simonpj@microsoft.com [Tue, 19 Oct 2010 09:01:00 +0000 (09:01 +0000)]
Clean up the debugger code
In particular there is much less fiddly skolemisation now
Things are not *quite* right (break001 and 006 still fail),
but they are *much* better than before.
simonpj@microsoft.com [Tue, 19 Oct 2010 08:56:09 +0000 (08:56 +0000)]
Add new VarEnv functions minusVarEnv, intersectsVarEnv, unionInScope
dimitris@microsoft.com [Mon, 18 Oct 2010 15:15:10 +0000 (15:15 +0000)]
Major pass through type checker:(1) prioritizing equalities, (2) improved Derived mechanism, (3) bugfixes
dimitris@microsoft.com [Fri, 15 Oct 2010 16:44:21 +0000 (16:44 +0000)]
(1) Caching FD improvements for efficiency, (2) preventing cascading deriveds from entering the inert, (3) Fixing bugs in the creation of FlexiTcS variables
dimitris@microsoft.com [Thu, 14 Oct 2010 14:38:11 +0000 (14:38 +0000)]
Midstream changes for performance improvement related to superclasses and functional dependencies.
dimitris@microsoft.com [Tue, 12 Oct 2010 08:09:51 +0000 (08:09 +0000)]
Minor
dimitris@microsoft.com [Mon, 11 Oct 2010 14:22:35 +0000 (14:22 +0000)]
Commentary changes
dimitris@microsoft.com [Fri, 8 Oct 2010 17:37:00 +0000 (17:37 +0000)]
Kind checking bugfix (#4356) and preventing wanteds from rewriting wanteds
Ian Lynagh [Tue, 19 Oct 2010 13:27:27 +0000 (13:27 +0000)]
Fix a retainer profiling segfault
The bitmap type wasn't big enough to hold large bitmaps on 64 bit
platforms. Profiling GHC was segfaulting when retainStack was handling a
size 33 bitmap.
Ian Lynagh [Mon, 18 Oct 2010 15:39:57 +0000 (15:39 +0000)]
Fix -auto-all: Add SCCs to IDs which have a monotype too
simonpj@microsoft.com [Mon, 18 Oct 2010 13:57:46 +0000 (13:57 +0000)]
Define SpecConstrAnnotation in GHC.Exts, and import it from there
Reason: avoid having to link the entire ghc package in modules
that use compile-time annotations:
import GHC.Exts( SpecConstrAnnotation )
{-# ANN type T ForceSpecConstr #-}
It's a kind of bug that the package exporting SpecConstrAnnotation
is linked even though it is only needed at compile time, but putting
the data type declaration in GHC.Exts is a simple way to sidestep
the problem
See See Note [SpecConstrAnnotation] in SpecConstr
David Terei [Thu, 7 Oct 2010 14:35:59 +0000 (14:35 +0000)]
Fix warnings in AsmCodeGen
David Terei [Mon, 4 Oct 2010 15:38:43 +0000 (15:38 +0000)]
LLVM: Fix compilation of writebarrier, #4308
Ian Lynagh [Sun, 17 Oct 2010 12:23:52 +0000 (12:23 +0000)]
Change how the OS X installer's create-links finds the versin number
It now gets created by configure, rather than trying to work out the
version number at runtime.
Ian Lynagh [Sun, 17 Oct 2010 11:26:48 +0000 (11:26 +0000)]
Add more quoting to distrib/MacOS/installer-scripts/create-links
Evan Laforge [Wed, 29 Sep 2010 23:45:38 +0000 (23:45 +0000)]
change os x installer to allow multiple installed versions
This puts the ghc version into the package name so they are considered separate
packages.
Ian Lynagh [Sat, 16 Oct 2010 18:00:31 +0000 (18:00 +0000)]
Only put the boot packages in the haddock contents/index
We don't install dph etc, so don't put them in the doc index.
Ian Lynagh [Sat, 16 Oct 2010 12:34:21 +0000 (12:34 +0000)]
Correct the regexp used to search for extra packages
We weren't ignoring comment lines
pho@cielonegro.org [Mon, 27 Sep 2010 22:41:45 +0000 (22:41 +0000)]
New member "archiveMemberName" for struct _ObjectCode
struct _ObjectCode should be able to retain the name of archive members.
Though currently the only use of those names are for debugging outputs.
Ian Lynagh [Fri, 15 Oct 2010 17:27:46 +0000 (17:27 +0000)]
Add a -fghci-sandbox flag so that we can en/disable the ghci sandbox
It's on by default (which matches the previous behaviour).
Motivation:
GLUT on OS X needs to run on the main thread. If you
try to use it from another thread then you just get a
white rectangle rendered. For this, or anything else
with such restrictions, you can turn the GHCi sandbox off
and things will be run in the main thread.
Ian Lynagh [Fri, 15 Oct 2010 16:45:49 +0000 (16:45 +0000)]
Fix boot; it was failing if darcs-all or validate were missing
(which is the case in sdists)
simonpj@microsoft.com [Fri, 15 Oct 2010 13:19:24 +0000 (13:19 +0000)]
Comments and layout
simonpj@microsoft.com [Fri, 15 Oct 2010 13:18:57 +0000 (13:18 +0000)]
Make (Located a) an instance of Eq, Ord
Fulfils Trac #4369
simonpj@microsoft.com [Fri, 15 Oct 2010 13:18:14 +0000 (13:18 +0000)]
Give user-defined rules precedence over built-in rules
This fixes Trac #4397. See comments with 'isMoreSpecific'.
simonpj@microsoft.com [Fri, 15 Oct 2010 13:08:18 +0000 (13:08 +0000)]
Fix Trac #4401: meta-tyvars allocated by the constraint solver are always touchable
See Note [Touchable meta type variables] in TcSMonad
simonpj@microsoft.com [Wed, 13 Oct 2010 09:11:07 +0000 (09:11 +0000)]
Remove GHC.extendGlobalRdrScope, GHC.extendGlobalTypeScope
These functions were added by
Tue Apr 18 03:36:06 BST 2006 Lemmih <lemmih@gmail.com>
* Make the initial rdr and type scope available in the ghc-api
The are extremely dubious, because they extend the Rdr and Type
env for every compilation. The right thing to do is to use
the InteractiveContext for temporary extensions.
So far as we know, no one uses them. And if they are being used
it's probably a mistake. So we're backing them out.
Simon Marlow [Fri, 15 Oct 2010 09:49:25 +0000 (09:49 +0000)]
InlinePrag needs an arity only for INLINE, not INLINABLE
This doesn't fix anything (we think), but it's morally correct.
Simon Marlow [Fri, 15 Oct 2010 09:48:36 +0000 (09:48 +0000)]
Fix #4346 (INLINABLE pragma not behaving consistently)
Debugged thanks to lots of help from Simon PJ: we weren't updating the
UnfoldingGuidance when the unfolding changed.
Also, a bit of refactoring and additinoal comments.
Ian Lynagh [Thu, 14 Oct 2010 14:05:56 +0000 (14:05 +0000)]
Have boot check that we have the dph packages when validating
Simon Marlow [Thu, 14 Oct 2010 08:42:53 +0000 (08:42 +0000)]
Add more documentation for interruptible foreign calls
Simon Marlow [Sun, 26 Sep 2010 10:58:19 +0000 (10:58 +0000)]
minor refactoring
Simon Marlow [Sat, 25 Sep 2010 19:34:42 +0000 (19:34 +0000)]
Fix for interruptible FFI handling
Set tso->why_blocked before calling maybePerformBlockedException(), so
that throwToSingleThreaded() doesn't try to unblock the current thread
(it is already unblocked).
Simon Marlow [Sat, 25 Sep 2010 19:33:17 +0000 (19:33 +0000)]
interruptible FFI: more robust handling of the exception case in the interpreter
Edward Z. Yang [Sat, 25 Sep 2010 03:30:26 +0000 (03:30 +0000)]
Don't interrupt when task blocks exceptions, don't immediately start exception.
Edward Z. Yang [Sun, 19 Sep 2010 00:29:05 +0000 (00:29 +0000)]
Interruptible FFI calls with pthread_kill and CancelSynchronousIO. v4
This is patch that adds support for interruptible FFI calls in the form
of a new foreign import keyword 'interruptible', which can be used
instead of 'safe' or 'unsafe'. Interruptible FFI calls act like safe
FFI calls, except that the worker thread they run on may be interrupted.
Internally, it replaces BlockedOnCCall_NoUnblockEx with
BlockedOnCCall_Interruptible, and changes the behavior of the RTS
to not modify the TSO_ flags on the event of an FFI call from
a thread that was interruptible. It also modifies the bytecode
format for foreign call, adding an extra Word16 to indicate
interruptibility.
The semantics of interruption vary from platform to platform, but the
intent is that any blocking system calls are aborted with an error code.
This is most useful for making function calls to system library
functions that support interrupting. There is no support for pre-Vista
Windows.
There is a partner testsuite patch which adds several tests for this
functionality.
Ian Lynagh [Wed, 13 Oct 2010 19:43:56 +0000 (19:43 +0000)]
Remove ghc-pkg's dependency on haskell98
Ian Lynagh [Wed, 13 Oct 2010 18:27:59 +0000 (18:27 +0000)]
Build haskell98 and haskell2010 with stage2
Stops us accidentally depending on them
Ian Lynagh [Wed, 13 Oct 2010 14:12:24 +0000 (14:12 +0000)]
Fix warning: Remove unused import
benl@ouroborus.net [Wed, 13 Oct 2010 04:03:35 +0000 (04:03 +0000)]
Fix warnings
benl@ouroborus.net [Wed, 13 Oct 2010 01:54:14 +0000 (01:54 +0000)]
RegAlloc: Track slot liveness over jumps in spill cleaner
Ian Lynagh [Tue, 12 Oct 2010 15:45:28 +0000 (15:45 +0000)]
Bump Cabal dep
Ian Lynagh [Tue, 12 Oct 2010 13:47:00 +0000 (13:47 +0000)]
Remove __HASKELL1__, __HASKELL98__, __CONCURRENT_HASKELL__
We used to define these CPP symbols, but nothing on hackage uses them
and the first 2 are no longer correct (as we support multiple Haskell
versions).
Ian Lynagh [Tue, 12 Oct 2010 13:05:38 +0000 (13:05 +0000)]
Follow Cabal changes: Cabal no longer has a docbook userguide
For now we don't build the Cabal userguide, but we should add markdown
support so that we can do so.
Ian Lynagh [Tue, 12 Oct 2010 11:21:11 +0000 (11:21 +0000)]
Fix build on Windows: ghc-pkg/Main.hs needs ForeignFunctionInterface
Ian Lynagh [Sun, 10 Oct 2010 22:22:31 +0000 (22:22 +0000)]
Remove unnecessary import
Ian Lynagh [Thu, 7 Oct 2010 00:43:27 +0000 (00:43 +0000)]
Make "./validate --slow" run the full testsuite
Ian Lynagh [Wed, 6 Oct 2010 16:06:56 +0000 (16:06 +0000)]
Fix build following haskell98 and -fglasgow-exts changes
Ian Lynagh [Wed, 6 Oct 2010 13:02:35 +0000 (13:02 +0000)]
Don't automatically link the haskell98 package
The default language is now Haskell2010, so this was a little odd.
Also, --make is now on by default, so this was largely irrelevant.
Ian Lynagh [Wed, 6 Oct 2010 12:44:13 +0000 (12:44 +0000)]
Deprecate -fglasgow-exts
Ian Lynagh [Wed, 6 Oct 2010 12:20:00 +0000 (12:20 +0000)]
Remove Opt_GADTs and Opt_TypeFamilies from -fglasgow-exts
This means most code doesn't get caught by monomorphic local bindings.
simonpj@microsoft.com [Fri, 8 Oct 2010 13:57:47 +0000 (13:57 +0000)]
Fix Trac #4360: omitted case in combineCtLoc
simonpj@microsoft.com [Fri, 8 Oct 2010 13:37:51 +0000 (13:37 +0000)]
Beautiful new approach to the skolem-escape check and untouchable
Instead of keeping a *set* of untouchable variables in each
implication contraints, we keep a *range* of uniques for the
*touchable* variables of an implication. This are precisely
the ones we would call the "existentials" if we were French.
It turns out that the code is more efficient, and vastly easier
to get right, than the set-based approach.
Fixes Trac #4355 among others
simonpj@microsoft.com [Fri, 8 Oct 2010 13:35:42 +0000 (13:35 +0000)]
Do less simplification when doing let-generalisation
This fixes Trac #4361. In a rather delicate way, but
no more delicate than before. A more remoseless typechecker
would reject #4361 altogether.
See Note [Avoid unecessary constraint simplification]
simonpj@microsoft.com [Fri, 8 Oct 2010 11:13:18 +0000 (11:13 +0000)]
Suppress ambiguity errors if any other errors occur
simonpj@microsoft.com [Fri, 8 Oct 2010 11:12:27 +0000 (11:12 +0000)]
Fix Trac #4361: be more discerning when inferring types
Note [Avoid unecessary constraint simplification] in TcSimplify
Simon Marlow [Fri, 8 Oct 2010 09:27:09 +0000 (09:27 +0000)]
Float out partial applications
This fixes at least one case of performance regression in 7.0, and
is nice win on nofib:
Program Size Allocs Runtime Elapsed
Min +0.3% -63.0% -38.5% -38.7%
Max +1.2% +0.2% +0.9% +0.9%
Geometric Mean +0.6% -3.0% -6.4% -6.6%
simonpj@microsoft.com [Fri, 8 Oct 2010 09:43:48 +0000 (09:43 +0000)]
Suppress knock-on typechecker errors
The error cascade caused puzzling errors in T4093b, and
suppressing some seems like a good plan. Very few test
outputs change.
simonpj@microsoft.com [Thu, 7 Oct 2010 16:35:00 +0000 (16:35 +0000)]
Some refactoring and simplification in TcInteract.occurCheck
simonpj@microsoft.com [Thu, 7 Oct 2010 13:03:01 +0000 (13:03 +0000)]
Comments only
simonpj@microsoft.com [Thu, 7 Oct 2010 11:10:51 +0000 (11:10 +0000)]
Implement auto-specialisation of imported Ids
This big-ish patch arranges that if an Id 'f' is
* Type-class overloaded
f :: Ord a => [a] -> [a]
* Defined with an INLINABLE pragma
{-# INLINABLE f #-}
* Exported from its defining module 'D'
then in any module 'U' that imports D
1. Any call of 'f' at a fixed type will generate
(a) a specialised version of f in U
(b) a RULE that rewrites unspecialised calls to the
specialised on
e.g. if the call is (f Int dOrdInt xs) then the
specialiser will generate
$sfInt :: [Int] -> [Int]
$sfInt = <code for f, imported from D, specialised>
{-# RULE forall d. f Int d = $sfInt #-}
2. In addition, you can give an explicit {-# SPECIALISE -#}
pragma for the imported Id
{-# SPECIALISE f :: [Bool] -> [Bool] #-}
This too generates a local specialised definition,
and the corresponding RULE
The new RULES are exported from module 'U', so that any module
importing U will see the specialised versions of 'f', and will
not re-specialise them.
There's a flag -fwarn-auto-orphan that warns you if the auto-generated
RULES are orphan rules. It's not in -Wall, mainly to avoid lots of
error messages with existing packages.
Main implementation changes
- A new flag on a CoreRule to say if it was auto-generated.
This is persisted across interface files, so there's a small
change in interface file format.
- Quite a bit of fiddling with plumbing, to get the
{-# SPECIALISE #-} pragmas for imported Ids. In particular, a
new field tgc_imp_specs in TcGblEnv, to keep the specialise
pragmas for imported Ids between the typechecker and the desugarer.
- Some new code (although surprisingly little) in Specialise,
to deal with calls of imported Ids
simonpj@microsoft.com [Thu, 7 Oct 2010 10:46:38 +0000 (10:46 +0000)]
Make NameEnv back into type NameEnv a = UniqFM a
I don't think the type distinction of declaring NameEnv with a newtype
(as it was) is really useful to us. Moreover, VarEnv is a UniqFM, and
I do sometimes want to build an envt with Ids and look up with Names.
This may not be the last word on the subject.
simonpj@microsoft.com [Thu, 7 Oct 2010 10:37:00 +0000 (10:37 +0000)]
Improve the rule-matcher
Previously it was rejecting the match
Template: forall s t. map s t
Actual: map Int t
which should obviously be fine. It turns out that this kind of match
comes up when specialising. By freshening that t we could avoid the
difficulty, but morally the (forall t) binds t and the rule should
be alpha-equivalent regardless of the forall'd variables.
This patch makes it so, and incidentally makes matching a little
more efficient. See Note [Eta expansion] in VarEnv.
simonpj@microsoft.com [Thu, 7 Oct 2010 10:27:20 +0000 (10:27 +0000)]
Fix Trac #4345: simplifier bug
This is another long-standing bug, in which there was a possibility
that a loop-breaker could lose its loop-breaker-hood OccInfo,
and then the simplifer re-simplified the expression. Result, either
non-termination or, in the case of #4345, an unbound identifier.
The fix is very simple, in Id.transferPolyIdInfo.
See Note [transferPolyIdInfo].
simonpj@microsoft.com [Thu, 7 Oct 2010 09:59:35 +0000 (09:59 +0000)]
Avoid redundant simplification
When adding specialisation for imported Ids, I noticed that the
Glorious Simplifier was repeatedly (and fruitlessly) simplifying the
same term. It turned out to be easy to fix this, because I already
had a flag in the ApplyTo and Select constructors of SimplUtils.SimplCont.
See Note [Avoid redundant simplification]
simonpj@microsoft.com [Thu, 7 Oct 2010 09:41:00 +0000 (09:41 +0000)]
Make the occurrence analyser deal correctly with RULES for imported Ids
This patch fixes a long-standing lurking bug, but it surfaced when I
was adding specialisation for imported Ids.
See Note [ImpRuleUsage], which explains the issue. The solution
seems more complicated than the problem really deserves, but I
could not think of a simpler way, so I just bit the bullet and
wrote the code. Improvements welcome.
simonpj@microsoft.com [Thu, 7 Oct 2010 09:20:07 +0000 (09:20 +0000)]
Make warning-free
simonpj@microsoft.com [Thu, 7 Oct 2010 09:16:18 +0000 (09:16 +0000)]
This is just white-space and layout
(At least, I don't think there is anything else.)
simonpj@microsoft.com [Thu, 7 Oct 2010 09:13:27 +0000 (09:13 +0000)]
Fix an ASSERT failure in FamInstEnv
I added a lot of comments too, to explain the preconditions;
esp Note [FamInstEnv]
simonpj@microsoft.com [Thu, 7 Oct 2010 08:41:04 +0000 (08:41 +0000)]
Fix a looping bug in the new occur-check code
simonpj@microsoft.com [Wed, 6 Oct 2010 15:52:23 +0000 (15:52 +0000)]
Fix test T4235 with -O
The tag2Enum rule wasn't doing the right thing for
enumerations with a phantom type parameter, like
data T a = A | B
simonpj@microsoft.com [Wed, 6 Oct 2010 15:50:33 +0000 (15:50 +0000)]
Make warning-free
dimitris@microsoft.com [Wed, 6 Oct 2010 15:28:54 +0000 (15:28 +0000)]
Major bugfixing pass through the type checker
dimitris@microsoft.com [Mon, 4 Oct 2010 13:02:00 +0000 (13:02 +0000)]
Typechecker performance fixes and flatten skolem bugfixing
dimitris@microsoft.com [Thu, 23 Sep 2010 14:39:18 +0000 (14:39 +0000)]
Performance bug fixes
simonpj@microsoft.com [Wed, 6 Oct 2010 11:53:16 +0000 (11:53 +0000)]
Fix Trac #4371: matching of view patterns
simonpj@microsoft.com [Wed, 6 Oct 2010 11:52:51 +0000 (11:52 +0000)]
Remove unused NoMatchContext construtor
simonpj@microsoft.com [Wed, 6 Oct 2010 10:28:30 +0000 (10:28 +0000)]
Refactoring: mainly rename ic_env_tvs to ic_untch
Plus remember to zonk the free_tvs in TcUnify.newImplication
Simon Marlow [Sat, 2 Oct 2010 19:51:18 +0000 (19:51 +0000)]
remove unnecessary/broken definition of mask_
Simon Marlow [Sat, 2 Oct 2010 19:51:00 +0000 (19:51 +0000)]
-fwarn-tabs: add "Warning" to the message
Simon Marlow [Wed, 29 Sep 2010 21:29:16 +0000 (21:29 +0000)]
give a better error message in the non-threaded RTS for out-of-range FDs
# ./aw
aw: file descriptor 1027 out of range for select (0--1024).
Recompile with -threaded to work around this.
Simon Marlow [Tue, 5 Oct 2010 14:47:35 +0000 (14:47 +0000)]
Fix a very rare crash in GHCi
When a BCO with a zero-length bitmap was right at the edge of
allocated memory, we were reading a word of non-existent memory.
This showed up as a segfault in T789(ghci) for me, but the crash was
extremely sensitive and went away with most changes.
Also, optimised scavenge_large_bitmap a bit while I was in there.
Simon Marlow [Fri, 24 Sep 2010 15:24:45 +0000 (15:24 +0000)]
Using 'stdcall' when it is not supported is only a warning now (#3336)
Simon Marlow [Fri, 24 Sep 2010 15:02:02 +0000 (15:02 +0000)]
remove unnecessary stg_noForceIO (#3508)
Ian Lynagh [Sun, 3 Oct 2010 12:57:07 +0000 (12:57 +0000)]
Replace an outputStr with putStrLn calls; fixes #4332
Ian Lynagh [Thu, 30 Sep 2010 22:47:41 +0000 (22:47 +0000)]
make test and fulltest targets in the main Makefile; fixes #4297
You can now run "make test" in the root, and the fast testsuite will be
run with cleaning enabled. It will also put the summary in
testsuite_summary.txt.
Ian Lynagh [Thu, 30 Sep 2010 21:01:28 +0000 (21:01 +0000)]
Don't show the loaded packages in ":show packages"; fixes #4300
It's never worked properly, and the information is in ":show linker".
Ian Lynagh [Thu, 30 Sep 2010 19:25:52 +0000 (19:25 +0000)]
Handle EXTRA_LIBRARIES when building programs
And set hp2ps's EXTRA_LIBRARIES. Based on a patch from Sergei Trofimovich.
Ian Lynagh [Wed, 29 Sep 2010 13:33:28 +0000 (13:33 +0000)]
Fix the doc directory on Windows
Ian Lynagh [Wed, 29 Sep 2010 00:00:24 +0000 (00:00 +0000)]
Remove an unused import on Windows
Ian Lynagh [Tue, 28 Sep 2010 23:58:44 +0000 (23:58 +0000)]
Use showCommandForUser when showing tracing commands
Ian Lynagh [Tue, 28 Sep 2010 20:19:38 +0000 (20:19 +0000)]
Fix hsc2hs docs: 'gcc' is now the default compiler, not 'ghc'; fixes #4341
Simon Marlow [Sat, 25 Sep 2010 19:35:48 +0000 (19:35 +0000)]
Use an empty signal handler for SIGPIPE instead of SIG_IGN
This is so that the SIGPIPE handler gets reset to the default
automatically on exec().
Ian Lynagh [Sat, 25 Sep 2010 21:00:29 +0000 (21:00 +0000)]
Fix the TH deps
Ian Lynagh [Fri, 24 Sep 2010 19:18:58 +0000 (19:18 +0000)]
Check inplace doesn't exist before we try to create it
This fixes rerunning configure in a tree which already has an inplace
directory. Edward Z Yang ran into this; I guess whether it actually
fails depends on details of your installation, or we'd have run into
it sooner.
simonpj@microsoft.com [Fri, 24 Sep 2010 15:58:15 +0000 (15:58 +0000)]
Fix an egregious bug: INLINE pragmas on monomorphic Ids were being ignored
I had do to some refactoring to make this work nicely
but now it does. I can't think how this escaped our
attention for so long!
simonpj@microsoft.com [Fri, 24 Sep 2010 15:57:07 +0000 (15:57 +0000)]
Eta expand only lambdas that bind a non-dictionary Id
See Note [When to eta expand]. The idea is that dictionary
lambdas are invisible to the user, so we shouldn't eta
expand them.
simonpj@microsoft.com [Fri, 24 Sep 2010 15:56:20 +0000 (15:56 +0000)]
Add a comment
simonpj@microsoft.com [Fri, 24 Sep 2010 15:56:14 +0000 (15:56 +0000)]
Add a debug print
simonpj@microsoft.com [Fri, 24 Sep 2010 15:56:00 +0000 (15:56 +0000)]
Just moving comments around
Simon Marlow [Fri, 24 Sep 2010 13:31:54 +0000 (13:31 +0000)]
use putStrLn instead of Haskeline's outputStrLn
use of the latter caused problems for Claus Reinke's macros that
redirect stdout.