Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 23:24:23 +0000 (23:24 +0000)]
SPARC NCG: Also do misaligned reads
Ian Lynagh [Wed, 21 Jan 2009 14:17:06 +0000 (14:17 +0000)]
When converting TH syntax to GHC syntax, need to put sections in parentheses
Fixes trac #2956
Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 05:23:34 +0000 (05:23 +0000)]
SPARC NCG: Add a SPARC version of rts_mkInt64 that handles misaligned closure payloads.
Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 03:47:16 +0000 (03:47 +0000)]
SPARC NCG: Reenable out of line 32 bit float ops
Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 02:55:49 +0000 (02:55 +0000)]
SPARC NCG: Clean up formatting and add comments in genCCall
Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 01:26:24 +0000 (01:26 +0000)]
SPARC NCG: Fix format problem when converting float to int
Ben.Lippmeier@anu.edu.au [Wed, 21 Jan 2009 00:37:29 +0000 (00:37 +0000)]
SPARC NCG: fill branch delay slot after tabled jump (doh!)
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 21:49:14 +0000 (21:49 +0000)]
SPARC NCG: Add tabled switch
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 09:06:17 +0000 (09:06 +0000)]
SPARC NCG: Fix 64bit integers returned from ccalls
'Jose Pedro Magalhaes [Mon, 19 Jan 2009 11:23:21 +0000 (11:23 +0000)]
#2875: Correct SYB's representation of Char
'Jose Pedro Magalhaes [Tue, 9 Dec 2008 12:55:51 +0000 (12:55 +0000)]
Fix #2759: add ability to serialize Rational
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 07:51:00 +0000 (07:51 +0000)]
SPARC NCG: Fix warnings
Ben.Lippmeier@anu.edu.au [Fri, 16 Jan 2009 01:18:53 +0000 (01:18 +0000)]
SPARC NCG: Remove a comment that was confusing haddock
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 07:40:00 +0000 (07:40 +0000)]
SPARC NCG: ppr 64 bit store sizes
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 07:15:36 +0000 (07:15 +0000)]
SPARC NCG: Fix generation of 64 bit ops on 32 bit sparc
Ben.Lippmeier@anu.edu.au [Tue, 20 Jan 2009 05:21:13 +0000 (05:21 +0000)]
SPARC NCG: Add support for hardware divide
Ben.Lippmeier@anu.edu.au [Thu, 15 Jan 2009 08:41:05 +0000 (08:41 +0000)]
SPARC NCG: Redo code for integer sign extension
Ben.Lippmeier@anu.edu.au [Thu, 15 Jan 2009 08:00:04 +0000 (08:00 +0000)]
SPARC NCG: Fix signed/unsigned operand format bug
Ben.Lippmeier@anu.edu.au [Thu, 15 Jan 2009 05:57:27 +0000 (05:57 +0000)]
More fixes to the SPARC native code generator
* Fix loading of 64bit floats
* Put SRT and other read only static data in the .text segment
Ben.Lippmeier@anu.edu.au [Wed, 14 Jan 2009 05:44:16 +0000 (05:44 +0000)]
Start fixing the SPARC native code generator
* Use BlockIds in branch instructions instead of Imms.
* Assign FP values returned from C calls to the right regs
* Fix loading of F32s
* Add a SPARC version of the FreeRegs map to the linear allcator.
Ben.Lippmeier@anu.edu.au [Mon, 12 Jan 2009 06:33:10 +0000 (06:33 +0000)]
Fix some holes in the SPARC native code generator.
This makes about half the tests in codeGen/should_run work.
Ben.Lippmeier@anu.edu.au [Mon, 12 Jan 2009 05:34:21 +0000 (05:34 +0000)]
Untag closure pointers before trying to print them.
In RTS tracing code, need to untag the pointer before trying
to load the info table in printClosure()
Ben.Lippmeier@anu.edu.au [Mon, 12 Jan 2009 00:56:25 +0000 (00:56 +0000)]
Add missing documention of -Da DEBUG: apply flag to RTS help.
Ben.Lippmeier@anu.edu.au [Sat, 10 Jan 2009 02:58:02 +0000 (02:58 +0000)]
Better panic message in RegAllocLinear
Ben.Lippmeier@anu.edu.au [Sat, 10 Jan 2009 01:44:18 +0000 (01:44 +0000)]
Make the SPARC NCG compile again - it's still broken though.
Ben.Lippmeier@anu.edu.au [Mon, 5 Jan 2009 07:04:29 +0000 (07:04 +0000)]
Enable the native code generator for SPARC
Ian Lynagh [Sun, 18 Jan 2009 19:33:28 +0000 (19:33 +0000)]
Always use PTHREAD_MUTEX_ERRORCHECK to create mutexes when -DDEBUG
Linux defines PTHREAD_MUTEX_ERRORCHECK = PTHREAD_MUTEX_ERRORCHECK_NP
anyway, so there's no need to special-case it.
Ian Lynagh [Sat, 17 Jan 2009 21:59:31 +0000 (21:59 +0000)]
Use error-checking mutexes on all platforms when DEBUG is on
Otherwise ASSERT_LOCK_HELD will cause deadlocks
Ian Lynagh [Sat, 17 Jan 2009 20:09:55 +0000 (20:09 +0000)]
Reinstate: Always check the result of pthread_mutex_lock() and pthread_mutex_unlock().
Sun Jan 4 19:24:43 GMT 2009 Matthias Kilian <kili@outback.escape.de>
Don't check pthread_mutex_*lock() only on Linux and/or only if DEBUG
is defined. The return values of those functions are well defined
and should be supported on all operation systems with pthreads. The
checks are cheap enough to do them even in the default build (without
-DDEBUG).
While here, recycle an unused macro ASSERT_LOCK_NOTHELD, and let
the debugBelch part enabled with -DLOCK_DEBUG work independently
of -DDEBUG.
Ian Lynagh [Sat, 17 Jan 2009 02:39:47 +0000 (02:39 +0000)]
Initialise and deinitialise the file_lock_mutex
Ian Lynagh [Sun, 18 Jan 2009 11:06:54 +0000 (11:06 +0000)]
Create package.conf when installing a bindist
Simon Marlow [Wed, 14 Jan 2009 16:28:34 +0000 (16:28 +0000)]
validate fix on 32-bit
Simon Marlow [Wed, 14 Jan 2009 15:58:42 +0000 (15:58 +0000)]
fix validate on Windows
Simon Marlow [Fri, 16 Jan 2009 11:43:39 +0000 (11:43 +0000)]
UNDO: Always check the result of pthread_mutex_lock() and pthread_mutex_unlock().
This patch caused problems on Mac OS X, undoing until we can do it better.
rolling back:
Sun Jan 4 19:24:43 GMT 2009 Matthias Kilian <kili@outback.escape.de>
* Always check the result of pthread_mutex_lock() and pthread_mutex_unlock().
Don't check pthread_mutex_*lock() only on Linux and/or only if DEBUG
is defined. The return values of those functions are well defined
and should be supported on all operation systems with pthreads. The
checks are cheap enough to do them even in the default build (without
-DDEBUG).
While here, recycle an unused macro ASSERT_LOCK_NOTHELD, and let
the debugBelch part enabled with -DLOCK_DEBUG work independently
of -DDEBUG.
M ./includes/OSThreads.h -30 +10
Simon Marlow [Fri, 16 Jan 2009 09:53:11 +0000 (09:53 +0000)]
Update config.guess, config.sub and install.sh from automake-1.10.2
In particular, config.guess should now know about x86_64-pc-solaris2
Simon Marlow [Thu, 15 Jan 2009 12:25:24 +0000 (12:25 +0000)]
More useful error message when a package .hi file cannot be found:
> import System.Process
Could not find module `System.Process':
There are files missing in the process-1.0.1.1 package,
try running 'ghc-pkg check'.
Use -v to see a list of the files searched for.
Simon Marlow [Thu, 15 Jan 2009 12:21:43 +0000 (12:21 +0000)]
soup-up "ghc-pkg check"
We now look for missing files (including .hi files), and report all
the packages that are transitively broken.
$ ghc-pkg check
There are problems in package syb-0.1.0.0:
dependency foo-4.0.0.0 doesn't exist
There are problems in package process-1.0.1.1:
file System/Process.hi is missing
The following packages are broken, either because they have a problem
listed above, or because they depend on a broken package.
syb-0.1.0.0
process-1.0.1.1
base-3.0.3.0
Cabal-1.7.0
haskell98-1.0.1.0
haddock-2.4.2
ghc-6.11
Simon Marlow [Thu, 15 Jan 2009 09:39:22 +0000 (09:39 +0000)]
document -feager-blackholing
Tim Chevalier [Thu, 15 Jan 2009 00:26:12 +0000 (00:26 +0000)]
External Core: re-add code I removed mistakenly in last commit
Tim Chevalier [Wed, 14 Jan 2009 22:44:28 +0000 (22:44 +0000)]
External Core lib: lots of cleanup
- Factor out code for applying newtypes from Check into CoreUtils
- Use this code in Prep, which allowed for some simplification
- Change Merge and ElimDeadCode to not flatten top-level binds
- Add a flag for elimDeadCode to tell it whether to keep
exported bindings or not.
- Other things.
Tim Chevalier [Wed, 14 Jan 2009 22:17:34 +0000 (22:17 +0000)]
External Core: print out more precise dependency info
Print out the same recursive/non-recursive binding groups
that existed in internal Core in an External Core file,
rather than dumping everything into one big recursive group.
Ian Lynagh [Wed, 14 Jan 2009 20:59:45 +0000 (20:59 +0000)]
Fix "make install": Put "[]" in the install package.conf
Ian Lynagh [Wed, 14 Jan 2009 18:19:37 +0000 (18:19 +0000)]
Remove redundant import; spotted by Thorkil Naur
Ian Lynagh [Wed, 14 Jan 2009 18:12:12 +0000 (18:12 +0000)]
Remove a redundant import
Simon Marlow [Wed, 14 Jan 2009 12:15:26 +0000 (12:15 +0000)]
Detect when a C finalizer calls back to Haskell
This is illegal now, after the fix for #1364, but it turns out that
the existing check for dodgy callbacks doesn't catch finalizers
calling back, so we need another test. This will be particularly
important for 6.10.2, because the behaviour has changed.
simonpj@microsoft.com [Tue, 13 Jan 2009 17:09:48 +0000 (17:09 +0000)]
Fix Trac #2931
Fix the lexical analyser when it encounters 'x<EOF> and Template
Haskell is on.
simonpj@microsoft.com [Tue, 13 Jan 2009 16:49:53 +0000 (16:49 +0000)]
Rewrite CorePrep and improve eta expansion
This patch does two main things
a) Rewrite most of CorePrep to be much easier to understand (I hope!).
The invariants established by CorePrep are now written out, and
the code is more perspicuous. It is surpringly hard to get right,
and the old code had become quite incomprehensible.
b) Rewrite the eta-expander so that it does a bit of simplifying
on-the-fly, and thereby guarantees to maintain the CorePrep
invariants. This make it much easier to use from CorePrep, and
is a generally good thing anyway.
A couple of pieces of re-structuring:
* I moved the eta-expander and arity analysis stuff into a new
module coreSyn/CoreArity.
Max will find that the type CoreArity.EtaInfo looks strangely
familiar.
* I moved a bunch of comments from Simplify to OccurAnal; that's
why it looks as though there's a lot of lines changed in those
modules.
On the way I fixed various things
- Function arguments are eta expanded
f (map g) ===> let s = \x. map g x in f s
- Trac #2368
The result is a modest performance gain, I think mainly due
to the first of these changes:
--------------------------------------------------------------------------------
Program Size Allocs Runtime Elapsed
--------------------------------------------------------------------------------
Min -1.0% -17.4% -19.1% -46.4%
Max +0.3% +0.5% +5.4% +53.8%
Geometric Mean -0.1% -0.3% -7.0% -10.2%
simonpj@microsoft.com [Tue, 13 Jan 2009 16:48:04 +0000 (16:48 +0000)]
Rename isIdentityCoercion to isIdentityCoI; add Coercion.isIdentityCoercion
simonpj@microsoft.com [Tue, 13 Jan 2009 16:46:24 +0000 (16:46 +0000)]
Spelling in comment only
simonpj@microsoft.com [Tue, 13 Jan 2009 16:41:33 +0000 (16:41 +0000)]
Make the ASSERT more informative
simonpj@microsoft.com [Tue, 13 Jan 2009 16:40:51 +0000 (16:40 +0000)]
Export mapOL from OrdList
simonpj@microsoft.com [Tue, 13 Jan 2009 16:40:20 +0000 (16:40 +0000)]
Improve error messages slightly
simonpj@microsoft.com [Tue, 13 Jan 2009 16:29:18 +0000 (16:29 +0000)]
Do not do SpecConstr on functions that unconditionally diverge
There is no point in specialising a function that is guaranteed to
diverge, and doing so screwed up arity stuff.
See Note [Do not specialise diverging functions].
simonpj@microsoft.com [Tue, 13 Jan 2009 16:27:16 +0000 (16:27 +0000)]
Make -XTypeFamilies imply -XRelaxedPolyRec (Trac #2944)
simonpj@microsoft.com [Tue, 13 Jan 2009 15:32:17 +0000 (15:32 +0000)]
Fix Trac #2937: deserialising assoicated type definitions
The deserialiser (TcIface) for associated type definitions wasn't
taking into account that the class decl brings into scope some
type variables that scope over the data/type family declaration.
Easy to fix: the new function is TcIface.bindIfaceTyVars_AT
Matthias Kilian [Sun, 4 Jan 2009 19:24:43 +0000 (19:24 +0000)]
Always check the result of pthread_mutex_lock() and pthread_mutex_unlock().
Don't check pthread_mutex_*lock() only on Linux and/or only if DEBUG
is defined. The return values of those functions are well defined
and should be supported on all operation systems with pthreads. The
checks are cheap enough to do them even in the default build (without
-DDEBUG).
While here, recycle an unused macro ASSERT_LOCK_NOTHELD, and let
the debugBelch part enabled with -DLOCK_DEBUG work independently
of -DDEBUG.
Simon Marlow [Mon, 12 Jan 2009 12:10:42 +0000 (12:10 +0000)]
sanity checking fixes
Simon Marlow [Mon, 12 Jan 2009 12:10:24 +0000 (12:10 +0000)]
Keep the remembered sets local to each thread during parallel GC
This turns out to be quite vital for parallel programs:
- The way we discover which threads to traverse is by finding
dirty threads via the remembered sets (aka mutable lists).
- A dirty thread will be on the remembered set of the capability
that was running it, and we really want to traverse that thread's
stack using the GC thread for the capability, because it is in
that CPU's cache. If we get this wrong, we get penalised badly by
the memory system.
Previously we had per-capability mutable lists but they were
aggregated before GC and traversed by just one of the GC threads.
This resulted in very poor performance particularly for parallel
programs with deep stacks.
Now we keep per-capability remembered sets throughout GC, which also
removes a lock (recordMutableGen_sync).
Simon Marlow [Wed, 7 Jan 2009 15:14:49 +0000 (15:14 +0000)]
indicate which TSOs are dirty in the printAllThreads() output
David Waern [Fri, 9 Jan 2009 19:17:13 +0000 (19:17 +0000)]
Fix Trac #2584: Pretty printing of types with HsDocTy
The pretty printing clause for HsDocTy was wrong, causing brackets to be left
out. We now print Haddock comments on types as if they were postfix type
operators.
Ian Lynagh [Fri, 9 Jan 2009 16:04:54 +0000 (16:04 +0000)]
Add "Word size" to the +RTS --info output
Ian Lynagh [Fri, 9 Jan 2009 15:10:06 +0000 (15:10 +0000)]
Check that make supports eval
Ian Lynagh [Thu, 8 Jan 2009 13:11:01 +0000 (13:11 +0000)]
Add some more fields to +RTS --info
Simon Marlow [Fri, 9 Jan 2009 09:06:58 +0000 (09:06 +0000)]
FIX BUILD on Windows (fix for #2873 broke it)
Simon Marlow [Thu, 8 Jan 2009 15:53:41 +0000 (15:53 +0000)]
when calling mmap() with MAP_ANON, the fd argument should be -1
might fix #2925
simonpj@microsoft.com [Thu, 8 Jan 2009 12:41:18 +0000 (12:41 +0000)]
Fix Trac #2914: record wild cards and assoicated types
Simon Marlow [Thu, 8 Jan 2009 09:56:28 +0000 (09:56 +0000)]
Fix #2873: should fail if a package DB desn't exist
We allowed non-existence before because the user DB is allowed to not
exist, so now we have an explicit exception for that case.
Simon Marlow [Wed, 7 Jan 2009 14:05:07 +0000 (14:05 +0000)]
Close the races between throwTo and thread completion
Any threads we missed were being caught by the GC (possibly the idle
GC if the system was otherwise inactive), but that's not ideal. The
fix (from Bertram Felgenhauer) is to use lockTSO to synchronise,
imposing an unconditional lockTSO on thread exit. I couldn't measure
any performance overhead from doing this, so it seems reasonable.
Simon Marlow [Wed, 7 Jan 2009 12:11:42 +0000 (12:11 +0000)]
add comment
Bertram Felgenhauer [Wed, 7 Jan 2009 12:08:08 +0000 (12:08 +0000)]
Fix two more locking issues in throwTo()
Simon Marlow [Wed, 7 Jan 2009 12:07:34 +0000 (12:07 +0000)]
maybePerformBlockedException() should handle ThreadComplete/ThreadKilled
Part of the fix for #2910
Simon Marlow [Wed, 7 Jan 2009 12:06:52 +0000 (12:06 +0000)]
fix a race where the timer signal could remain turned off, leading to deadlock
Simon Marlow [Wed, 7 Jan 2009 11:20:26 +0000 (11:20 +0000)]
putMVar and takeMVar: add write_barrier() to fix race with throwTo
Simon Marlow [Tue, 6 Jan 2009 15:44:08 +0000 (15:44 +0000)]
cruft removal
Simon Marlow [Tue, 6 Jan 2009 15:32:54 +0000 (15:32 +0000)]
wake up the blocked exception queue on ThreadFinished; fixes #2910
Simon Marlow [Fri, 19 Dec 2008 11:22:11 +0000 (11:22 +0000)]
bump GHC's max stack size to 512M
To accomodate compiling very long static lists (#2002)
Tim Chevalier [Mon, 5 Jan 2009 19:27:57 +0000 (19:27 +0000)]
ext-core: change .cabal file so we can build with either GHC 6.8 or 6.10
Tim Chevalier [Mon, 5 Jan 2009 19:27:34 +0000 (19:27 +0000)]
ext-core: fix some Prep bugs
Tim Chevalier [Mon, 5 Jan 2009 19:26:45 +0000 (19:26 +0000)]
ext-core: use shorter names when combining modules
Tim Chevalier [Mon, 5 Jan 2009 19:24:34 +0000 (19:24 +0000)]
ext-core: twiddle primitive things
Ben.Lippmeier@anu.edu.au [Mon, 5 Jan 2009 03:07:58 +0000 (03:07 +0000)]
Don't pin a register for gc_thread on SPARC.
This makes the build work again.
Ian Lynagh [Sun, 4 Jan 2009 21:46:47 +0000 (21:46 +0000)]
Require HsColour by default
This should stop us ending up without HsColour'ed sources on some
platforms.
We also now tell Cabal where to find HsColour, rather than it finding
it itself.
Ian Lynagh [Sun, 4 Jan 2009 21:18:10 +0000 (21:18 +0000)]
Fix build
Ori Avtalion [Tue, 9 Dec 2008 19:42:10 +0000 (19:42 +0000)]
Add GHCi completions to :set and :show
Ian Lynagh [Sun, 4 Jan 2009 18:46:52 +0000 (18:46 +0000)]
Fix sync-all: Check for --complete/partial before --<anything>
Patch from megacz in trac #2857
Ian Lynagh [Sun, 4 Jan 2009 11:55:09 +0000 (11:55 +0000)]
Remove time from extralibs at request of maintainer
Ian Lynagh [Sat, 3 Jan 2009 15:47:54 +0000 (15:47 +0000)]
validate fix: InteractiveEval no longer needs to import IdInfo
Ian Lynagh [Sat, 3 Jan 2009 15:36:24 +0000 (15:36 +0000)]
Fix validate: strs is no longer used in IfaceSyn
Ian Lynagh [Mon, 29 Dec 2008 19:17:27 +0000 (19:17 +0000)]
Remove trailing whitespace from HaddockUtils
Ian Lynagh [Mon, 29 Dec 2008 19:16:57 +0000 (19:16 +0000)]
Fix warnings in HaddockUtils
Ian Lynagh [Mon, 29 Dec 2008 19:14:30 +0000 (19:14 +0000)]
Remove dead code from HaddockUtils
simonpj@microsoft.com [Fri, 2 Jan 2009 14:28:51 +0000 (14:28 +0000)]
Make record selectors into ordinary functions
This biggish patch addresses Trac #2670. The main effect is to make
record selectors into ordinary functions, whose unfoldings appear in
interface files, in contrast to their previous existence as magic
"implicit Ids". This means that the usual machinery of optimisation,
analysis, and inlining applies to them, which was failing before when
the selector was somewhat complicated. (Which it can be when
strictness annotations, unboxing annotations, and GADTs are involved.)
The change involves the following points
* Changes in Var.lhs to the representation of Var. Now a LocalId can
have an IdDetails as well as a GlobalId. In particular, the
information that an Id is a record selector is kept in the
IdDetails. While compiling the current module, the record selector
*must* be a LocalId, so that it participates properly in compilation
(free variables etc).
This led me to change the (hidden) representation of Var, so that there
is now only one constructor for Id, not two.
* The IdDetails is persisted into interface files, so that an
importing module can see which Ids are records selectors.
* In TcTyClDecls, we generate the record-selector bindings in renamed,
but not typechecked form. In this way, we can get the typechecker
to add all the types and so on, which is jolly helpful especially
when GADTs or type families are involved. Just like derived
instance declarations.
This is the big new chunk of 180 lines of code (much of which is
commentary). A call to the same function, mkAuxBinds, is needed in
TcInstDcls for associated types.
* The typechecker therefore has to pin the correct IdDetails on to
the record selector, when it typechecks it. There was a neat way
to do this, by adding a new sort of signature to HsBinds.Sig, namely
IdSig. This contains an Id (with the correct Name, Type, and IdDetails);
the type checker uses it as the binder for the final binding. This
worked out rather easily.
* Record selectors are no longer "implicit ids", which entails changes to
IfaceSyn.ifaceDeclSubBndrs
HscTypes.implicitTyThings
TidyPgm.getImplicitBinds
(These three functions must agree.)
* MkId.mkRecordSelectorId is deleted entirely, some 300+ lines (incl
comments) of very error prone code. Happy days.
* A TyCon no longer contains the list of record selectors:
algTcSelIds is gone
The renamer is unaffected, including the way that import and export of
record selectors is handled.
Other small things
* IfaceSyn.ifaceDeclSubBndrs had a fragile test for whether a data
constructor had a wrapper. I've replaced that with an explicit flag
in the interface file. More robust I hope.
* I renamed isIdVar to isId, which touched a few otherwise-unrelated files.
simonpj@microsoft.com [Wed, 31 Dec 2008 16:43:00 +0000 (16:43 +0000)]
Fix Trac #2721: reject newtype deriving if the class has associated types
simonpj@microsoft.com [Wed, 31 Dec 2008 15:25:17 +0000 (15:25 +0000)]
-XImpredicativeTypes implies -XRankNTypes, and improve error msg in TcMType
If you are going for impredicative types you almost certainly want RankN
too. The change to TcMType improves the error when you say
T (forall a. blah)
where T is a type synonym. This doesn't necessarily need impredicativity,
if you have LiberalTypeSynonyms.
simonpj@microsoft.com [Wed, 31 Dec 2008 14:41:51 +0000 (14:41 +0000)]
Fix Trac #2856: make deriving work for type families
Darn, but TcDeriv is complicated, when type families get in on
the act! This patch makes GeneralisedNewtypeDeriving work
properly for type families. I think.
In order to do so, I found that GeneralisedNewtypeDeriving can
work for recursive newtypes too -- and since families are conservatively
marked recursive, that's a crucial part of the fix, and useful too.
See Note [Recursive newtypes] in TcDeriv.
simonpj@microsoft.com [Wed, 31 Dec 2008 14:41:31 +0000 (14:41 +0000)]
White space and spelling in comments
simonpj@microsoft.com [Wed, 31 Dec 2008 14:40:06 +0000 (14:40 +0000)]
Remove -XImpredicativeTypes from -fglasgow-exts
See Trac #2846: impredicative types are far from stable, so
-fglasgow-exts should not imply them. Maybe we should merge
this into 6.10?
simonpj@microsoft.com [Wed, 31 Dec 2008 14:35:21 +0000 (14:35 +0000)]
Improve error reporting for 'deriving'
a) Improve the extra suggested fix when there's a "no instance"
error in a deriving clause.
b) Improve error location recording in tcInstDecl2
Many of the changes in tcInstDecl2 are simple reformatting.
simonpj@microsoft.com [Tue, 30 Dec 2008 16:59:06 +0000 (16:59 +0000)]
Improve error message in deriving (fix Trac #2851)
simonpj@microsoft.com [Tue, 30 Dec 2008 16:44:32 +0000 (16:44 +0000)]
Avoid nasty name clash with associated data types (fixes Trac #2888)
The main bug was in TcHsType; see Note [Avoid name clashes for
associated data types]. However I did a bit of re-factoring while
I was abouut it.
I'm still a but unhappy with the use of TyCon.setTyConArgPoss; it'd
be better to construct the TyCon correctly in the first place. But
that means passing an extra parameter to tcTyDecl1... maybe we should
do this.