ghc-hetmet.git
16 years agoImprove pretty-printing for Insts
simonpj@microsoft.com [Wed, 28 Nov 2007 17:31:25 +0000 (17:31 +0000)]
Improve pretty-printing for Insts

16 years agoReorganise TcSimplify (again); FIX Trac #1919
simonpj@microsoft.com [Wed, 28 Nov 2007 17:31:46 +0000 (17:31 +0000)]
Reorganise TcSimplify (again); FIX Trac #1919

This was a bit tricky.  We had a "given" dict like (d7:Eq a); then it got
supplied to reduceImplication, which did some zonking, and emerged with
a "needed given" (d7:Eq Int). That got everything confused.

I found a way to simplify matters significantly.  Now reduceContext
- first deals with methods/literals/dictionaries
- then deals with implications
Separating things in this way not only made the bug go away, but
eliminated the need for the recently-added "needed-givens" results returned
by checkLoop.  Hurrah.

It's still a swamp.  But it's a bit better.

16 years agoFIX #1914: GHCi forgot all the modules that were loaded before an error
Simon Marlow [Fri, 30 Nov 2007 13:07:34 +0000 (13:07 +0000)]
FIX #1914: GHCi forgot all the modules that were loaded before an error

16 years agoFIX #1744: ignore the byte-order mark at the beginning of a file
Simon Marlow [Fri, 30 Nov 2007 10:11:00 +0000 (10:11 +0000)]
FIX #1744: ignore the byte-order mark at the beginning of a file

16 years agoFIX Trac #1935: generate superclass constraints for derived classes
simonpj@microsoft.com [Wed, 28 Nov 2007 15:05:41 +0000 (15:05 +0000)]
FIX Trac #1935: generate superclass constraints for derived classes

This bug only reports a problem with phantom types, but actually
there was quite a long-standing and significant omission in the
constraint generation for derived classes.  See
Note [Superclasses of derived instance] in TcDeriv.

The test deriving-1935 tests both cases.

16 years agoPrint a bit more info in VarBinds (no need to merge)
simonpj@microsoft.com [Wed, 28 Nov 2007 15:03:54 +0000 (15:03 +0000)]
Print a bit more info in VarBinds (no need to merge)

16 years agoCheck for duplicate bindings in CoreLint
simonpj@microsoft.com [Wed, 28 Nov 2007 15:02:14 +0000 (15:02 +0000)]
Check for duplicate bindings in CoreLint

16 years agoadd comment
Simon Marlow [Wed, 28 Nov 2007 11:14:17 +0000 (11:14 +0000)]
add comment

16 years agoFIX #1916: don't try to convert float constants to int in CMM optimizer
Bertram Felgenhauer [Thu, 22 Nov 2007 09:55:13 +0000 (09:55 +0000)]
FIX #1916: don't try to convert float constants to int in CMM optimizer

16 years agogive a more useful message when the static flags have not been initialised (#1938)
Simon Marlow [Tue, 27 Nov 2007 13:54:35 +0000 (13:54 +0000)]
give a more useful message when the static flags have not been initialised (#1938)

16 years agoRebuild utils with the stage1 compiler when making a bindist; fixes trac #1860
Ian Lynagh [Tue, 27 Nov 2007 20:39:59 +0000 (20:39 +0000)]
Rebuild utils with the stage1 compiler when making a bindist; fixes trac #1860
This is a bit unpleasant, as "make binary-dist" really shouldn't actually
build anything, but it works.

16 years agoRemove the --print-docdir flag
Ian Lynagh [Tue, 27 Nov 2007 19:56:05 +0000 (19:56 +0000)]
Remove the --print-docdir flag
It wasn't doing the right thing for bindists. Let's rethink...

16 years agoFIX #1925: the interpreter was not maintaining tag bits correctly
Simon Marlow [Tue, 27 Nov 2007 12:26:14 +0000 (12:26 +0000)]
FIX #1925: the interpreter was not maintaining tag bits correctly
See comment for details

16 years agoadd missing instruction: ALLOC_AP_NOUPD
Simon Marlow [Tue, 27 Nov 2007 12:26:04 +0000 (12:26 +0000)]
add missing instruction: ALLOC_AP_NOUPD

16 years agoCheck tag bits on the fun pointer of a PAP
Simon Marlow [Mon, 26 Nov 2007 16:04:20 +0000 (16:04 +0000)]
Check tag bits on the fun pointer of a PAP

16 years agocanonicalise the path to HsColour
Simon Marlow [Mon, 26 Nov 2007 14:16:14 +0000 (14:16 +0000)]
canonicalise the path to HsColour

16 years agoConsistently put www. on the front of haskell.org in URLs
Ian Lynagh [Mon, 26 Nov 2007 21:52:56 +0000 (21:52 +0000)]
Consistently put  on the front of haskell.org in URLs

16 years agoFix some more URLs
Ian Lynagh [Mon, 26 Nov 2007 21:41:47 +0000 (21:41 +0000)]
Fix some more URLs

16 years agoTweak some URLs
Ian Lynagh [Mon, 26 Nov 2007 19:41:48 +0000 (19:41 +0000)]
Tweak some URLs

16 years agoFix some links
Ian Lynagh [Mon, 26 Nov 2007 18:44:06 +0000 (18:44 +0000)]
Fix some links

16 years agoCopy gmp stamps into bindists, so we don't try and rebuild gmp
Ian Lynagh [Sun, 25 Nov 2007 21:19:19 +0000 (21:19 +0000)]
Copy gmp stamps into bindists, so we don't try and rebuild gmp

16 years agoOn Windows, Delete the CriticalSection's we Initialize
Ian Lynagh [Sun, 25 Nov 2007 12:58:45 +0000 (12:58 +0000)]
On Windows, Delete the CriticalSection's we Initialize

16 years agoOn Windows, add a start menu link to the flag reference
Ian Lynagh [Sun, 25 Nov 2007 12:44:29 +0000 (12:44 +0000)]
On Windows, add a start menu link to the flag reference

16 years agoRemove html/ from the paths we put in the start menu on Windows
Ian Lynagh [Sun, 25 Nov 2007 12:41:50 +0000 (12:41 +0000)]
Remove html/ from the paths we put in the start menu on Windows

16 years agoMERGED: Make ":" in GHCi repeat the last command
Ian Lynagh [Sun, 25 Nov 2007 12:20:20 +0000 (12:20 +0000)]
MERGED: Make ":" in GHCi repeat the last command
Ian Lynagh <igloo@earth.li>**20071124231857
 It used to be a synonym for ":r" in 6.6.1, but this wasn't documented or
 known about by the developers. In 6.8.1 it was accidentally broken.
 This patch brings it back, but as "repeat the last command", similar to
 pressing enter in gdb. This is almost as good for people who want it to
 reload, and means that it can also be used to repeat commands like :step.

16 years agoMERGED: Put library docs in a $pkg, rather than $pkgid, directory; fixes trac #1864
Ian Lynagh [Sat, 24 Nov 2007 21:23:05 +0000 (21:23 +0000)]
MERGED: Put library docs in a $pkg, rather than $pkgid, directory; fixes trac #1864
Ian Lynagh <igloo@earth.li>**20071124171220

16 years agoDon't make a library documentation prologue
Ian Lynagh [Sat, 24 Nov 2007 21:19:43 +0000 (21:19 +0000)]
Don't make a library documentation prologue
It's far too large now, and no-one complained when 6.8.1 didn't have one.

16 years agoDon't put package version numbers in links in index.html
Ian Lynagh [Sat, 24 Nov 2007 21:16:29 +0000 (21:16 +0000)]
Don't put package version numbers in links in index.html

16 years agoDefine install-strip in Makefile
Ian Lynagh [Sat, 24 Nov 2007 20:50:37 +0000 (20:50 +0000)]
Define install-strip in Makefile

16 years agoDefine install-strip in distrib/Makefile
Ian Lynagh [Sat, 24 Nov 2007 20:48:03 +0000 (20:48 +0000)]
Define install-strip in distrib/Makefile

16 years agoInstall gmp from bindists; fixes trac #1848
Ian Lynagh [Sat, 24 Nov 2007 18:52:40 +0000 (18:52 +0000)]
Install gmp from bindists; fixes trac #1848

16 years ago(native gen) fix code generated for GDTOI on x86_32
Bertram Felgenhauer [Wed, 21 Nov 2007 06:39:42 +0000 (06:39 +0000)]
(native gen) fix code generated for GDTOI on x86_32
See trac #1910.

16 years agoCopy the INSTALL hack from mk/config.mk.in into distrib/Makefile-bin-vars.in
Ian Lynagh [Sat, 24 Nov 2007 16:30:28 +0000 (16:30 +0000)]
Copy the INSTALL hack from mk/config.mk.in into distrib/Makefile-bin-vars.in
configure will set INSTALL to ./install-sh if it can't find it in the path,
so we need to replace the . with the path to our root.

16 years agoMake install-sh executable /before/ we try to find it
Ian Lynagh [Sat, 24 Nov 2007 16:24:50 +0000 (16:24 +0000)]
Make install-sh executable /before/ we try to find it

16 years agoDocument --info in the +RTS -? help
Ian Lynagh [Fri, 23 Nov 2007 20:43:52 +0000 (20:43 +0000)]
Document --info in the +RTS -? help

16 years agoMERGED: If we have hscolour then make source code links in teh haddock docs
Ian Lynagh [Fri, 23 Nov 2007 23:31:13 +0000 (23:31 +0000)]
MERGED: If we have hscolour then make source code links in teh haddock docs
Fri Nov 23 13:15:59 PST 2007  Ian Lynagh <igloo@earth.li>

16 years agoTidy and trim the type environment in mkBootModDetails
simonpj@microsoft.com [Fri, 23 Nov 2007 15:35:19 +0000 (15:35 +0000)]
Tidy and trim the type environment in mkBootModDetails

Should fix Trac #1833

We were failing to trim the type envt in mkBootModDetails, so several
functions all called (*), for example, were getting into the interface.
Result chaos.  It only actually bites when we do the retyping-loop thing,
which is why it's gone so long without a fix.

16 years agorefactor: HscNothing and boot modules do not need desugaring
Simon Marlow [Fri, 23 Nov 2007 13:52:37 +0000 (13:52 +0000)]
refactor: HscNothing and boot modules do not need desugaring

16 years agoFIX #1910: fix code generated for GDTOI on x86_32
Bertram Felgenhauer [Wed, 21 Nov 2007 10:26:27 +0000 (10:26 +0000)]
FIX #1910: fix code generated for GDTOI on x86_32

16 years agoProperly ppr InstEqs in wanteds of implication constraints
Manuel M T Chakravarty [Thu, 22 Nov 2007 09:30:02 +0000 (09:30 +0000)]
Properly ppr InstEqs in wanteds of implication constraints

16 years agoFIX #1910: fix code generated for GDTOI on x86_32
Bertram Felgenhauer [Wed, 21 Nov 2007 10:26:27 +0000 (10:26 +0000)]
FIX #1910: fix code generated for GDTOI on x86_32

16 years agoAdd built-in Double operations to vectorisation
Roman Leshchinskiy [Thu, 22 Nov 2007 00:25:17 +0000 (00:25 +0000)]
Add built-in Double operations to vectorisation

16 years agoTeach vectorisation about Double
Roman Leshchinskiy [Wed, 21 Nov 2007 05:49:32 +0000 (05:49 +0000)]
Teach vectorisation about Double

16 years agoVectorise polyexprs with notes
Roman Leshchinskiy [Wed, 21 Nov 2007 05:31:02 +0000 (05:31 +0000)]
Vectorise polyexprs with notes

16 years agoMake rebindable do-notation behave as advertised
simonpj@microsoft.com [Wed, 21 Nov 2007 17:49:14 +0000 (17:49 +0000)]
Make rebindable do-notation behave as advertised

Adopt Trac #1537.  The patch ended up a bit bigger than I expected,
so I suggest we do not merge this into the 6.8 branch.  But there
is no funadamental reason why not.

With this patch, rebindable do-notation really does type as if you
had written the original (>>) and (>>=) operations in desguared form.

I ended up refactoring some of the (rather complicated) error-context
stuff in TcUnify, by pushing an InstOrigin into tcSubExp and its
various calls. That means we could get rid of tcFunResTy, and the
SubCtxt type.  This should improve error messages slightly
in complicated situations, because we have an Origin to hand
to instCall (in the (isSigmaTy actual_ty) case of tc_sub1).

Thanks to Pepe for the first draft of the patch.

16 years agoAdd DEBUG-only flag -dsuppress-uniques to suppress printing of uniques
simonpj@microsoft.com [Fri, 16 Nov 2007 15:24:46 +0000 (15:24 +0000)]
Add DEBUG-only flag -dsuppress-uniques to suppress printing of uniques

This is intended only for debugging use: it makes it easier to
compare two variants without the variations between uniques mattering.

(Of course, you can't actually feed the output to the C compiler
or assembler and expect anything sensible to happen!)

16 years agoAdd -dcore-lint when validating libraries
simonpj@microsoft.com [Mon, 5 Nov 2007 16:47:33 +0000 (16:47 +0000)]
Add -dcore-lint when validating libraries

16 years agoFix Trac #1913: check data const for derived types are in scope
simonpj@microsoft.com [Wed, 21 Nov 2007 15:14:28 +0000 (15:14 +0000)]
Fix Trac #1913: check data const for derived types are in scope

When deriving an instance, the data constructors should all be in scope.
This patch checks the condition.

16 years agoFix Trac #1909: type of map in docs
simonpj@microsoft.com [Tue, 20 Nov 2007 16:01:52 +0000 (16:01 +0000)]
Fix Trac #1909: type of map in docs

16 years agoMove file locking into the RTS, fixing #629, #1109
Simon Marlow [Tue, 20 Nov 2007 14:08:59 +0000 (14:08 +0000)]
Move file locking into the RTS, fixing #629, #1109
File locking (of the Haskell 98 variety) was previously done using a
static table with linear search, which had two problems: the array had
a fixed size and was sometimes too small (#1109), and performance of
lockFile/unlockFile was suboptimal due to the linear search.
Also the algorithm failed to count readers as required by Haskell 98
(#629).

Now it's done using a hash table (provided by the RTS).  Furthermore I
avoided the extra fstat() for every open file by passing the dev_t and
ino_t into lockFile.  This and the improvements to the locking
algorithm result in a healthy 20% or so performance increase for
opening/closing files (see openFile008 test).

16 years agoFIX Trac #1825: standalone deriving Typeable
simonpj@microsoft.com [Tue, 20 Nov 2007 12:57:32 +0000 (12:57 +0000)]
FIX Trac #1825: standalone deriving Typeable

Standalone deriving of typeable now requires you to say
instance Typeable1 Maybe
which is exactly the shape of instance decl that is generated
by a 'deriving( Typeable )' clause on the data type decl.

This is a bit horrid, but it's the only consistent way, at least
for now.  If you say something else, the error messages are helpful.

MERGE to 6.8 branch

16 years agoFIX #1715: egregious bug in ifaceDeclSubBndrs
simonpj@microsoft.com [Tue, 20 Nov 2007 11:17:23 +0000 (11:17 +0000)]
FIX #1715: egregious bug in ifaceDeclSubBndrs

ifaceDeclSubBndrs didn't have an IfaceSyn case; but with type
families an IfaceSyn can introduce subordinate binders.  Result:
chaos.

The fix is easy though.  Merge to 6.8 branch.

16 years agoAlways do 'setup makefile' before building each library
Simon Marlow [Tue, 20 Nov 2007 10:33:29 +0000 (10:33 +0000)]
Always do 'setup makefile' before building each library
This forces preprocessing to happen, which is necessary if any of the
.hsc files have been modified.  Without this change, a 'setup
makefile' would be required by hand after a .hsc file changed.
Fortunately 'setup makefile' isn't much extra work, and I've made it
not overwrite GNUmakefile if it hasn't changed, which avoids
recalculating the dependencies each time.

16 years agoFIX #1847 (improve :browse! docs, fix unqual)
claus.reinke@talk21.com [Thu, 8 Nov 2007 01:31:47 +0000 (01:31 +0000)]
FIX #1847 (improve :browse! docs, fix unqual)

- add example to docs, explain how to interpret
  output of `:browse! Data.Maybe`
- print unqualified names according to current
  context, not the context of the target module

16 years agoTrack changes to package ndp
Roman Leshchinskiy [Tue, 20 Nov 2007 03:37:16 +0000 (03:37 +0000)]
Track changes to package ndp

16 years agoTemporary hack for passing PArrays from unvectorised to vectorised code
Roman Leshchinskiy [Tue, 20 Nov 2007 02:45:45 +0000 (02:45 +0000)]
Temporary hack for passing PArrays from unvectorised to vectorised code

16 years agoBind NDP stuff to [:.:] arrays
Roman Leshchinskiy [Mon, 19 Nov 2007 02:03:02 +0000 (02:03 +0000)]
Bind NDP stuff to [:.:] arrays

16 years agoDon't treat enumerations specially during vectorisation for the moment
Roman Leshchinskiy [Mon, 19 Nov 2007 01:37:29 +0000 (01:37 +0000)]
Don't treat enumerations specially during vectorisation for the moment

16 years agoFix bugs in vectorisation of case expressions
Roman Leshchinskiy [Mon, 19 Nov 2007 01:37:14 +0000 (01:37 +0000)]
Fix bugs in vectorisation of case expressions

16 years agoMore built-in NDP combinators
Roman Leshchinskiy [Mon, 19 Nov 2007 01:22:05 +0000 (01:22 +0000)]
More built-in NDP combinators

16 years agoNew vectorisation built-ins
Roman Leshchinskiy [Sun, 18 Nov 2007 05:19:40 +0000 (05:19 +0000)]
New vectorisation built-ins

16 years agoFix bug in conversion unvect/vect
Roman Leshchinskiy [Sun, 18 Nov 2007 05:19:26 +0000 (05:19 +0000)]
Fix bug in conversion unvect/vect

16 years agoExtend built-in vectorisation environments
Roman Leshchinskiy [Sun, 18 Nov 2007 04:52:19 +0000 (04:52 +0000)]
Extend built-in vectorisation environments

16 years agoFix bug in generation of environments for vectorisation
Roman Leshchinskiy [Sun, 18 Nov 2007 04:52:03 +0000 (04:52 +0000)]
Fix bug in generation of environments for vectorisation

16 years agoAdd builtin var->var mapping to vectorisation
Roman Leshchinskiy [Sun, 18 Nov 2007 04:26:05 +0000 (04:26 +0000)]
Add builtin var->var mapping to vectorisation

16 years agoExtend vectorisation built-in mappings with datacons
Roman Leshchinskiy [Sun, 18 Nov 2007 03:43:51 +0000 (03:43 +0000)]
Extend vectorisation built-in mappings with datacons

16 years agoChange representation of parallel arrays of enumerations
Roman Leshchinskiy [Sun, 18 Nov 2007 03:33:55 +0000 (03:33 +0000)]
Change representation of parallel arrays of enumerations

16 years agoAdd vectorisation-related builtin
Roman Leshchinskiy [Sun, 18 Nov 2007 03:15:13 +0000 (03:15 +0000)]
Add vectorisation-related builtin

16 years agoTeach vectorisation about Bool
Roman Leshchinskiy [Sat, 17 Nov 2007 04:27:14 +0000 (04:27 +0000)]
Teach vectorisation about Bool

16 years agoIncomplete support for boxing during vectorisation
Roman Leshchinskiy [Sat, 17 Nov 2007 04:07:39 +0000 (04:07 +0000)]
Incomplete support for boxing during vectorisation

16 years agoMake sure some TyCons always vectorise to themselves
Roman Leshchinskiy [Sat, 17 Nov 2007 04:05:37 +0000 (04:05 +0000)]
Make sure some TyCons always vectorise to themselves

16 years agoSimple conversion vectorised -> unvectorised
Roman Leshchinskiy [Sat, 17 Nov 2007 02:30:29 +0000 (02:30 +0000)]
Simple conversion vectorised -> unvectorised

16 years agoFix bug in case vectorisation
Roman Leshchinskiy [Sat, 17 Nov 2007 01:50:14 +0000 (01:50 +0000)]
Fix bug in case vectorisation

16 years agoVectorisation of algebraic case expressions
Roman Leshchinskiy [Fri, 16 Nov 2007 07:48:14 +0000 (07:48 +0000)]
Vectorisation of algebraic case expressions

16 years agoMore vectorisation-related built-ins
Roman Leshchinskiy [Fri, 16 Nov 2007 06:18:31 +0000 (06:18 +0000)]
More vectorisation-related built-ins

16 years agoVectorisation utilities
Roman Leshchinskiy [Fri, 16 Nov 2007 05:10:37 +0000 (05:10 +0000)]
Vectorisation utilities

16 years agoAdd vectorisation built-ins
Roman Leshchinskiy [Fri, 16 Nov 2007 05:09:59 +0000 (05:09 +0000)]
Add vectorisation built-ins

16 years agoFix vectorisation of binders in case expressions
Roman Leshchinskiy [Fri, 16 Nov 2007 02:18:33 +0000 (02:18 +0000)]
Fix vectorisation of binders in case expressions

16 years agoTwo small typos in the flags summary (merge to 6.8 branch)
simonpj@microsoft.com [Mon, 19 Nov 2007 13:46:39 +0000 (13:46 +0000)]
Two small typos in the flags summary (merge to 6.8 branch)

16 years agoImprove the situation for Trac #959: civilised warning instead of a trace msg
simonpj@microsoft.com [Mon, 19 Nov 2007 12:29:38 +0000 (12:29 +0000)]
Improve the situation for Trac #959: civilised warning instead of a trace msg

This doesn't fix the root cause of the bug, but it makes the report
more civilised, and points to further info.

16 years agoFIX Trac #1806: test for correct arity for datacon in infix pattern patch
simonpj@microsoft.com [Mon, 19 Nov 2007 11:43:01 +0000 (11:43 +0000)]
FIX Trac #1806: test for correct arity for datacon in infix pattern patch

Happily the fix is easy; pls merge

16 years agoAccept x86_64-*-freebsd* as well as amd64-*-freebsd* in configure.ac
Ian Lynagh [Sat, 17 Nov 2007 15:45:02 +0000 (15:45 +0000)]
Accept x86_64-*-freebsd* as well as amd64-*-freebsd* in configure.ac
Patch from Brian P. O'Hanlon

16 years agoAttempt at fixing #1873, #1360
Simon Marlow [Fri, 16 Nov 2007 15:21:48 +0000 (15:21 +0000)]
Attempt at fixing #1873, #1360

I think I figured out a reasonable way to manage the GHCi context,
comments welcome.

Rule 1: external package modules in the context are persistent.  That
is, when you say 'import Data.Maybe' it survives over :load, :add,
:reload and :cd.

Rule 2: :load and :add remove all home-package modules from the
context and add the rightmost target, as a *-module if possible.  This
is as before, and makes sense for :load because we're starting a new
program; the old home-package modules don't make sense any more.  For
:add, it usually does what you want, because the new target will
become the context.

Rule 3: any modules from the context that fail to load during a
:reload are remembered, and re-added to the context at the next
successful :reload.

Claus' suggestion about adding the "remembered" modules to the prompt
prefixed with a ! is implemented but commented out.  I couldn't
decide whether it was useful or confusing.

One difference that people might notice is that after a :reload where
there were errors, GHCi would previously dump you in the most recent
module that it loaded.  Now it dumps you in whatever subset of the
current context still makes sense, and in the common case that will
probably be {Prelude}.

16 years agoWibble to fix Trac #1901 (shorten messsage slightly)
simonpj@microsoft.com [Fri, 16 Nov 2007 15:03:41 +0000 (15:03 +0000)]
Wibble to fix Trac #1901 (shorten messsage slightly)

16 years agoImprove links from flag reference to the relevant section; and improve doc of RankN...
simonpj@microsoft.com [Fri, 16 Nov 2007 14:58:16 +0000 (14:58 +0000)]
Improve links from flag reference to the relevant section; and improve doc of RankN flags

16 years agoFIX Trac #1901: check no existential context in H98 mode
simonpj@microsoft.com [Fri, 16 Nov 2007 14:56:09 +0000 (14:56 +0000)]
FIX Trac #1901: check no existential context in H98 mode

16 years agoImprove documentation of data type declarations (Trac #1901)
simonpj@microsoft.com [Fri, 16 Nov 2007 08:18:41 +0000 (08:18 +0000)]
Improve documentation of data type declarations (Trac #1901)

16 years agoChange the command-line semantics for query commands
Simon Marlow [Fri, 16 Nov 2007 13:20:46 +0000 (13:20 +0000)]
Change the command-line semantics for query commands

From the help text:

  Commands that query the package database (list, latest, describe,
  field) operate on the list of databases specified by the flags
  --user, --global, and --package-conf.  If none of these flags are
  given, the default is --global --user.

This makes it possible to query just a single database (e.g. the
global one without the user one), which needed tricks to accomplish
before.

16 years agouse "ghc-pkg latest --global" instead of "ghc-pkg list --simple-output"
Simon Marlow [Fri, 16 Nov 2007 12:20:18 +0000 (12:20 +0000)]
use "ghc-pkg latest --global" instead of "ghc-pkg list --simple-output"
The former now does the right thing: it uses the global database only,
and picks the most recent package with the given name.

16 years agoDisallow installing packages whose names differ in case only.
Simon Marlow [Fri, 16 Nov 2007 12:11:53 +0000 (12:11 +0000)]
Disallow installing packages whose names  differ in case only.
--force overrides.  Requested by Duncan Coutts, with a view to
treating package names as case-insensitive in the future.

16 years agoFIX BUILD (with GHC 6.2.x): update .hi-boot file
Simon Marlow [Fri, 16 Nov 2007 10:12:27 +0000 (10:12 +0000)]
FIX BUILD (with GHC 6.2.x): update .hi-boot file

16 years agoFIX #1828: installing to a patch with spaces in
Simon Marlow [Thu, 15 Nov 2007 15:57:47 +0000 (15:57 +0000)]
FIX #1828: installing to a patch with spaces in
We have to pass the path to gcc when calling windres, which itself
might have spaces in.  Furthermore, we have to pass the path to gcc's
tools to gcc.  This means getting the quoting right, and after much
experimentation and reading of the windres sources I found something
that works: passing --use-temp-files to windres makes it use its own
implementation of quoting instead of popen(), and this does what we
want.  Sigh.

16 years agoon Windows, install to a directory with spaces (test for #1828)
Simon Marlow [Thu, 15 Nov 2007 15:53:27 +0000 (15:53 +0000)]
on Windows, install to a directory with spaces (test for #1828)

16 years agoFIX #1679: crash on returning from a foreign call
Simon Marlow [Thu, 15 Nov 2007 13:16:35 +0000 (13:16 +0000)]
FIX #1679: crash on returning from a foreign call
We forgot to save a pointer to the BCO over the foreign call.  Doing
enough allocation and GC during the call could provoke a crash.

16 years agoAvoid the use of unversioned package dependencies
Simon Marlow [Thu, 15 Nov 2007 10:32:49 +0000 (10:32 +0000)]
Avoid the use of unversioned package dependencies
Fortunately "ghc-pkg list $pkg --simple-output" is a good way to add
the version number.

16 years agoFIX #1596 (remove deprecated --define-name)
Simon Marlow [Wed, 14 Nov 2007 16:53:23 +0000 (16:53 +0000)]
FIX #1596 (remove deprecated --define-name)
Also remove the old command-line syntax for ghc-pkg, which was not
documented.  Do not merge.

16 years agoFIX #1837: remove deprecated support for unversioned dependencies (do not merge)
Simon Marlow [Wed, 14 Nov 2007 16:10:44 +0000 (16:10 +0000)]
FIX #1837: remove deprecated support for unversioned dependencies (do not merge)

16 years agowibble
Pepe Iborra [Wed, 14 Nov 2007 23:33:56 +0000 (23:33 +0000)]
wibble

16 years agoMake pprNameLoc more robust in absence of loc information
Pepe Iborra [Wed, 14 Nov 2007 23:33:43 +0000 (23:33 +0000)]
Make pprNameLoc more robust in absence of loc information

16 years agoTry to manage the size of the text rendered for ':show bindings'
Pepe Iborra [Wed, 14 Nov 2007 23:16:01 +0000 (23:16 +0000)]
Try to manage the size of the text rendered for ':show bindings'