ghc-hetmet.git
24 years ago[project @ 2000-04-06 13:37:30 by simonmar]
simonmar [Thu, 6 Apr 2000 13:37:30 +0000 (13:37 +0000)]
[project @ 2000-04-06 13:37:30 by simonmar]
a few updates

24 years ago[project @ 2000-04-06 13:37:10 by simonmar]
simonmar [Thu, 6 Apr 2000 13:37:10 +0000 (13:37 +0000)]
[project @ 2000-04-06 13:37:10 by simonmar]
fix ccall argument ordering

24 years ago[project @ 2000-04-06 13:31:09 by simonmar]
simonmar [Thu, 6 Apr 2000 13:31:09 +0000 (13:31 +0000)]
[project @ 2000-04-06 13:31:09 by simonmar]
remove instance for CCallable [Char]

24 years ago[project @ 2000-04-06 10:45:11 by simonmar]
simonmar [Thu, 6 Apr 2000 10:45:11 +0000 (10:45 +0000)]
[project @ 2000-04-06 10:45:11 by simonmar]
add a note about POSIX compliance.

24 years ago[project @ 2000-04-06 10:33:06 by rrt]
rrt [Thu, 6 Apr 2000 10:33:07 +0000 (10:33 +0000)]
[project @ 2000-04-06 10:33:06 by rrt]
K&R is long dead
So why do its prototypes still live?

24 years ago[project @ 2000-04-06 10:26:09 by rrt]
rrt [Thu, 6 Apr 2000 10:26:09 +0000 (10:26 +0000)]
[project @ 2000-04-06 10:26:09 by rrt]
POSIX has not heard
of S_IWRITE
It uses S_IWSR.

24 years ago[project @ 2000-04-06 00:36:12 by andy]
andy [Thu, 6 Apr 2000 00:36:12 +0000 (00:36 +0000)]
[project @ 2000-04-06 00:36:12 by andy]
Removing the "import privileged" implementation from hugs.

24 years ago[project @ 2000-04-06 00:28:31 by andy]
andy [Thu, 6 Apr 2000 00:28:31 +0000 (00:28 +0000)]
[project @ 2000-04-06 00:28:31 by andy]
Version of PrelPrim with all the primitives listed.

24 years ago[project @ 2000-04-06 00:04:01 by andy]
andy [Thu, 6 Apr 2000 00:04:01 +0000 (00:04 +0000)]
[project @ 2000-04-06 00:04:01 by andy]
Changing PrimPrel to PrelPrim

24 years ago[project @ 2000-04-06 00:02:09 by andy]
andy [Thu, 6 Apr 2000 00:02:09 +0000 (00:02 +0000)]
[project @ 2000-04-06 00:02:09 by andy]
wibble.

24 years ago[project @ 2000-04-06 00:01:26 by andy]
andy [Thu, 6 Apr 2000 00:01:27 +0000 (00:01 +0000)]
[project @ 2000-04-06 00:01:26 by andy]
Changing PrimPrel to PrelPrim.

24 years ago[project @ 2000-04-05 17:05:50 by rrt]
rrt [Wed, 5 Apr 2000 17:05:50 +0000 (17:05 +0000)]
[project @ 2000-04-05 17:05:50 by rrt]
Cut a lot of crappy code,
Now it works in Bill's abode.

24 years ago[project @ 2000-04-05 16:57:18 by sewardj]
sewardj [Wed, 5 Apr 2000 16:57:18 +0000 (16:57 +0000)]
[project @ 2000-04-05 16:57:18 by sewardj]
Clean up the storage manager a little, and reinstate the compile time
garbage collector.  Then pray.

24 years ago[project @ 2000-04-05 16:25:51 by simonpj]
simonpj [Wed, 5 Apr 2000 16:25:54 +0000 (16:25 +0000)]
[project @ 2000-04-05 16:25:51 by simonpj]
* Add new flag -fddump-minimal-imports, which dumps a file
  M.imports that contains the (allegedly) minimal bunch of
  imports that make the system work.
  It's done by Rename.printMinimalImports

* Extend foreign import/export to handle
* Booleans
* newtypes
  as requested by the FFI team

* Tidy up DsCCall quite a bit
  Remove maybeBoxedPrimTy from TcHsSyn

24 years ago[project @ 2000-04-05 15:32:08 by simonmar]
simonmar [Wed, 5 Apr 2000 15:32:08 +0000 (15:32 +0000)]
[project @ 2000-04-05 15:32:08 by simonmar]
Preliminary support for heap profiles by module, type, or closure
description (ie. +RTS -h{Y,D,M}).

The hash tables are dynamically expanding, but lookup seems fairly
slow so the implementation may need to change at some point.

24 years ago[project @ 2000-04-05 15:28:59 by simonmar]
simonmar [Wed, 5 Apr 2000 15:28:59 +0000 (15:28 +0000)]
[project @ 2000-04-05 15:28:59 by simonmar]
TSOs now belong to CCS_SYSTEM.

24 years ago[project @ 2000-04-05 15:27:59 by simonmar]
simonmar [Wed, 5 Apr 2000 15:27:59 +0000 (15:27 +0000)]
[project @ 2000-04-05 15:27:59 by simonmar]
Include the closure description and type description fields in the
info table, for PROFILING.

24 years ago[project @ 2000-04-05 15:21:28 by simonmar]
simonmar [Wed, 5 Apr 2000 15:21:28 +0000 (15:21 +0000)]
[project @ 2000-04-05 15:21:28 by simonmar]
Ignore system CCSs in the generated XML profile, since the browser
isn't doing this yet.

24 years ago[project @ 2000-04-05 15:20:41 by panne]
panne [Wed, 5 Apr 2000 15:20:41 +0000 (15:20 +0000)]
[project @ 2000-04-05 15:20:41 by panne]
Imported sources for the much-debated HsFFI.h

24 years ago[project @ 2000-04-05 15:20:13 by simonmar]
simonmar [Wed, 5 Apr 2000 15:20:13 +0000 (15:20 +0000)]
[project @ 2000-04-05 15:20:13 by simonmar]
Fill in the closure description and type description fields for the
heap-resident closures in this file (affects profiling only).

24 years ago[project @ 2000-04-05 15:18:29 by simonmar]
simonmar [Wed, 5 Apr 2000 15:18:29 +0000 (15:18 +0000)]
[project @ 2000-04-05 15:18:29 by simonmar]
Use the decoded string, not the z-encoded string, for the closure
description.

24 years ago[project @ 2000-04-05 15:17:38 by simonmar]
simonmar [Wed, 5 Apr 2000 15:17:38 +0000 (15:17 +0000)]
[project @ 2000-04-05 15:17:38 by simonmar]
generate the correct closureTypeDescr for an LFCon.

24 years ago[project @ 2000-04-05 14:26:31 by panne]
panne [Wed, 5 Apr 2000 14:26:31 +0000 (14:26 +0000)]
[project @ 2000-04-05 14:26:31 by panne]
Changed a bunch of `#endif FOO' to `#endif /* FOO */', the former is
not strictly ANSI (don't know if the latter is, but `gcc -Wall -ansi
-pedantic' is silent then).

24 years ago[project @ 2000-04-05 14:14:51 by sewardj]
sewardj [Wed, 5 Apr 2000 14:14:51 +0000 (14:14 +0000)]
[project @ 2000-04-05 14:14:51 by sewardj]
Nuke inexplicable calls to moduleThisScript.

24 years ago[project @ 2000-04-05 14:13:58 by sewardj]
sewardj [Wed, 5 Apr 2000 14:13:58 +0000 (14:13 +0000)]
[project @ 2000-04-05 14:13:58 by sewardj]
Restore +Q and -Q operation, so we can run nofib again.

24 years ago[project @ 2000-04-05 13:59:05 by rrt]
rrt [Wed, 5 Apr 2000 13:59:05 +0000 (13:59 +0000)]
[project @ 2000-04-05 13:59:05 by rrt]
A bracket lost
is sense restored!

24 years ago[project @ 2000-04-05 13:56:30 by panne]
panne [Wed, 5 Apr 2000 13:56:30 +0000 (13:56 +0000)]
[project @ 2000-04-05 13:56:30 by panne]
Ooops, forgot that one in my previous commit (don't mention
HsFalse/HsTrue anymore).

24 years ago[project @ 2000-04-05 13:53:33 by panne]
panne [Wed, 5 Apr 2000 13:53:33 +0000 (13:53 +0000)]
[project @ 2000-04-05 13:53:33 by panne]
Removed comments about HsFalse/HsTrue, we guarantee 0/1.

24 years ago[project @ 2000-04-05 11:39:07 by sewardj]
sewardj [Wed, 5 Apr 2000 11:39:07 +0000 (11:39 +0000)]
[project @ 2000-04-05 11:39:07 by sewardj]
* Change showsPrec on ErrorCall to include "error: "
* Go back to old defn of length; new version requires O(N) stack :-(

24 years ago[project @ 2000-04-05 10:28:02 by simonmar]
simonmar [Wed, 5 Apr 2000 10:28:02 +0000 (10:28 +0000)]
[project @ 2000-04-05 10:28:02 by simonmar]
Clean up the error conditions a bit.

24 years ago[project @ 2000-04-05 10:27:10 by simonmar]
simonmar [Wed, 5 Apr 2000 10:27:10 +0000 (10:27 +0000)]
[project @ 2000-04-05 10:27:10 by simonmar]
add icons subdir

24 years ago[project @ 2000-04-05 10:25:08 by sewardj]
sewardj [Wed, 5 Apr 2000 10:25:09 +0000 (10:25 +0000)]
[project @ 2000-04-05 10:25:08 by sewardj]
Correctly handle constructors with strict fields, which was broken by
overenthusiastic constructor inlining some time back:
* notice if a constructor has strict fields, and set name(n).hasStrict,
  both for source modules and interfaces
* if a constr has strict fields, do not inline applications of it

24 years ago[project @ 2000-04-05 10:11:55 by simonmar]
simonmar [Wed, 5 Apr 2000 10:11:55 +0000 (10:11 +0000)]
[project @ 2000-04-05 10:11:55 by simonmar]
install the icons.

24 years ago[project @ 2000-04-05 10:06:36 by simonmar]
simonmar [Wed, 5 Apr 2000 10:06:36 +0000 (10:06 +0000)]
[project @ 2000-04-05 10:06:36 by simonmar]
Add new profiling tool (probably non-working so far).

24 years ago[project @ 2000-04-05 09:22:28 by sewardj]
sewardj [Wed, 5 Apr 2000 09:22:28 +0000 (09:22 +0000)]
[project @ 2000-04-05 09:22:28 by sewardj]
processInterfaces: use dupOnto to create constructor_list, not appendOnto.
Ignore strictness annotations on iface constructor sigs.  (Probably wrong).

24 years ago[project @ 2000-04-04 18:27:46 by andy]
andy [Tue, 4 Apr 2000 18:27:46 +0000 (18:27 +0000)]
[project @ 2000-04-04 18:27:46 by andy]
Observe wibble.

24 years ago[project @ 2000-04-04 17:35:04 by sewardj]
sewardj [Tue, 4 Apr 2000 17:35:04 +0000 (17:35 +0000)]
[project @ 2000-04-04 17:35:04 by sewardj]
Restore ability to load a script which doesn't have a module header,
ie doesn't start   module M where ...

24 years ago[project @ 2000-04-04 17:07:15 by sewardj]
sewardj [Tue, 4 Apr 2000 17:07:15 +0000 (17:07 +0000)]
[project @ 2000-04-04 17:07:15 by sewardj]
Fix various small bugs in module chasing & error handling.
Report a static analysis error if a module's name does not match
its file name.

24 years ago[project @ 2000-04-04 15:41:56 by sewardj]
sewardj [Tue, 4 Apr 2000 15:41:56 +0000 (15:41 +0000)]
[project @ 2000-04-04 15:41:56 by sewardj]
* Enforce downward closure rule (first attempt :-)
  If both object and source of a module are available, be simple and
  choose the more recent.  If that causes a subsequent violation of the
  DCR, complain to the user at the time the violation is detected.
  The alternative is to have a clever algorithm which makes clever
  choices now to avoid conflicts later, but that looks complicated to
  do, and I think it would also confuse users.

* As a side effect of the above, enforce the rule that the Prelude
  must be all source or all object, but not a combination.

* Rationalise signature and semantics for findFilesForModule, so as to
  make client code simpler.

24 years ago[project @ 2000-04-04 15:22:58 by simonmar]
simonmar [Tue, 4 Apr 2000 15:22:58 +0000 (15:22 +0000)]
[project @ 2000-04-04 15:22:58 by simonmar]
a few corrections.

24 years ago[project @ 2000-04-04 15:22:39 by rrt]
rrt [Tue, 4 Apr 2000 15:22:39 +0000 (15:22 +0000)]
[project @ 2000-04-04 15:22:39 by rrt]
find changed to FIND, that Windows users may
ignore DOS find when PATH has gone astray.

24 years ago[project @ 2000-04-04 15:14:18 by rrt]
rrt [Tue, 4 Apr 2000 15:14:19 +0000 (15:14 +0000)]
[project @ 2000-04-04 15:14:18 by rrt]
Added FIND as a symbol so that it can be changed on OSes which have
different ideas as to what "find" should do.

24 years ago[project @ 2000-04-04 15:02:02 by simonmar]
simonmar [Tue, 4 Apr 2000 15:02:02 +0000 (15:02 +0000)]
[project @ 2000-04-04 15:02:02 by simonmar]
Restore debugging support that disappeared when the latest parallel
stuff was committed.

24 years ago[project @ 2000-04-04 13:40:27 by panne]
panne [Tue, 4 Apr 2000 13:40:27 +0000 (13:40 +0000)]
[project @ 2000-04-04 13:40:27 by panne]
Enable Bool in foreign import/export

24 years ago[project @ 2000-04-04 11:24:48 by sewardj]
sewardj [Tue, 4 Apr 2000 11:24:48 +0000 (11:24 +0000)]
[project @ 2000-04-04 11:24:48 by sewardj]
identToStr(): use whatIs(), not fst in switch.

24 years ago[project @ 2000-04-04 11:01:33 by simonmar]
simonmar [Tue, 4 Apr 2000 11:01:33 +0000 (11:01 +0000)]
[project @ 2000-04-04 11:01:33 by simonmar]
Fix the EOF condition in readChunk.

24 years ago[project @ 2000-04-04 10:04:47 by simonmar]
simonmar [Tue, 4 Apr 2000 10:04:47 +0000 (10:04 +0000)]
[project @ 2000-04-04 10:04:47 by simonmar]
- make the second ^C kill the program immediately (the first one tries
  to interrupt it safely by killing all the threads, running
  finalizers etc.).

- don't ignore SIGPIPE by default, the program can do this itself.

24 years ago[project @ 2000-04-04 07:55:24 by andy]
andy [Tue, 4 Apr 2000 07:55:24 +0000 (07:55 +0000)]
[project @ 2000-04-04 07:55:24 by andy]
Adding names of various primitives, as required by the extended libs.

24 years ago[project @ 2000-04-04 02:20:30 by andy]
andy [Tue, 4 Apr 2000 02:20:30 +0000 (02:20 +0000)]
[project @ 2000-04-04 02:20:30 by andy]
Using the new Prelude structure to handle prims.

24 years ago[project @ 2000-04-04 01:19:07 by andy]
andy [Tue, 4 Apr 2000 01:19:07 +0000 (01:19 +0000)]
[project @ 2000-04-04 01:19:07 by andy]
Making expressions inport Prelude, not PrimPrel,
fixing the implicit import Prelude back to the Haskell semantics.

24 years ago[project @ 2000-04-04 01:07:49 by andy]
andy [Tue, 4 Apr 2000 01:07:50 +0000 (01:07 +0000)]
[project @ 2000-04-04 01:07:49 by andy]
Adding in support for split Hugs Prelude.
There are now two preludes.
(1) PrimPrel - the Prelude defintions, and the extra magic datatypes.
(2) Prelude  - the external interface for Prelude.

24 years ago[project @ 2000-04-03 23:43:13 by andy]
andy [Mon, 3 Apr 2000 23:43:13 +0000 (23:43 +0000)]
[project @ 2000-04-03 23:43:13 by andy]
Fixing problem withn "Prelude> take.P" which was causing Hugs to crash.
The new abstraction for the Text segment was not being observed.

24 years ago[project @ 2000-04-03 19:39:17 by panne]
panne [Mon, 3 Apr 2000 19:39:17 +0000 (19:39 +0000)]
[project @ 2000-04-03 19:39:17 by panne]
Nuke evil #include

24 years ago[project @ 2000-04-03 17:27:10 by sewardj]
sewardj [Mon, 3 Apr 2000 17:27:10 +0000 (17:27 +0000)]
[project @ 2000-04-03 17:27:10 by sewardj]
Track recent renaming of True_static_closure ---> True_closure, etc,
and avoid segfault with filename extension handling.

24 years ago[project @ 2000-04-03 16:46:41 by simonpj]
simonpj [Mon, 3 Apr 2000 16:46:42 +0000 (16:46 +0000)]
[project @ 2000-04-03 16:46:41 by simonpj]
* Minor wibble to type checker error message

* Make error messages come out to stderr (I'd switched
  to stdout temporarily when fighting the Dreaded Stderr Bug
  and forgot to change back)

24 years ago[project @ 2000-04-03 16:28:08 by simonmar]
simonmar [Mon, 3 Apr 2000 16:28:08 +0000 (16:28 +0000)]
[project @ 2000-04-03 16:28:08 by simonmar]
Don't use MainRegTable to communicate Sp to stg_init, use a
dynamically-allocated Capability instead (works with SMP).

24 years ago[project @ 2000-04-03 16:26:01 by simonmar]
simonmar [Mon, 3 Apr 2000 16:26:01 +0000 (16:26 +0000)]
[project @ 2000-04-03 16:26:01 by simonmar]
eliminate warning

24 years ago[project @ 2000-04-03 16:22:51 by sewardj]
sewardj [Mon, 3 Apr 2000 16:22:51 +0000 (16:22 +0000)]
[project @ 2000-04-03 16:22:51 by sewardj]
Spotted this piece of ancient history and decided to delete it.

24 years ago[project @ 2000-04-03 15:56:51 by simonmar]
simonmar [Mon, 3 Apr 2000 15:56:51 +0000 (15:56 +0000)]
[project @ 2000-04-03 15:56:51 by simonmar]
remove this file, its contents have been moved into StgProf.h

24 years ago[project @ 2000-04-03 15:55:26 by simonmar]
simonmar [Mon, 3 Apr 2000 15:55:26 +0000 (15:55 +0000)]
[project @ 2000-04-03 15:55:26 by simonmar]
misc cleanups

24 years ago[project @ 2000-04-03 15:54:49 by simonmar]
simonmar [Mon, 3 Apr 2000 15:54:50 +0000 (15:54 +0000)]
[project @ 2000-04-03 15:54:49 by simonmar]
misc profiling cleanups

24 years ago[project @ 2000-04-03 15:52:53 by simonmar]
simonmar [Mon, 3 Apr 2000 15:52:53 +0000 (15:52 +0000)]
[project @ 2000-04-03 15:52:53 by simonmar]
Remove unnecessary #include.

24 years ago[project @ 2000-04-03 15:24:21 by rrt]
rrt [Mon, 3 Apr 2000 15:24:21 +0000 (15:24 +0000)]
[project @ 2000-04-03 15:24:21 by rrt]
Stopped getourtimeofday() being called under mingwin.
MBlock.c: changed one instance of 128 * 1024 * 1024 into SIZE_RESERVED_POOL
and added a TODO comment (bug in is_heap_alloced(), which won't work if
more than one megablock is ever allocated).

24 years ago[project @ 2000-04-03 13:48:37 by sewardj]
sewardj [Mon, 3 Apr 2000 13:48:37 +0000 (13:48 +0000)]
[project @ 2000-04-03 13:48:37 by sewardj]
Deal with MachWords, which recently have started appearing for unknown reasons.

24 years ago[project @ 2000-04-03 13:44:47 by simonmar]
simonmar [Mon, 3 Apr 2000 13:44:47 +0000 (13:44 +0000)]
[project @ 2000-04-03 13:44:47 by simonmar]
call __init_PrelMain instead of __init_Main.

24 years ago[project @ 2000-04-03 13:05:52 by sewardj]
sewardj [Mon, 3 Apr 2000 13:05:52 +0000 (13:05 +0000)]
[project @ 2000-04-03 13:05:52 by sewardj]
When generating native code, print the final result with printForAsm
rather than printForUser.  Why the former ever worked beats me.

24 years ago[project @ 2000-04-03 12:50:25 by simonmar]
simonmar [Mon, 3 Apr 2000 12:50:26 +0000 (12:50 +0000)]
[project @ 2000-04-03 12:50:25 by simonmar]
Finally retire the update analyser.

24 years ago[project @ 2000-04-03 12:47:08 by simonmar]
simonmar [Mon, 3 Apr 2000 12:47:57 +0000 (12:47 +0000)]
[project @ 2000-04-03 12:47:08 by simonmar]
don't mention non-existent RTS options.

24 years ago[project @ 2000-04-03 10:10:22 by panne]
panne [Mon, 3 Apr 2000 10:10:22 +0000 (10:10 +0000)]
[project @ 2000-04-03 10:10:22 by panne]
Declare `plusAddr` as infixl 5, allowing sensible expressions like
addr `plusAddr` 47 * 11     or      addr `plusAddr` 08 + 15

24 years ago[project @ 2000-04-03 09:52:28 by simonpj]
simonpj [Mon, 3 Apr 2000 09:52:30 +0000 (09:52 +0000)]
[project @ 2000-04-03 09:52:28 by simonpj]
* Make it so that recursive newtype declarations don't send
  GHC into an infinite loop.

newtype T = MkT T

  This happened because Type.repType looked throught newtypes,
  and that never stopped!  Now TcTyDecls.mkNewTyConRep does the job
  more carefully, and the result is cached in the TyCon itself.

* Improve the handling of type signatures & pragmas.  Previously a
  mis-placed (say) SPECIALISE instance pragmas could be silently
  ignored.

Both these changes involved moving quite a lot of stuff between modules.

24 years ago[project @ 2000-04-01 22:30:27 by andy]
andy [Sat, 1 Apr 2000 22:30:27 +0000 (22:30 +0000)]
[project @ 2000-04-01 22:30:27 by andy]
Initial revision

24 years ago[project @ 2000-03-31 04:13:27 by andy]
andy [Fri, 31 Mar 2000 04:13:27 +0000 (04:13 +0000)]
[project @ 2000-03-31 04:13:27 by andy]
Fixing a couple of problems with Quantified identifiers.

(1) The string building routine for Qid's was not using
    the new abstraction for showing names. The old abstraction
    worked most of the time in the new system, so was only
    getting tickled sometimes.
(2) Fixing the local module that top level expressions
    evaluate in. By importing quantified Prelude,
    this allows top level expressions like Prelude.take

And a trivial Makefile change.

24 years ago[project @ 2000-03-31 03:09:35 by hwloidl]
hwloidl [Fri, 31 Mar 2000 03:09:38 +0000 (03:09 +0000)]
[project @ 2000-03-31 03:09:35 by hwloidl]
Numerous changes in the RTS to get GUM-4.06 working (currently works with
parfib-ish programs). Most changes are isolated in the rts/parallel dir.

rts/parallel/:
  The most important changes are a rewrite of the (un-)packing code (Pack.c)
  and changes in LAGA, GALA table operations (Global.c) expecially in
  rebuilding the tables during GC.

rts/:
  Minor changes in Schedule.c, GC.c (interface to par specific root marking
  and evacuation), and lots of additions to Sanity.c (surprise ;-)
  Main.c change for startup: I use a new function rts_evalNothing to
  start non-main-PEs in a PAR || SMP setup (RtsAPI.c)

includes/:
  Updated GranSim macros in PrimOps.h.

lib/std:
  Few changes in PrelHandle.c etc replacing ForeignObj by Addr in a PAR
  setup (we still don't support ForeignObjs or WeakPtrs in GUM).
  Typically use
    #define FILE_OBJECT     Addr
  when dealing with files.

hslibs/lang/:
  Same as above (in Foreign(Obj).lhs, Weak.lhs, IOExts.lhs etc).

-- HWL

24 years ago[project @ 2000-03-30 16:23:56 by simonpj]
simonpj [Thu, 30 Mar 2000 16:23:57 +0000 (16:23 +0000)]
[project @ 2000-03-30 16:23:56 by simonpj]
* Remove the unnecessary CPR parameter to mkUnfolding and friends

* Make sure that even trivial wrappers have a __inline__
  (this was causing lots of 'substWorker' DEBUG messages)

* Nuke demand info when the unfolding is a value
  (see notes with IdInfo.setUnfoldingInfo)

* Add an update-in-place test to the 'interesting context'
  predicate in SimplUtils.

24 years ago[project @ 2000-03-30 16:07:53 by simonmar]
simonmar [Thu, 30 Mar 2000 16:07:53 +0000 (16:07 +0000)]
[project @ 2000-03-30 16:07:53 by simonmar]
Support AP_UPDs/PAPs larger than a block in size.  It turned out not
to be too hard to do this.  This fixes George Russell's bug ("hsc:
fatal error: scavenge: stack frame").

24 years ago[project @ 2000-03-30 15:35:13 by simonmar]
simonmar [Thu, 30 Mar 2000 15:35:13 +0000 (15:35 +0000)]
[project @ 2000-03-30 15:35:13 by simonmar]
Add an ASSERT to catch a known bug: when allocating an AP_UPD in
raiseAsync, if the object is larger than a block in size then all
sorts of things can go wrong.

24 years ago[project @ 2000-03-30 14:12:42 by rrt]
rrt [Thu, 30 Mar 2000 14:12:42 +0000 (14:12 +0000)]
[project @ 2000-03-30 14:12:42 by rrt]
Filter out PrelHugs in all ways for building DLLs.

24 years ago[project @ 2000-03-30 14:11:34 by rrt]
rrt [Thu, 30 Mar 2000 14:11:34 +0000 (14:11 +0000)]
[project @ 2000-03-30 14:11:34 by rrt]
Added subl clause to epilogue munging.

24 years ago[project @ 2000-03-30 14:09:44 by rrt]
rrt [Thu, 30 Mar 2000 14:09:44 +0000 (14:09 +0000)]
[project @ 2000-03-30 14:09:44 by rrt]
Added rules for building GMP as a DLL. N.B. GMP is not actually used
yet as a DLL; this remains to be added to the build system.

24 years ago[project @ 2000-03-30 14:08:43 by rrt]
rrt [Thu, 30 Mar 2000 14:08:43 +0000 (14:08 +0000)]
[project @ 2000-03-30 14:08:43 by rrt]
Removed __init_PrelMain, which shouldn't be in here.

24 years ago[project @ 2000-03-30 13:46:54 by simonpj]
simonpj [Thu, 30 Mar 2000 13:46:54 +0000 (13:46 +0000)]
[project @ 2000-03-30 13:46:54 by simonpj]
Document hoisting foralls

24 years ago[project @ 2000-03-30 12:04:13 by simonmar]
simonmar [Thu, 30 Mar 2000 12:04:13 +0000 (12:04 +0000)]
[project @ 2000-03-30 12:04:13 by simonmar]
update for extra parameter to startupHaskell().

24 years ago[project @ 2000-03-30 12:03:30 by simonmar]
simonmar [Thu, 30 Mar 2000 12:03:31 +0000 (12:03 +0000)]
[project @ 2000-03-30 12:03:30 by simonmar]
HEADS UP!!!

change the type of startupHaskell():

   void startupHaskell  ( int argc, char *argv[], void *init_root );

the extra parameter is a pointer to the initialisation function for
the root module in the program.  eg., Main.c now passes __init_Main for
this parameter.  It can be left as NULL if there is no root module.

This interface may need to be revised, since in some circumstances
there may be more than one "root module".

Sigbjorn: H/Direct will need some changes to stay in sync here.

24 years ago[project @ 2000-03-30 10:37:09 by simonmar]
simonmar [Thu, 30 Mar 2000 10:37:09 +0000 (10:37 +0000)]
[project @ 2000-03-30 10:37:09 by simonmar]
Remove this file, it moved to ghc/rts.  (sigh, I thought I removed it
already, but apparently not).

24 years ago[project @ 2000-03-30 10:36:15 by simonmar]
simonmar [Thu, 30 Mar 2000 10:36:15 +0000 (10:36 +0000)]
[project @ 2000-03-30 10:36:15 by simonmar]
Replace _static_closures with _closures.

24 years ago[project @ 2000-03-30 09:04:34 by andy]
andy [Thu, 30 Mar 2000 09:04:34 +0000 (09:04 +0000)]
[project @ 2000-03-30 09:04:34 by andy]
Wibble One.

24 years ago[project @ 2000-03-30 09:02:12 by andy]
andy [Thu, 30 Mar 2000 09:02:12 +0000 (09:02 +0000)]
[project @ 2000-03-30 09:02:12 by andy]
Fixing bug with import privileged clashing with import hidden.
This caused a rather nasty name-leak, where catch from the prelude
was being given the type of catch from Exceptions.

Now, when you use import privileged Prelude (...) you also
need to do import Prelude, allowing you the option of
import Prelude hiding (...).

A bucket load of wibbles will follow in various libraries,
implementing this restriction.

24 years ago[project @ 2000-03-29 08:59:28 by simonmar]
simonmar [Wed, 29 Mar 2000 08:59:28 +0000 (08:59 +0000)]
[project @ 2000-03-29 08:59:28 by simonmar]
freezeDoubleArray ==> freezeByteArray

24 years ago[project @ 2000-03-29 08:58:04 by simonmar]
simonmar [Wed, 29 Mar 2000 08:58:04 +0000 (08:58 +0000)]
[project @ 2000-03-29 08:58:04 by simonmar]
freeze<blah>Array ==> freezeByteArray

24 years ago[project @ 2000-03-28 14:29:13 by simonmar]
simonmar [Tue, 28 Mar 2000 14:29:13 +0000 (14:29 +0000)]
[project @ 2000-03-28 14:29:13 by simonmar]
The new itimer stuff is NON_POSIX_SOURCE, and further protect it with
HAVE_SETITIMER.  Fixes building on Solaris again.

24 years ago[project @ 2000-03-28 13:19:28 by simonmar]
simonmar [Tue, 28 Mar 2000 13:19:28 +0000 (13:19 +0000)]
[project @ 2000-03-28 13:19:28 by simonmar]
Remove extra \end{code}

24 years ago[project @ 2000-03-28 13:18:49 by simonmar]
simonmar [Tue, 28 Mar 2000 13:18:49 +0000 (13:18 +0000)]
[project @ 2000-03-28 13:18:49 by simonmar]
Don't use "ccall" as a variable name; it confuses the interface file
parser in 4.06.  This bug has been fixed in 4.07.

24 years ago[project @ 2000-03-28 11:49:06 by rrt]
rrt [Tue, 28 Mar 2000 11:49:06 +0000 (11:49 +0000)]
[project @ 2000-03-28 11:49:06 by rrt]
Replaced $(HC_OPTS) in mkdependhs rule whence I had inadvertently removed
it along with the odir filter in which it sat.

24 years ago[project @ 2000-03-28 11:21:23 by rrt]
rrt [Tue, 28 Mar 2000 11:21:23 +0000 (11:21 +0000)]
[project @ 2000-03-28 11:21:23 by rrt]
Used new SplitObjs setting, and removed patsubst hack (no longer needed,
thank goodness).

24 years ago[project @ 2000-03-28 11:09:31 by rrt]
rrt [Tue, 28 Mar 2000 11:09:31 +0000 (11:09 +0000)]
[project @ 2000-03-28 11:09:31 by rrt]
Added SplitObjs for when -split-objs is being used. Removed -split-objs and
-odir from GhcLibHcOpts. -odir shouldn't be needed any more in any case.

Changed order of recursive makes so that recursive make happens before make
in the current directory. This helps when building hslibs as DLLs, where
the cbits DLLs need to be made before the main DLLs.

24 years ago[project @ 2000-03-28 11:07:19 by rrt]
rrt [Tue, 28 Mar 2000 11:07:19 +0000 (11:07 +0000)]
[project @ 2000-03-28 11:07:19 by rrt]
Added SplitObjs and removed -split-objs from GhcLibHcOpts.

24 years ago[project @ 2000-03-28 10:20:55 by andy]
andy [Tue, 28 Mar 2000 10:20:56 +0000 (10:20 +0000)]
[project @ 2000-03-28 10:20:55 by andy]
Restoring Hugs ability to accept the -98/+98 and -h<heap> options.

24 years ago[project @ 2000-03-28 10:10:17 by andy]
andy [Tue, 28 Mar 2000 10:10:17 +0000 (10:10 +0000)]
[project @ 2000-03-28 10:10:17 by andy]
Making STG Hugs compile again after the recent RTS checking.

24 years ago[project @ 2000-03-28 08:52:28 by simonmar]
simonmar [Tue, 28 Mar 2000 08:52:29 +0000 (08:52 +0000)]
[project @ 2000-03-28 08:52:28 by simonmar]
Replace freeze{Char,Int,Word,Float,Double}Array with freezeByteArray
(using sizeofByteArray and a foreign import of C's memcpy()).

24 years ago[project @ 2000-03-28 08:51:09 by simonmar]
simonmar [Tue, 28 Mar 2000 08:51:09 +0000 (08:51 +0000)]
[project @ 2000-03-28 08:51:09 by simonmar]
remove redundant import.