ghc-hetmet.git
23 years ago[project @ 2000-11-21 12:32:55 by rrt]
rrt [Tue, 21 Nov 2000 12:32:55 +0000 (12:32 +0000)]
[project @ 2000-11-21 12:32:55 by rrt]
Added rules for GreenCard.

23 years ago[project @ 2000-11-21 11:31:21 by simonmar]
simonmar [Tue, 21 Nov 2000 11:31:21 +0000 (11:31 +0000)]
[project @ 2000-11-21 11:31:21 by simonmar]
make ":l hello" load hello.hs if it exists.

23 years ago[project @ 2000-11-21 11:22:25 by sewardj]
sewardj [Tue, 21 Nov 2000 11:22:26 +0000 (11:22 +0000)]
[project @ 2000-11-21 11:22:25 by sewardj]
Clean up CM's messages a bit.

23 years ago[project @ 2000-11-21 10:48:20 by simonmar]
simonmar [Tue, 21 Nov 2000 10:48:20 +0000 (10:48 +0000)]
[project @ 2000-11-21 10:48:20 by simonmar]
- add default default settings for typecheckExpr
- bugfixes in the interpreter

print (1+2) now works!!

23 years ago[project @ 2000-11-21 10:46:01 by simonmar]
simonmar [Tue, 21 Nov 2000 10:46:01 +0000 (10:46 +0000)]
[project @ 2000-11-21 10:46:01 by simonmar]
fix some cut-n-pastos

23 years ago[project @ 2000-11-21 10:29:54 by simonpj]
simonpj [Tue, 21 Nov 2000 10:29:54 +0000 (10:29 +0000)]
[project @ 2000-11-21 10:29:54 by simonpj]
Fix field name checking

23 years ago[project @ 2000-11-21 10:00:44 by simonmar]
simonmar [Tue, 21 Nov 2000 10:00:44 +0000 (10:00 +0000)]
[project @ 2000-11-21 10:00:44 by simonmar]
add \$Id\$

23 years ago[project @ 2000-11-21 09:57:02 by sewardj]
sewardj [Tue, 21 Nov 2000 09:57:02 +0000 (09:57 +0000)]
[project @ 2000-11-21 09:57:02 by sewardj]
rm debugging printfs

23 years ago[project @ 2000-11-21 09:55:47 by simonmar]
simonmar [Tue, 21 Nov 2000 09:55:47 +0000 (09:55 +0000)]
[project @ 2000-11-21 09:55:47 by simonmar]
add missing case

23 years ago[project @ 2000-11-21 09:30:16 by simonpj]
simonpj [Tue, 21 Nov 2000 09:30:17 +0000 (09:30 +0000)]
[project @ 2000-11-21 09:30:16 by simonpj]
Fix renamer bugs

23 years ago[project @ 2000-11-20 17:42:00 by sewardj]
sewardj [Mon, 20 Nov 2000 17:42:01 +0000 (17:42 +0000)]
[project @ 2000-11-20 17:42:00 by sewardj]
Split HscResult into HscFail | HscNoRecomp | HscRecomp, and clean up
producers and consumers of such.  In particular, if no recompilation
happens, the resulting iface is put into the HIT instead of being
thrown away.

Also (trivial) unify functions *ModuleInThisPackage with *HomeModule.

23 years ago[project @ 2000-11-20 16:51:35 by simonmar]
simonmar [Mon, 20 Nov 2000 16:51:35 +0000 (16:51 +0000)]
[project @ 2000-11-20 16:51:35 by simonmar]
wibble

23 years ago[project @ 2000-11-20 16:40:12 by rrt]
rrt [Mon, 20 Nov 2000 16:40:12 +0000 (16:40 +0000)]
[project @ 2000-11-20 16:40:12 by rrt]
Tag-o.

23 years ago[project @ 2000-11-20 16:37:42 by sewardj]
sewardj [Mon, 20 Nov 2000 16:37:42 +0000 (16:37 +0000)]
[project @ 2000-11-20 16:37:42 by sewardj]
Fix to -cpp not working?

23 years ago[project @ 2000-11-20 16:28:29 by simonmar]
simonmar [Mon, 20 Nov 2000 16:28:32 +0000 (16:28 +0000)]
[project @ 2000-11-20 16:28:29 by simonmar]
Allow the root of the module tree to have a filename which is
different from its module name.  The argument to cmLoadModule is now a
filename.

23 years ago[project @ 2000-11-20 16:27:36 by simonmar]
simonmar [Mon, 20 Nov 2000 16:27:36 +0000 (16:27 +0000)]
[project @ 2000-11-20 16:27:36 by simonmar]
Don't return the old iface from addVersionInfo, it might not have
reasonable mi_globals.

23 years ago[project @ 2000-11-20 16:12:51 by rrt]
rrt [Mon, 20 Nov 2000 16:12:51 +0000 (16:12 +0000)]
[project @ 2000-11-20 16:12:51 by rrt]
Simplified installation instructions (thanks to Christian Lescher) which
allow you to install Cygwin where you like.

23 years ago[project @ 2000-11-20 16:08:13 by simonmar]
simonmar [Mon, 20 Nov 2000 16:08:13 +0000 (16:08 +0000)]
[project @ 2000-11-20 16:08:13 by simonmar]
fix compile errors in hscExpr

23 years ago[project @ 2000-11-20 16:07:12 by simonpj]
simonpj [Mon, 20 Nov 2000 16:07:13 +0000 (16:07 +0000)]
[project @ 2000-11-20 16:07:12 by simonpj]
Remember local decls when no recompilation is required

23 years ago[project @ 2000-11-20 15:54:27 by sewardj]
sewardj [Mon, 20 Nov 2000 15:54:27 +0000 (15:54 +0000)]
[project @ 2000-11-20 15:54:27 by sewardj]
Summarise and mkDependHS faster, by doing fewer passes over the
module's text.

23 years ago[project @ 2000-11-20 15:40:54 by simonmar]
simonmar [Mon, 20 Nov 2000 15:40:54 +0000 (15:40 +0000)]
[project @ 2000-11-20 15:40:54 by simonmar]
fix bogus implementation of splitFilename3 (!!!)

23 years ago[project @ 2000-11-20 14:48:52 by simonpj]
simonpj [Mon, 20 Nov 2000 14:48:59 +0000 (14:48 +0000)]
[project @ 2000-11-20 14:48:52 by simonpj]
When renaming, typechecking an expression from the user
interface, we may suck in declarations from interface
files (e.g. the Prelude).  This commit takes account of that.

To do so, I did some significant restructuring in TcModule,
with consequential changes and tidy ups elsewhere in the type
checker.  I think there should be fewer lines in total than before.

23 years ago[project @ 2000-11-20 14:26:27 by simonmar]
simonmar [Mon, 20 Nov 2000 14:26:27 +0000 (14:26 +0000)]
[project @ 2000-11-20 14:26:27 by simonmar]
wibble

23 years ago[project @ 2000-11-20 13:43:19 by sewardj]
sewardj [Mon, 20 Nov 2000 13:43:19 +0000 (13:43 +0000)]
[project @ 2000-11-20 13:43:19 by sewardj]
Delete dead code resulting from redoing of source-up-to-date logic.

23 years ago[project @ 2000-11-20 13:39:26 by sewardj]
sewardj [Mon, 20 Nov 2000 13:39:26 +0000 (13:39 +0000)]
[project @ 2000-11-20 13:39:26 by sewardj]
Redo the source-up-to-date logic (in CompManager.upsweep_mod).

23 years ago[project @ 2000-11-20 11:39:57 by sewardj]
sewardj [Mon, 20 Nov 2000 11:39:57 +0000 (11:39 +0000)]
[project @ 2000-11-20 11:39:57 by sewardj]
* (CompManager) recompile if in interactive mode and no old linkable exists
* (HscMain) don't write interface files in interactive mode
* (everywhere) switch arg order to unJust for PAP purposes

23 years ago[project @ 2000-11-20 11:19:21 by simonmar]
simonmar [Mon, 20 Nov 2000 11:19:21 +0000 (11:19 +0000)]
[project @ 2000-11-20 11:19:21 by simonmar]
Don't put a label as the last thing in a statement block (ANSI C fix)

23 years ago[project @ 2000-11-20 11:15:41 by simonmar]
simonmar [Mon, 20 Nov 2000 11:15:41 +0000 (11:15 +0000)]
[project @ 2000-11-20 11:15:41 by simonmar]
remove #warn

23 years ago[project @ 2000-11-19 19:40:07 by simonmar]
simonmar [Sun, 19 Nov 2000 19:40:08 +0000 (19:40 +0000)]
[project @ 2000-11-19 19:40:07 by simonmar]
Cleanup sweep.

Includes code to get -H working again, #ifdefed out for the time being
since it needs support in the RTS.

23 years ago[project @ 2000-11-19 18:39:53 by simonmar]
simonmar [Sun, 19 Nov 2000 18:39:53 +0000 (18:39 +0000)]
[project @ 2000-11-19 18:39:53 by simonmar]
small fixes/cleanups

23 years ago[project @ 2000-11-19 18:32:56 by simonmar]
simonmar [Sun, 19 Nov 2000 18:32:56 +0000 (18:32 +0000)]
[project @ 2000-11-19 18:32:56 by simonmar]
Fix use of non-standard "+" in sed regexp (darn GNU "extensions" :-)

23 years ago[project @ 2000-11-17 17:03:17 by simonmar]
simonmar [Fri, 17 Nov 2000 17:03:17 +0000 (17:03 +0000)]
[project @ 2000-11-17 17:03:17 by simonmar]
add a few missing cases

23 years ago[project @ 2000-11-17 16:53:27 by simonmar]
simonmar [Fri, 17 Nov 2000 16:53:28 +0000 (16:53 +0000)]
[project @ 2000-11-17 16:53:27 by simonmar]
Results of today's hacking:

  - We can now execute expressions from the GHCi prompt.  However,
    a problem with the typechecker environment means that identifiers
    from outside the current module aren't resolved :-(

  - loading up a multi-module program in the interpreter seems to
    work.  Interpreting is kinda slow (ok, very slow), but I'm hoping
    it'll get better when I compile the interpreter w/ optimisation.

  - :set sort of works - you can do ":set -dshow-passes", for example

  - lots of bugfixes, etc.

23 years ago[project @ 2000-11-17 13:33:17 by sewardj]
sewardj [Fri, 17 Nov 2000 13:33:17 +0000 (13:33 +0000)]
[project @ 2000-11-17 13:33:17 by sewardj]
Facilitate handling the situation where name of module /= name of file.
This is handled in CompManager.summarise, which can inspect the source
to see what the _real_ module name is, and act accordingly.
All a bit messy -- needs a rethink.

23 years ago[project @ 2000-11-17 11:59:59 by sewardj]
sewardj [Fri, 17 Nov 2000 11:59:59 +0000 (11:59 +0000)]
[project @ 2000-11-17 11:59:59 by sewardj]
Filter the HST and HIT passed to compile to only contain the downwards
closure of the module being compiled.  This means that the instance-env
construction machinery cannot pick up out-of-scope instances from HST.
Don't know if it's necessary to filter HIT -- perhaps not.

Instance env construction still doesn't work right, due to duplicates
being picked up from interface files.

23 years ago[project @ 2000-11-17 10:13:21 by sewardj]
sewardj [Fri, 17 Nov 2000 10:13:22 +0000 (10:13 +0000)]
[project @ 2000-11-17 10:13:21 by sewardj]
Make file up-to-dateness checking work in batch mode.

23 years ago[project @ 2000-11-16 16:54:36 by simonmar]
simonmar [Thu, 16 Nov 2000 16:54:36 +0000 (16:54 +0000)]
[project @ 2000-11-16 16:54:36 by simonmar]
WOOHOO!  GHCi interprets "Hello World":

 _____  __   __  ____         _________________________________________________
(|      ||   || (|  |)        GHC Interactive, version 5.00
||  __  ||___|| ||     ()     For Haskell 98.
||   |) ||---|| ||     ||     http://www.haskell.org/ghc
||   || ||   || ||     (|     Bug reports to: glasgow-haskell-bugs@haskell.org
(|___|| ||   || (|__|) \\______________________________________________________

Loading package std ... resolving ... done.
Prelude> :l Main
cmLoadModule: downsweep begins
getSummary: Main
after tsort:

NONREC
   ModSummary { ms_mod = Main, ms_imps = [] ms_srcimps = [] }
CHECKING OLD IFACE for hs = Just "./Main.hs", hspp = Just "/tmp/ghc25011.hspp"
COMPILATION IS REQUIRED
UPSWEEP COMPLETELY SUCCESSFUL
CmLink.link: linkables are ...
LinkableM Main [Trees [Main.main = letP stg_cIM = (@PP Native PrelShow.$fShow[] Native PrelShow.$fShowChar)
                                   in
                                   letP stg_cIL = (@IP Native PrelBase.unpackCString# 141613256#)
                                   in
                                   (@PP (@PP Native PrelIO.print stg_cIM) stg_cIL)]]
Main.main
CmLink.link: done
Prelude> Main.main
Run expression: Main.main
evalI:
141613256#

"Hello, world!"
done.

23 years ago[project @ 2000-11-16 16:23:03 by sewardj]
sewardj [Thu, 16 Nov 2000 16:23:04 +0000 (16:23 +0000)]
[project @ 2000-11-16 16:23:03 by sewardj]
* Move along the source-changed checkery.
* Make the driver put object files in the right place when using CM.
* Don't do hscNoRecomp in one-shot mode.

23 years ago[project @ 2000-11-16 15:57:05 by simonmar]
simonmar [Thu, 16 Nov 2000 15:57:06 +0000 (15:57 +0000)]
[project @ 2000-11-16 15:57:05 by simonmar]
Moving things around a bit to avoid cycles.

Further progress on interactive linker.

23 years ago[project @ 2000-11-16 14:43:05 by simonpj]
simonpj [Thu, 16 Nov 2000 14:43:06 +0000 (14:43 +0000)]
[project @ 2000-11-16 14:43:05 by simonpj]
Add stuff to support hscExpr

23 years ago[project @ 2000-11-16 12:49:05 by simonmar]
simonmar [Thu, 16 Nov 2000 12:49:05 +0000 (12:49 +0000)]
[project @ 2000-11-16 12:49:05 by simonmar]
oops, undo part of rev 1.57 that was commited by mistake.

23 years ago[project @ 2000-11-16 11:39:36 by simonmar]
simonmar [Thu, 16 Nov 2000 11:39:37 +0000 (11:39 +0000)]
[project @ 2000-11-16 11:39:36 by simonmar]
Current state of the interactive system; can load packages (in theory).

23 years ago[project @ 2000-11-16 11:34:42 by simonmar]
simonmar [Thu, 16 Nov 2000 11:34:42 +0000 (11:34 +0000)]
[project @ 2000-11-16 11:34:42 by simonmar]
Fix the order in which we look for .hi-boot files.

23 years ago[project @ 2000-11-16 10:48:22 by simonmar]
simonmar [Thu, 16 Nov 2000 10:48:22 +0000 (10:48 +0000)]
[project @ 2000-11-16 10:48:22 by simonmar]
on second thoughts, add this somewhere more sensible

23 years ago[project @ 2000-11-16 10:32:30 by simonmar]
simonmar [Thu, 16 Nov 2000 10:32:30 +0000 (10:32 +0000)]
[project @ 2000-11-16 10:32:30 by simonmar]
merge rev. 1.78.2.4 (applyTy bugfix)

23 years ago[project @ 2000-11-16 10:00:01 by simonmar]
simonmar [Thu, 16 Nov 2000 10:00:01 +0000 (10:00 +0000)]
[project @ 2000-11-16 10:00:01 by simonmar]
Add test for applyTy bug in the simplifier.

23 years ago[project @ 2000-11-16 09:51:31 by simonmar]
simonmar [Thu, 16 Nov 2000 09:51:31 +0000 (09:51 +0000)]
[project @ 2000-11-16 09:51:31 by simonmar]
merge rev. 1.66.2.6 (add splitMethodTy)

23 years ago[project @ 2000-11-15 17:07:34 by simonpj]
simonpj [Wed, 15 Nov 2000 17:07:36 +0000 (17:07 +0000)]
[project @ 2000-11-15 17:07:34 by simonpj]
I finally got tired of not having
splitTyConApp
tyConAppTyCon
tyConAppArgs

(Previously we called splitTyConApp_maybe,
 but it's a pain in the neck.)

23 years ago[project @ 2000-11-15 17:07:17 by simonpj]
simonpj [Wed, 15 Nov 2000 17:07:17 +0000 (17:07 +0000)]
[project @ 2000-11-15 17:07:17 by simonpj]
The main thing in this commit is to make sure the desugarer
looks a the right TyCon when desugaring a record update.
Again, it's because in GHCI-land, it's possible that the
some versions of the TyCon won't have the constructors;
but the one in the FieldLabel will.

23 years ago[project @ 2000-11-15 15:43:30 by sewardj]
sewardj [Wed, 15 Nov 2000 15:43:31 +0000 (15:43 +0000)]
[project @ 2000-11-15 15:43:30 by sewardj]
Overhaul of CM, reducing the complexity of recursive module machinery.
Also, don't compute package dependencies at all, and don't pass them
to the linker.

23 years ago[project @ 2000-11-15 14:37:08 by simonpj]
simonpj [Wed, 15 Nov 2000 14:37:10 +0000 (14:37 +0000)]
[project @ 2000-11-15 14:37:08 by simonpj]
The main thing in this commit is to change StgAlts so that
it carries a TyCon, and not a Type.  Furthermore, the TyCon
is derived from the alternatives, so it should have its
constructors etc, even if there's a module loop involved, so that
some versions of the TyCon don't have the constructors visible.

There's a comment in StgSyn.lhs, with the type decl for StgAlts

Also: a start on hscExpr in HscMain.

23 years ago[project @ 2000-11-15 10:49:53 by sewardj]
sewardj [Wed, 15 Nov 2000 10:49:54 +0000 (10:49 +0000)]
[project @ 2000-11-15 10:49:53 by sewardj]
Fix up the Batch vs Interactive plumbing.

23 years ago[project @ 2000-11-15 09:58:00 by sewardj]
sewardj [Wed, 15 Nov 2000 09:58:00 +0000 (09:58 +0000)]
[project @ 2000-11-15 09:58:00 by sewardj]
Fix z-encoding problem in findDependency (moduleNameString ->
moduleNameUserString).

23 years ago[project @ 2000-11-14 17:41:04 by sewardj]
sewardj [Tue, 14 Nov 2000 17:41:05 +0000 (17:41 +0000)]
[project @ 2000-11-14 17:41:04 by sewardj]
Fixes to do with CM and module cycles.  Also to do with OPTIONS pragmas.

23 years ago[project @ 2000-11-14 16:28:38 by simonmar]
simonmar [Tue, 14 Nov 2000 16:28:38 +0000 (16:28 +0000)]
[project @ 2000-11-14 16:28:38 by simonmar]
Make -fvia-C and -fasm-XXX into dynamic flags.  The HscLang handling
is somewhat cleaned up.

23 years ago[project @ 2000-11-14 15:04:15 by sewardj]
sewardj [Tue, 14 Nov 2000 15:04:15 +0000 (15:04 +0000)]
[project @ 2000-11-14 15:04:15 by sewardj]
Stop the world in the Approved Manner if the finder can't find a module.

23 years ago[project @ 2000-11-14 15:03:46 by sewardj]
sewardj [Tue, 14 Nov 2000 15:03:46 +0000 (15:03 +0000)]
[project @ 2000-11-14 15:03:46 by sewardj]
Don't forget to prepend paths to source module names.

23 years ago[project @ 2000-11-14 14:47:23 by simonmar]
simonmar [Tue, 14 Nov 2000 14:47:23 +0000 (14:47 +0000)]
[project @ 2000-11-14 14:47:23 by simonmar]
Don't redefine int8_t and friends, even if we didn't find a definition
of INT8_MIN.  Some /usr/includes, eg. cygwin it seems, like to define
the types but not the constants.

23 years ago[project @ 2000-11-14 14:34:06 by simonmar]
simonmar [Tue, 14 Nov 2000 14:34:06 +0000 (14:34 +0000)]
[project @ 2000-11-14 14:34:06 by simonmar]
remove all the module-specific -H settings.

23 years ago[project @ 2000-11-14 14:33:02 by simonmar]
simonmar [Tue, 14 Nov 2000 14:33:02 +0000 (14:33 +0000)]
[project @ 2000-11-14 14:33:02 by simonmar]
don't hardcode -H options in the Makefile

23 years ago[project @ 2000-11-14 14:30:40 by simonmar]
simonmar [Tue, 14 Nov 2000 14:30:40 +0000 (14:30 +0000)]
[project @ 2000-11-14 14:30:40 by simonmar]
don't prepend current_dir to the name of the hifile twice.

23 years ago[project @ 2000-11-14 13:59:44 by sewardj]
sewardj [Tue, 14 Nov 2000 13:59:44 +0000 (13:59 +0000)]
[project @ 2000-11-14 13:59:44 by sewardj]
Only do batch linking if the upsweep succeeds and someone exports Main.main.

23 years ago[project @ 2000-11-14 12:53:29 by simonmar]
simonmar [Tue, 14 Nov 2000 12:53:29 +0000 (12:53 +0000)]
[project @ 2000-11-14 12:53:29 by simonmar]
wibble

23 years ago[project @ 2000-11-14 12:52:55 by simonmar]
simonmar [Tue, 14 Nov 2000 12:52:55 +0000 (12:52 +0000)]
[project @ 2000-11-14 12:52:55 by simonmar]
catch_info ==> stg_catch_info
raise_entry ==> stg_raise_entry

23 years ago[project @ 2000-11-14 12:51:51 by simonmar]
simonmar [Tue, 14 Nov 2000 12:51:51 +0000 (12:51 +0000)]
[project @ 2000-11-14 12:51:51 by simonmar]
more renamings

23 years ago[project @ 2000-11-14 12:49:57 by simonmar]
simonmar [Tue, 14 Nov 2000 12:49:57 +0000 (12:49 +0000)]
[project @ 2000-11-14 12:49:57 by simonmar]
some more missing changes in the RTS renaming.

23 years ago[project @ 2000-11-14 12:47:38 by simonmar]
simonmar [Tue, 14 Nov 2000 12:47:38 +0000 (12:47 +0000)]
[project @ 2000-11-14 12:47:38 by simonmar]
Missing change in the recent renaming of RTS symbols.

23 years ago[project @ 2000-11-14 12:07:44 by simonpj]
simonpj [Tue, 14 Nov 2000 12:07:44 +0000 (12:07 +0000)]
[project @ 2000-11-14 12:07:44 by simonpj]
Fix another naming bug

23 years ago[project @ 2000-11-14 11:25:22 by simonpj]
simonpj [Tue, 14 Nov 2000 11:25:22 +0000 (11:25 +0000)]
[project @ 2000-11-14 11:25:22 by simonpj]
vanilla -> constantIdInfo

23 years ago[project @ 2000-11-14 10:56:08 by simonmar]
simonmar [Tue, 14 Nov 2000 10:56:08 +0000 (10:56 +0000)]
[project @ 2000-11-14 10:56:08 by simonmar]
Fix the non-GTKish build.

23 years ago[project @ 2000-11-14 10:46:39 by simonpj]
simonpj [Tue, 14 Nov 2000 10:46:41 +0000 (10:46 +0000)]
[project @ 2000-11-14 10:46:39 by simonpj]
Compiles now

23 years ago[project @ 2000-11-14 08:07:11 by simonpj]
simonpj [Tue, 14 Nov 2000 08:07:14 +0000 (08:07 +0000)]
[project @ 2000-11-14 08:07:11 by simonpj]
Changing the way we know whether
something is exported.

THIS COMMIT WON'T EVEN COMPILE
(I'm doing it to transfer from my laptop.)
Wait till later today before updating.

23 years ago[project @ 2000-11-13 17:17:40 by simonmar]
simonmar [Mon, 13 Nov 2000 17:17:40 +0000 (17:17 +0000)]
[project @ 2000-11-13 17:17:40 by simonmar]
Pull in inttypes.h if we have it, as the second-favourite option after
stdint.h.  This should fix problems building the readline cbits on Solaris.

23 years ago[project @ 2000-11-13 17:13:51 by simonmar]
simonmar [Mon, 13 Nov 2000 17:13:51 +0000 (17:13 +0000)]
[project @ 2000-11-13 17:13:51 by simonmar]
Test for inttypes.h: Solaris' version of stdint.h

23 years ago[project @ 2000-11-13 17:12:37 by sewardj]
sewardj [Mon, 13 Nov 2000 17:12:37 +0000 (17:12 +0000)]
[project @ 2000-11-13 17:12:37 by sewardj]
Temporarily turn off the source-unchanged? test in DriverPipeline.compile
whilst I figure out what's really supposed to happen :-)

23 years ago[project @ 2000-11-13 17:08:36 by sewardj]
sewardj [Mon, 13 Nov 2000 17:08:36 +0000 (17:08 +0000)]
[project @ 2000-11-13 17:08:36 by sewardj]
First shot at batch linking.  Does not attempt linking if upsweep
was not completely successful.  Always attempts linking if upsweep
successful, even if there's no 'main' to be found anywhere.

23 years ago[project @ 2000-11-13 16:17:05 by sewardj]
sewardj [Mon, 13 Nov 2000 16:17:05 +0000 (16:17 +0000)]
[project @ 2000-11-13 16:17:05 by sewardj]
Changes so that the new hsc-inplace can build the prelude unassisted.

23 years ago[project @ 2000-11-13 16:16:05 by sewardj]
sewardj [Mon, 13 Nov 2000 16:16:05 +0000 (16:16 +0000)]
[project @ 2000-11-13 16:16:05 by sewardj]
Implement -ddump-hi.

23 years ago[project @ 2000-11-13 14:53:27 by simonmar]
simonmar [Mon, 13 Nov 2000 14:53:27 +0000 (14:53 +0000)]
[project @ 2000-11-13 14:53:27 by simonmar]
merged from ghc-4-07-branch: don't save %i7 across the call.

23 years ago[project @ 2000-11-13 14:45:22 by sewardj]
sewardj [Mon, 13 Nov 2000 14:45:22 +0000 (14:45 +0000)]
[project @ 2000-11-13 14:45:22 by sewardj]
Deal correctly with module names containing underscores.

23 years ago[project @ 2000-11-13 14:42:16 by simonmar]
simonmar [Mon, 13 Nov 2000 14:42:16 +0000 (14:42 +0000)]
[project @ 2000-11-13 14:42:16 by simonmar]
Set tso->link to END_TSO_QUEUE in resumeThread.  This is probably
harmless, but it can trigger an assertion failure later on in the
scheduler.

23 years ago[project @ 2000-11-13 14:41:13 by simonmar]
simonmar [Mon, 13 Nov 2000 14:41:13 +0000 (14:41 +0000)]
[project @ 2000-11-13 14:41:13 by simonmar]
common up INTLIKE and CHARLIKE closures during GC.

23 years ago[project @ 2000-11-13 14:40:36 by simonmar]
simonmar [Mon, 13 Nov 2000 14:40:38 +0000 (14:40 +0000)]
[project @ 2000-11-13 14:40:36 by simonmar]
Change the names of several RTS symbols so they don't potentially
clash with user code.  All the symbols that may clash now have an
"stg_" prefix.

23 years ago[project @ 2000-11-13 14:34:37 by sewardj]
sewardj [Mon, 13 Nov 2000 14:34:37 +0000 (14:34 +0000)]
[project @ 2000-11-13 14:34:37 by sewardj]
-E now puts results in a <file>.hspp, doesn't dump to stdout.

23 years ago[project @ 2000-11-13 14:34:03 by sewardj]
sewardj [Mon, 13 Nov 2000 14:34:03 +0000 (14:34 +0000)]
[project @ 2000-11-13 14:34:03 by sewardj]
Document change in behaviour of -E.

23 years ago[project @ 2000-11-13 13:57:31 by sewardj]
sewardj [Mon, 13 Nov 2000 13:57:31 +0000 (13:57 +0000)]
[project @ 2000-11-13 13:57:31 by sewardj]
No longer needed.

23 years ago[project @ 2000-11-13 12:43:20 by sewardj]
sewardj [Mon, 13 Nov 2000 12:43:20 +0000 (12:43 +0000)]
[project @ 2000-11-13 12:43:20 by sewardj]
First shot at wiring up 'ghc --make'.

23 years ago[project @ 2000-11-13 11:04:06 by simonmar]
simonmar [Mon, 13 Nov 2000 11:04:07 +0000 (11:04 +0000)]
[project @ 2000-11-13 11:04:06 by simonmar]
Merge register allocation fix for idiv from before-ghci-branch:

  1.59.2.1  +2 -37     fptools/ghc/compiler/nativeGen/MachCode.lhs
  1.40.2.1  +15 -4     fptools/ghc/compiler/nativeGen/MachMisc.lhs
  1.43.2.1  +20 -1     fptools/ghc/compiler/nativeGen/PprMach.lhs
  1.26.2.1  +8 -15     fptools/ghc/compiler/nativeGen/RegAllocInfo.lhs

23 years ago[project @ 2000-11-13 10:34:52 by simonmar]
simonmar [Mon, 13 Nov 2000 10:34:52 +0000 (10:34 +0000)]
[project @ 2000-11-13 10:34:52 by simonmar]
merge rev. 1.46.2.3 from ghc-4-07-branch (fix line numbers in
defaulting warnings).

23 years ago[project @ 2000-11-13 10:25:00 by simonmar]
simonmar [Mon, 13 Nov 2000 10:25:00 +0000 (10:25 +0000)]
[project @ 2000-11-13 10:25:00 by simonmar]
merge rev. 1.45.2.4 from ghc-4.07-branch (fix location in warning)

23 years ago[project @ 2000-11-10 17:04:01 by simonpj]
simonpj [Fri, 10 Nov 2000 17:04:01 +0000 (17:04 +0000)]
[project @ 2000-11-10 17:04:01 by simonpj]
Fix printing of deprecations

23 years ago[project @ 2000-11-10 15:12:50 by simonpj]
simonpj [Fri, 10 Nov 2000 15:12:55 +0000 (15:12 +0000)]
[project @ 2000-11-10 15:12:50 by simonpj]
1. Outputable.PprStyle now carries a bit more information
In particular, the printing style tells whether to print
a name in unqualified form.  This used to be embedded in
a Name, but since Names now outlive a single compilation unit,
that's no longer appropriate.

So now the print-unqualified predicate is passed in the printing
style, not embedded in the Name.

   2. I tidied up HscMain a little.  Many of the showPass messages
have migraged into the repective pass drivers

23 years ago[project @ 2000-11-10 14:29:20 by simonmar]
simonmar [Fri, 10 Nov 2000 14:29:21 +0000 (14:29 +0000)]
[project @ 2000-11-10 14:29:20 by simonmar]
cleanups; print the full version number of the building compiler
(inc. the patchlevel) in verbose mode.

23 years ago[project @ 2000-11-10 09:51:45 by simonmar]
simonmar [Fri, 10 Nov 2000 09:51:45 +0000 (09:51 +0000)]
[project @ 2000-11-10 09:51:45 by simonmar]
Add test illustrating the BlockedOnDeadMVar exception.

23 years ago[project @ 2000-11-09 12:54:08 by simonmar]
simonmar [Thu, 9 Nov 2000 12:54:09 +0000 (12:54 +0000)]
[project @ 2000-11-09 12:54:08 by simonmar]
A few doLink fixes (merged from driver).

23 years ago[project @ 2000-11-09 12:19:02 by rrt]
rrt [Thu, 9 Nov 2000 12:19:02 +0000 (12:19 +0000)]
[project @ 2000-11-09 12:19:02 by rrt]
...that's about it.

23 years ago[project @ 2000-11-09 12:14:56 by rrt]
rrt [Thu, 9 Nov 2000 12:14:56 +0000 (12:14 +0000)]
[project @ 2000-11-09 12:14:56 by rrt]
...

23 years ago[project @ 2000-11-09 11:54:36 by rrt]
rrt [Thu, 9 Nov 2000 11:54:36 +0000 (11:54 +0000)]
[project @ 2000-11-09 11:54:36 by rrt]
Removed spurious file.

23 years ago[project @ 2000-11-09 11:53:41 by rrt]
rrt [Thu, 9 Nov 2000 11:53:41 +0000 (11:53 +0000)]
[project @ 2000-11-09 11:53:41 by rrt]
Spurious file.

23 years ago[project @ 2000-11-09 11:48:14 by rrt]
rrt [Thu, 9 Nov 2000 11:51:38 +0000 (11:51 +0000)]
[project @ 2000-11-09 11:48:14 by rrt]
...