add a note about POSIX compliance.

K&R is long dead
So why do its prototypes still live?

POSIX has not heard
of S_IWRITE
It uses S_IWSR.

Removing the "import privileged" implementation from hugs.

Version of PrelPrim with all the primitives listed.

Changing PrimPrel to PrelPrim

wibble.

Changing PrimPrel to PrelPrim.

Cut a lot of crappy code,
Now it works in Bill's abode.

Clean up the storage manager a little, and reinstate the compile time
garbage collector.  Then pray.

* Add new flag -fddump-minimal-imports, which dumps a file
  M.imports that contains the (allegedly) minimal bunch of
  imports that make the system work.
  It's done by Rename.printMinimalImports

* Extend foreign import/export to handle
	* Booleans
	* newtypes
  as requested by the FFI team

* Tidy up DsCCall quite a bit
  Remove maybeBoxedPrimTy from TcHsSyn

Preliminary support for heap profiles by module, type, or closure
description (ie. +RTS -h{Y,D,M}).

The hash tables are dynamically expanding, but lookup seems fairly
slow so the implementation may need to change at some point.

TSOs now belong to CCS_SYSTEM.

Include the closure description and type description fields in the
info table, for PROFILING.

Ignore system CCSs in the generated XML profile, since the browser
isn't doing this yet.

Imported sources for the much-debated HsFFI.h

Fill in the closure description and type description fields for the
heap-resident closures in this file (affects profiling only).

Use the decoded string, not the z-encoded string, for the closure
description.

generate the correct closureTypeDescr for an LFCon.

Changed a bunch of `#endif FOO' to `#endif /* FOO */', the former is
not strictly ANSI (don't know if the latter is, but `gcc -Wall -ansi
-pedantic' is silent then).

Nuke inexplicable calls to moduleThisScript.

Restore +Q and -Q operation, so we can run nofib again.

A bracket lost
is sense restored!

Ooops, forgot that one in my previous commit (don't mention
HsFalse/HsTrue anymore).

Removed comments about HsFalse/HsTrue, we guarantee 0/1.

* Change showsPrec on ErrorCall to include "error: "
* Go back to old defn of length; new version requires O(N) stack :-(

Clean up the error conditions a bit.

add icons subdir

Correctly handle constructors with strict fields, which was broken by
overenthusiastic constructor inlining some time back:
* notice if a constructor has strict fields, and set name(n).hasStrict,
  both for source modules and interfaces
* if a constr has strict fields, do not inline applications of it

install the icons.

Add new profiling tool (probably non-working so far).

processInterfaces: use dupOnto to create constructor_list, not appendOnto.
Ignore strictness annotations on iface constructor sigs.  (Probably wrong).

Observe wibble.

Restore ability to load a script which doesn't have a module header,
ie doesn't start   module M where ...

Fix various small bugs in module chasing & error handling.
Report a static analysis error if a module's name does not match
its file name.

* Enforce downward closure rule (first attempt :-)
  If both object and source of a module are available, be simple and
  choose the more recent.  If that causes a subsequent violation of the
  DCR, complain to the user at the time the violation is detected.
  The alternative is to have a clever algorithm which makes clever
  choices now to avoid conflicts later, but that looks complicated to
  do, and I think it would also confuse users.

* As a side effect of the above, enforce the rule that the Prelude
  must be all source or all object, but not a combination.

* Rationalise signature and semantics for findFilesForModule, so as to
  make client code simpler.

a few corrections.

find changed to FIND, that Windows users may
ignore DOS find when PATH has gone astray.

Added FIND as a symbol so that it can be changed on OSes which have
different ideas as to what "find" should do.

Restore debugging support that disappeared when the latest parallel
stuff was committed.