GitLab

Implemented gettimeofday on Windows by calling GetTickCount(). This
only seems to have a resolution of 1/100s, but that's just about OK
for threadDelay, which only needs 50 ticks per second.

Pass --target to configure of GMP, so that building for mingwin under
cygwin works properly.

Make sure windows.h is always checked for, not just when HOpenGL is present.

Revised GHC installer size downwards.

Removed Windows FAQ about -static no longer being supported (it is).

First shot at a STG interpreter for GHCI.  Translates Stg syntax into a
form convenient for interpretation, and can then run that.  Most of the
translation stuff is there and works.  The interpreter framework is there
and partly filled in, and seems to work.  There are still quite a lot of
cases, etc, to fill in, but this should be straightforward given that
the framework exists.  This interpreter cannot handle (yet?) unboxed
tuples, but can deal with more or less everything else, including standard
unboxed Int, Double, etc, code.

Initial primop support for the metacircular interpreter (GHCI).
Only appears if you compile with -DGHCI; if not, the world is
unchanged.

new primops:
   indexPtrOffClosure#
   indexWordOffClosure#

modified:
   dataToTag#   -- now dereferences indirections before extracting tag

new entry code
   mci_constr_entry          and
   mci_constr[1..8]entry
being the direct and vectored return code fragments for interpreter
created constructors.  Support for static constructors is not yet
done.

New handwritten .hc functions:
   mci_make_constr*
being code to create various flavours of constructors from args
on the stack.  An interface file to describe these will follow in
a later commit.

HP-PA fixes from Eric Schweitz <schweitz@nortelnetworks.com>

Remove redundant setNoDiscardId call from Specialise.newIdSM

First (awkward) steps towards an HOpenGL integration

<program> should be <command>

Fix the loop in SimplUtils.interestingArg

oops, forgot to remove an instance of REALLY_HASKELL_1_3

A list of simplifier-related stuff, triggered
	by looking at GHC's performance.

	I don't guarantee that this lot will lead to
	a uniform improvement over 4.08, but it it should
	be a bit better.  More work probably required.


* Make the simplifier's Stop continuation record whether the expression being
  simplified is the RHS of a thunk, or (say) the body of a lambda or case RHS.
  In the thunk case we want to be a bit keener about inlining if the type of
  the thunk is amenable to update in place.

* Fix interestingArg, which was being too liberal, and hence doing
  too much inlining.

* Extended CoreUtils.exprIsCheap to make two more things cheap:
    - 	case (coerce x) of ...
    -   let x = y +# z
  This makes a bit more eta expansion happen.  It was provoked by
  a program of Marcin's.

* MkIface.ifaceBinds.   Make sure that we emit rules for things
  (like class operations) that don't get a top-level binding in the
  interface file.  Previously such rules were silently forgotten.

* Move transformRhs to *after* simplification, which makes it a
  little easier to do, and means that the arity it computes is
  readily available to completeBinding.  This gets much better
  arities.

* Do coerce splitting in completeBinding. This gets good code for
	newtype CInt = CInt Int

	test:: CInt -> Int
	test x = case x of
	      	   1 -> 2
	      	   2 -> 4
	      	   3 -> 8
	      	   4 -> 16
	      	   _ -> 0

* Modify the meaning of "arity" so that during compilation it means
  "if you apply this function to fewer args, it will do virtually
  no work".   So, for example
	f = coerce t (\x -> e)
  has arity at least 1.  When a function is exported, it's arity becomes
  the number of exposed, top-level lambdas, which is subtly different.
  But that's ok.

  I removed CoreUtils.exprArity altogether: it looked only at the exposed
  lambdas.  Instead, we use exprEtaExpandArity exclusively.

  All of this makes I/O programs work much better.

* The simplifier used to glom together all the top-level bindings into
  a single Rec every time it was invoked.  The reason for this is explained
  in SimplCore.lhs, but for at least one simple program it meant that the
  simplifier never got around to unravelling the recursive group into
  non-recursive pieces.  So I've put the glomming under explicit flag
  control with a -fglom-binds simplifier pass.   A side benefit is
  that because it happens less often, the (expensive) SCC algorithm
  runs less often.

Omit unnecessary import

Do the begin-pass/end-pass stuff like the other core passes

Add comment

Add an access function substEnvEnv

Get the worker and wrapper in the right order

Primitives can return ThreadId# values

1) Fix a bad bug in Subst.lhs that made uniqAway go into an
   infinite loop when the 'hash code' in the in-scope set was
   zero.

2) Rename BasicTypes.isFragileOccInfo to isFragileOcc
   Add isDeadOcc to BasisTypes


(2) is just a tidy-up.  I have to commit it now because of (1), which
is a bad bug.  I hope that I've committed all the files involved in (2).

Added support for building gmp.dll again.

Gave version of RedHat in DocBook note.

Added note on how to fix RedHat 6.2 DocBook installation.

Improve -ddump-types

* Make the desugarer use string equality for string literal
  patterns longer than 1 character.  And put a specialised
  eqString into PrelBase, with a suitable specialisation rule.
  This makes a huge difference to the size of the code generated
  by deriving(Read) notably in Time.lhs

revert accidental parts of previous commit

add test for typedef-conflict with f.i.d.

Generate a new unique to be used in the typedef for a f.i.d., rather
than re-using the one from the function call, which might conflict if
there are two similar calls in the same module.

Document -unreg and unregisterised compilation.  Untested since
our DocBook installation is currently broken.

When compiling code for a case where the scrutinee is a primitve
comparison operator, we used to place the tag in a variable whose
unique was always the same: `mkPseudoUnique1 1'.  This was mostly
harmless but confused the Stix inliner in the NCG into generating
slightly less efficient code when the variable was used twice in a
basic block.

This patch fixes the problem by generating a new unique by just
changing the "tag" of an existing unique, namely the case binder.

Corrected URL for release notes.

Made message for exceeding 128Mb of heap on Windows more helpful, and
only abort when this limit is exceeded, not on first memory allocation!

Add codegen test for returning MutVar#.

Add new PrimRep, namely PrimPtrRep, as a catch-all for the various
boxed primitive types that currently don't have their own PrimReps.
Use this for MVar# and MutVar#.  This fixes a crash in the code
generator when a function returns one of these types.

recode primOpOkForSpeculation in terms of primOpIsCheap and primOpCanFail.

Add tests for newtypes in foreign declarations.

update output

Use newtype Ptr instead of just Addr, test newtypes in foreign decls.