ghc-hetmet.git
20 years ago[project @ 2003-05-23 08:49:22 by simonmar]
simonmar [Fri, 23 May 2003 08:49:22 +0000 (08:49 +0000)]
[project @ 2003-05-23 08:49:22 by simonmar]
Make hackery to give $(GhcCanonVersion) the right value even when we
have a single-digit $(GhcMinVersion).

20 years ago[project @ 2003-05-23 08:28:48 by simonmar]
simonmar [Fri, 23 May 2003 08:28:48 +0000 (08:28 +0000)]
[project @ 2003-05-23 08:28:48 by simonmar]
Remove ancient vestiges of StgHugs: the COMPILER #define.

20 years ago[project @ 2003-05-22 15:36:08 by simonmar]
simonmar [Thu, 22 May 2003 15:36:08 +0000 (15:36 +0000)]
[project @ 2003-05-22 15:36:08 by simonmar]
First hack at supporting standalone packages.

What this means is that a non-GHC fptools project which sets
$(PACKAGE) in a Makefile should do something reasonable.  It works by
testing whether $(ProjectNameShort) == "ghc", and if it isn't, then:

 - it turns off all the machinery that deals with installing the
   package for use with the in-place compiler and a (future)
   installation of GHC.

 - instead, it enables a rule that installs the package using
   the installed $(GHC_PKG).  The configure script still needs
   updating to detect this, so 'make install' won't work yet.

 - $(GHC) is used instead of $(GHC_INPLACE) for compiling Haskell
   code.

Apart from the detection of ghc-pkg in configure, this should be all
that's needed to get source distributions and 'make install' going.

20 years ago[project @ 2003-05-22 13:22:39 by simonmar]
simonmar [Thu, 22 May 2003 13:22:39 +0000 (13:22 +0000)]
[project @ 2003-05-22 13:22:39 by simonmar]
Comment only

20 years ago[project @ 2003-05-22 10:53:53 by simonmar]
simonmar [Thu, 22 May 2003 10:53:53 +0000 (10:53 +0000)]
[project @ 2003-05-22 10:53:53 by simonmar]
Fix obscure bug in GHCi: when generating code for tag2enum#, we were
wrongly using the source name for the DataCons rather than the worker
name, which lead to spurious link errors.

This fixes galois_raytrace(ghci).

20 years ago[project @ 2003-05-22 10:27:52 by ross]
ross [Thu, 22 May 2003 10:27:52 +0000 (10:27 +0000)]
[project @ 2003-05-22 10:27:52 by ross]
Make it work with newer versions of flex, which produce an extern
declaration for yywrap() unless YY_SKIP_YYWRAP is defined.

20 years ago[project @ 2003-05-21 23:40:08 by igloo]
igloo [Wed, 21 May 2003 23:40:10 +0000 (23:40 +0000)]
[project @ 2003-05-21 23:40:08 by igloo]
Rename and reorder the internals for unique ids etc. Also fixed a couple
of THSyntax names.

20 years ago[project @ 2003-05-21 18:07:13 by igloo]
igloo [Wed, 21 May 2003 18:07:30 +0000 (18:07 +0000)]
[project @ 2003-05-21 18:07:13 by igloo]
The Great Renaming. I hope I've kept everything in sync - and all the tests
pass. Now datatypes follow the

data Foo = <a kind of Foo>Foo
         | <another kind of Foo>Foo

convention and the smart constructors can be uniformly created by
lowercasing the first letter of the constructor.

20 years ago[project @ 2003-05-21 16:02:42 by stolz]
stolz [Wed, 21 May 2003 16:02:44 +0000 (16:02 +0000)]
[project @ 2003-05-21 16:02:42 by stolz]
Solaris2 needs _POSIX_PTHREAD_SEMANTICS for the getpw*_r() prototypes.
Make libraries/unix/Makefile use a new variable unix_SRC_HSC2HS_OPTS
which we configure in mk/config.mk.

20 years ago[project @ 2003-05-21 15:49:54 by simonpj]
simonpj [Wed, 21 May 2003 15:49:54 +0000 (15:49 +0000)]
[project @ 2003-05-21 15:49:54 by simonpj]
Add a microscopic amount of documentation for functional dependencies; who would like to volunteer to write more?

20 years ago[project @ 2003-05-21 15:20:37 by simonmar]
simonmar [Wed, 21 May 2003 15:20:37 +0000 (15:20 +0000)]
[project @ 2003-05-21 15:20:37 by simonmar]
The 6.0 release notes.

20 years ago[project @ 2003-05-21 13:22:46 by stolz]
stolz [Wed, 21 May 2003 13:22:46 +0000 (13:22 +0000)]
[project @ 2003-05-21 13:22:46 by stolz]
Call strerror() instead of sys_errlist[errno] (deprecated & smells bad)
(Broke Solaris-build here)

20 years ago[project @ 2003-05-21 13:05:49 by simonmar]
simonmar [Wed, 21 May 2003 13:05:49 +0000 (13:05 +0000)]
[project @ 2003-05-21 13:05:49 by simonmar]
Restore the correct partitioning of command-line arguments into
objects and compilation-manager inputs for --make mode, and restore a
comment explaining the behaviour.

Rev. 1.120 made some subtle changes to the semantics, in particular in
cases where a file given on the command line is neither a source nor
an object file.

I believe the behaviour for one-shot mode has not changed.  The
behaviour for GHCi mode is now the same as --make mode (previous to
rev. 1.120 it was subtly broken in this respect).

20 years ago[project @ 2003-05-21 12:46:19 by simonmar]
simonmar [Wed, 21 May 2003 12:46:19 +0000 (12:46 +0000)]
[project @ 2003-05-21 12:46:19 by simonmar]
revert rev. 1.16; looks like this was an experiment that got
accidentally committed.

20 years ago[project @ 2003-05-21 12:38:36 by simonmar]
simonmar [Wed, 21 May 2003 12:38:37 +0000 (12:38 +0000)]
[project @ 2003-05-21 12:38:36 by simonmar]
Put the call to saveDynFlags back in one place.

20 years ago[project @ 2003-05-21 10:50:55 by reid]
reid [Wed, 21 May 2003 10:51:31 +0000 (10:51 +0000)]
[project @ 2003-05-21 10:50:55 by reid]
Improved greencard-related definitions

20 years ago[project @ 2003-05-21 02:58:39 by igloo]
igloo [Wed, 21 May 2003 02:58:40 +0000 (02:58 +0000)]
[project @ 2003-05-21 02:58:39 by igloo]
Added support for newtypes to TH and altered a test for them.

20 years ago[project @ 2003-05-21 02:48:56 by igloo]
igloo [Wed, 21 May 2003 02:48:56 +0000 (02:48 +0000)]
[project @ 2003-05-21 02:48:56 by igloo]
When --make is used we need to save the initial dynamic flags too

20 years ago[project @ 2003-05-20 22:39:33 by igloo]
igloo [Tue, 20 May 2003 22:39:33 +0000 (22:39 +0000)]
[project @ 2003-05-20 22:39:33 by igloo]
Fix building on 5.02

20 years ago[project @ 2003-05-20 11:07:54 by stolz]
stolz [Tue, 20 May 2003 11:07:54 +0000 (11:07 +0000)]
[project @ 2003-05-20 11:07:54 by stolz]
Generate
  int main (int argc, char *argv [])
instead of
  int main (void)

Suggested by: Nicolas Oury

20 years ago[project @ 2003-05-20 10:03:25 by simonmar]
simonmar [Tue, 20 May 2003 10:03:25 +0000 (10:03 +0000)]
[project @ 2003-05-20 10:03:25 by simonmar]
Skeleton 6.0 release notes.

20 years ago[project @ 2003-05-20 10:03:06 by simonmar]
simonmar [Tue, 20 May 2003 10:03:06 +0000 (10:03 +0000)]
[project @ 2003-05-20 10:03:06 by simonmar]
defaultsHook is now ghc_rts_opts.

20 years ago[project @ 2003-05-20 10:02:13 by simonmar]
simonmar [Tue, 20 May 2003 10:02:13 +0000 (10:02 +0000)]
[project @ 2003-05-20 10:02:13 by simonmar]
Update version numbering policy

20 years ago[project @ 2003-05-20 09:54:42 by simonmar]
simonmar [Tue, 20 May 2003 09:56:58 +0000 (09:56 +0000)]
[project @ 2003-05-20 09:54:42 by simonmar]
Version 6.0

20 years ago[project @ 2003-05-20 03:46:48 by sof]
sof [Tue, 20 May 2003 03:46:48 +0000 (03:46 +0000)]
[project @ 2003-05-20 03:46:48 by sof]
--with-gcc=<path>: for mingw32 plat, canonicalise path using cygpath

20 years ago[project @ 2003-05-19 15:39:17 by simonpj]
simonpj [Mon, 19 May 2003 15:39:18 +0000 (15:39 +0000)]
[project @ 2003-05-19 15:39:17 by simonpj]
---------------------------------
Improve the dynamic-linking story
---------------------------------

Arrange proper initialisation for the dynamic linker.  Whenever the dynamic linker does
anything (Linker.linkExpr), it first initialises itself, arranging to only do so once
of course.

"Initialising itself" includes loading any .o files, libraries, and packages specified
on the command line.  The main effect of all this is to fix a Template Haskell problem,
which happened when a TH link needed some C library that it couldn't link.  Now it does.

While I was at it, I tidied up main/Main.hs quite a bit.  This is a delicate area (handling
the command line arguments), but I don't believe I broke anything!  All the libraries
build, and lots of tests run.

Wolfgang: I also jiggled the darwin_TARGET_OS stuff a little in Linker.lhs, but again
I think the net effect is zero. You might want to check.

20 years ago[project @ 2003-05-19 15:11:15 by simonpj]
simonpj [Mon, 19 May 2003 15:11:15 +0000 (15:11 +0000)]
[project @ 2003-05-19 15:11:15 by simonpj]
Comments only

20 years ago[project @ 2003-05-19 15:10:40 by simonpj]
simonpj [Mon, 19 May 2003 15:10:41 +0000 (15:10 +0000)]
[project @ 2003-05-19 15:10:40 by simonpj]
--------------------------
Minor Template Haskell bug
--------------------------

This bug meant that spliced-in class declarations yielded a 'op not in scope',
where op was the class operation.  Thanks to Andre Pang for spotting this.

Some consequential tidying up in parsing too.

20 years ago[project @ 2003-05-19 13:29:15 by simonmar]
simonmar [Mon, 19 May 2003 13:29:15 +0000 (13:29 +0000)]
[project @ 2003-05-19 13:29:15 by simonmar]
- rearrange the file a little

- build a binary distribution from stage 2 by default, unless
  overriden from the command line.

20 years ago[project @ 2003-05-19 13:01:23 by simonmar]
simonmar [Mon, 19 May 2003 13:01:23 +0000 (13:01 +0000)]
[project @ 2003-05-19 13:01:23 by simonmar]
Remove obsolete references to DriFT, DtdToHaskell and Xtract.

20 years ago[project @ 2003-05-19 12:11:14 by simonmar]
simonmar [Mon, 19 May 2003 12:11:14 +0000 (12:11 +0000)]
[project @ 2003-05-19 12:11:14 by simonmar]
tryM again: make it work with GHC 5.02.x (I hope).  Seems we changed
the representation of UserErrors several times in the last few
releases...

20 years ago[project @ 2003-05-19 11:37:07 by simonmar]
simonmar [Mon, 19 May 2003 11:37:07 +0000 (11:37 +0000)]
[project @ 2003-05-19 11:37:07 by simonmar]
Fix a pair of off-by-one errors that meant that BCOs were normally two
words larger than they needed to be.

20 years ago[project @ 2003-05-19 11:33:06 by simonmar]
simonmar [Mon, 19 May 2003 11:33:06 +0000 (11:33 +0000)]
[project @ 2003-05-19 11:33:06 by simonmar]
Eeek!  intsToReverseBitmap was borked for bitmaps with more than
wORD_SIZE_IN_BITS entries.

20 years ago[project @ 2003-05-19 08:31:11 by simonpj]
simonpj [Mon, 19 May 2003 08:31:11 +0000 (08:31 +0000)]
[project @ 2003-05-19 08:31:11 by simonpj]
More Win32 help

20 years ago[project @ 2003-05-17 14:49:45 by reid]
reid [Sat, 17 May 2003 14:51:28 +0000 (14:51 +0000)]
[project @ 2003-05-17 14:49:45 by reid]
Add checks for ghc-pkg and green-card 3.00 (used in x11 lib)

20 years ago[project @ 2003-05-17 00:11:27 by ross]
ross [Sat, 17 May 2003 00:11:30 +0000 (00:11 +0000)]
[project @ 2003-05-17 00:11:27 by ross]
Rename per-package configuration files from $(PACKAGE).conf.* to
package.conf.*, making them easier to find (since each package is
in a separate directory anyway).

20 years ago[project @ 2003-05-16 14:39:29 by simonmar]
simonmar [Fri, 16 May 2003 14:39:29 +0000 (14:39 +0000)]
[project @ 2003-05-16 14:39:29 by simonmar]
Make retainer profiling work again after the recent SRT changes.

20 years ago[project @ 2003-05-16 14:16:53 by simonmar]
simonmar [Fri, 16 May 2003 14:16:53 +0000 (14:16 +0000)]
[project @ 2003-05-16 14:16:53 by simonmar]
heapCensus should grok IND_OLDGEN objects, because compacting GC
doesn't always eliminate them (perhaps it should).

20 years ago[project @ 2003-05-16 10:14:20 by simonmar]
simonmar [Fri, 16 May 2003 10:14:24 +0000 (10:14 +0000)]
[project @ 2003-05-16 10:14:20 by simonmar]
Now that we have auto packages, it makes sense to keep all the
interfaces for hierarchical libraries in the same directory tree.  So
now, instead of putting interfaces for package P in $libdir/imports/P,
we put them all in $libdir/imports.

Interfaces for old non-auto non-hierarchical packages now go in
$libdir/hslibs-imports/P for package P.

20 years ago[project @ 2003-05-16 08:38:59 by simonmar]
simonmar [Fri, 16 May 2003 08:38:59 +0000 (08:38 +0000)]
[project @ 2003-05-16 08:38:59 by simonmar]
Doc the -L ghc-pkg option.

20 years ago[project @ 2003-05-16 08:35:49 by simonmar]
simonmar [Fri, 16 May 2003 08:35:49 +0000 (08:35 +0000)]
[project @ 2003-05-16 08:35:49 by simonmar]
Add -L (--list-packages-local) flag which lists packages in the
topmost configuration file only.

20 years ago[project @ 2003-05-15 12:28:29 by ross]
ross [Thu, 15 May 2003 12:28:29 +0000 (12:28 +0000)]
[project @ 2003-05-15 12:28:29 by ross]
update, including a short note on Nhc98.

20 years ago[project @ 2003-05-14 11:55:11 by simonmar]
simonmar [Wed, 14 May 2003 11:55:11 +0000 (11:55 +0000)]
[project @ 2003-05-14 11:55:11 by simonmar]
Should have been committed with recent SRT changes (see rev. 1.39 of StgSyn.lhs).

20 years ago[project @ 2003-05-14 09:13:52 by simonmar]
simonmar [Wed, 14 May 2003 09:14:02 +0000 (09:14 +0000)]
[project @ 2003-05-14 09:13:52 by simonmar]
Change the way SRTs are represented:

Previously, the SRT associated with a function or thunk would be a
sub-list of the enclosing top-level function's SRT.  But this approach
can lead to lots of duplication: if a CAF is referenced in several
different thunks, then it may appear several times in the SRT.
Let-no-escapes compound the problem, because the occurrence of a
let-no-escape-bound variable would expand to all the CAFs referred to
by the let-no-escape.

The new way is to describe the SRT associated with a function or thunk
as a (pointer+offset,bitmap) pair, where the pointer+offset points
into some SRT table (the enclosing function's SRT), and the bitmap
indicates which entries in this table are "live" for this closure.
The bitmap is stored in the 16 bits previously used for the length
field, but this rarely overflows.  When it does overflow, we store the
bitmap externally in a new "SRT descriptor".

Now the enclosing SRT can be a set, hence eliminating the duplicates.

Also, we now have one SRT per top-level function in a recursive group,
where previously we used to have one SRT for the whole group.  This
helps keep the size of SRTs down.

Bottom line: very little difference most of the time.  GHC itself got
slightly smaller.  One bad case of a module in GHC which had a huge
SRT has gone away.

While I was in the area:

  - Several parts of the back-end require bitmaps.  Functions for
    creating bitmaps are now centralised in the Bitmap module.

  - We were trying to be independent of word-size in a couple of
    places in the back end, but we've now abandoned that strategy so I
    simplified things a bit.

20 years ago[project @ 2003-05-14 09:11:49 by simonmar]
simonmar [Wed, 14 May 2003 09:11:49 +0000 (09:11 +0000)]
[project @ 2003-05-14 09:11:49 by simonmar]
Comments only

20 years ago[project @ 2003-05-14 09:11:06 by simonmar]
simonmar [Wed, 14 May 2003 09:11:06 +0000 (09:11 +0000)]
[project @ 2003-05-14 09:11:06 by simonmar]
Calling mmap() with a size > 3Gig results in EINVAL on Linux.  Catch
this case and report it as an out of memory condition instead of a
panic.

20 years ago[project @ 2003-05-14 09:07:28 by simonmar]
simonmar [Wed, 14 May 2003 09:07:28 +0000 (09:07 +0000)]
[project @ 2003-05-14 09:07:28 by simonmar]
Comment only

20 years ago[project @ 2003-05-13 15:37:23 by simonpj]
simonpj [Tue, 13 May 2003 15:37:23 +0000 (15:37 +0000)]
[project @ 2003-05-13 15:37:23 by simonpj]
Add optimistic-eval page; not linked yet

20 years ago[project @ 2003-05-13 08:17:13 by simonpj]
simonpj [Tue, 13 May 2003 08:17:13 +0000 (08:17 +0000)]
[project @ 2003-05-13 08:17:13 by simonpj]
Improve notes on Cygwin installation

20 years ago[project @ 2003-05-12 14:12:52 by simonmar]
simonmar [Mon, 12 May 2003 14:12:52 +0000 (14:12 +0000)]
[project @ 2003-05-12 14:12:52 by simonmar]
Another bugfix to tryM, this time for GHC 5.00--5.04

20 years ago[project @ 2003-05-12 09:22:35 by simonmar]
simonmar [Mon, 12 May 2003 09:22:35 +0000 (09:22 +0000)]
[project @ 2003-05-12 09:22:35 by simonmar]
Updates:

  - Loading module Main without a main binding is no longer an error.

  - Clarify the section on libraries to note that it means object-code
    libraries, not source code.

  - Update the section on packages now that most packages are
    auto-loaded.

20 years ago[project @ 2003-05-07 16:20:21 by simonpj]
simonpj [Wed, 7 May 2003 16:20:21 +0000 (16:20 +0000)]
[project @ 2003-05-07 16:20:21 by simonpj]
Comment only

20 years ago[project @ 2003-05-07 09:30:09 by simonpj]
simonpj [Wed, 7 May 2003 09:30:09 +0000 (09:30 +0000)]
[project @ 2003-05-07 09:30:09 by simonpj]
Allow deriving(Show) for data types with unboxed fields

20 years ago[project @ 2003-05-07 08:30:08 by simonpj]
simonpj [Wed, 7 May 2003 08:30:08 +0000 (08:30 +0000)]
[project @ 2003-05-07 08:30:08 by simonpj]
Print type of data constructors correctly in GHCi

20 years ago[project @ 2003-05-07 08:29:42 by simonpj]
simonpj [Wed, 7 May 2003 08:29:48 +0000 (08:29 +0000)]
[project @ 2003-05-07 08:29:42 by simonpj]
Comments and formatting

20 years ago[project @ 2003-05-06 14:48:12 by simonpj]
simonpj [Tue, 6 May 2003 14:48:12 +0000 (14:48 +0000)]
[project @ 2003-05-06 14:48:12 by simonpj]
Note about context in data type decls

20 years ago[project @ 2003-05-06 10:36:17 by simonmar]
simonmar [Tue, 6 May 2003 10:36:17 +0000 (10:36 +0000)]
[project @ 2003-05-06 10:36:17 by simonmar]
Update: we now follow Haskell 98 w.r.t. exports from module Main.

20 years ago[project @ 2003-05-06 10:28:32 by simonpj]
simonpj [Tue, 6 May 2003 10:28:33 +0000 (10:28 +0000)]
[project @ 2003-05-06 10:28:32 by simonpj]
-------------------------------------
        Implement deriving( Data )
-------------------------------------

Implements deriving( Data ), where the Data class is defined
in Data.Generics; its the "scrap your boilerplate" Term class.

Ralf is still converging on the exact definition of the Data class,
so the details may change.

20 years ago[project @ 2003-05-06 10:26:48 by simonpj]
simonpj [Tue, 6 May 2003 10:26:48 +0000 (10:26 +0000)]
[project @ 2003-05-06 10:26:48 by simonpj]
Import wibbles

20 years ago[project @ 2003-05-06 10:25:32 by simonpj]
simonpj [Tue, 6 May 2003 10:26:29 +0000 (10:26 +0000)]
[project @ 2003-05-06 10:25:32 by simonpj]
-------------------------------------
Main module exports
-------------------------------------

Make it so that

        module Main where ....

exports everything defined in Main, as the report says it should.

20 years ago[project @ 2003-05-06 10:23:16 by simonpj]
simonpj [Tue, 6 May 2003 10:23:16 +0000 (10:23 +0000)]
[project @ 2003-05-06 10:23:16 by simonpj]
Formatting only

20 years ago[project @ 2003-05-06 10:22:54 by simonpj]
simonpj [Tue, 6 May 2003 10:22:54 +0000 (10:22 +0000)]
[project @ 2003-05-06 10:22:54 by simonpj]
Comments about eta expansion

20 years ago[project @ 2003-05-04 13:21:48 by igloo]
igloo [Sun, 4 May 2003 13:21:49 +0000 (13:21 +0000)]
[project @ 2003-05-04 13:21:48 by igloo]
Add support for unboxed Ints, Floats and Doubles to Template Haskell.

21 years ago[project @ 2003-05-01 10:46:19 by ross]
ross [Thu, 1 May 2003 10:46:19 +0000 (10:46 +0000)]
[project @ 2003-05-01 10:46:19 by ross]
Make -t option to the script override the template file, just as -c
overrides the compiler, as the the documentation implies.  (Formerly -t
added an extra template file, which is not so useful.)

21 years ago[project @ 2003-05-01 10:25:38 by simonmar]
simonmar [Thu, 1 May 2003 10:25:38 +0000 (10:25 +0000)]
[project @ 2003-05-01 10:25:38 by simonmar]
Use the DocBook 4.1 DTD instead of 3.1.  Seems to work fine...

21 years ago[project @ 2003-05-01 08:43:54 by simonmar]
simonmar [Thu, 1 May 2003 08:43:54 +0000 (08:43 +0000)]
[project @ 2003-05-01 08:43:54 by simonmar]
Remove special cases for the setting of $(libdir) and $(datadir) on
mingw32, after discussion with Sigbjorn.

21 years ago[project @ 2003-04-29 21:37:31 by wolfgang]
wolfgang [Tue, 29 Apr 2003 21:37:31 +0000 (21:37 +0000)]
[project @ 2003-04-29 21:37:31 by wolfgang]
Unbreak Mac OS X build (#ifdef mixup)

21 years ago[project @ 2003-04-29 09:32:20 by simonmar]
simonmar [Tue, 29 Apr 2003 09:32:20 +0000 (09:32 +0000)]
[project @ 2003-04-29 09:32:20 by simonmar]
update aLL_NON_PTRS for the new RET_DYN stack frame layout.

21 years ago[project @ 2003-04-28 10:02:15 by simonmar]
simonmar [Mon, 28 Apr 2003 10:02:15 +0000 (10:02 +0000)]
[project @ 2003-04-28 10:02:15 by simonmar]
Add a compile-time sanity check that RESERVED_STACK_WORDS is right.

21 years ago[project @ 2003-04-28 09:57:12 by simonmar]
simonmar [Mon, 28 Apr 2003 09:57:12 +0000 (09:57 +0000)]
[project @ 2003-04-28 09:57:12 by simonmar]
Comments to note the relationship between the RET_DYN stack frame size
and RESERVED_STACK_WORDS.

21 years ago[project @ 2003-04-28 09:55:20 by simonmar]
simonmar [Mon, 28 Apr 2003 09:55:20 +0000 (09:55 +0000)]
[project @ 2003-04-28 09:55:20 by simonmar]
Following the recent change to the layout of the StgRetDyn frame, we
now need to bump RESERVED_STACK_WORDS because this governs the amount
of room which is guaranteed to be available on the stack in the event
of a stack check failure.

This accounts for at least one cause of recent crashes in the HEAD.

21 years ago[project @ 2003-04-27 10:37:01 by qrczak]
qrczak [Sun, 27 Apr 2003 10:37:01 +0000 (10:37 +0000)]
[project @ 2003-04-27 10:37:01 by qrczak]
Strings in perl are compared by eq, not ==.
Fixes creating "nul" files on Unix.

21 years ago[project @ 2003-04-27 10:30:13 by qrczak]
qrczak [Sun, 27 Apr 2003 10:30:13 +0000 (10:30 +0000)]
[project @ 2003-04-27 10:30:13 by qrczak]
s/symbol link/symbolic link/

21 years ago[project @ 2003-04-25 20:58:09 by panne]
panne [Fri, 25 Apr 2003 20:58:09 +0000 (20:58 +0000)]
[project @ 2003-04-25 20:58:09 by panne]
Unbreak 2nd stage

21 years ago[project @ 2003-04-25 09:17:05 by simonmar]
simonmar [Fri, 25 Apr 2003 09:17:05 +0000 (09:17 +0000)]
[project @ 2003-04-25 09:17:05 by simonmar]
- Add a note about the incorrect handling of the 'module Main where'
  header.

- While I'm here: fix out of date stuff, reformat and generally tidy up.

21 years ago[project @ 2003-04-23 10:42:36 by simonmar]
simonmar [Wed, 23 Apr 2003 10:42:36 +0000 (10:42 +0000)]
[project @ 2003-04-23 10:42:36 by simonmar]
Make a leading '~' expand to $HOME in commands that load source files
(:load, :add) in addition to :cd which already had this behaviour.

21 years ago[project @ 2003-04-23 08:54:45 by simonmar]
simonmar [Wed, 23 Apr 2003 08:54:45 +0000 (08:54 +0000)]
[project @ 2003-04-23 08:54:45 by simonmar]
Oops, forgot to update this file after the RET_DYN changes.

21 years ago[project @ 2003-04-22 20:39:59 by igloo]
igloo [Tue, 22 Apr 2003 20:40:00 +0000 (20:40 +0000)]
[project @ 2003-04-22 20:39:59 by igloo]
Order declarations in reifications in order of source line number.
The bugs still there but it bites less often now...

Also remove the type parameterisation and do some type renaming as
discussed on the template-haskell list.

21 years ago[project @ 2003-04-22 16:25:08 by simonmar]
simonmar [Tue, 22 Apr 2003 16:25:12 +0000 (16:25 +0000)]
[project @ 2003-04-22 16:25:08 by simonmar]
Fix an obscure bug: the most general kind of heap check,
HEAP_CHECK_GEN(), is supposed to save the contents of *every* register
known to the STG machine (used in cases where we either can't figure
out which ones are live, or doing so would be too much hassle).  The
problem is that it wasn't saving the L1 register.

A slight complication arose in that saving the L1 register pushed the
size of the frame over the 16 words allowed for the size of the bitmap
stored in the frame, so I changed the layout of the frame a bit.
Describing all the registers using a single bitmap is overkill when
only 8 of them can actually be pointers, so now the bitmap is only 8
bits long and we always skip over a fixed number of non-ptr words to
account for all the non-ptr regs.  This is all described in StgMacros.h.

21 years ago[project @ 2003-04-22 11:44:17 by simonpj]
simonpj [Tue, 22 Apr 2003 11:44:17 +0000 (11:44 +0000)]
[project @ 2003-04-22 11:44:17 by simonpj]
Fix the context for derived Typeable instances

21 years ago[project @ 2003-04-22 09:30:52 by simonpj]
simonpj [Tue, 22 Apr 2003 09:30:53 +0000 (09:30 +0000)]
[project @ 2003-04-22 09:30:52 by simonpj]
Stage-2 wibbles to the Expected type changes

21 years ago[project @ 2003-04-21 14:45:28 by sof]
sof [Mon, 21 Apr 2003 14:45:28 +0000 (14:45 +0000)]
[project @ 2003-04-21 14:45:28 by sof]
drop curious *_debug_opts_prefix tables; no longer used

21 years ago[project @ 2003-04-18 09:45:32 by simonmar]
simonmar [Fri, 18 Apr 2003 09:45:32 +0000 (09:45 +0000)]
[project @ 2003-04-18 09:45:32 by simonmar]
Add a comment about the entry convention to stg_ap_* and why it's done
this way (thanks to Andy Moran for the prodding).

21 years ago[project @ 2003-04-18 09:40:10 by simonmar]
simonmar [Fri, 18 Apr 2003 09:40:10 +0000 (09:40 +0000)]
[project @ 2003-04-18 09:40:10 by simonmar]
Just noticed that all the stack checks in stg_ap_[1-8]_entry are too
low by one word.  Thanks to Andy Moran for (unwittingly) causing me peer
at this code.

21 years ago[project @ 2003-04-17 15:23:32 by simonpj]
simonpj [Thu, 17 Apr 2003 15:23:37 +0000 (15:23 +0000)]
[project @ 2003-04-17 15:23:32 by simonpj]
----------------------------------
Implement Typeable properly
----------------------------------

1.  Add 'deriving' for Typeable class. So you can say

data T a b = .... deriving( Typeable )

    At the moment you only get this if you ask for it. If you say
    nothing you get nothing.

2.  Implement Typeable better, with proper O(1) comparison of
    type representations

3.  Add the 'cast' operation described in 'Scrap your boilerplate'
    and use it.

4.  Consequence: need to move the definition of IOArray from
    Data.Array.IO.Internals to GHC.IOBase, where it joins IORef.
    This is necssary so that HashTable can be low down in the compilation
    hierarchy, and hence so can Dynamic.

WARNING: I'm not certain the imports in HashTable and Dynamic
 will all be right for Hugs and NHC. I hope you can
    fix them up.

21 years ago[project @ 2003-04-17 15:22:38 by simonpj]
simonpj [Thu, 17 Apr 2003 15:22:41 +0000 (15:22 +0000)]
[project @ 2003-04-17 15:22:38 by simonpj]
Comments

21 years ago[project @ 2003-04-17 14:12:15 by simonpj]
simonpj [Thu, 17 Apr 2003 14:12:15 +0000 (14:12 +0000)]
[project @ 2003-04-17 14:12:15 by simonpj]
Comments

21 years ago[project @ 2003-04-17 12:00:58 by simonmar]
simonmar [Thu, 17 Apr 2003 12:00:58 +0000 (12:00 +0000)]
[project @ 2003-04-17 12:00:58 by simonmar]
Comment fix

21 years ago[project @ 2003-04-17 07:13:54 by simonpj]
simonpj [Thu, 17 Apr 2003 07:13:54 +0000 (07:13 +0000)]
[project @ 2003-04-17 07:13:54 by simonpj]
Update hi-boot files

21 years ago[project @ 2003-04-16 13:34:13 by simonpj]
simonpj [Wed, 16 Apr 2003 13:34:17 +0000 (13:34 +0000)]
[project @ 2003-04-16 13:34:13 by simonpj]
----------------------------------
 Use the Infer/Check idea for typechecking higher-rank types
----------------------------------

The main idea is that

data Expected ty = Infer (TcRef ty) | Check ty

tcMonoExpr :: Expr -> Expected TcRhoType -> TcM Expra

This "Expected" type tells tcMonoExpr whether it's doing inference or
checking.  It replaces the "HoleTv" flavour of type variable.

This actually leads to slightly more lines of code, but it's much
clearer, and the new type distinctions showed up several subtle bugs
in the previous implementation.  It all arose out of writing the
prototype implementation for the paper.

Error messages wibble around a little bit.  I'm not quite certain why!  But the
changes look like improvements to me.

21 years ago[project @ 2003-04-15 14:37:12 by simonmar]
simonmar [Tue, 15 Apr 2003 14:37:12 +0000 (14:37 +0000)]
[project @ 2003-04-15 14:37:12 by simonmar]
Fix bug in forkProcess: it should be using RET_N rather than returing
directly.  Makes a difference for unregisterised only.

Remove dead code in atomicModifyMVarzh_fast.

21 years ago[project @ 2003-04-14 13:36:04 by simonpj]
simonpj [Mon, 14 Apr 2003 13:36:04 +0000 (13:36 +0000)]
[project @ 2003-04-14 13:36:04 by simonpj]
Guidance about tags on windows

21 years ago[project @ 2003-04-14 13:15:52 by simonpj]
simonpj [Mon, 14 Apr 2003 13:15:52 +0000 (13:15 +0000)]
[project @ 2003-04-14 13:15:52 by simonpj]
Remove -- option for HSTAGS for tags target, now that we are using hasktags instead

21 years ago[project @ 2003-04-14 08:52:07 by simonmar]
simonmar [Mon, 14 Apr 2003 08:52:07 +0000 (08:52 +0000)]
[project @ 2003-04-14 08:52:07 by simonmar]
We support hs_init() and friends now, so remove the point that says we
don't.

21 years ago[project @ 2003-04-12 16:38:23 by panne]
panne [Sat, 12 Apr 2003 16:38:23 +0000 (16:38 +0000)]
[project @ 2003-04-12 16:38:23 by panne]
Typo :-}

21 years ago[project @ 2003-04-12 16:27:24 by panne]
panne [Sat, 12 Apr 2003 16:27:24 +0000 (16:27 +0000)]
[project @ 2003-04-12 16:27:24 by panne]
Unbreak 2nd stage: GHC.Posix => System.Posix.Internals

21 years ago[project @ 2003-04-12 09:39:03 by panne]
panne [Sat, 12 Apr 2003 09:39:03 +0000 (09:39 +0000)]
[project @ 2003-04-12 09:39:03 by panne]
Fixed the -print arg to a $(FIND) command, so as to be more compatible
with shells.  :-)

21 years ago[project @ 2003-04-11 13:09:44 by simonmar]
simonmar [Fri, 11 Apr 2003 13:09:44 +0000 (13:09 +0000)]
[project @ 2003-04-11 13:09:44 by simonmar]
Add -print to some $(FIND) commands, so as to be more compatible with
non-GNU finds.

21 years ago[project @ 2003-04-11 13:04:37 by simonmar]
simonmar [Fri, 11 Apr 2003 13:04:37 +0000 (13:04 +0000)]
[project @ 2003-04-11 13:04:37 by simonmar]
The Id for a foreign import should be a LocalId, not a GlobalId.
Making it a GlobalId violates the invariant that all Ids defined in
the current module should be LocalIds until CoreTidy.

21 years ago[project @ 2003-04-11 12:51:45 by simonmar]
simonmar [Fri, 11 Apr 2003 12:51:45 +0000 (12:51 +0000)]
[project @ 2003-04-11 12:51:45 by simonmar]
Add an ASSERT to confirm an invariant: all Ids defined in this module
should be LocalIds until the CoreTidy phase.