Don't forget to set the CCCS from the thunk when entering std thunks.

Better typechecker error message when a type-signature variable is unified with a type

Allow a worker to be present for functions with no strictness info in
an interface file.  This is useful for functions which get hit by CPR
but not the strictness analyser.

Fixes to the unbox-strict-fields stuff for existential constructors.

(keving)

Sigh,  No point w/w pass storing the worker id in wrapper's info
field,  because later simplifications may change the worker id. I've
restored the original code which looks in the wrapper function and
guesses the worker id.  With the addition of CPR transf. this looks a
bit fragile. thinks, thinks, ....

The {Int,Word}{8,16,32,64} types are no longer 'wired-in', just names
with a known key.

Adjusted Id import to avoid module cycle; removed some unused helpers

(keving)

Big Bang introduction of CPR Analysis Pass.  Note that now
-fstrictness only does the strictness analysis phase,  it is necessary
to follow this with -fworker-wrapper to actually do the required Core
transformations. The -O option in the ghc driver script has been
modified appropriately.

For now,  CPR analysis is turned off.  To try it,  insert a
-fcpr_analyse between the -fstrictness and the -fworker-wrapper
options.

Misc. comments:

- The worker flag has been removed from an ID's StrictnessInfo field.
Now the worker info is an extra field in the Id's prag info.

- We do a nested CPR analysis,  but worker-wrapper only looks at the
info for the outermost constructor,  else laziness can be lost.

- Id's CPR Info in traces and interfaces file follows __M

- Worker-wrappery transformation now accounts for both strictness and
CPR analysis results.

(keving) The CPR Analysis Pass Module

Added Haskell wrapper for freeHaskellFunctionPtr()

changed to generate tail-jumping code for ccall thunks

(keving)

Test Update: Re-instate support for -fprint-core

Profiling fixes:

Function closures which are inside a lambda now *set* the CCCS,
instead of possibly appending to it.

In Simplify.lhs: allow inlining imported functions when profiling.
What we really want to do is allow any top-level binding to be
inlined, but there doesn't seem to be an easy way to tell whether a
binding is top-level or not.

Don't set the cost centre when entering a PAP.  Either:

	-  it's fully applied, in which case we're about to
	   set the CCS in the function anyway, or

	-  we're about to build another PAP in which case the
	   costs get attributed to the current CCS.

This avoids appending CCSs twice in some cases.

GHC 3.02 doesn't have Maybe.isNothing; switch to using isJust instead.

-funbox-strict-fields test.

Document -funbox-strict-fields.

Add -funbox-strict-fields

Add -funbox-strict-fields flag.

Hack from Simon: put abstract type decls (i.e. from an hi-boot file)
on the defferred list, just in case we happen to pull in the real one
as well.

On Win32 platforms, provide a (IOError.ComError Int) constructor for
handling HRESULTs conveniently.

The bugfix that time forgot

New compiler option -fignore-asserts: Causes

    (PrelGHC.assert pred expr)

to be rewritten to (expr).

Useful comment added.
(this is really a CVS checkin test, please ignore...)  --keithw

Fixed to allow desugaring of lit-lits with unboxed types.

Overhauled by Sven Panne.

updated to reflect what lib now provides

Generalised the {read,write}{Int,Word}{8,16,32,64}Array operations
to ST, plus the arrays are indexed using Ix.

Look inside build tree too for happy binary

Acknowledge the existence of both literal-literals and assertions.

assertError: raise an AssertionFailed exception

re-export PrelGHC.assert

Fix for parsing __scc expressions.

profiling-related sanity checker fixes.

Fix bug in allocGroup() when allocating an entire megablock in one go.

More profiling fixes.

Profiling fixes.

	- top-level CAF CCSs now *append* themselves to the
	  current CCS when called.

	- remove DICT stuff.

	- fixes to the auto-scc annotating in the desugarer.

Remove #ifdef __CONCURRENT_HASKELL__, update for H98.

enable __CONCURRENT_HASKELL__

RET_DYN doesn't have a proper header, just an info pointer.