ghc-hetmet.git
16 years agoFix build with external gmp library.
gwright@antiope.com [Mon, 13 Aug 2007 12:42:11 +0000 (12:42 +0000)]
Fix build with external gmp library.

ghc fails to build if you use an external gmp library. This is because
ghc requires the header file gmp.h, which used to be provided by the
internal gmp source code.  The file gmp.h is no longer part of the
gmp source code, but is generated as part of the build procedure.
If an external gmp is specified, the internal gmp is not build and the
gmp.h file never gets generated.

Of course, it was a bad idea anyway to use a header file from a potentially
different version of the library.

The patch sets HAVE_LIB_GMP if the gmp library is found during configuration
and conditionalizes including the library header file on it.

16 years agoUpdate config.sub and config.guess
Ian Lynagh [Tue, 14 Aug 2007 14:30:32 +0000 (14:30 +0000)]
Update config.sub and config.guess

16 years agoUse lookupOrig to find built-in NDP-related names
Roman Leshchinskiy [Wed, 15 Aug 2007 03:06:05 +0000 (03:06 +0000)]
Use lookupOrig to find built-in NDP-related names

16 years agoImplicitly import NDP stuff when vectorising
Roman Leshchinskiy [Tue, 7 Aug 2007 04:39:07 +0000 (04:39 +0000)]
Implicitly import NDP stuff when vectorising

We add the following declaration to all vectorised modules:

  import qualified <NDP interface module> as :NDP

This allows us to refer to NDP stuff by RdrName later on but does not expose
NDP names to the user's code.

16 years agofix typo in ghci.xml
Bertram Felgenhauer [Sun, 12 Aug 2007 13:22:01 +0000 (13:22 +0000)]
fix typo in ghci.xml

16 years agoAdd bytestring to core-packages
Ian Lynagh [Sun, 12 Aug 2007 16:36:45 +0000 (16:36 +0000)]
Add bytestring to core-packages

16 years agoFix installation (we need to use the right ghc-pkg)
Ian Lynagh [Sun, 12 Aug 2007 12:48:11 +0000 (12:48 +0000)]
Fix installation (we need to use the right ghc-pkg)

16 years agoGet closer to GhcCompilerWays=p working
Ian Lynagh [Sun, 12 Aug 2007 12:27:34 +0000 (12:27 +0000)]
Get closer to GhcCompilerWays=p working
We also now have GhcThreaded rather than GhcNotThreaded.

16 years agoAdd a type sig
Ian Lynagh [Sat, 11 Aug 2007 00:34:24 +0000 (00:34 +0000)]
Add a type sig

16 years agoSplit off packedstring and array packages
Ian Lynagh [Wed, 1 Aug 2007 23:53:18 +0000 (23:53 +0000)]
Split off packedstring and array packages

16 years agoAdd a containers package
Ian Lynagh [Wed, 1 Aug 2007 22:39:45 +0000 (22:39 +0000)]
Add a containers package

16 years agoTweak package build order
Ian Lynagh [Sun, 29 Jul 2007 21:51:41 +0000 (21:51 +0000)]
Tweak package build order

16 years agobytestring is now in its own package
Ian Lynagh [Sun, 29 Jul 2007 11:53:47 +0000 (11:53 +0000)]
bytestring is now in its own package

16 years agomore cmpFS refactoring
Isaac Dupree [Tue, 7 Aug 2007 00:12:48 +0000 (00:12 +0000)]
more cmpFS refactoring

16 years agouse 'compare' when using the law of trichotomy
Isaac Dupree [Mon, 6 Aug 2007 23:52:43 +0000 (23:52 +0000)]
use 'compare' when using the law of trichotomy

16 years agouse {-# UNPACK #-} !Int rather than explicit Int# (#1405)
Isaac Dupree [Mon, 6 Aug 2007 23:44:28 +0000 (23:44 +0000)]
use {-# UNPACK #-} !Int rather than explicit Int# (#1405)

16 years agoRemove a duplicate case
Ian Lynagh [Fri, 10 Aug 2007 11:57:44 +0000 (11:57 +0000)]
Remove a duplicate case

16 years agoChange standalone deriving syntax and semantics; fixes trac #1481
Ian Lynagh [Fri, 10 Aug 2007 22:37:28 +0000 (22:37 +0000)]
Change standalone deriving syntax and semantics; fixes trac #1481
You now say
    deriving instance Cxt => Head

16 years agoProperly guard imports because they have to be precise on Windows and Darwin sets...
Clemens Fruhwirth [Fri, 10 Aug 2007 15:10:45 +0000 (15:10 +0000)]
Properly guard imports because they have to be precise on Windows and Darwin sets __PIC__ automatically

16 years agoAdd explicit imports for RTS-external variables
Clemens Fruhwirth [Mon, 6 Aug 2007 08:49:38 +0000 (08:49 +0000)]
Add explicit imports for RTS-external variables

16 years agoImprove error message (part of Trac #1606)
simonpj@microsoft.com [Fri, 10 Aug 2007 08:33:12 +0000 (08:33 +0000)]
Improve error message (part of Trac #1606)

16 years agoPrint fewer parens when pretty-printing HsSyn
simonpj@microsoft.com [Thu, 9 Aug 2007 22:33:29 +0000 (22:33 +0000)]
Print fewer parens when pretty-printing HsSyn

16 years agoPrint more "..." in long lists of declarations
simonpj@microsoft.com [Thu, 9 Aug 2007 22:32:28 +0000 (22:32 +0000)]
Print more "..." in long lists of declarations

16 years agoBuild RTS as dynamic library
Clemens Fruhwirth [Wed, 8 Aug 2007 06:50:43 +0000 (06:50 +0000)]
Build RTS as dynamic library

16 years agoIntroduce new class for external symbols in Linker.c that use __imp__<sym> instead...
Clemens Fruhwirth [Mon, 6 Aug 2007 08:44:49 +0000 (08:44 +0000)]
Introduce new class for external symbols in Linker.c that use __imp__<sym> instead of &<sym>

16 years agoMake the new functions accessible from the vectorisation monad
Roman Leshchinskiy [Wed, 8 Aug 2007 06:56:21 +0000 (06:56 +0000)]
Make the new functions accessible from the vectorisation monad

16 years agoAdd more NDP functions to PrelNames
Roman Leshchinskiy [Wed, 8 Aug 2007 06:55:59 +0000 (06:55 +0000)]
Add more NDP functions to PrelNames

16 years agoPA functions are no longer methods
Roman Leshchinskiy [Wed, 8 Aug 2007 06:48:34 +0000 (06:48 +0000)]
PA functions are no longer methods

16 years agoVectorise Case on products
Roman Leshchinskiy [Wed, 8 Aug 2007 05:04:33 +0000 (05:04 +0000)]
Vectorise Case on products

16 years agoUtility functions for accessing parallel array representations
Roman Leshchinskiy [Wed, 8 Aug 2007 04:10:32 +0000 (04:10 +0000)]
Utility functions for accessing parallel array representations

16 years agoFix bug in vectorised DataCon worker generation
Roman Leshchinskiy [Wed, 8 Aug 2007 02:02:57 +0000 (02:02 +0000)]
Fix bug in vectorised DataCon worker generation

16 years agoUse dataConTag in flattened array representation
Roman Leshchinskiy [Wed, 8 Aug 2007 01:49:16 +0000 (01:49 +0000)]
Use dataConTag in flattened array representation

16 years agoAdd built-in PA dictionary for closures
Roman Leshchinskiy [Wed, 8 Aug 2007 01:48:59 +0000 (01:48 +0000)]
Add built-in PA dictionary for closures

16 years agoModify PA dictionary generation to the new record-based scheme
Roman Leshchinskiy [Tue, 7 Aug 2007 06:21:01 +0000 (06:21 +0000)]
Modify PA dictionary generation to the new record-based scheme

16 years agoChange DataCon worker vectorisation to use PA records
Roman Leshchinskiy [Tue, 7 Aug 2007 06:05:53 +0000 (06:05 +0000)]
Change DataCon worker vectorisation to use PA records

16 years agoAdd built-in PA dictionaries for boxed tuples
Roman Leshchinskiy [Tue, 7 Aug 2007 06:05:24 +0000 (06:05 +0000)]
Add built-in PA dictionaries for boxed tuples

16 years agoTrace more vectorisation failures
Roman Leshchinskiy [Tue, 7 Aug 2007 05:31:45 +0000 (05:31 +0000)]
Trace more vectorisation failures

16 years agoMove code
Roman Leshchinskiy [Tue, 7 Aug 2007 05:24:29 +0000 (05:24 +0000)]
Move code

16 years agoRemove dead code
Roman Leshchinskiy [Tue, 7 Aug 2007 05:21:36 +0000 (05:21 +0000)]
Remove dead code

16 years agoAssociate vectorised tycons with their PA dfuns
Roman Leshchinskiy [Tue, 7 Aug 2007 05:20:32 +0000 (05:20 +0000)]
Associate vectorised tycons with their PA dfuns

16 years agoPA is now an explicit record instead of a typeclass
Roman Leshchinskiy [Tue, 7 Aug 2007 05:10:52 +0000 (05:10 +0000)]
PA is now an explicit record instead of a typeclass

16 years agoImplicitly import NDP stuff when vectorising
Roman Leshchinskiy [Tue, 7 Aug 2007 04:39:07 +0000 (04:39 +0000)]
Implicitly import NDP stuff when vectorising

We add the following declaration to all vectorised modules:

  import qualified <NDP interface module> as :NDP

This allows us to refer to NDP stuff by RdrName later on but does not expose
NDP names to the user's code.

16 years agoSupport for using built-in PA dictionaries for some types
Roman Leshchinskiy [Tue, 7 Aug 2007 02:31:54 +0000 (02:31 +0000)]
Support for using built-in PA dictionaries for some types

16 years agoUtility functions for accessing the RdrEnv during vectorisation
Roman Leshchinskiy [Tue, 7 Aug 2007 02:28:18 +0000 (02:28 +0000)]
Utility functions for accessing the RdrEnv during vectorisation

16 years agoAdd RdrEnv to vectorisation state
Roman Leshchinskiy [Tue, 7 Aug 2007 02:27:18 +0000 (02:27 +0000)]
Add RdrEnv to vectorisation state

16 years agoAdd utility functions for accessing PA info during vectorisation
Roman Leshchinskiy [Fri, 3 Aug 2007 03:02:37 +0000 (03:02 +0000)]
Add utility functions for accessing PA info during vectorisation

16 years agoMake sure PA dfuns are keyed on the vectorised tycon in VectInfo
Roman Leshchinskiy [Fri, 3 Aug 2007 03:00:31 +0000 (03:00 +0000)]
Make sure PA dfuns are keyed on the vectorised tycon in VectInfo

16 years agoAdd PA dfuns to VectMonad state
Roman Leshchinskiy [Fri, 3 Aug 2007 02:50:14 +0000 (02:50 +0000)]
Add PA dfuns to VectMonad state

16 years agoFix grouping by module in the mi_exports, for indexed data type families
simonpj@microsoft.com [Thu, 9 Aug 2007 16:47:46 +0000 (16:47 +0000)]
Fix grouping by module in the mi_exports, for indexed data type families

This is a little tricky.  See Note [Original module] in MkIface.

16 years agoTrim imports
simonpj@microsoft.com [Thu, 9 Aug 2007 15:37:55 +0000 (15:37 +0000)]
Trim imports

16 years agoRename a constructor CmmForeignCall to CmmCallee, and tidy Cmm code
simonpj@microsoft.com [Thu, 9 Aug 2007 15:37:37 +0000 (15:37 +0000)]
Rename a constructor CmmForeignCall to CmmCallee, and tidy Cmm code

This patch should have no effect; it's mainly comments, layout,
plus this contructor name change.

16 years agoTidy up the treatment of newtypes, refactor, and fix Trac #736
simonpj@microsoft.com [Thu, 9 Aug 2007 15:34:37 +0000 (15:34 +0000)]
Tidy up the treatment of newtypes, refactor, and fix Trac #736

I've forgotten the precise details already, but this patch
significantly refactors the way newtypes are handled, fixes
the foreign-export problem Trac #736 (which concerned newtypes),
and gets rid of a bogus unsafeCoerce in the foreign export
desugaring.

16 years agoImprovd documentation for overlapping instances
simonpj@microsoft.com [Thu, 9 Aug 2007 15:18:39 +0000 (15:18 +0000)]
Improvd documentation for overlapping instances

16 years agoBetter scoring for loop breakers; fixes simplifier loop in nofib/minimax
simonpj@microsoft.com [Thu, 9 Aug 2007 15:14:57 +0000 (15:14 +0000)]
Better scoring for loop breakers; fixes simplifier loop in nofib/minimax

See Note [Inline candidates] in OccurAnal.  We were getting a recursive
loop exposed, which led to infinite inlinings.  Doesn't bite much, but
was obviously wrong.

I've change the "scoring order" for loop breakers, which could possibly
have a performance impact on other programs.  A full nofib run exposed
a 0.00% change in allocation in any nofib program, so I don't think it's
likely, but keep an eye out.

16 years agoRemove dead code
simonpj@microsoft.com [Fri, 3 Aug 2007 09:31:07 +0000 (09:31 +0000)]
Remove dead code

16 years agoCabal depends on -package unix when GHC <= 6.2
Simon Marlow [Thu, 9 Aug 2007 10:01:39 +0000 (10:01 +0000)]
Cabal depends on -package unix when GHC <= 6.2

16 years agojoinToTargets to emit fixup code even when movement graph contains cycles
Clemens Fruhwirth [Sat, 14 Jul 2007 08:32:41 +0000 (08:32 +0000)]
joinToTargets to emit fixup code even when movement graph contains cycles

First, cycles can only start of with registers and their destination
must involve a register location. This is because memory locations are
allocated exclusively for a virtual register and hence can never cause
a conflict in the assignment, hence need no fixup code. Therefore, we
only have to deal with InReg -> InReg, or InReg -> InReg/InMem
movements.

The strategy is to take the first cycle element, which is guaranteed
to start with a register, spill it to a fresh memory location, compute
the fixup for the rest, and restore from the spill slot to its
destinations. The "rest" will degenerate into an acyclic scc, so we do
not need take care of the empty list case in CyclicScc.
 ***END OF DESCRIPTION***

Place the long patch description above the ***END OF DESCRIPTION*** marker.
The first line of this file will be the patch name.

This patch contains the following changes:

M ./compiler/nativeGen/RegisterAlloc.hs -6 +27

16 years agoMake SpecConstr specialise for constant arguments again
simonpj@microsoft.com [Sun, 5 Aug 2007 22:03:35 +0000 (22:03 +0000)]
Make SpecConstr specialise for constant arguments again

Consider
  lvl = Just True

  foo :: Maybe Bool -> Int -> Int
  foo (Just True) i = i
  foo _           i = foo lvl i

SpecConstr should specialise foo, but it wasn't doing so (spotted
by Roman).

Reason: lvl's unfolding wasn't in the cloned version of lvl.
Solution: extend the value environment to record top-level bindings too

At the same time I made it work if 'lvl' is a lambda, in which case it
is again worth specialisg.  This meant renaming ConEnv to ValueEnv,
and adding a case for 'LambdaVal'.

(To make specialisation on lambdas work properly, we have to do lambda
lifting as well, but this gets part of the way, and fixes a bug too.)

16 years agoComments only: explain why StgConApp is still useful
simonpj@microsoft.com [Sun, 5 Aug 2007 21:44:04 +0000 (21:44 +0000)]
Comments only: explain why StgConApp is still useful

16 years agoAdd missing case to sizeofPrimRep
simonpj@microsoft.com [Sat, 4 Aug 2007 17:47:59 +0000 (17:47 +0000)]
Add missing case to sizeofPrimRep

16 years agoFix Trac #1037
simonpj@microsoft.com [Sat, 4 Aug 2007 17:44:20 +0000 (17:44 +0000)]
Fix Trac #1037

It makes *sense* for a foreign import to have a zero-sized return, thus

  foreign import ccall foo :: Int -> State# RealWorld

but it's not clear that it's useful, and it requires some back-end (a
Hint for void types) to make it go right through. It's not clear that
we really want this facility, so rather than fixing the code
generator, I'm just making the construct illegal for now.

16 years agoIn GHCi, filter instances by what is in scope, not just by what is in scope unqualified
simonpj@microsoft.com [Sat, 4 Aug 2007 17:35:39 +0000 (17:35 +0000)]
In GHCi, filter instances by what is in scope, not just by what is in scope unqualified

Trac #1581 was doing too much filtering; it even filtered out intances
defined in this very module!  The new rule shows more instances, but
hopefully not to many.

Furthermore I have moved the filtering out of TcRnDriver (where it does
not belong) to InteractiveEval. And I've added a note to the documentation.

16 years agoImprove pretty-printing of 'foreign' declarations
simonpj@microsoft.com [Sat, 4 Aug 2007 17:21:51 +0000 (17:21 +0000)]
Improve pretty-printing of 'foreign' declarations

16 years agoFIX read030 (error message wibble)
Simon Marlow [Thu, 9 Aug 2007 08:58:17 +0000 (08:58 +0000)]
FIX read030 (error message wibble)

16 years agoFix the use of emptyCopyFlags
Ian Lynagh [Wed, 8 Aug 2007 23:16:20 +0000 (23:16 +0000)]
Fix the use of emptyCopyFlags

16 years agoRemoved a dependency on Cabal that caused a build failure when Cabal changed
Magnus Jonsson [Wed, 8 Aug 2007 22:10:29 +0000 (22:10 +0000)]
Removed a dependency on Cabal that caused a build failure when Cabal changed

16 years agoAdd blurb in the user guide re stdout buffering differing between GHC and GHCi
Ian Lynagh [Tue, 7 Aug 2007 19:31:39 +0000 (19:31 +0000)]
Add blurb in the user guide re stdout buffering differing between GHC and GHCi

16 years agoUse emptyRegisterFlags and emptyCopyFlags for reduced fragility
Duncan Coutts [Mon, 6 Aug 2007 15:28:24 +0000 (15:28 +0000)]
Use emptyRegisterFlags and emptyCopyFlags for reduced fragility
So we will inherit default values for any new flags that get added,
so things should not break if new flags are added.

16 years agoremove gratuitous usage of Double in favor of Rational
Isaac Dupree [Tue, 7 Aug 2007 12:22:34 +0000 (12:22 +0000)]
remove gratuitous usage of Double in favor of Rational

16 years agoWarning police: eliminate all defaulting within stage1
Isaac Dupree [Tue, 7 Aug 2007 12:14:54 +0000 (12:14 +0000)]
Warning police: eliminate all defaulting within stage1
Defaulting makes compilation of multiple modules more complicated (re: #1405)
Although it was all locally within functions, not because of the module
monomorphism-restriction... but it's better to be clear what's meant, anyway.
I changed some that were defaulting to Integer, to explicit Int, where Int
seemed appropriate rather than Integer.

16 years agoNOINLINE the global FastString.string_table
Isaac Dupree [Tue, 7 Aug 2007 00:28:59 +0000 (00:28 +0000)]
NOINLINE the global FastString.string_table

16 years agocomment FastMutInt possibilities
Isaac Dupree [Mon, 6 Aug 2007 21:19:14 +0000 (21:19 +0000)]
comment FastMutInt possibilities

16 years agoFollow Cabal changes
Ian Lynagh [Sun, 5 Aug 2007 17:51:16 +0000 (17:51 +0000)]
Follow Cabal changes

16 years agoAdd a --supported-languages flag
Ian Lynagh [Sun, 5 Aug 2007 15:33:41 +0000 (15:33 +0000)]
Add a --supported-languages flag
Print the list of strings that are accepted in a LANGUAGE pragma,
or as a -XFoo flag. (No can be prepended to any of the strings).

16 years agoWhitespace
Ian Lynagh [Sun, 5 Aug 2007 15:03:02 +0000 (15:03 +0000)]
Whitespace

16 years agoAdd Distribution.System to compat
Ian Lynagh [Sun, 5 Aug 2007 13:28:56 +0000 (13:28 +0000)]
Add Distribution.System to compat

16 years agoCanonicalise flags in error message
Ian Lynagh [Sat, 4 Aug 2007 21:14:15 +0000 (21:14 +0000)]
Canonicalise flags in error message

16 years agoCanonicalise the flag format in the user guide
Ian Lynagh [Sat, 4 Aug 2007 21:01:45 +0000 (21:01 +0000)]
Canonicalise the flag format in the user guide

16 years agoWe now need to pass a version to extensionsToGHCFlag
Ian Lynagh [Sat, 4 Aug 2007 19:24:31 +0000 (19:24 +0000)]
We now need to pass a version to extensionsToGHCFlag

16 years agoFix -fallow* flags
Ian Lynagh [Sat, 4 Aug 2007 16:49:30 +0000 (16:49 +0000)]
Fix -fallow* flags

16 years agoAdd more of Cabal to compat
Ian Lynagh [Sat, 4 Aug 2007 16:22:41 +0000 (16:22 +0000)]
Add more of Cabal to compat

16 years agoRename Opt_Allow* to Opt_* to match the language names
Ian Lynagh [Sat, 4 Aug 2007 15:54:25 +0000 (15:54 +0000)]
Rename Opt_Allow* to Opt_* to match the language names

16 years agoRename Opt_TH to Opt_TemplateHaskell to match the language name
Ian Lynagh [Sat, 4 Aug 2007 15:48:05 +0000 (15:48 +0000)]
Rename Opt_TH to Opt_TemplateHaskell to match the language name

16 years agoAdd -fparr deprecated flag, and change -XParr to -XPArr
Ian Lynagh [Sat, 4 Aug 2007 15:45:05 +0000 (15:45 +0000)]
Add -fparr deprecated flag, and change -XParr to -XPArr
-fparr did exist before, it just wasn't documented.

16 years agoRename Opt_FFI to Opt_ForeignFunctionInterface to match the language name
Ian Lynagh [Sat, 4 Aug 2007 15:42:34 +0000 (15:42 +0000)]
Rename Opt_FFI to Opt_ForeignFunctionInterface to match the language name

16 years agoRemove many of the new flag variants permitted
Ian Lynagh [Sat, 4 Aug 2007 15:34:12 +0000 (15:34 +0000)]
Remove many of the new flag variants permitted
Now we only allow -XFooBar syntax, not alternate case, hyphens or -f.
There are some deprecated -f flags accordingly.

16 years agoAdd a comment about when maybePrefixMatch can be removed
Ian Lynagh [Sat, 4 Aug 2007 15:33:33 +0000 (15:33 +0000)]
Add a comment about when maybePrefixMatch can be removed

16 years agoTypo in a comment
Ian Lynagh [Sat, 4 Aug 2007 02:01:56 +0000 (02:01 +0000)]
Typo in a comment

16 years agoUpdate the boring file
Ian Lynagh [Fri, 3 Aug 2007 18:38:35 +0000 (18:38 +0000)]
Update the boring file

16 years agoCatch exceptions on Windows, to stop it popping up dialog boxes
Ian Lynagh [Fri, 3 Aug 2007 18:32:29 +0000 (18:32 +0000)]
Catch exceptions on Windows, to stop it popping up dialog boxes
Adaptated from code from Sigbjorn Finne

16 years agoFIX -ddump-cmm by temporarily disabling the gc_target part (#1582)
Simon Marlow [Fri, 3 Aug 2007 10:21:43 +0000 (10:21 +0000)]
FIX -ddump-cmm by temporarily disabling the gc_target part (#1582)

16 years agomkLiveness: when calculating the size of a parameter, round up not down.
Simon Marlow [Fri, 3 Aug 2007 09:32:29 +0000 (09:32 +0000)]
mkLiveness: when calculating the size of a parameter, round up not down.
Fixes read002 on x86_64 (and maybe others).  The stg_ap_f_info info
table had the wrong liveness on it, because the float had been treated
as having zero size.

16 years agoAdd PA dfuns to VectInfo
Roman Leshchinskiy [Fri, 3 Aug 2007 02:09:36 +0000 (02:09 +0000)]
Add PA dfuns to VectInfo

16 years agoAdd a --ignore-failure flag to push-all script
Ian Lynagh [Thu, 2 Aug 2007 21:33:00 +0000 (21:33 +0000)]
Add a --ignore-failure flag to push-all script

16 years agoDump renamer output with -ddump-rn for command-line entry in ghci
simonpj@microsoft.com [Thu, 2 Aug 2007 16:51:10 +0000 (16:51 +0000)]
Dump renamer output with -ddump-rn for command-line entry in ghci

16 years agoFix Trac #1525
simonpj@microsoft.com [Thu, 2 Aug 2007 11:12:04 +0000 (11:12 +0000)]
Fix Trac #1525

A Name used to have a Parent, but no longer has.  When we want to
print info about data type T and data constructor MkT, the info about
T already describes MkT so we want to discard the latter.  Now that
Names don't have a Parent, we must do that in a different way,
using implicitTyThings

Test is ghci011

16 years agoWibbble to last fix to generalisation (again!); fixes Trac #1575
simonpj@microsoft.com [Thu, 2 Aug 2007 08:51:36 +0000 (08:51 +0000)]
Wibbble to last fix to generalisation (again!); fixes Trac #1575

16 years agoSmall error-message change
simonpj@microsoft.com [Wed, 1 Aug 2007 14:13:39 +0000 (14:13 +0000)]
Small error-message change

16 years agoVectorise DataCon workers
Roman Leshchinskiy [Thu, 2 Aug 2007 04:37:12 +0000 (04:37 +0000)]
Vectorise DataCon workers

16 years agoInclude original tycon in PAInstance
Roman Leshchinskiy [Thu, 2 Aug 2007 04:29:38 +0000 (04:29 +0000)]
Include original tycon in PAInstance