width and element type.

SIMD primops are now polymorphic in vector size and element type, but
only internally to the compiler. More specifically, utils/genprimopcode
has been extended so that it "knows" about SIMD vectors. This allows us
to, for example, write a single definition for the "add two vectors"
primop in primops.txt.pp and have it instantiated at many vector types.
This generates a primop in GHC.Prim for each vector type at which "add
two vectors" is instantiated, but only one data constructor for the
PrimOp data type, so the code generator is much, much simpler.

GHC.PrimopWrappers is only used by GHCi, which cannot evaluate LLVM-only
primops in any case.

gmainland authored Aug 08, 2013 and gmainlan@microsoft.com committed Sep 22, 2013

Note that this will only work with the LLVM back end pending LLVM patches to
change the GHC calling convention appropriately.

gmainland authored Aug 08, 2013 and gmainlan@microsoft.com committed Sep 22, 2013

On x86-32, the C calling convention specifies that when SSE2 is enabled, vector
arguments are passed in xmm* registers; however, float and double arguments are
still passed on the stack. This patch allows us to make the same choice for
GHC. Even when SSE2 is enabled, we don't want to pass Float and Double arguments
in registers because this would change the ABI and break the ability to link
with code that was compiled without -msse2.

The next patch will enable passing vector arguments in xmm registers on x86-32.

GHCi now runs constructors for linked libraries.

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

This commit exposes GHC's internal compiler pipeline through a `Hooks`
module in the GHC API. It currently allows you to hook:

 * Foreign import/exports declarations
 * The frontend up to type checking
 * The one shot compilation mode
 * Core compilation, and the module iface
 * Linking and the phases in DriverPhases.hs
 * Quasiquotation

Authored-by: Luite Stegeman <stegeman@gmail.com>
Authored-by: Edsko de Vries <edsko@well-typed.com>
Signed-off-by: Austin Seipp <austin@well-typed.com>

Ben Gamari authored Sep 20, 2013 and Krzysztof Gogolewski committed Sep 22, 2013

Also a small formatting change in GHCi :help

Thanks to Krzysztof Gogolewski (monoidal) for the first draft of this patch

The trace was being over-eager, which disturbs the typecheckers
careful knot-tying

There was a bit of clutter (tryTc stuff) caused by the
fact that tcLookupImported didn't return a MaybeErr.
Now it does.  That finishes up Trac #8004.

This fixes Trac #8333

Thanks to 'klao', who identified the cause and provided the patch

wordTyCon was treated as wired-in, but
  * It didn't have a WiredInName
  * It didn't appear in the list of wiredInTyCons

I'm not sure how anything worked!

In 6579a6c7 we removed existing comparison primops and introduced new ones
returning Int# instead of Bool. This commit (and associated commits in
array, base, dph, ghc-prim, integer-gmp, integer-simple, primitive, testsuite and
template-haskell) restores old names of primops. This allows us to keep
our API cleaner at the price of not having backwards compatibility.

This patch also temporalily disables fix for #8317 (optimization of
tagToEnum# at Core level). We need to fix #8326 first, otherwise
our primops code will be very slow.

It's much easier (and more efficient) to pattern match on
the HsOverLit than on the desugared version!

See Note [Optimising tagToEnum#] in Simplify

If we have
   class (F a ~ b) => C a b
then we can produce *derived* CFunEqCans.  These were not being
treated properly in two places:

a) in TcMType.zonkFlats (Trac #8134)
b) in TcSMonad.prepareInertsForImplications (Trac #8129)

This patch fixes both.

* Make isCFunEqCan_maybe return Maybe (TyCon, [Type])
* Add insertTM, deleteTM to TrieMap

It is off by default, which is meant to be a workaround for #8275.
Once #8275 is fixed we will enable this option by default.

If someone also feels that credits should better go into the
release announcement, just nuke the last line of that paragraph.

twanvl authored Sep 11, 2013 and Herbert Valerio Riedel committed Sep 18, 2013

This commit adds a `{-# MINIMAL #-}` pragma, which defines the possible
minimal complete definitions for a class. The body of the pragma is a
boolean formula of names.

The old warning for missing methods is replaced with this new one.

Note: The interface file format is changed to store the minimal complete
definition.

Authored-by: Twan van Laarhoven <twanvl@gmail.com>
Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This is temporary until new bool primops have been pushed.

Signed-off-by: Edward Z. Yang <ezyang@mit.edu>