GitLab

---------------------------------------------
	    Calculate the free vars of a type 'right'
	---------------------------------------------

	type C u a = u

Question: is 'a' free in 'C u a'?

I think the answer should be 'no'; see typecheck/should_compile/tc157.hs
for an example of why it matters.  This commit makes it so, and adds
comments to explain a dark corner in the zonking code.

Dont use the isomorphic-newtype deriving trick for recursive newtypes

bugfix

Add $(INSTALL_IFACES) and $(INSTALL_IFACES_WITH_DIRS) for installing
interfaces (the latter replaces $(INSTALL_DATAS_WITH_DIRS).

Add support for building & installing Haddock docs.

Also minor cleanups: use $(ifacedir) rather than overriding $(datadir)
when installing interfaces.

Add:
	$(ifacedir)		where we install interfaces
	$(HADDOCK_INPLACE)	how to invoke Haddock

Fix validity checking of an instance-decl context

Teach code gen about some extra no-ops

When handling external call-ins (via the RTS API) in
the multi-threaded case, add the StgMainThread that
the external thread is going to block waiting on
to the main_threads list prior to scheduling the new
worker thread.

Do this by having the scheduler provide a new entry
point, scheduleWaitThread().

Fixes a bug/race condition spotted by Wolfgang Thaller
(see scheduleWaitThread() comment) + enables a little
tidier interface between RtsAPI and Schedule.

Include directories called 'tests' in source distributions

Fix wibble in previous commit to Weak.c.  Should fix problems with
--enable-threaded-rts.

clarification in multi-threading docs

Comments only

typo

---------------------------------------
	    Rehash the handling of SeqOp
	---------------------------------------

See the comments in the commentary (Cunning Prelude Code).

* Expunge SeqOp altogether

* Add GHC.Base.lazy :: a -> a
  to GHC.Base

* Add GHC.Base.lazy
  to basicTypes/MkId.  The idea is that this defn will over-ride
  the info from GHC.Base.hi, thereby hiding strictness and
  unfolding

* Make stranal/WorkWrap do a "manual inlining" for GHC.Base.lazy
  This happens nicely after the strictness analyser has run.

* Expunge the SeqOp/ParOp magic in CorePrep

* Expunge the RULE for seq in PrelRules

* Change the defns of pseq/par in GHC.Conc to:

	{-# INLINE pseq  #-}
       	pseq :: a -> b -> b
       	pseq  x y = x `seq` lazy y

       	{-# INLINE par  #-}
       	par :: a -> b -> b
       	par  x y = case (par# x) of { _ -> lazy y }

Notes about par/seq

Wibble on printing outmost parens

Ignore fewer type errors in tcSimplifyTop; fixes tc106

Fix HnType parenthesisation; fixes rnfail020, tcfail057

Record correct location for fixity errors

* Use the encoded module name for the C stub of "foreign import
  wrapper", so e.g. the '.' in hierarchical module names gets mangled.
* Nuke unused parameter in mkFExportCBits

Documentation for Darwin/MacOS X-Specific options/packacke.conf entries
(-framework, -framework-path; framework_dirs, extra_frameworks)

remove RTS_DARWIN_ONLY_SYMBOLS, use RTS_EXTRA_SYMBOLS instead.

all: restore old behaviour of exiting upon 'boot' failure in a SUBDIR

Accept empty package lists.

Another HsParTy wibble; cures rn018 failure

Make count_lines use the in-place unlit

---------------------------------------
	Utterly expunge the tyGenInfo field of
			an IdInfo
	---------------------------------------

tyGenInfo was a relic of a previous version of Keith's usage
analyser.  It's just dead code, so I've nuked it.

Wibble in Generics; fixes str002

---------------------------------------
	Add a target "count" which counts lines
	       in user source files
	---------------------------------------

Make it compile

added a clarifying comment to the PowerPC version

Fix a logic bug which caused some dead weak pointers to not get
stubbed out with DEAD_WEAK.

Initial support for GHCi on MacOS X.
The code still has problems with global variables in compiled C code, but
it _seems_ to work OK for GHC-compiled code. Many error checks are still
missing.

Added support for Frameworks on MacOS X.
*) On MacOS X, two additional command-line options are supported:
-framework <FRAMEWORK>    link with framework, gets passed on to ld
-framework-path <PATH>    gets passed on to ld as "-F<PATH>". (-F is already taken for GHC).
*) Two corresponding additional options for package.conf files:
framework_dirs
extra_frameworks
These options are allowed on any platform. They are ignored everywhere except on MacOS X.
*) ghc-pkg no longer uses Read. Instead, it uses a Happy parser. ghc/utils/ghc-pkg/ParsePkgConfLite.y is basically a copy of ghc/compiler/main/ParsePkgConf.y. "Light" refers to the fact that it doesn't depend on other GHC modules and has less sophisticated error reporting.
*) Frameworks will need some additional work for GHCi.

(powerpc-darwin) Add a missing newline to the prologue.

#ifdef use of EM_IA_64

Fix nuking of warnings, and nuke another one.

Un-break previous commit and nuke a few warnings

Build cbits (not necessarily RTS) with -G0 on IA64, so GHCI doesn't have
to deal with sdata sections