12 years agoFix a pair of classic, but tricky, bugs in the type matcher; exposed by a program... [Fri, 2 Feb 2007 23:36:34 +0000 (23:36 +0000)] 
Fix a pair of classic, but tricky, bugs in the type matcher; exposed by a program from Roman

12 years agoOne more wibble to FloatOut, fixes HEAD breakage (I hope) [Fri, 2 Feb 2007 22:53:48 +0000 (22:53 +0000)] 
One more wibble to FloatOut, fixes HEAD breakage (I hope)

12 years agoclarify that -main-is sets main module and function [Wed, 31 Jan 2007 22:39:21 +0000 (22:39 +0000)] 
clarify that -main-is sets main module and function

12 years agoDon't dump floated bindings just outside a lambda [Fri, 2 Feb 2007 15:54:52 +0000 (15:54 +0000)] 
Don't dump floated bindings just outside a lambda

We do not want the FloatOut pass to transform
f = \x. e
f = let lvl = ... in \x.e

The arity pinned on f isn't right any more; and see
Note [Floating out of RHSs].

Core Lint is now spotting the arity lossage (for a letrec), which is
how I spotted this bug.

I also re-jigged the code around floatBind; it's a bit tidier now.

12 years agoReset the demand info on bindings going to top level (since they cannot be strict) [Fri, 2 Feb 2007 15:12:06 +0000 (15:12 +0000)] 
Reset the demand info on bindings going to top level (since they cannot be strict)

12 years agoRecord arity of the worker, to maintain consistency of arity and strictness informati... [Fri, 2 Feb 2007 15:10:51 +0000 (15:10 +0000)] 
Record arity of the worker, to maintain consistency of arity and strictness information (checked by Lint)

12 years agoExtend the local bindings at a breakpoint with one for the wrapped expression
Pepe Iborra [Fri, 2 Feb 2007 11:26:29 +0000 (11:26 +0000)] 
Extend the local bindings at a breakpoint with one for the wrapped expression

By popular request, in a breakpoint it is possible now to inspect the result of the expression wrapped by the breakpoint.

The user interface for this is right now preliminar; there is a new binding called '_result' at every breakpoint. Suggestions are welcome!

12 years agoRefactoring of DsBreakpoint.hs
Pepe Iborra [Fri, 2 Feb 2007 10:59:07 +0000 (10:59 +0000)] 
Refactoring of DsBreakpoint.hs

12 years agoFix imports
Pepe Iborra [Fri, 2 Feb 2007 10:05:02 +0000 (10:05 +0000)] 
Fix imports

12 years agoWarning police: Avoid warning about unused variable [Thu, 1 Feb 2007 15:08:39 +0000 (15:08 +0000)] 
Warning police: Avoid warning about unused variable

12 years agoAdhere to the new GNU Coding Standards, avoiding a warning with autoconf 2.59c and... [Thu, 1 Feb 2007 13:42:43 +0000 (13:42 +0000)] 
Adhere to the new GNU Coding Standards, avoiding a warning with autoconf 2.59c and later

12 years agochmod +x install-sh, see bug #978
Simon Marlow [Thu, 1 Feb 2007 13:11:25 +0000 (13:11 +0000)] 
chmod +x install-sh, see bug #978

12 years agosome dynamic flags cannot be used with OPTIONS_GHC, e.g. -i, -package
Simon Marlow [Thu, 1 Feb 2007 12:40:58 +0000 (12:40 +0000)] 
some dynamic flags cannot be used with OPTIONS_GHC, e.g. -i, -package
fixes #1108

12 years agoPartial fix for #926
Simon Marlow [Thu, 1 Feb 2007 11:40:47 +0000 (11:40 +0000)] 
Partial fix for #926
It seems that when a program exits with open DLLs on Windows, the
system attempts to shut down the DLLs, but it also terminates (some
of?) the running threads.  The RTS isn't prepared for threads to die
unexpectedly, so it sits around waiting for its workers to finish.
This bites in two places: ShutdownIOManager() in the the unthreaded
RTS, and shutdownCapability() in the threaded RTS.  So far I've
modified the latter to notice when worker threads have died
unexpectedly and continue shutting down.  It seems a bit trickier to
fix the unthreaded RTS, so for now the workaround for #926 is to use
the threaded RTS.

12 years agoAlways pay attention to -keep-tmp-files when we want to delete files
Ian Lynagh [Wed, 31 Jan 2007 15:10:58 +0000 (15:10 +0000)] 
Always pay attention to -keep-tmp-files when we want to delete files

12 years agoSlight refactoring for overloaded strings [Wed, 31 Jan 2007 15:07:01 +0000 (15:07 +0000)] 
Slight refactoring for overloaded strings

12 years agoUse Id.isStrictId [Wed, 31 Jan 2007 15:06:16 +0000 (15:06 +0000)] 
Use Id.isStrictId

12 years agoTwo new warnings: arity differing from demand type, and strict IDs at top level
Kirsten Chevalier [Mon, 29 Jan 2007 22:55:33 +0000 (22:55 +0000)] 
Two new warnings: arity differing from demand type, and strict IDs at top level

I added two new Core Lint checks in lintSingleBinding:

1. Check that the id's arity is equal to the
   number of arguments in its demand type, if it has a demand type
   at all (i.e., if demand analysis already happened).

2. Check that top-level or recursive binders aren't demanded.

12 years agoFix typo causing the PowerPC OS X build to fail
Ian Lynagh [Wed, 31 Jan 2007 12:01:53 +0000 (12:01 +0000)] 
Fix typo causing the PowerPC OS X build to fail

12 years agoFix typo
Ian Lynagh [Wed, 31 Jan 2007 11:27:24 +0000 (11:27 +0000)] 
Fix typo

12 years agoBugFix: do not insert breakpoints around expressions with unlifted kind
Pepe Iborra [Wed, 31 Jan 2007 10:28:21 +0000 (10:28 +0000)] 
BugFix: do not insert breakpoints around expressions with unlifted kind

  I have added a check, and while there removed a few kludges in my code.
  Kudos to -dcore-lint for uncovering this.

  I think that this restriction could be lifted, if GHC.Base.breakpoint could have kind ?? -> ??. But is this a legal type? Does not look so to me.

12 years agoFix a typo
Pepe Iborra [Wed, 31 Jan 2007 10:40:01 +0000 (10:40 +0000)] 
Fix a typo

12 years agoAdd note about Template Haskell and mutual recursion [Wed, 31 Jan 2007 09:14:51 +0000 (09:14 +0000)] 
Add note about Template Haskell and mutual recursion

12 years agoFix Trac #1122; spot absence of TyCon in data/newtype header [Tue, 30 Jan 2007 16:13:30 +0000 (16:13 +0000)] 
Fix Trac #1122; spot absence of TyCon in data/newtype header

12 years agoFix for #936
Simon Marlow [Tue, 30 Jan 2007 10:13:06 +0000 (10:13 +0000)] 
Fix for #936
We must traverse dependencies recursively if we encounter any [boot]
modules in the dependencies.

12 years agofix bug in GetLastError/SetLastError patch
Simon Marlow [Mon, 29 Jan 2007 12:45:59 +0000 (12:45 +0000)] 
fix bug in GetLastError/SetLastError patch

12 years agoFix GHCi on PowerPC OS X
Ian Lynagh [Sun, 28 Jan 2007 21:32:16 +0000 (21:32 +0000)] 
Fix GHCi on PowerPC OS X
David Kirkman and Peter Tanski noticed that a line had been removed during
a patch merge which meant that oc->image pointed to the wrong place and
ultimately caused an error from realloc.

12 years agoSave the Win32 error code where necessary
Simon Marlow [Fri, 26 Jan 2007 16:26:20 +0000 (16:26 +0000)] 
Save the Win32 error code where necessary
Similarly to the way we save errno across context switches and
suspendThread/resumeThread, we must save and restore the Win32 error
code via GetLastError()/SetLastError().  Fixes #896.

12 years agoComments and layout only [Thu, 25 Jan 2007 12:54:48 +0000 (12:54 +0000)] 
Comments and layout only

12 years agoAdd -stdout-binary and -stderr-binary flags to bypass dos2unix
Simon Marlow [Tue, 23 Jan 2007 12:38:43 +0000 (12:38 +0000)] 
Add -stdout-binary and -stderr-binary flags to bypass dos2unix
Apparently some versions of dos2unix can detect a binary file and
ignore it, but others can't, so provide a way to tell runstdtest that
the output is in binary.

12 years agoDocument the overloaded string extension. [Sat, 20 Jan 2007 16:14:36 +0000 (16:14 +0000)] 
Document the overloaded string extension.

12 years agoAdd support for overloaded string literals. [Thu, 21 Dec 2006 21:01:47 +0000 (21:01 +0000)] 
Add support for overloaded string literals.

The class is named IsString with the single method fromString.
Overloaded strings work the same way as overloaded numeric literals.
In expressions a string literals gets a fromString applied to it.
In a pattern there will be an equality comparison with the fromString:ed literal.

Use -foverloaded-strings to enable this extension.

12 years agoImprove presentation of :print output
Pepe Iborra [Sun, 21 Jan 2007 11:06:10 +0000 (11:06 +0000)] 
Improve presentation of :print output

12 years agoComments only
Pepe Iborra [Sun, 21 Jan 2007 11:03:44 +0000 (11:03 +0000)] 
Comments only

12 years agoRefactoring of Debugger.hs
Pepe Iborra [Sun, 21 Jan 2007 10:59:09 +0000 (10:59 +0000)] 
Refactoring of Debugger.hs

A big motivation to start with it was getting several independently useful functions out of the Ghci monad and into the IO monad instead. Working in debugger integration for Emacs via the ghc-api is helping me to improve reusability..
      While I was there, I tried to make the code less tangled, easier to understand, switched from implicit Exceptions to explicit Eithers in the bkptTable code, etc.

12 years agoFix a bug in the closure viewer
Pepe Iborra [Sat, 20 Jan 2007 20:11:05 +0000 (20:11 +0000)] 
Fix a bug in the closure viewer

12 years agoAdd obtainTerm1 to the GHC api
Pepe Iborra [Sat, 20 Jan 2007 18:52:49 +0000 (18:52 +0000)] 
Add obtainTerm1 to the GHC api

12 years agoReorganizing my mess a bit
Pepe Iborra [Sun, 7 Jan 2007 21:12:40 +0000 (21:12 +0000)] 
Reorganizing my mess a bit

12 years agoTweak hc-build
Ian Lynagh [Fri, 19 Jan 2007 22:08:50 +0000 (22:08 +0000)] 
Tweak hc-build

Touch all the object files at the end of hc-build, as otherwise make
gets upset as the .h files are newer than them and tries to rebuild
stuff (and fails).

Print a nice message when we're finished.

12 years agoUser's guide: length is not a good consumer anymore.
Spencer Janssen [Thu, 4 Jan 2007 07:52:43 +0000 (07:52 +0000)] 
User's guide: length is not a good consumer anymore.

12 years agoDon't start a GHC session before answering --version/--numeric-version
Ian Lynagh [Fri, 19 Jan 2007 15:52:31 +0000 (15:52 +0000)] 
Don't start a GHC session before answering --version/--numeric-version
This also fixes bootstrapping issues, as when bootstrapping we don't
have a package.conf when we run configure, so it fails to work out
what version of GHC it is which causes breakage later on.

12 years agoprettifying Cmm: print MachOps as identifiers by replacing ' ' with '_' [Wed, 17 Jan 2007 13:32:34 +0000 (13:32 +0000)] 
prettifying Cmm: print MachOps as identifiers by replacing ' ' with '_'

12 years agoprettifying Cmm a bit: convert (x + (-N)) to (x - N) [Wed, 17 Jan 2007 13:00:37 +0000 (13:00 +0000)] 
prettifying Cmm a bit: convert (x + (-N)) to (x - N)

12 years agoaddition to "Eagerly raise a blocked exception" to fix unreg case
Simon Marlow [Wed, 17 Jan 2007 09:59:44 +0000 (09:59 +0000)] 
addition to "Eagerly raise a blocked exception" to fix unreg case

12 years agoRemove special lambda unicode character, it didn't work anyway
Simon Marlow [Tue, 16 Jan 2007 16:11:00 +0000 (16:11 +0000)] 
Remove special lambda unicode character, it didn't work anyway
Since lambda is a lower-case letter, it's debatable whether we want to
steal it to mean lambda in Haskell source.  However if we did, then we
would probably want to make it a "special" symbol, not just a reserved
symbol, otherwise writing \x->... (using unicode characters of course)
wouldn't work, because \x would be treated as a single identifier,
you'd need a space.

12 years agoallow 'ghci -threaded' (fixes #1101)
Simon Marlow [Tue, 16 Jan 2007 12:04:09 +0000 (12:04 +0000)] 
allow 'ghci -threaded' (fixes #1101)

12 years agoCheck with sysconf _POSIX_THREAD_CPUTIME is available before we try to use it.
Ian Lynagh [Tue, 16 Jan 2007 01:29:07 +0000 (01:29 +0000)] 
Check with sysconf _POSIX_THREAD_CPUTIME is available before we try to use it.

Calling clock_gettime(CLOCK_THREAD_CPUTIME_ID,_) regardless was
causing a segfault (trac #1030).

12 years agoGive -fwrapv to gcc when it supports it
Ian Lynagh [Mon, 15 Jan 2007 19:42:50 +0000 (19:42 +0000)] 
Give -fwrapv to gcc when it supports it
Fixes trac #952: Haskell requires consistent overflow behaviour, which
gcc doesn't give without this flag.

12 years agoFix ghc-pkg now showError has been removed
Ian Lynagh [Mon, 15 Jan 2007 12:44:41 +0000 (12:44 +0000)] 
Fix ghc-pkg now showError has been removed

12 years agoDocumentation for the new ':breakpoint continue' command
Pepe Iborra [Mon, 15 Jan 2007 09:56:43 +0000 (09:56 +0000)] 
Documentation for the new ':breakpoint continue' command

12 years agoAdd TcGadt to package.conf
Pepe Iborra [Sun, 14 Jan 2007 13:08:53 +0000 (13:08 +0000)] 
Add TcGadt to package.conf

I'm assuming that this was overlooked during the System Fc overhaul
If TcGadt is not supposed to be in package.conf please rollback this patch

12 years agoAdd a header for GhciMonad.hs
Pepe Iborra [Sun, 14 Jan 2007 13:08:15 +0000 (13:08 +0000)] 
Add a header for GhciMonad.hs

12 years agoRemove a false assertion
Pepe Iborra [Fri, 12 Jan 2007 14:40:40 +0000 (14:40 +0000)] 
Remove a false assertion

12 years agoToo agressive breakpoint coalescing politics
Pepe Iborra [Thu, 11 Jan 2007 17:25:55 +0000 (17:25 +0000)] 
Too agressive breakpoint coalescing politics

12 years agoFix an obscure bug in rule-matching [Sat, 13 Jan 2007 22:04:25 +0000 (22:04 +0000)] 
Fix an obscure bug in rule-matching

This bug is the cause of Trac #1092.  The fix is easy
by making the RnEnv2 implementation do the right thing.
See Note [rnBndrLR] in VarEnv.

Test case is simplCore/should_compile/rule1

12 years agoAdd a warning for tabs in source files
Ian Lynagh [Fri, 12 Jan 2007 16:47:32 +0000 (16:47 +0000)] 
Add a warning for tabs in source files

12 years agoexpand $topdir in the output of 'ghc-pkg field'
Simon Marlow [Tue, 9 Jan 2007 17:00:19 +0000 (17:00 +0000)] 
expand $topdir in the output of 'ghc-pkg field'
this fixed #937, and gets us further towards 'setup haddock' working
for Cabal on Windows.

12 years agoRemove bogus assertion in getCallMethod
Kirsten Chevalier [Thu, 11 Jan 2007 17:25:01 +0000 (17:25 +0000)] 
Remove bogus assertion in getCallMethod

With my as-yet-uncommitted changes to the demand analyzer, code got
 generated for some programs that caused this assertion to fail.  The
 transformation I was doing was correct; it was the assertion that
 wasn't. So, the assertion is removed.

 This is actually Simon PJ's patch rather than mine, but I noticed that
 it wasn't checked in and it seems completely safe to do so.

12 years agoFixed .spec file (no building guide anymore) [Thu, 11 Jan 2007 16:29:56 +0000 (16:29 +0000)] 
Fixed .spec file (no building guide anymore)

12 years agoAn object code module in ghci is not and cannot be handled by the GHCi debugger
Pepe Iborra [Thu, 11 Jan 2007 14:56:34 +0000 (14:56 +0000)] 
An object code module in ghci is not and cannot be handled by the GHCi debugger

This patch should provide more informative messages to the user

12 years agoComments only
Pepe Iborra [Thu, 11 Jan 2007 13:14:31 +0000 (13:14 +0000)] 
Comments only

12 years agoAdded the new :breakpoint continue option
Pepe Iborra [Thu, 11 Jan 2007 13:13:59 +0000 (13:13 +0000)] 
Added the new :breakpoint continue option

Previously, when in a breakpoint, :quit was used to continue execution.
This is not the right thing to do, so this patch restores :quit to its
original meaning whether or not ghci is in an inferior session.

The continue behavior is now provided by ":breakpoint continue".
I added a synonim command in :continue because it is much shorter,
but this is optional

12 years ago:cd is a top level only command, that is, not allowed in inferior ghci sessions
Pepe Iborra [Thu, 11 Jan 2007 13:11:56 +0000 (13:11 +0000)] 
:cd is a top level only command, that is, not allowed in inferior ghci sessions

12 years agoWibble [Thu, 11 Jan 2007 13:14:23 +0000 (13:14 +0000)] 

12 years agoSlightly improve -ddump-hi-diffs output [Thu, 11 Jan 2007 13:12:26 +0000 (13:12 +0000)] 
Slightly improve -ddump-hi-diffs output

12 years agoAvoid duplicate defns of 'main' in module Main [Thu, 11 Jan 2007 13:10:29 +0000 (13:10 +0000)] 
Avoid duplicate defns of 'main' in module Main

12 years agoAdd -ddump-rule-firings [Thu, 11 Jan 2007 13:08:04 +0000 (13:08 +0000)] 
Add -ddump-rule-firings

12 years agoComments [Thu, 11 Jan 2007 08:53:31 +0000 (08:53 +0000)] 

12 years agoCorrect spelling [Wed, 10 Jan 2007 13:33:24 +0000 (13:33 +0000)] 
Correct spelling

12 years agoSort rules and instances lexicographically in interface files [Wed, 10 Jan 2007 12:43:27 +0000 (12:43 +0000)] 
Sort rules and instances lexicographically in interface files

We should sort rules and instances lexicographically, not by Unique,
in interface files, else we get unnecessary "rules changed" and hence
unnecessary recompilation.

This bug has been there since the interface-file upheaval that put Names
into IfaceSyn.

12 years agoThis patch teaches ghci to announce that a module has beeen loaded in debugging mode
Pepe Iborra [Thu, 11 Jan 2007 11:23:51 +0000 (11:23 +0000)] 
This patch teaches ghci to announce that a module has beeen loaded in debugging mode

Usually ghci announces when a module has been loaded interpreted. Now it will also announce it when the module has been instrumented for debugging

12 years agoFix some text in the GHCi help message that was going over 80 columns
Pepe Iborra [Thu, 11 Jan 2007 11:17:42 +0000 (11:17 +0000)] 
Fix some text in the GHCi help message that was going over 80 columns

12 years agoComments only
Pepe Iborra [Thu, 11 Jan 2007 11:17:07 +0000 (11:17 +0000)] 
Comments only

12 years agoComments only [Thu, 11 Jan 2007 09:16:23 +0000 (09:16 +0000)] 
Comments only

12 years agoMake the LiberateCase transformation understand associated types [Thu, 11 Jan 2007 09:15:33 +0000 (09:15 +0000)] 
Make the LiberateCase transformation understand associated types

Consider this FC program:
data family AT a :: *
data instance AT Int = T1 Int Int

f :: AT Int -> Int
f t = case t of DEFAULT -> <body>

We'd like to replace the DEFAULT by a use of T1, so that if
we scrutinise t inside <body> we share the evaluation:

f t = case (t `cast` co) of T1 x y -> <body>

I decided to do this as part of the liberate-case transformation,
which is already trying to avoid redundant evals.

The new transformation requires knowledge of the family instance
environment, so I had to extend ModGuts to carry the fam_inst_env,
and put that envt into the liberate-case environment.

Otherwise it's all pretty straightforward.

12 years agoAdd comments about invariants [Thu, 11 Jan 2007 09:10:03 +0000 (09:10 +0000)] 
Add comments about invariants

12 years agoComments [Thu, 11 Jan 2007 09:09:22 +0000 (09:09 +0000)] 

12 years agoComments [Thu, 11 Jan 2007 09:08:37 +0000 (09:08 +0000)] 

12 years agoAdd the function TypeRep.pprTypeApp, and use it [Thu, 11 Jan 2007 09:07:04 +0000 (09:07 +0000)] 
Add the function TypeRep.pprTypeApp, and use it

  pprTypeApp :: SDoc -> [Type] -> SDoc
  pprTypeApp pp tys = hang pp 2 (sep (map pprParendType tys))

12 years agoLog message for: Fix a nasty recursive loop in typechecking interface files [Thu, 11 Jan 2007 09:02:52 +0000 (09:02 +0000)] 
Log message for: Fix a nasty recursive loop in typechecking interface files

(Alas, Darcs failed to record my log-message for the above patch,
so this patch is an attempt to add the log message retrospectively.)

Roman found a case where the type-checker for interface files
would go into a loop.  Here it is:

  module BarAT where
    class Foo a where
      data FooT a :: *
      int :: FooT a -> Int

  module Baz where
    import BarAT
    foo :: FooT Int -> Int
    foo = foo

The trouble turned out to be that Foo gives rise to a *newtype*,
and using a newtpe caused a little bit too much strictness in
BuildTyCl.mkNewTyConRhs.  Specifically, mkNewTypeCoercion did pattern
matching, which forced the call to eta_reduce in mkNewTyConRhs.

This is all too delicate really.  But for now I've fixed the bug,
and added an explanatory comment.  I'll add a test for it, in

12 years agoLog-message for: Improve command-line parser (add OptIntSuffix); make -fliberate... [Thu, 11 Jan 2007 09:01:23 +0000 (09:01 +0000)] 
Log-message for: Improve command-line parser (add OptIntSuffix); make -fliberate-case-threshold dynamic

(Alas, Darcs failed to record my log-message for the above patch,
so this patch is an attempt to add the log message retrospectively.)

Parsing options numeric arguments, such as
for dynamic flags wasn't being handled well.  I elaborated the
command-line parser to have a new constructor, OptIntSuffix, for this
case, and did the consequential changes.

This patch also makes the -fliberate-case-threshold flag into a
dynamic flag.

12 years agoImprove command-line parser (add OptIntSuffix); make -fliberate-case-threshold dynamic [Thu, 11 Jan 2007 08:48:51 +0000 (08:48 +0000)] 
Improve command-line parser (add OptIntSuffix); make -fliberate-case-threshold dynamic

12 years agoRemove redundant import [Thu, 11 Jan 2007 08:47:02 +0000 (08:47 +0000)] 
Remove redundant import

12 years agoAdd comments (AT-related) [Thu, 11 Jan 2007 08:46:45 +0000 (08:46 +0000)] 
Add comments (AT-related)

12 years agoSpelling correction only [Thu, 11 Jan 2007 08:45:21 +0000 (08:45 +0000)] 
Spelling correction only

12 years agoFix a nasty recursive loop in typechecking interface files [Thu, 11 Jan 2007 08:29:50 +0000 (08:29 +0000)] 
Fix a nasty recursive loop in typechecking interface files

12 years agoUpdate comments [Thu, 11 Jan 2007 08:29:07 +0000 (08:29 +0000)] 
Update comments

12 years agoChanging bootstrapping for pipes between debugger and debuggee [Wed, 10 Jan 2007 20:26:35 +0000 (20:26 +0000)] 
Changing bootstrapping for pipes between debugger and debuggee

 - Now we use two pipes and a fork rather than use named pipes
 - We use the HPCRIX env var to send the pipe numbers in %d:%d format
 - We now always breakpoint for special events
 - We check to see we are the original process before each breakpoint event
   sub processes are not debugged; this lets us debug the debugger.

12 years agoDynamically allocating the tix box breakpoint array [Tue, 9 Jan 2007 19:46:31 +0000 (19:46 +0000)] 
Dynamically allocating the tix box breakpoint array

12 years agoAdding command channel for the hpc debugger to the hpc part of the RTS [Tue, 9 Jan 2007 19:40:11 +0000 (19:40 +0000)] 
Adding command channel for the hpc debugger to the hpc part of the RTS

12 years agocheckTauTvUpdate: take synonym families into account
Manuel M T Chakravarty [Wed, 10 Jan 2007 18:44:51 +0000 (18:44 +0000)] 
checkTauTvUpdate: take synonym families into account

12 years agoremove old building guide, change links to point to the wiki
Simon Marlow [Wed, 10 Jan 2007 16:11:12 +0000 (16:11 +0000)] 
remove old building guide, change links to point to the wiki

12 years agoFix apparently-long-standing bug in FloatIn [Wed, 10 Jan 2007 11:13:44 +0000 (11:13 +0000)] 
Fix apparently-long-standing bug in FloatIn

The float-in pass wasn't doing the right thing when you have

let x{rule mentions y} = rhs in body

It allowed a binding mentioning y to float into the body, which is
obviously wrong.  I think this bug has been there a long time; I don't
really know why it has not come up before.

It showed up when compiling Text.Regex.Base.Context with WAY=p in
package regex-base.

12 years agoComments only [Wed, 10 Jan 2007 11:13:31 +0000 (11:13 +0000)] 
Comments only

12 years agoHandle synonym families in check_tau_type
Manuel M T Chakravarty [Tue, 9 Jan 2007 21:55:41 +0000 (21:55 +0000)] 
Handle synonym families in check_tau_type

12 years agoPreserve rigidity for GADTs when typechecking explicit tuples [Tue, 9 Jan 2007 13:53:14 +0000 (13:53 +0000)] 
Preserve rigidity for GADTs when typechecking explicit tuples

12 years agoFix egregious bug in implication constraints [Tue, 9 Jan 2007 13:47:19 +0000 (13:47 +0000)] 
Fix egregious bug in implication constraints

Build an implication constraint if there is a non-trivial
refinement, even if there are no other 'given' constraints.

Test = gadt/set.hs

12 years agoreorganise text in the "using packages" section
Simon Marlow [Tue, 9 Jan 2007 09:17:58 +0000 (09:17 +0000)] 
reorganise text in the "using packages" section

12 years agoUnbreak usage of ASSERTM
Pepe Iborra [Tue, 9 Jan 2007 07:42:58 +0000 (07:42 +0000)] 
Unbreak usage of ASSERTM

12 years agoWhen setting stdout and stderr to NoBuffering in GHCi, do stdin too.
Ian Lynagh [Mon, 8 Jan 2007 16:28:38 +0000 (16:28 +0000)] 
When setting stdout and stderr to NoBuffering in GHCi, do stdin too.
Fixes trac #929.
Merge to 6.6 branch.