Cut'n'paste the latest version of rawSystem, not the horrible old
past-its-sell-by-date version I had in my smelly old tree.

Dearie dearie me, we are in a tizz today.

Put some imports needed for all platforms where they belong, and not in a
mingwin-only ifdef.

correct the help information about :info

Implement the :info command for GHCi.

Cut'n'paste rawSystem from hslibs rather than messing around trying to copy
SystemExts.lhs into compiler/main, which makes for Makefile pain.

Added note about daily updated online version.

Nuke these two hangovers from Stg Hugs daze.

Some Byzantine hacks from Reuben to make the main/SystemExts business
work as intended.

Add a couple of #ifdef ILX.

Fix updating of HS_SRCS for SystemExts.lhs

Add way i for ILX, plus ILX2IL and ILASM

argc and argv no longer need faking; System.lhs copes

This one-line file must have been here by mistake.

Now, commit ILX changes (mostly disablements)

Back out last commit, which overrode a lot of changes

Reverse the effects of rev 1.54 (which accidentally disabled the strictness
analyser.)

Fix peekArray for size==0.

Disable much ILXage owing to new driver support

Remove much gunk thanks to new driver support

Tidy

Many excellent fixes

Add ILX2IL and ILASM to Config.hs

Driver support for ILX compilation

sigh, I'll make this work one day

more wibbles

.NET can't cope with foreign labels

Fake argc & argv

Add ILX2IL_OPTS and ILASM_OPTS

wibble

Use the trick from unpackCStringLenIO to make peekArray run in
constant stack space: read the array from back to front using an
accumulating parameter.  Sadly this doesn't work for peekArray0.

Add getProgArgv, setProgArgv

Pass "-x c" as "-x" "c"; gcc-mingw gets confused by the former.

DocBook catalog: Only prefix hardtop if we end up using a glafp-utils/docbook CATALOG

Don't use 'foreign label' to get at prog_arg{v,c}, use
the RtsAPI-provided getProgArgv().

Add TyCon.mkLiftedPrimTyCon, solely for RealWorld which is the only
lifted primitive TyCon.

More wibbles in checking type validity

Wibbles to the checking-types commit

Yet another bug in lub

Change the story about POSIX headers in C compilation.

Until now, all C code in the RTS and library cbits has by default been
compiled with settings for POSIXness enabled, that is:
   #define _POSIX_SOURCE   1
   #define _POSIX_C_SOURCE 199309L
   #define _ISOC9X_SOURCE
If you wanted to negate this, you'd have to define NON_POSIX_SOURCE
before including headers.

This scheme has some bad effects:

* It means that ccall-unfoldings exported via interfaces from a
  module compiled with -DNON_POSIX_SOURCE may not compile when
  imported into a module which does not -DNON_POSIX_SOURCE.

* It overlaps with the feature tests we do with autoconf.

* It seems to have caused borkage in the Solaris builds for some
  considerable period of time.

The New Way is:

* The default changes to not-being-in-Posix mode.

* If you want to force a C file into Posix mode, #include as
  the **first** include the new file ghc/includes/PosixSource.h.
  Most of the RTS C sources have this include now.

* NON_POSIX_SOURCE is almost totally expunged.  Unfortunately
  we have to retain some vestiges of it in ghc/compiler so that
  modules compiled via C on Solaris using older compilers don't
  break.

Fix example code in the generics section.