
 
          (wordsize(32),  515672240, 5)   # Put in your value here if you hit this
          # 2016-04-06    515672240     (x86/Linux) initial value
@@ -857,7 +865,7 @@ test('T9233',
 test('T10370',
      [ only_ways(['optasm']),
        compiler_stats_num_field('max_bytes_used', # Note [residency]
-          [(wordsize(64), 38221184, 15),
+          [(wordsize(64), 51126304, 15),
           # 2015-10-22    19548720
           # 2016-02-24    22823976   Changing Levity to RuntimeRep; not sure why this regresses though, even after some analysis
           # 2016-04-14    28256896   final demand analyzer run
@@ -870,15 +878,17 @@ test('T10370',
           #     were identical, so I think it's just GC noise.
           # 2016-10-20    38221184   Allow top-level string literals.
           #                          See the comment 16 on #8472.
+          # 2017-02-17    51126304   Type-indexed Typeawble
            (wordsize(32), 11371496, 15),
           # 2015-10-22    11371496
           ]),
        compiler_stats_num_field('peak_megabytes_allocated', # Note [residency]
-          [(wordsize(64), 146, 15),
+          [(wordsize(64), 187, 15),
           # 2015-10-22     76
           # 2016-04-14    101 final demand analyzer run
           # 2016-08-08    121 see above
           # 2017-01-18    146 Allow top-level string literals in Core
+          # 2017-02-17    187 Type-indexed Typeawble
            (wordsize(32),  39, 15),
           # 2015-10-22     39
           ]),
@@ -916,9 +926,10 @@ test('T12227',
 test('T12425',
      [ only_ways(['optasm']),
        compiler_stats_num_field('bytes allocated',
-          [(wordsize(64), 133380960, 5),
+          [(wordsize(64), 173257664, 5),
           # initial:    125831400
           # 2017-01-18: 133380960  Allow top-level string literals in Core
+          # 2017-02-17: 173257664  Type-indexed Typeable
           ]),
      ],
      compile,
@@ -929,11 +940,12 @@ test('T12234',
        compiler_stats_num_field('bytes allocated',
           [(platform('x86_64-unknown-mingw32'), 77949232, 5),
           # initial:      77949232
-           (wordsize(64), 74374440, 5),
+           (wordsize(64), 86525344, 5),
           # initial:      72958288
           # 2016-01-17:   76848856  (x86-64, Linux. drift?)
           # 2017-02-01:   80882208  (Use superclass instances when solving)
           # 2017-02-05:   74374440  (Probably OccAnal fixes)
+          # 2017-02-17:   86525344  (Type-indexed Typeable)
           ]),
      ],
      compile,
@@ -942,10 +954,11 @@ test('T12234',
 test('T13035',
      [ only_ways(['normal']),
        compiler_stats_num_field('bytes allocated',
-          [(wordsize(64), 88806416, 5),
-          # 2017-01-05   90595208 initial
-          # 2017-01-19   95269000 Allow top-level string literals in Core
-          # 2017-02-05   88806416 Probably OccAnal fixes
+          [(wordsize(64), 103890200, 5),
+          # 2017-01-05   90595208  initial
+          # 2017-01-19   95269000  Allow top-level string literals in Core
+          # 2017-02-05   88806416  Probably OccAnal fixes
+          # 2017-02-17   103890200 Type-indexed Typeable
           ]),
      ],
      compile,
diff --git a/testsuite/tests/perf/haddock/all.T b/testsuite/tests/perf/haddock/all.T
index 4c641d5828..a148b712d2 100644
--- a/testsuite/tests/perf/haddock/all.T
+++ b/testsuite/tests/perf/haddock/all.T
@@ -5,7 +5,7 @@
 test('haddock.base',
      [unless(in_tree_compiler(), skip), req_haddock
      ,stats_num_field('bytes allocated',
-          [(wordsize(64), 38425793776, 5)
+          [(wordsize(64), 25592972912, 5)
             # 2012-08-14:  5920822352 (amd64/Linux)
             # 2012-09-20:  5829972376 (amd64/Linux)
             # 2012-10-08:  5902601224 (amd64/Linux)
@@ -34,6 +34,7 @@ test('haddock.base',
             # 2017-02-11: 34819979936 (x86_64/Linux) - OccurAnal / One-Shot  (#13227)
             # 2017-02-16: 32695562088 Better Lint for join points
             # 2017-02-17: 38425793776 (x86_64/Linux) - Generalize kind of (->)
+            # 2017-02-12: 25592972912 (x86_64/Linux) - Type-indexed Typeable
 
           ,(platform('i386-unknown-mingw32'), 4434804940, 5)
             # 2013-02-10:                     3358693084 (x86/Windows)
@@ -56,7 +57,7 @@ test('haddock.base',
 test('haddock.Cabal',
      [unless(in_tree_compiler(), skip), req_haddock
      ,stats_num_field('bytes allocated',
-          [(wordsize(64), 27784875792, 5)
+          [(wordsize(64), 18865432648, 5)
             # 2012-08-14:  3255435248 (amd64/Linux)
             # 2012-08-29:  3324606664 (amd64/Linux, new codegen)
             # 2012-10-08:  3373401360 (amd64/Linux)
@@ -100,6 +101,7 @@ test('haddock.Cabal',
             # 2017-02-11: 25533642168 (amd64/Linux) - OccurAnal / One-Shot  (#13227)
             # 2017-02-16: 23867276992  Better Lint for join points
             # 2017-02-17: 27784875792 (amd64/Linux) - Generalize kind of (->)
+            # 2017-02-12: 18865432648 (amd64/Linux) - Type-indexed Typeable
 
           ,(platform('i386-unknown-mingw32'), 3293415576, 5)
             # 2012-10-30:                     1733638168 (x86/Windows)
diff --git a/testsuite/tests/perf/should_run/all.T b/testsuite/tests/perf/should_run/all.T
index 4bd75f70de..27d8df87b8 100644
--- a/testsuite/tests/perf/should_run/all.T
+++ b/testsuite/tests/perf/should_run/all.T
@@ -73,7 +73,7 @@ test('lazy-bs-alloc',
      [stats_num_field('peak_megabytes_allocated', (2, 1)),
                                  # expected value: 2 (amd64/Linux)
       stats_num_field('bytes allocated',
-          [(wordsize(64), 444720, 5),
+          [(wordsize(64), 421792, 5),
             #             489776 (amd64/Linux)
             # 2013-02-07: 429744 (amd64/Linux)
             # 2013-12-12: 425400 (amd64/Linux)
@@ -81,6 +81,7 @@ test('lazy-bs-alloc',
             # 2015-08-15: 431500 (Windows not good enough. avg of Windows&Linux)
             # 2015-12-15: 444720 (amd64/Linux, D1616)
             # 2015-12-17: 444720 (widen 3->5%, Windows is at 462688)
+            # 2017-01-30: 421792 (amd64/Linux, strangely Type-indexed Typeable)
            (wordsize(32), 429760, 2)]),
             # 2013-02-10: 421296 (x86/Windows)
             # 2013-02-10: 414180 (x86/OSX)
diff --git a/testsuite/tests/perf/space_leaks/all.T b/testsuite/tests/perf/space_leaks/all.T
index a58ae2c97b..76ad7a7606 100644
--- a/testsuite/tests/perf/space_leaks/all.T
+++ b/testsuite/tests/perf/space_leaks/all.T
@@ -58,13 +58,14 @@ test('T4018',
 
 test('T4029',
      [stats_num_field('peak_megabytes_allocated',
-          [(wordsize(64), 71, 10)]),
+          [(wordsize(64), 80, 10)]),
             # 2016-02-26: 66 (amd64/Linux)           INITIAL
             # 2016-05-23: 82 (amd64/Linux)           Use -G1
             # 2016-07-13: 92 (amd64/Linux)           Changes to tidyType
             # 2016-09-01: 71 (amd64/Linux)           Restore w/w limit (#11565)
+            # 2017-02-12: 80 (amd64/Linux)           Type-indexed Typeable
       stats_num_field('max_bytes_used',
-          [(wordsize(64), 22770352, 5)]),
+          [(wordsize(64), 24151096, 5)]),
             # 2016-02-26: 24071720 (amd64/Linux)     INITIAL
             # 2016-04-21: 25542832 (amd64/Linux)
             # 2016-05-23: 25247216 (amd64/Linux)     Use -G1
@@ -75,6 +76,7 @@ test('T4029',
             # 2016-11-14: 21387048 (amd64/Linux)     Creep back upwards :(
             # 2017-01-18: 21670448 (amd64/Linux)     Float string literals to toplevel
             # 2017-02-07: 22770352 (amd64/Linux)     It is unclear
+            # 2017-02-12: 24151096 (amd64/Linux)     Type-indexed Typeable
       extra_hc_opts('+RTS -G1 -RTS' ),
       ],
      ghci_script,