Merge before-ghci -> before-ghci-branch-merged in the non-ghc parts of
the tree.

Cleaning up the configuration/build process.

  - New build.mk option: GhcWithInterpreter. It has reasonable
    defaults, and shouldn't need to be overriden (you get the interpreter
    if you're bootstrapping with ghc 4.09+ on an ELF architecture, at
    the moment).

  - compilation manager now lives in compiler/compMan.  compiler/ghci
    contains only interpreter-related files.

  - WithGhcHc has gone, it now defaults to $(GHC).  This is so that
    we can reliably determine the version of $(GHC) using the stuff that
    configure tells us.  configure gets a new --with-ghc option so you
    can specify which ghc to use.

Just for fun, perl v5.6.0 version info has a different syntax.

First (awkward) steps towards an HOpenGL integration

more changes to FPTOOLS_GHC_VERSION:
* fixed bug in too-many-args case
* reworked regexp for versions

updated comments

New autoconf tests for ghc version from Michael Weber, with minor
modifications by me.

Check the version of the installed ghc, and set GhcVersion,
GhcMajVersion, GhcMinVersion and GhcPatchLevel make variables.

Add test for GNU regex in libc, which we'll use to decide whether to
compile our own copy of the library in package text.

Quick workaround for Reuben's M$ configuration problems

autoconf hackery for CTypes{,ISO} (continued)

Autoconf magic to determine a working DocBook CATALOG file (if any).  Makes
the makefile variable DOCBOOK_CATALOG refer to this CATALOG file (absolute
path).

Changed Perl test so that it simply allows vv5 & 6 and disallows any other;
also hardwire path to perl to /bin/perl under Windows, to avoid clashing with
other weird Perls that may be on the path.

Hopefully banish bogus "Happy version 1.6 or later is required to
compile GHC" messages.  The new rule is: if you don't have an
installed copy of Happy, but have a Happy source tree, we point the
build system at the latter (and pull the version number out of
happy/mk/version.mk).

Fix wibble in setting of HappyCmd.

Only require happy if compiling ghc, and ask for version 1.6

Don't attempt to discover the exact location of cpp, instead use 'gcc
-E'.  With the right combination of flags, it seems we can get the
same behaviour as calling cpp directly.

Use $CC -E instead of gcc.

Workarond apparent bug in FreeBSD 'sh'.

Look inside build tree too for happy binary

Use 'data_start' as the end-of-text marker if it exists.  This includes
the read-only text section on Linux/ELF, which is what we want.

Added `timezone' config test.

Prepared to support i?86-*-mingw32

Fixed caching bug in FPTOOLS_END_{TEXT,DATA}_SECTION.

Move 4.01 onto the main trunk.

Merge from HEAD

Re-install hack which fixes 'leading underscore' feature test under cygwin32

Distinguish between defn and use of etext and end syms; they're not
the same on platforms where we have to resort to asm labels.

Added tests required by Hugs

fix a couple of bugs in the text/data segment end symbol tests.

Add support for FreeBSD 3.0 (ELFish).
Rename i386-unknown-freebsd to i386-unknown-freebsd2, and add
i386-unknown-freebsd3.

NOTE: this won't bootstrap an ELF GHC using an a.out GHC, some laying
on of hands is required for that.

Added feature tests for etext and end plus text vs. data section placement in executable

O_BINARY test: protect fcntl.h include

Add feature test for O_BINARY

remove old test for GLIBC.

Fix leading underscore test.

Added long long feature test; macro namespace cleanup

define LEADING_UNDERSCORE in config.h.

GHC_CHECK_ALIGNMENT changed to check natural alignment restrictions

Added alignment tests