Make object file splitting simpler, in preparation for conversion to
the new driver.

The "inject split markers" phase is now omitted, instead we generate
the split markers directly.

Driver: also removed now-defunct -fpedantic-bottoms flag.

Fix a bug in previous commit, some .globls were getting thrown away.

Fix -monly-3-regs problem.

Crude allocation-counting extension to ticky-ticky profiling.

Allocations are counted against the closest lexically enclosing
function closure, so you need to map the output back to the STG code.

Change the convention for cost-centre labels to be <name>_cc and
cost-centre stacks to be <name>_ccs.  This makes cost-centre labels
more consistent with our other naming conventions, and fixes some
problems caused by cost-centre labels being misinterpreted by the
mangler.

This fixes one cause of profiled programs crashing; if you're seeing
this symptom then this patch may help.

Add NetBSD/x86 support.

- generalise the per-module initialisation stubs so that we use it
  in normal (non-profiled) code too.  The initialisation stubs are
  now called '__init_<module>' rather than '_reg<module>'.

- Register foreign exported functions as stable pointers in the
  initialisation code for the module.  This fixes the foreign export
  problems reported by several people.

- remove the concept of "module groups" from the profiling subsystem.

- change the profiling semantics slightly; it should be unnecessary
  to use '-caf-all' to get reasonable profiles now.

sparc-sun-solaris2 uses .LLC<n> to refer to local string constants.

mangler fix.

fixup problem with '.' at the front of a local label.

egcs sometimes leaves a 'popl %ecx' in the epilogue, it seems.

Reduce alignment for info tables on x86 to 1 word.  It was previously
4 words, which meant that the entry code was wrongly aligned all the
time (because info tables are either 2 or 3 words).  At least this way
we've got a 1 in 4 chance of being properly aligned.

Move 4.01 onto the main trunk.

Miraculously, the same bug exists on sparc-sun-solaris2.

HPPA tweaks.

Fix splitting on i386-unknwon-linux: string constants were being
duplicated in each object, leading to large increases in binary
sizes.

Fix for egcs on sparc.

Changes to unbox the state in the ST and IO monads.

ST now has type

	newtype ST s a = ST (State# s -> STret s a)
	data STret s a = STret (State# s) a

IO now has type

	newtype IO a = IO (State# RealWorld -> IOResult a)
	data IOResult a = IOok   (State# RealWorld) a
	                | IOfail (State# RealWorld) IOError

So ST should be slightly more efficient, and IO should be nearly as
efficient as ST.

Bring in rev. 1.22.2.1 - Don't apply -monly-n-reg patches to _regBlah
functions (profiling cost-centre register things)

Another sparc/egcs fix.

- support the .p2align directive on x86 a.out architectures.  This
  might be required for others, too.

Added subl clause to epilogue munging.

partain 1.3 changes to 960626

Initial revision

Bulk of final changes for 2.01

Alleged post-Sansom 0.27+ code

After Andy Gill's late-95 changes

Merged GUM-4-04 branch into the main trunk. In particular merged GUM and
SMP code. Most of the GranSim code in GUM-4-04 still has to be carried over.

RS6000/AIX patches

RS6000 prof & conc patches

Parallel 2.04 patches

AIX/RS6000 patches

Fix: mkPiType has moved.

Alter to match new CoreLint.

A few things here:

  * fix some comment typos
  * alter CoreLint architecture to permit warnings as well as errors
  * add `endPassWithRules' to permit printing of rules as well as binds
  * move mkPiType to CoreUtils (for UsageSP reasons)

printf => showFFloat. Now waugh_neural fails with
   Fail: Prelude.(!!): negative index,
but that's better than before.  :-)

Synched with output required by H98 report

*nix-ified the expected output. M$-people are more used to error
messages, and one has to lose...  >:-)

"Non-exhaustive guards" => "Non-exhaustive patterns"