Don't just pick the back edges ({-# SOURCE #-} imports) when
collecting the list of imports for module initialisation.  It's
entirely possible that a module is only reachable from Main via {-#
SOURCE #-} imports.

oops, revert part of previous commit that sneaked in by accident

Make the TSO structure shorter by 2 words:

	- remove splim.  It was always equal to &(tso->stack) +
	  RESERVED_STACK_WORDS.

	- make two enum fields 16 bits each instead of 32.

this test had random end conditions; make it a bit more deterministic

fail properly in ghcExit, now that we don't call error any more.

Minor cleanups to a few closure types: we don't use StgPtr in
Closures.h anymore.

Add -fvia-C to cg026 (which it needs).

Updated error message for "Compilation had errors" so as not to break tests.

remove unused imports

Reinstated isLocallyDefinedName test in isDllName, because nameModule can't
cope with local names.

Changed use of error for "Compilation failed" message to hPutStr stderr;
seemed better to SWM, and stops error messages being different on Windows
where RTS hooks don't work with DLLs.

Changes to Windows docs; specifically, remove advice to use -static.

Updates to Windows docs to cope with changes in new Cygwin and head of tree
GHC.

Clarified remarks about memory usage, and zapped bits specifically about
resources required to compile GHC (this is not ANNOUNCE material).

remove unused import

remove unused imports

avoid some repetition of filenames in the primop rules.

build genprimopcode at boot time.

werid -> weird

make this compile with pre-4.09 GHCs.

Fix so that libHS_imp_stub is made before HSrts.dll.

Changed DLL_IMPORT_RTS to DLL_IMPORT (I had it the wrong way around).

Give the included bits of Haskell their own rules in the Makefile, so
that make boot/depend work properly.

Now Char, Char#, StgChar have 31 bits (physically 32).
"foo"# is still an array of bytes.

CharRep represents 32 bits (on a 64-bit arch too). There is also
Int8Rep, used in those places where bytes were originally meant.
readCharArray, indexCharOffAddr etc. still use bytes. Storable and
{I,M}Array use wide Chars.

In future perhaps all sized integers should be primitive types. Then
some usages of indexing primops scattered through the code could
be changed to then-available Int8 ones, and then Char variants of
primops could be made wide (other usages that handle text should use
conversion that will be provided later).

I/O and _ccall_ arguments assume ISO-8859-1. UTF-8 is internally used
for string literals (only).

Z-encoding is ready for Unicode identifiers.

Ranges of intlike and charlike closures are more easily configurable.

I've probably broken nativeGen/MachCode.lhs:chrCode for Alpha but I
don't know the Alpha assembler to fix it (what is zapnot?). Generally
I'm not sure if I've done the NCG changes right.

This commit breaks the binary compatibility (of course).

TODO:
* is* and to{Lower,Upper} in Char (in progress).
* Libraries for text conversion (in design / experiments),
  to be plugged to I/O and a higher level foreign library.
* PackedString.
* StringBuffer and accepting source in encodings other than ISO-8859-1.

Use \{ \} in regex instead of { } for literal { }.
Hope it should be OK in theory; needed in practice here.

Add COMPILING_STDLIB symbol analagous to COMPILING_RTS for the extern vars
in stgio.h.

Add gmp.h to boot target so that it gets copied during booting (otherwise
make boot breaks down).

Use $(MV) not mv.

This is the program which generates bits of primop-related Haskell
(and, soon, C) code from ghc/compiler/prelude/primops.txt.

Reorganise the way primops are done.  Most of the information about
primops, their types and relevant attributes is in prelude/primops.txt.
A supporting program in fptools/ghc/utils/genprimopcode reads this
file and generates various bits of code which are #include'd into
prelude/PrimOp.lhs.  Eventually this mechanism will be extended to
generate PrelGHC.hi and C code for primops in the bytecode evaluator.

Also, add a few primops for creating, reading and writing BCOs.

Fixed conversion of MachWord to MachInt for large literals. Although
completely untested, I believe this fixes bug #111021.

Replaced `catchIO justIoErrors' with `catchJust ioErrors'

Port to OpenBSD from Thomas Nordin.  Please merge with 4.08 (you'll also
need the new config.guess and config.sub from the previous commit).

Pulled in latest versions from autoconf 2.13.  Necessary for OpenBSD port.

wibble

Another attempt at getting the pipeline stuff right.  Fixed at least
one bug.

-split-objs is entirely reasonable in conjunction with -fasm

wibble

- Implement the mkdepend -X<dir> and --exclude-directory=<dir> options.
- some small cleanups

Removed isLocallyDefinedName test in isDllName, as SPJ's with SM's suggestion
that it's redundant.