ghc-hetmet.git
21 years ago[project @ 2002-10-27 10:38:32 by mthomas]
mthomas [Sun, 27 Oct 2002 10:38:33 +0000 (10:38 +0000)]
[project @ 2002-10-27 10:38:32 by mthomas]
Remove Win32 library dependency.

21 years ago[project @ 2002-10-25 21:59:51 by simonpj]
simonpj [Fri, 25 Oct 2002 21:59:52 +0000 (21:59 +0000)]
[project @ 2002-10-25 21:59:51 by simonpj]
Put findLinkable in Finder.lhs, where it can be found in non-ghci stage1 compilers!

21 years ago[project @ 2002-10-25 21:58:24 by simonpj]
simonpj [Fri, 25 Oct 2002 21:58:24 +0000 (21:58 +0000)]
[project @ 2002-10-25 21:58:24 by simonpj]
Import wibbles

21 years ago[project @ 2002-10-25 16:54:55 by simonpj]
simonpj [Fri, 25 Oct 2002 16:55:00 +0000 (16:55 +0000)]
[project @ 2002-10-25 16:54:55 by simonpj]
Import wibbles

21 years ago[project @ 2002-10-25 15:57:03 by simonpj]
simonpj [Fri, 25 Oct 2002 15:57:03 +0000 (15:57 +0000)]
[project @ 2002-10-25 15:57:03 by simonpj]
Reduce exports, and add comments

21 years ago[project @ 2002-10-25 15:23:03 by simonpj]
simonpj [Fri, 25 Oct 2002 15:23:07 +0000 (15:23 +0000)]
[project @ 2002-10-25 15:23:03 by simonpj]
------------------------
More dependency fiddling
------------------------

WARNING: Interface file format has changed (again)
 You need to 'make clean' in all library code

* Orphan modules are now kept separately
  Home-package dependencies now contain only home-package dependencies!
  See HscTypes.Dependencies

* Linker now uses the dependencies to do dynamic linking

Result: Template Haskell should work even without --make (not yet tested)

21 years ago[project @ 2002-10-25 12:56:34 by simonmar]
simonmar [Fri, 25 Oct 2002 12:56:34 +0000 (12:56 +0000)]
[project @ 2002-10-25 12:56:34 by simonmar]
We shouldn't be using MAP_FIXED on solaris (fixes last night's crash).

Unfortunately on Solaris the heuristics don't work too well.  Solaris
seems to ignore the addr argument to mmap when MAP_FIXED is not
specified, and starts handing out memory from the top of the address
space.  So we allocate 2M each time and munmap() ends to leave an
aligned 1M chunk.  This will eventually leave the address space
completely full of 1M holes.  I don't have a good solution to this at
the moment.

MERGE TO STABLE

21 years ago[project @ 2002-10-25 12:34:12 by michaelw]
michaelw [Fri, 25 Oct 2002 12:34:12 +0000 (12:34 +0000)]
[project @ 2002-10-25 12:34:12 by michaelw]
* typo (dl -> dnl)

21 years ago[project @ 2002-10-25 11:33:27 by simonpj]
simonpj [Fri, 25 Oct 2002 11:33:27 +0000 (11:33 +0000)]
[project @ 2002-10-25 11:33:27 by simonpj]
Wibble; cures failure in stage2 build

21 years ago[project @ 2002-10-25 09:40:47 by simonmar]
simonmar [Fri, 25 Oct 2002 09:40:47 +0000 (09:40 +0000)]
[project @ 2002-10-25 09:40:47 by simonmar]
In eval_thunk_selector(), don't follow IND_STATICs because they might
lead us into to-space.  Fixes a case of "EVACUATED object entered!".

Also, add an assertion to catch this bug earlier.

MERGE TO STABLE

21 years ago[project @ 2002-10-24 16:54:19 by simonpj]
simonpj [Thu, 24 Oct 2002 16:54:19 +0000 (16:54 +0000)]
[project @ 2002-10-24 16:54:19 by simonpj]
Uh oh; got the versioning stuff a bit wrong in the last commit

21 years ago[project @ 2002-10-24 14:17:46 by simonpj]
simonpj [Thu, 24 Oct 2002 14:17:53 +0000 (14:17 +0000)]
[project @ 2002-10-24 14:17:46 by simonpj]
------------------------------------------
1. New try and module and package dependencies
2. OrigNameCache always contains final info
------------------------------------------

These things nearly complete sorting out the incremental
linking problem that started us off!

1. This commit separates two kinds of information:

  (a) HscTypes.Dependencies:
What (i)  home-package modules, and
     (ii) other packages
      this module depends on, transitively.

      That is, to link the module, it should be enough
      to link the dependent modules and packages (plus
      any C stubs etc).

      Along with this info we record whether the dependent module
      is (a) a boot interface or (b) an orphan module.  So in
      fact (i) can contain non-home-package modules, namely the
      orphan ones in other packages (sigh).

  (b) HscTypes.Usage:
      What version of imported things were used to
      actually compile the module.  This info is used for
      recompilation control only.

2. The Finder now returns a correct Module (incl package indicator)
first time, so we can install the absolutely final Name in the
OrigNameCache when we first come across an occurrence of that name,
even if it's only an occurrence in an unfolding in some other interface
file.  This is much tidier.

As a result Module.lhs is much cleaner
No DunnoYet
No mkVanillaModule
ALl very joyful stuff.

21 years ago[project @ 2002-10-24 13:08:35 by simonmar]
simonmar [Thu, 24 Oct 2002 13:08:35 +0000 (13:08 +0000)]
[project @ 2002-10-24 13:08:35 by simonmar]
In --make mode, consider everything that doesn't contain a '.' to be a
compilation manager input.  This fixes cases like

  ghc --make test37

where test37.hs exists.  I broke this in the previous commit, causing
a number of tests to break.

21 years ago[project @ 2002-10-23 15:56:39 by simonpj]
simonpj [Wed, 23 Oct 2002 15:56:39 +0000 (15:56 +0000)]
[project @ 2002-10-23 15:56:39 by simonpj]
Document implicit parameter bindings

21 years ago[project @ 2002-10-23 14:30:00 by simonpj]
simonpj [Wed, 23 Oct 2002 14:30:03 +0000 (14:30 +0000)]
[project @ 2002-10-23 14:30:00 by simonpj]
------------------------------------------------
Allow implicit-parameter bindings anywhere that
a normal binding group is allowed.
------------------------------------------------

That is, you can have implicit parameters

* in a let binding
* in a where clause (but then you can't have non-implicit
  ones as well)
* in a let group in a list comprehension or monad do-notation

The implementation is simple: just add IPBinds to the allowable forms of HsBinds,
and remove the HsWith expression form altogether.   (It now comes in via the
HsLet form.)

It'a a nice generalisation really.  Needs a bit of documentation, which I'll do next.

21 years ago[project @ 2002-10-23 12:26:11 by mthomas]
mthomas [Wed, 23 Oct 2002 12:26:11 +0000 (12:26 +0000)]
[project @ 2002-10-23 12:26:11 by mthomas]
Declare loop index i.

21 years ago[project @ 2002-10-23 08:52:26 by simonmar]
simonmar [Wed, 23 Oct 2002 08:52:26 +0000 (08:52 +0000)]
[project @ 2002-10-23 08:52:26 by simonmar]
Add atomicModifyIORefzh_fast symbol (should have been done with the
rest of the atomicModifyIORef# commit, thanks to Manuel Chakravarty
for pointing out the problem).

21 years ago[project @ 2002-10-23 02:36:37 by chak]
chak [Wed, 23 Oct 2002 02:36:37 +0000 (02:36 +0000)]
[project @ 2002-10-23 02:36:37 by chak]
Added an import to make TH compile.

21 years ago[project @ 2002-10-22 13:36:56 by simonmar]
simonmar [Tue, 22 Oct 2002 13:36:56 +0000 (13:36 +0000)]
[project @ 2002-10-22 13:36:56 by simonmar]
Fix recent FreeBSD breakage in the mangler: the "Prologue Junk" test
was a little bit too strict about whitespace.

21 years ago[project @ 2002-10-22 11:01:18 by simonmar]
simonmar [Tue, 22 Oct 2002 11:01:20 +0000 (11:01 +0000)]
[project @ 2002-10-22 11:01:18 by simonmar]
change the types of cmp_thread, rts_getThreadId, and labelThread to
take StgPtr rather than StgTSO *, since the compiler now has no
distinction between these two types in the back end.

I also noticed that labelThread need not be a primitive: it could just
as well be a normal C function called by the FFI, but I haven't made
that change.

21 years ago[project @ 2002-10-22 10:31:21 by simonpj]
simonpj [Tue, 22 Oct 2002 10:31:21 +0000 (10:31 +0000)]
[project @ 2002-10-22 10:31:21 by simonpj]
Import wibbles

21 years ago[project @ 2002-10-21 12:42:31 by simonmar]
simonmar [Mon, 21 Oct 2002 12:42:31 +0000 (12:42 +0000)]
[project @ 2002-10-21 12:42:31 by simonmar]
Turn an instance of 0xfff00000 into ~MBLOCK_MASK in the Win32 code.

21 years ago[project @ 2002-10-21 11:38:53 by simonmar]
simonmar [Mon, 21 Oct 2002 11:38:54 +0000 (11:38 +0000)]
[project @ 2002-10-21 11:38:53 by simonmar]
Bite the bullet and generalise the central memory allocation scheme.
Previously we tried to allocate memory starting from a fixed address,
which was set for each architecture (0x5000000 was a common one), and
to decide whether a particular address was in the heap or not we would
do a simple comparison against this address.

This doesn't work too well, because:

 - if we dynamically-load some objects above the boundary, the
   heap-allocated test becomes invalid

 - on windows we have less control, and the heap might be
   split into multiple sections

 - it turns out that on some Linux kernels we don't get memory where
   we asked for it.  This might be a bug in those kernels, but it
   exposes the fragility of our allocation scheme.

The solution is to bite the bullet and maintain a table mapping
addresses to a value indicating whether that address is in the heap or
not.  Since we normally allocate heap in chunks of 1Mb, the table is
quite small: 4k on a 32-bit machine, using one byte for each 1Mb
block.  Testing an address for heap residency now involves a memory
access, but the table is normally cache-resident.  I didn't manage to
measure any slowdown after making the change.

On a 64-bit machine, we'll need to use a 2-level table; I haven't
implemented that yet.

Now we can generalise the procedure used to grab memory from the OS.
In the general case, we allocate one megablock more than we need to,
and trim off the slop around the allocation to leave an aligned chunk.
The next time around, however, we try to allocate memory right after
the last chunk allocated, on the grounds that it is aligned and
probably free: if this doesn't work, we have to back off to the
general mechanism (it seems to work most of the time).

This cleans up the Windows story too: is_heap_alloced() has gone, and
we should be able to handle more than 256M of memory (or whatever the
arbitrary limit was before).

MERGE TO STABLE (after lots of testing)

21 years ago[project @ 2002-10-19 10:42:23 by stolz]
stolz [Sat, 19 Oct 2002 10:42:24 +0000 (10:42 +0000)]
[project @ 2002-10-19 10:42:23 by stolz]
Retro Police: There are hosts w/o in_addr_t (e.g. SuSE 7.0)

21 years ago[project @ 2002-10-18 13:41:50 by simonpj]
simonpj [Fri, 18 Oct 2002 13:41:55 +0000 (13:41 +0000)]
[project @ 2002-10-18 13:41:50 by simonpj]
--------------------------------
   Fix a serious error in the "newtype deriving" feature
--------------------------------

The "newtype deriving" feature lets you derive arbitrary classes for
a newtype, not just the built-in ones (Read, Show, Ix etc).  It's very
cool, but Hal Duame discovered that it did utterly the Wrong Thing
for superclasses.  E.g.

newtype Foo = MkFoo Int deriving( Show, Num, Eq )

You'd get a Num instance for Foo that was *identical* to the
Num instance for Int, *including* the Show superclass. So the
superclass in the Num dictionary would show a Foo just like an
Int, which is wrong... it should show as "Foo n".

This commit fixes the problem, by building a new dictionary every time,
but using the methods from the dictionary for the representation type.

I also fixed a bug that prevented it working altogether when the
representation type was not the application of a type constructor.
For example, this now works

newtype Foo a = MkFoo a deriving( Num, Eq, Show )

I also made it a bit more efficient in the case where the type is
not parameterised.  Then the "dfun" doesn't need to be a function.

21 years ago[project @ 2002-10-18 13:36:17 by simonpj]
simonpj [Fri, 18 Oct 2002 13:36:17 +0000 (13:36 +0000)]
[project @ 2002-10-18 13:36:17 by simonpj]
Add a trace

21 years ago[project @ 2002-10-18 13:35:46 by simonpj]
simonpj [Fri, 18 Oct 2002 13:35:47 +0000 (13:35 +0000)]
[project @ 2002-10-18 13:35:46 by simonpj]
Import wibbles

21 years ago[project @ 2002-10-18 10:16:04 by simonmar]
simonmar [Fri, 18 Oct 2002 10:16:04 +0000 (10:16 +0000)]
[project @ 2002-10-18 10:16:04 by simonmar]
Add the Haskell Web Server to the fptools repository so I can give it
a proper license.

21 years ago[project @ 2002-10-18 09:51:03 by simonmar]
simonmar [Fri, 18 Oct 2002 09:51:04 +0000 (09:51 +0000)]
[project @ 2002-10-18 09:51:03 by simonmar]
Add atomicModifyIORef, as discussed on the FFI list.

21 years ago[project @ 2002-10-18 09:36:21 by simonmar]
simonmar [Fri, 18 Oct 2002 09:36:21 +0000 (09:36 +0000)]
[project @ 2002-10-18 09:36:21 by simonmar]
Add a note about the profiling versions of the interface files in a package.

21 years ago[project @ 2002-10-17 14:49:52 by simonmar]
simonmar [Thu, 17 Oct 2002 14:49:52 +0000 (14:49 +0000)]
[project @ 2002-10-17 14:49:52 by simonmar]
- Don't flush the finder cache after adding a new package.  We'll
  assume that packages don't overlap.

- Add the extra command-line flags specified by a package when we
  add a package from the GHCi prompt.

21 years ago[project @ 2002-10-17 14:26:16 by simonmar]
simonmar [Thu, 17 Oct 2002 14:26:19 +0000 (14:26 +0000)]
[project @ 2002-10-17 14:26:16 by simonmar]
Finder overhaul.

The finder had got pretty complicated; this commit is mainly a
cleanup, with one new feature:

  - the finder has a cache (again).  The cache may be flushed by
    calling flushFinderCache, which actually only flushes home modules
    from the cache, because package modules are assumed not to move.
    This change is apropos of some other changes which will result in
    the finder being called more often, so we think a cache is going
    to be worthwhile.

Also a couple of bugs were fixed:

  - the field ml_hi_file in a ModLocation is now *always* the name
    of the .hi file.  If you need a .hi-boot file, you have to make
    it up by changing the suffix of ml_hi_file.  (DriverMkDepend and
    RnHiFiles do this).  This was the cause of a bug, but I can't
    remember the details.

  - The -odir flag now works in a more reasonable way: hierarchical
    modules get put in subdirectories of the -odir directory.  eg.
    if your module is A.B.C, and -odir D is specified, then the object
    will be D/A/B/C.o; previously it would have been D/C.o.

21 years ago[project @ 2002-10-15 13:20:18 by simonmar]
simonmar [Tue, 15 Oct 2002 13:20:18 +0000 (13:20 +0000)]
[project @ 2002-10-15 13:20:18 by simonmar]
beginMake: separate out compilation-manager inputs from linkable
objects in a way that is more consistent with the one-shot compilation
mode.  Specifically, we now pass anything that looks like a module
name or a Haskell source filename to the compilation manager, and
everything else to the linker.

MERGE TO STABLE

21 years ago[project @ 2002-10-15 13:18:51 by simonmar]
simonmar [Tue, 15 Oct 2002 13:18:51 +0000 (13:18 +0000)]
[project @ 2002-10-15 13:18:51 by simonmar]
Remove looksLikeModuleName, it is defined in Util now.

21 years ago[project @ 2002-10-15 13:17:40 by simonmar]
simonmar [Tue, 15 Oct 2002 13:17:40 +0000 (13:17 +0000)]
[project @ 2002-10-15 13:17:40 by simonmar]
Move looksLikeModuleName here from InterativeUI, so we can use it elsewhere.

21 years ago[project @ 2002-10-15 12:56:03 by simonmar]
simonmar [Tue, 15 Oct 2002 12:56:03 +0000 (12:56 +0000)]
[project @ 2002-10-15 12:56:03 by simonmar]
wibble

21 years ago[project @ 2002-10-15 12:38:30 by simonmar]
simonmar [Tue, 15 Oct 2002 12:38:30 +0000 (12:38 +0000)]
[project @ 2002-10-15 12:38:30 by simonmar]
wibble

21 years ago[project @ 2002-10-15 12:04:05 by simonmar]
simonmar [Tue, 15 Oct 2002 12:04:05 +0000 (12:04 +0000)]
[project @ 2002-10-15 12:04:05 by simonmar]
Don't use stdPackage (it has gone)

21 years ago[project @ 2002-10-15 11:52:31 by simonpj]
simonpj [Tue, 15 Oct 2002 11:52:32 +0000 (11:52 +0000)]
[project @ 2002-10-15 11:52:31 by simonpj]
------------------------------------------------
Fix exports for a module with  no local bindings
------------------------------------------------

Fix the export-list generation for a module with no local bindings.
This one bug was breaking a bunch of modxxx tests, plus some renamer
test, plus a desugarer test.

21 years ago[project @ 2002-10-15 11:02:32 by simonmar]
simonmar [Tue, 15 Oct 2002 11:02:32 +0000 (11:02 +0000)]
[project @ 2002-10-15 11:02:32 by simonmar]
Slight fix to the allocated memory calculation

21 years ago[project @ 2002-10-15 11:01:29 by simonmar]
simonmar [Tue, 15 Oct 2002 11:01:29 +0000 (11:01 +0000)]
[project @ 2002-10-15 11:01:29 by simonmar]
Don't need to export Exception.try or Exception.throwDyn here.

21 years ago[project @ 2002-10-15 11:00:37 by simonmar]
simonmar [Tue, 15 Oct 2002 11:00:37 +0000 (11:00 +0000)]
[project @ 2002-10-15 11:00:37 by simonmar]
Use Panic.tryMost rather than Exception.try, since we don't want to
catch ^C and panic exceptions.

21 years ago[project @ 2002-10-15 10:59:38 by simonmar]
simonmar [Tue, 15 Oct 2002 10:59:38 +0000 (10:59 +0000)]
[project @ 2002-10-15 10:59:38 by simonmar]
Remove unused 'stdPackage'

21 years ago[project @ 2002-10-15 10:53:21 by simonmar]
simonmar [Tue, 15 Oct 2002 10:53:21 +0000 (10:53 +0000)]
[project @ 2002-10-15 10:53:21 by simonmar]
Remove the "contributors" section herein an point the interested
reader to the contributors list on the web site which is much more
up-to-date.

21 years ago[project @ 2002-10-15 08:56:50 by simonpj]
simonpj [Tue, 15 Oct 2002 08:56:55 +0000 (08:56 +0000)]
[project @ 2002-10-15 08:56:50 by simonpj]
Some reorganising.  Simon and I agreed to leave out most of the (inaccurate) section
on primitive operations, pointing people instead to the Real Truth in primops.txt
and the External Core document.

Also: bugs section added

21 years ago[project @ 2002-10-14 15:50:14 by sof]
sof [Mon, 14 Oct 2002 15:50:14 +0000 (15:50 +0000)]
[project @ 2002-10-14 15:50:14 by sof]
make it compile again (x86)

21 years ago[project @ 2002-10-14 14:54:16 by simonmar]
simonmar [Mon, 14 Oct 2002 14:54:16 +0000 (14:54 +0000)]
[project @ 2002-10-14 14:54:16 by simonmar]
Put stdin back into non-blocking mode after calling readline, because
readline sometimes clears the O_NONBLOCK flag (done using the
GHC.Posix interface, so it shouldn't break Win32).

21 years ago[project @ 2002-10-14 14:44:06 by sof]
sof [Mon, 14 Oct 2002 14:44:06 +0000 (14:44 +0000)]
[project @ 2002-10-14 14:44:06 by sof]
make it compile again with 4.08.x

21 years ago[project @ 2002-10-13 10:55:06 by wolfgang]
wolfgang [Sun, 13 Oct 2002 10:55:06 +0000 (10:55 +0000)]
[project @ 2002-10-13 10:55:06 by wolfgang]
Darwin/PowerPC: Don't generate PIC code by default
Non-PIC-code is slightly smaller and faster.
This means that GHC now requires GCC3 (Mac OS X 10.2 Jaguar).

MERGE TO STABLE

21 years ago[project @ 2002-10-12 23:28:48 by wolfgang]
wolfgang [Sat, 12 Oct 2002 23:28:51 +0000 (23:28 +0000)]
[project @ 2002-10-12 23:28:48 by wolfgang]
The Native Code Generator for PowerPC.
Still to be done:
*) Proper support of Floats and Doubles
   currently it seems to work, but it's just guesswork.
*) Some missing operations, only needed for -O, AFAICT.
*) Mach-O dynamic linker stub generation.
   (can't import foreign functions from dynamic libraries,
   and it might fail for big programs)

21 years ago[project @ 2002-10-12 23:19:54 by wolfgang]
wolfgang [Sat, 12 Oct 2002 23:19:54 +0000 (23:19 +0000)]
[project @ 2002-10-12 23:19:54 by wolfgang]
Object Splitting for Mac OS X.

MERGE TO STABLE (... but not today, I have to test it first)

21 years ago[project @ 2002-10-12 23:12:08 by wolfgang]
wolfgang [Sat, 12 Oct 2002 23:12:12 +0000 (23:12 +0000)]
[project @ 2002-10-12 23:12:08 by wolfgang]
Make the Mac OS X build use the HaskellSupport.framework (a MacOS-style "framework" that includes the required libraries libgmp and dlcompat) if it is present. The HaskellSupport.framework is not yet in CVS, but is available from me.

21 years ago[project @ 2002-10-12 18:16:11 by erkok]
erkok [Sat, 12 Oct 2002 18:16:11 +0000 (18:16 +0000)]
[project @ 2002-10-12 18:16:11 by erkok]
mdo wibbles

21 years ago[project @ 2002-10-11 16:56:09 by simonpj]
simonpj [Fri, 11 Oct 2002 16:56:09 +0000 (16:56 +0000)]
[project @ 2002-10-11 16:56:09 by simonpj]
Document infix type constructors

21 years ago[project @ 2002-10-11 16:49:56 by simonpj]
simonpj [Fri, 11 Oct 2002 16:49:56 +0000 (16:49 +0000)]
[project @ 2002-10-11 16:49:56 by simonpj]
title wibble

21 years ago[project @ 2002-10-11 16:45:16 by simonpj]
simonpj [Fri, 11 Oct 2002 16:45:20 +0000 (16:45 +0000)]
[project @ 2002-10-11 16:45:16 by simonpj]
More reification wibbling; and -ddump-splices

21 years ago[project @ 2002-10-11 16:42:44 by simonpj]
simonpj [Fri, 11 Oct 2002 16:42:44 +0000 (16:42 +0000)]
[project @ 2002-10-11 16:42:44 by simonpj]
Docs for Template Haskell

21 years ago[project @ 2002-10-11 15:45:05 by simonpj]
simonpj [Fri, 11 Oct 2002 15:45:05 +0000 (15:45 +0000)]
[project @ 2002-10-11 15:45:05 by simonpj]
Wibble

21 years ago[project @ 2002-10-11 14:46:02 by simonpj]
simonpj [Fri, 11 Oct 2002 14:46:09 +0000 (14:46 +0000)]
[project @ 2002-10-11 14:46:02 by simonpj]
------------------------------------------
Implement reification for Template Haskell
------------------------------------------

This is entirely un-tested, but I don't think it'll break non-TH stuff.

Implements
reifyDecl T :: Dec -- Data type T
reifyDecl C :: Dec -- Class C
reifyType f :: Typ -- Function f

I hope.

21 years ago[project @ 2002-10-11 14:25:24 by stolz]
stolz [Fri, 11 Oct 2002 14:25:25 +0000 (14:25 +0000)]
[project @ 2002-10-11 14:25:24 by stolz]
'usleep' nightmare: Sometimes return type is void, sometimes int.

21 years ago[project @ 2002-10-11 08:48:13 by simonpj]
simonpj [Fri, 11 Oct 2002 08:48:13 +0000 (08:48 +0000)]
[project @ 2002-10-11 08:48:13 by simonpj]
wibbles

21 years ago[project @ 2002-10-11 08:47:12 by simonpj]
simonpj [Fri, 11 Oct 2002 08:47:13 +0000 (08:47 +0000)]
[project @ 2002-10-11 08:47:12 by simonpj]
Fix mdo so that it works with polymorphic functions

21 years ago[project @ 2002-10-11 08:46:26 by simonpj]
simonpj [Fri, 11 Oct 2002 08:46:26 +0000 (08:46 +0000)]
[project @ 2002-10-11 08:46:26 by simonpj]
Remove usage-analysis guff in applyTys

21 years ago[project @ 2002-10-11 08:45:43 by simonpj]
simonpj [Fri, 11 Oct 2002 08:45:44 +0000 (08:45 +0000)]
[project @ 2002-10-11 08:45:43 by simonpj]
Compatibility fixes for Exception.try

21 years ago[project @ 2002-10-11 08:04:55 by simonpj]
simonpj [Fri, 11 Oct 2002 08:04:55 +0000 (08:04 +0000)]
[project @ 2002-10-11 08:04:55 by simonpj]
Add an example about -osuf etc

21 years ago[project @ 2002-10-11 08:04:29 by simonpj]
simonpj [Fri, 11 Oct 2002 08:04:29 +0000 (08:04 +0000)]
[project @ 2002-10-11 08:04:29 by simonpj]
Add a note about existential data constructors

21 years ago[project @ 2002-10-11 07:22:05 by simonpj]
simonpj [Fri, 11 Oct 2002 07:22:05 +0000 (07:22 +0000)]
[project @ 2002-10-11 07:22:05 by simonpj]
Fix two separate egregious errors in RdrHsSyn, which I
heavily modified when re-doing the top level plumbing.

One had the effect of throwing away fixity decls if there
was also a class decl

The other threw away all but the first and last equation of
a function definition.

Sorry about having so utterly broken the head, guys.

21 years ago[project @ 2002-10-10 15:14:37 by sof]
sof [Thu, 10 Oct 2002 15:14:37 +0000 (15:14 +0000)]
[project @ 2002-10-10 15:14:37 by sof]
rules: permit trailing semi-colon(s) again (inserted by the lexer & needed when compiling Base.lhs)

21 years ago[project @ 2002-10-10 14:35:55 by sof]
sof [Thu, 10 Oct 2002 14:35:55 +0000 (14:35 +0000)]
[project @ 2002-10-10 14:35:55 by sof]
make it compile with 4.08 (module author may very reasonably want to abstract away the ifdef this commit intros.)

21 years ago[project @ 2002-10-10 14:20:54 by sof]
sof [Thu, 10 Oct 2002 14:20:54 +0000 (14:20 +0000)]
[project @ 2002-10-10 14:20:54 by sof]
make it compile in non-DEBUG mode

21 years ago[project @ 2002-10-10 13:58:18 by sof]
sof [Thu, 10 Oct 2002 13:58:18 +0000 (13:58 +0000)]
[project @ 2002-10-10 13:58:18 by sof]
partial undo of prev commit (AC_CHECK_HEADERS([foo],..) isn't equal to AC_CHECK_HEADER([foo],..). Fixes HEAD breakage (cf. _SC_CLK_TCK undefinedness)

21 years ago[project @ 2002-10-09 16:53:10 by simonpj]
simonpj [Wed, 9 Oct 2002 16:53:13 +0000 (16:53 +0000)]
[project @ 2002-10-09 16:53:10 by simonpj]
Fix to mdo, plus SrcLocs on splices and brackets

21 years ago[project @ 2002-10-09 16:35:56 by simonpj]
simonpj [Wed, 9 Oct 2002 16:35:56 +0000 (16:35 +0000)]
[project @ 2002-10-09 16:35:56 by simonpj]
Import Exception.Try via HsVersions

21 years ago[project @ 2002-10-09 16:28:56 by simonpj]
simonpj [Wed, 9 Oct 2002 16:28:57 +0000 (16:28 +0000)]
[project @ 2002-10-09 16:28:56 by simonpj]
Fix and simplify renamer warnings about export lists

21 years ago[project @ 2002-10-09 15:51:43 by simonpj]
simonpj [Wed, 9 Oct 2002 15:51:43 +0000 (15:51 +0000)]
[project @ 2002-10-09 15:51:43 by simonpj]
Dont try to splice if compiler isnt bootstrapped

21 years ago[project @ 2002-10-09 15:37:57 by simonpj]
simonpj [Wed, 9 Oct 2002 15:37:57 +0000 (15:37 +0000)]
[project @ 2002-10-09 15:37:57 by simonpj]
Missing eqn in getMonoBind

21 years ago[project @ 2002-10-09 15:36:47 by simonpj]
simonpj [Wed, 9 Oct 2002 15:36:48 +0000 (15:36 +0000)]
[project @ 2002-10-09 15:36:47 by simonpj]
Remove circular dependencies

21 years ago[project @ 2002-10-09 15:16:53 by simonpj]
simonpj [Wed, 9 Oct 2002 15:16:53 +0000 (15:16 +0000)]
[project @ 2002-10-09 15:16:53 by simonpj]
Dont complain about unused things bound by deriving clause

21 years ago[project @ 2002-10-09 15:15:59 by simonpj]
simonpj [Wed, 9 Oct 2002 15:15:59 +0000 (15:15 +0000)]
[project @ 2002-10-09 15:15:59 by simonpj]
Bale out if duplicate binders

21 years ago[project @ 2002-10-09 15:05:05 by simonpj]
simonpj [Wed, 9 Oct 2002 15:05:05 +0000 (15:05 +0000)]
[project @ 2002-10-09 15:05:05 by simonpj]
---------------------
PS: remove ParseUtils
---------------------

I've combined ParseUtils into RdrHsSyn.  I could never
figure out which thing was defined in which module, and
they were both short.

21 years ago[project @ 2002-10-09 15:03:48 by simonpj]
simonpj [Wed, 9 Oct 2002 15:04:01 +0000 (15:04 +0000)]
[project @ 2002-10-09 15:03:48 by simonpj]
-----------------------------------
Lots more Template Haskell stuff
-----------------------------------

At last!  Top-level declaration splices work!
Syntax is

$(f x)

not "splice (f x)" as in the paper.

Lots jiggling around, particularly with the top-level plumbining.
Note the new data type HsDecls.HsGroup.

21 years ago[project @ 2002-10-08 08:03:01 by wolfgang]
wolfgang [Tue, 8 Oct 2002 08:03:02 +0000 (08:03 +0000)]
[project @ 2002-10-08 08:03:01 by wolfgang]
Make the new Posix bindings compile on Mac OS X.
Most notable, Mac OS X lacks
*) lchown
*) SIGPOLL
I don't know of a replacement of either, so they are just left out when
they are not detected by configure.

21 years ago[project @ 2002-10-06 12:45:41 by wolfgang]
wolfgang [Sun, 6 Oct 2002 12:45:41 +0000 (12:45 +0000)]
[project @ 2002-10-06 12:45:41 by wolfgang]
Make it compile for Mac OS X again.

21 years ago[project @ 2002-10-05 23:54:53 by panne]
panne [Sat, 5 Oct 2002 23:54:54 +0000 (23:54 +0000)]
[project @ 2002-10-05 23:54:53 by panne]
The never-ending docbook tragedy: Adapt the catalogs once more... >:-P
NOTE: I don't really have a clue what I'm doing here, but at least it
works with the current SuSE Linux distribution 8.1.

21 years ago[project @ 2002-10-05 22:31:04 by panne]
panne [Sat, 5 Oct 2002 22:31:04 +0000 (22:31 +0000)]
[project @ 2002-10-05 22:31:04 by panne]
Warning police #13: Sync fprintf format strings with args in an
architecture-neutral way.

21 years ago[project @ 2002-10-05 22:18:45 by panne]
panne [Sat, 5 Oct 2002 22:18:46 +0000 (22:18 +0000)]
[project @ 2002-10-05 22:18:45 by panne]
Warning police #12: Multi-line string literals are uncool.

21 years ago[project @ 2002-10-02 09:36:00 by wolfgang]
wolfgang [Wed, 2 Oct 2002 09:36:00 +0000 (09:36 +0000)]
[project @ 2002-10-02 09:36:00 by wolfgang]
PowerPC Mach-O:
Handle more kinds of relocations, so that non-PIC-code can be loaded

21 years ago[project @ 2002-10-02 09:08:44 by wolfgang]
wolfgang [Wed, 2 Oct 2002 09:08:44 +0000 (09:08 +0000)]
[project @ 2002-10-02 09:08:44 by wolfgang]
PowerPC-related changes:
Use proper register names for floating-point registers
Define SpillTmp registers for NCG

21 years ago[project @ 2002-10-02 02:04:44 by erkok]
erkok [Wed, 2 Oct 2002 02:04:44 +0000 (02:04 +0000)]
[project @ 2002-10-02 02:04:44 by erkok]
typos never end.

21 years ago[project @ 2002-10-01 18:10:00 by erkok]
erkok [Tue, 1 Oct 2002 18:10:00 +0000 (18:10 +0000)]
[project @ 2002-10-01 18:10:00 by erkok]
typo

21 years ago[project @ 2002-10-01 17:48:37 by erkok]
erkok [Tue, 1 Oct 2002 17:48:37 +0000 (17:48 +0000)]
[project @ 2002-10-01 17:48:37 by erkok]
wibbles on mdo-documentation

21 years ago[project @ 2002-10-01 16:28:20 by simonpj]
simonpj [Tue, 1 Oct 2002 16:28:20 +0000 (16:28 +0000)]
[project @ 2002-10-01 16:28:20 by simonpj]
Wibble in loading of boot interfaces

21 years ago[project @ 2002-10-01 15:59:03 by erkok]
erkok [Tue, 1 Oct 2002 15:59:03 +0000 (15:59 +0000)]
[project @ 2002-10-01 15:59:03 by erkok]
MonadRec => MonadFix changes

21 years ago[project @ 2002-10-01 15:58:42 by erkok]
erkok [Tue, 1 Oct 2002 15:58:42 +0000 (15:58 +0000)]
[project @ 2002-10-01 15:58:42 by erkok]
get mfix from Control.Monad.Fix

21 years ago[project @ 2002-10-01 09:55:38 by simonpj]
simonpj [Tue, 1 Oct 2002 09:55:38 +0000 (09:55 +0000)]
[project @ 2002-10-01 09:55:38 by simonpj]
Better derived Ord code

21 years ago[project @ 2002-10-01 08:59:43 by simonpj]
simonpj [Tue, 1 Oct 2002 08:59:43 +0000 (08:59 +0000)]
[project @ 2002-10-01 08:59:43 by simonpj]
Use the ModSummary fields in record construction

21 years ago[project @ 2002-10-01 08:53:14 by simonmar]
simonmar [Tue, 1 Oct 2002 08:53:14 +0000 (08:53 +0000)]
[project @ 2002-10-01 08:53:14 by simonmar]
The help text for the -i profiling option was wrong: it claimed the
argument was in milliseconds rather than seconds.

MERGE TO STABLE

21 years ago[project @ 2002-09-30 13:42:01 by simonpj]
simonpj [Mon, 30 Sep 2002 13:42:01 +0000 (13:42 +0000)]
[project @ 2002-09-30 13:42:01 by simonpj]
Move mdo section

21 years ago[project @ 2002-09-30 13:38:30 by simonpj]
simonpj [Mon, 30 Sep 2002 13:38:30 +0000 (13:38 +0000)]
[project @ 2002-09-30 13:38:30 by simonpj]
Minor wibble

21 years ago[project @ 2002-09-30 10:53:21 by simonmar]
simonmar [Mon, 30 Sep 2002 10:53:21 +0000 (10:53 +0000)]
[project @ 2002-09-30 10:53:21 by simonmar]
.PHONY: package-way-dist