Who knows how long the tests for containers have been broken. They
haven't bitrotted, however.

Signed-off-by: Austin Seipp <austin@well-typed.com>

There were two unrelated errors fixed here:
 1) Make sure that only the *result kind* is reified when reifying
    a type family. Previously, the whole kind was reified, which
    defies the TH spec.

 2) Omit kind patterns in equations.

We assume that library authors supply correct role annotations
for their types, and therefore we do not need to check for
the availability of data constructors in Safe mode. See
discussion in #8725. This effectively fixes #8827 and #8826.

This results in a 57% runtime decrease when allocating an array of 128
bytes on a 64-bit machine.

Fixes #8876.

Signed-off-by: Austin Seipp <austin@well-typed.com>

Signed-off-by: Austin Seipp <austin@well-typed.com>

While -O1 and -O2 both include -globalopt, the order in which the passes are
run means that aliases aren't resolved which then causes llc to fall over.
See GHC bug #8855.

Signed-off-by: Austin Seipp <austin@well-typed.com>

Signed-off-by: Austin Seipp <austin@well-typed.com>

Fixes #8839

Signed-off-by: Austin Seipp <austin@well-typed.com>

Signed-off-by: Austin Seipp <austin@well-typed.com>

The patch provided by Christian Maeder <Christian.Maeder@dfki.de>
Signed-off-by: Karel Gardas <karel.gardas@centrum.cz>
Signed-off-by: Austin Seipp <austin@well-typed.com>

The implementations of newArray# and newArrayArray#, stg_newArrayzh
and stg_newArrayArrayzh, had three issues:

 * The condition for the loop that fills the array with the initial
   element was incorrect. It would write into the card table as
   well. The condition for the loop that filled the card table was
   never executed, as its condition was also wrong. In the end this
   didn't lead to any disasters as the value of the card table doesn't
   matter for newly allocated arrays.

 * The card table was unnecessarily initialized. The card table is
   only used when the array isn't copied, which new arrays always
   are. By not writing the card table at all we save some cycles.

 * The ticky allocation accounting was wrong. The second argument to
   TICK_ALLOC_PRIM is the size of the closure excluding the header
   size, but the header size was incorrectly included.

Fixes #8867.

This adds role annotations to Map and Set and therefore addresses #8718

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Clarify the order of the arguments. Also, remove any use of # in the
comments, which would make the rest of that comment line disappear in
the docs, due to being treated as a comment by the preprocessor.

 * AnId
 * ACoAxiom
 * AConLike

The test is a bit crude; -ddump-simpl | grep '#'.

I'm concerned that the -ddump-simpl output may differ on 32 and 64-bit
platforms.  So far I've only put in output for 64-bit platforms.

I don't know why these constant-folding rules were implemented for
and/or/xor but not for 'not'.

Adding them is part of the fix for Trac #8832.
(The other part is in Data.Bits.)

Documentation in response to Johan's questions

Plus, don't export hpRel from StgCmmHeap, StgCmmLayout
  (it is only used locally in StgCmmLayout)

(Otherwise the analysis was wrong, as covered by the new test case.)

Also make sure allocHeapClosure updates profiling counters with the
memory allocated.

Simon Marlow authored 11 years ago and tibbe committed 11 years ago

- Move array representation knowledge into SMRep

- Separate out low-level heap-object allocation so that we can reuse
  it from doNewArrayOp

- remove card-table initialisation, we can safely ignore the card
  table for newly allocated arrays.

Simon Marlow authored 11 years ago and tibbe committed 11 years ago

I'd like to be able to pack together non-pointer fields that are less
than a word in size, and this is a necessary prerequisite.

This results in a 46% runtime decrease when allocating an array of 16
unit elements on a 64-bit machine.

In order to allow newArray# to have both an inline and an out-of-line
implementation, cgOpApp is refactored slightly. The new implementation
of cgOpApp should make it easier to add other primops with both inline
and out-of-line implementations in the future.

Sorry about that...

Nowadays SetLevels floats case expressions as well as let-bindings,
and case expressions bind type variables.  We need to clone all such
floated binders, to avoid accidental name capture.  But I'd forgotten
to substitute for the cloned type variables, causing #8714.  (In the
olden days only Ids were cloned, from let-bindings.)

This patch fixes the bug and does quite a bit of clean-up refactoring
as well, by putting the context level in the LvlEnv.

There is no effect on performance, except that nofib 'rewrite' improves
allocations by 3%.  On investigation I think it was a fluke to do with
loop-cutting in big letrec nests.  But at least it's a fluke in the
right direction.

        Program           Size    Allocs   Runtime   Elapsed  TotalMem
--------------------------------------------------------------------------------
            Min          -0.4%     -3.0%    -19.4%    -19.4%    -26.7%
            Max          -0.0%     +0.0%    +17.9%    +17.9%      0.0%
 Geometric Mean          -0.1%     -0.0%     -0.7%     -0.7%     -0.4%