Reduce noise if there's no in-tree version of Happy.

Fix a URL link

Comments only

Patches to configuration script for MSYS.  It's not clear that this'll
always work, because we generate paths like /c/mingw/bin, which MinGW
binaries won't understand -- but all of them seem to be fed to the
MSYS shell, which translates them to c:/mingw/bin etc.

Perhaps the Right Thing is to implement msyspath, by analogy with
cygpath, to convert /c/xxx to c:/xxx? 

Anyway, this patch seems to work for now.

Once more revised some details of the Data class.
Comitting the Data.Generics* library in a second.

New chapter on MSYS, filenames etc

Threaded RTS improvements:

  - Make the main_threads list doubly linked.  Have threads
    remove themselves from this list when they complete, rather
    than searching for completed main threads each time around
    the scheduler loop.  This removes an O(n) loop from the
    scheduler, but adds some new constraints (basically completed
    threads must remain on the run queue until dealt with, including
    threads which have been killed by an async exception).

  - Add a pointer from the TSO to the StgMainThread struct, for
    main threads.  This avoids a number of places where we had
    to traverse the list of main threads to find the right one,
    including one place in the scheduler loop.  Adding a field to
    a TSO is cheap.

  - taskStart: we should be resetting the startingWorkerThread flag
    in here.  Not sure why we aren't; maybe this got lost at some point.

  - Use the BlockedOnCCall flags in the non-threaded RTS too.  Q: what
    should happen if a thread does a foreign call which re-enters the
    RTS, and then sends an async exception to the original thread?
    Answer: it should deadlock, which it does in the threaded RTS, and
    this commit makes it do so in the non-threaded RTS too (see
    testsuite/tests/concurrent/should_run/conc040.hs).

Remove Deadlock_closure (not used any more).

add a few RTS ways (combinations) I missed out

Warning police, and omit a useless debugging msg in !RTS_SUPPORTS_THREADS.

wibble

Update version numbering policy: we now allow snapshots along the
stable branch (eg. 6.2.20040229).

Update docs for RTS options and GhcRtsThreaded

Add -fno-generics to SRC_HC_OPTS

build target should use '::' not ':'

make it compile (win32)

Revised dataTypeOf member in "deriving ( ... Data ...)".
This revised schemes relies on the updated modules
Data.Typeable and Data.Generics*.

cosmetic only

redo the condition-cache hack

Remove an unnecessary #ifdef

hs_exit(): call exitScheduler() *after* finalizeWeakPointersNow(),
because the latter needs to create/run threads.

scheduleWaitThread(): don't call THREAD_RUNNABLE because the new
thread is bound, and therefore only runnable by the current OS
thread.  Calling THREAD_RUNNABLE just wakes up workers and slows
things down.

waitCondition: use RELEASE_LOCK/ACQUIRE_LOCK because these check the
return value for errors.

Check result of SetEvent() for errors.

#include "RtsUtils.h" for barf prototype

Add the usual multi-slurp protection

Avoid signalling thread_ready_cond if rts_n_waiting_tasks==0.

merge rev. 1.6.2.3 from ghc-6-2-branch

thread_id_mutex isn't required (at least for non-SMP), since
createThread is under the Capability.

no need to declare thread_ready_cond

thread_ready_cond is initialised in two places

Back out hiding change -- hiding should hide the qualified name too

Fix a long-standing bug, whereby 'hiding' hid the qualified name as
well as the unqualified one.

This bug is not as easy to fix in the 6.2 branch, unfortunately; don't merge.

oops, fix the build

Feeble performance hack for the threaded RTS: instead of
allocating/releasing a new condition variable for each new call-in, we
just cache one in the RTS and re-use it for the next call.

On a little test I have here which does lots of call-ins on Windows,
this reduces the slowdown for using the threaded RTS from a factor of
7-8 down to a factor of 4-5.  I'm aiming for a factor of 2 or better...

ShutdownIOManager: don't free the IOManagerState, because the worker
threads haven't necessarily died yet and might reference it again.

Apparently fixes a strange bug: I have a Haskell program which on
Windows causes the shell to exit after I run it.  I don't know what
the real cause of this is, but this is a clear bug and fixing it stops
the shell from exiting.

MERGE TO STABLE

Add a -fno-full-laziness flag

- For each language option, document exactly the syntax that is
  stolen from Haskell 98.

- Fix up docrot at the beginning of the chapter.

Yet another fix to the -Onot optimisation that hides data type
representations in .hi files.

1.  Expose the representation if any fields are exposed

2.  Don't expose newtypes whose data-cons are abstract, unless the
    rep type is a FFI type.  (Previously we were more conservative
    and always exposed newtypes, just in case of a foreign decl.)

Document -threaded and -debug options.