Fix to build with 4.08.x

Fix some issues with hschooks.c, and hopefully clear up (or at least
document) the mess regarding which versions of Rts.h and RtsFlags.h
are included.

Sigh, I thought I could keep this file private to the RTS, but sadly
it's needed in order to #include RtsFlags.h, and we advertise
RtsFlags.h as a way to tweak flags through defaultsHook().  Oh well.

- Don't use non-portable autoconfery (this configure script was broken
  with newer autoconfs)

- Use FPTOOLS_SHEBANG_PERL from aclocal.m4 rather than local copy.

Also install <package>.haddock when we install the docs.

Use ioError instead of throw for IOErrors, in anticipation of
a change from IOError = Exception to IOError = IOException.

Make the new info-table construction code word without instance Bits
Int (i.e. GHC 4.08.x).  This probably makes it more correct, too.

Towards a complete binary package.

Towards a MinGW32 binary distribution with all accessories + compiler.

Back out the function tycon kind change (big comment to explain why)

Tidy up info table generation

Remove tests for text/data boundary symbols.

Remove tests for text/data section boundary symbols.

Merge the eval-apply-branch on to the HEAD
------------------------------------------

This is a change to GHC's evaluation model in order to ultimately make
GHC more portable and to reduce complexity in some areas.

At some point we'll update the commentary to describe the new state of
the RTS.  Pending that, the highlights of this change are:

  - No more Su.  The Su register is gone, update frames are one
    word smaller.

  - Slow-entry points and arg checks are gone.  Unknown function calls
    are handled by automatically-generated RTS entry points (AutoApply.hc,
    generated by the program in utils/genapply).

  - The stack layout is stricter: there are no "pending arguments" on
    the stack any more, the stack is always strictly a sequence of
    stack frames.

    This means that there's no need for LOOKS_LIKE_GHC_INFO() or
    LOOKS_LIKE_STATIC_CLOSURE() any more, and GHC doesn't need to know
    how to find the boundary between the text and data segments (BIG WIN!).

  - A couple of nasty hacks in the mangler caused by the neet to
    identify closure ptrs vs. info tables have gone away.

  - Info tables are a bit more complicated.  See InfoTables.h for the
    details.

  - As a side effect, GHCi can now deal with polymorphic seq.  Some bugs
    in GHCi which affected primitives and unboxed tuples are now
    fixed.

  - Binary sizes are reduced by about 7% on x86.  Performance is roughly
    similar, some programs get faster while some get slower.  I've seen
    GHCi perform worse on some examples, but haven't investigated
    further yet (GHCi performance *should* be about the same or better
    in theory).

  - Internally the code generator is rather better organised.  I've moved
    info-table generation from the NCG into the main codeGen where it is
    shared with the C back-end; info tables are now emitted as arrays
    of words in both back-ends.  The NCG is one step closer to being able
    to support profiling.

This has all been fairly thoroughly tested, but no doubt I've messed
up the commit in some way.

Handle bindings to just a pattern variable in cvtd.

Update config.guess and config.sub from GNU; a bug was reported in the
previous versions.

Improved foreign import conversion.

wibbles to External Core

Behave decently if there are NoStubs in ForeignStubs

Change interface to mkIfaceOrig

Better comparision for Exact RdrNames

Original names can occur in binding positions in External Core

Give -> the correct kind!

Use real datatypes for TH Clause and Match

Check for qualified names in binding positions in the parser instead
of the rename.  In External Core it's OK to have qualified names in
these places.

Parse Word literals

Remove checkGHCI

Use layout inside declarations quasi-quotes [d| ... |]

Wibble

Report TH errors better in stage 1

Fix a race condition/possible deadlock in the threaded rts:

If a callback into haskell finished before waitThread_() was called,
the signal was lost  ans waitThread_() waited indefinitely.

Solution: Don't release the sched_mutex between calls to scheduleThread_
and waitThread_.

Please note that the scheduler API function waitThread is still possibly
affected by this race condition. It's used in rts_mainEvalIO (I think that's
safe) and in finishAllThreads (this looks dangerous, but finishAllThreads is
never used).

Correctly convert expressions from TH datastructures to the internel Hs*
datastructures containing right infix operators left-parenthesised.

Mac OS X: fix a small bug that caused trouble with object splitting

config.guess says "m68k-apple-netbsd1.6", and that's fine.

Stub objects are called foo_stub.o, not foo.stub_o.  In other words,
findLinkable (aka maybe_getFileLinkable in the 5.04 branch) has never
worked.

A new rejig script which takes arguments rather than hardwired locations.

Add SA_RESETHAND (aka SA_ONESHOT) support.
Requested by: John Meacham

> module Main where

> import System.Posix.Signals

> main = do
>   installHandler sigUSR1 (Catch (print "usr1")) Nothing
>   installHandler sigUSR2 (CatchOnce (print "usr2")) Nothing
>   _ <- getLine
>   return ()

- remove old code for GHC < 3.03 (we don't support building GHC with
  less than 4.08).

- remove ErrorHdrHook and PatError, we now catch all exceptions at
  the top level and report panics (this will fix some test
  discrepancies with GHCi).

Don't put a recursive exception handler around the flushing of
stdout/stderr after running a command.  If the user closes stdout or
stderr, we now fail rather than going into an infinite loop.

Let configure know about perl v5.8