simonmar [Tue, 14 Nov 2000 12:53:29 +0000 (12:53 +0000)]
[project @ 2000-11-14 12:53:29 by simonmar]
wibble
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
simonmar [Tue, 14 Nov 2000 12:51:51 +0000 (12:51 +0000)]
[project @ 2000-11-14 12:51:51 by simonmar]
more renamings
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.
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.
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
simonpj [Tue, 14 Nov 2000 11:25:22 +0000 (11:25 +0000)]
[project @ 2000-11-14 11:25:22 by simonpj]
vanilla -> constantIdInfo
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.
simonpj [Tue, 14 Nov 2000 10:46:41 +0000 (10:46 +0000)]
[project @ 2000-11-14 10:46:39 by simonpj]
Compiles now
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.
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.
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
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 :-)
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.
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.
sewardj [Mon, 13 Nov 2000 16:16:05 +0000 (16:16 +0000)]
[project @ 2000-11-13 16:16:05 by sewardj]
Implement -ddump-hi.
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.
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.
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.
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.
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.
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.
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.
sewardj [Mon, 13 Nov 2000 13:57:31 +0000 (13:57 +0000)]
[project @ 2000-11-13 13:57:31 by sewardj]
No longer needed.
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'.
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
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).
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)
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
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
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.
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.
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).
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.
rrt [Thu, 9 Nov 2000 12:14:56 +0000 (12:14 +0000)]
[project @ 2000-11-09 12:14:56 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.
rrt [Thu, 9 Nov 2000 11:53:41 +0000 (11:53 +0000)]
[project @ 2000-11-09 11:53:41 by rrt]
Spurious file.
rrt [Thu, 9 Nov 2000 11:51:38 +0000 (11:51 +0000)]
[project @ 2000-11-09 11:48:14 by rrt]
...
rrt [Thu, 9 Nov 2000 11:44:44 +0000 (11:44 +0000)]
[project @ 2000-11-09 11:44:44 by rrt]
...
rrt [Thu, 9 Nov 2000 11:42:29 +0000 (11:42 +0000)]
[project @ 2000-11-09 11:42:29 by rrt]
IS install continues...
rrt [Thu, 9 Nov 2000 11:40:35 +0000 (11:40 +0000)]
[project @ 2000-11-09 11:40:35 by rrt]
More files added.
rrt [Thu, 9 Nov 2000 11:39:48 +0000 (11:39 +0000)]
[project @ 2000-11-09 11:39:48 by rrt]
Updated to reflect new, more positive, reality.
rrt [Thu, 9 Nov 2000 11:37:25 +0000 (11:37 +0000)]
[project @ 2000-11-09 11:37:24 by rrt]
Add more bits of InstallShield setup
rrt [Thu, 9 Nov 2000 11:28:23 +0000 (11:28 +0000)]
[project @ 2000-11-09 11:28:23 by rrt]
Second stab at getting the GHC InstallShield setup into CVS in a sensible
manner; the previous sticking point was having spaces in filenames.
simonpj [Thu, 9 Nov 2000 08:18:12 +0000 (08:18 +0000)]
[project @ 2000-11-09 08:18:11 by simonpj]
Make data constructors visible in unfoldings
simonmar [Wed, 8 Nov 2000 17:33:55 +0000 (17:33 +0000)]
[project @ 2000-11-08 17:33:55 by simonmar]
Fix a test that was backwards. Now we get unfoldings in the iface
files again.
simonmar [Wed, 8 Nov 2000 16:24:34 +0000 (16:24 +0000)]
[project @ 2000-11-08 16:24:34 by simonmar]
Fix a bug in object splitting which meant that no object files were
being created :-)
simonmar [Wed, 8 Nov 2000 16:24:01 +0000 (16:24 +0000)]
[project @ 2000-11-08 16:24:01 by simonmar]
remove traces
simonpj [Wed, 8 Nov 2000 15:54:05 +0000 (15:54 +0000)]
[project @ 2000-11-08 15:54:05 by simonpj]
Move fixity decl inside ifdef Hugs
simonmar [Wed, 8 Nov 2000 15:25:25 +0000 (15:25 +0000)]
[project @ 2000-11-08 15:25:25 by simonmar]
- Finder now copes with .hi-boot files
- some driver fixes: the -osuf flag turns out to mean "object suffix"
rather than "output suffix" (duh. it makes more sense that way after
all).
simonmar [Wed, 8 Nov 2000 15:14:03 +0000 (15:14 +0000)]
[project @ 2000-11-08 15:14:03 by simonmar]
fix heavy-handed merge-o :(
simonpj [Wed, 8 Nov 2000 14:52:08 +0000 (14:52 +0000)]
[project @ 2000-11-08 14:52:06 by simonpj]
Compiles most of the Prelude; versioning still not good
simonmar [Wed, 8 Nov 2000 13:51:58 +0000 (13:51 +0000)]
[project @ 2000-11-08 13:51:58 by simonmar]
Fixes for compiling w/ 4.08.1
simonmar [Wed, 8 Nov 2000 13:51:16 +0000 (13:51 +0000)]
[project @ 2000-11-08 13:51:16 by simonmar]
-fvia-C for Main.hs; it uses myThreadId
simonmar [Wed, 8 Nov 2000 10:22:03 +0000 (10:22 +0000)]
[project @ 2000-11-08 10:22:03 by simonmar]
document options to ./configure.
simonmar [Tue, 7 Nov 2000 18:00:26 +0000 (18:00 +0000)]
[project @ 2000-11-07 18:00:26 by simonmar]
URK! when scanning the heap we might find a DEAD_WEAK that either (a)
is the size of a WEAK, because we just overwrote its info table, or
(b) is the size of a DEAD_WEAK, because its been around at least one
GC. The code herein only accounted for case (a).
Fix: check the word at location p+sizeof(DEAD_WEAK), see whether it
looks like an info table ptr, and if so we're in case (b). This is
guaranteed to work, because we know that in case (a) the appropriate
word will be a closure pointer.
simonmar [Tue, 7 Nov 2000 17:05:47 +0000 (17:05 +0000)]
[project @ 2000-11-07 17:05:47 by simonmar]
Clean ups:
- reduce the namespace pollution of StgTypes.h, it doesn't define
the shorthand versions any more (W_, I_ etc.). These are moved into
Stg.h. StgTypes.h also defines StgClosure as an "opaque" struct.
- RtsAPI.h is now standalone, and includes HsFFI.h and thereby
config.h & StgTypes.h. Now we don't need to #include "Stg.h" in
*_stub.c.
- all the rts_mkXXXX and rts_getXXXX functions are defined in terms
of the HsXXXX types rather than random C types (this fixes some
potential bugs in our foreign export support).
- added HsWord type, to match StgWord. The Haskell version of this
type isn't "documented", but perhaps it should be.
simonmar [Tue, 7 Nov 2000 16:39:15 +0000 (16:39 +0000)]
[project @ 2000-11-07 16:39:15 by simonmar]
small fixes
simonmar [Tue, 7 Nov 2000 16:24:25 +0000 (16:24 +0000)]
[project @ 2000-11-07 16:24:25 by simonmar]
wibble
simonmar [Tue, 7 Nov 2000 16:03:38 +0000 (16:03 +0000)]
[project @ 2000-11-07 16:03:38 by simonmar]
Fix compilation with 4.08.1
simonmar [Tue, 7 Nov 2000 15:38:31 +0000 (15:38 +0000)]
[project @ 2000-11-07 15:38:31 by simonmar]
conditionalise imports of StgInterp etc.
simonmar [Tue, 7 Nov 2000 15:28:36 +0000 (15:28 +0000)]
[project @ 2000-11-07 15:28:36 by simonmar]
add hsc2hs, missed in the merge
simonmar [Tue, 7 Nov 2000 15:21:43 +0000 (15:21 +0000)]
[project @ 2000-11-07 15:21:38 by simonmar]
This commit completes the merge of compiler part
of the HEAD with the before-ghci-branch to
before-ghci-branch-merged.
simonmar [Tue, 7 Nov 2000 14:07:09 +0000 (14:07 +0000)]
[project @ 2000-11-07 14:07:09 by simonmar]
add -optc-DGHCI
simonmar [Tue, 7 Nov 2000 14:01:53 +0000 (14:01 +0000)]
[project @ 2000-11-07 14:01:53 by simonmar]
revert 1.109 (ghci isn't included in DIRS unless we're bootstrapping)
simonmar [Tue, 7 Nov 2000 13:30:41 +0000 (13:30 +0000)]
[project @ 2000-11-07 13:30:40 by simonmar]
Wave goodbye to FLOATS_AS_DOUBLES, it was a somewhat misconceived idea
which will cause trouble with the FFI on 64-bit machines.
simonpj [Tue, 7 Nov 2000 13:12:24 +0000 (13:12 +0000)]
[project @ 2000-11-07 13:12:21 by simonpj]
More small changes
simonmar [Tue, 7 Nov 2000 10:42:57 +0000 (10:42 +0000)]
[project @ 2000-11-07 10:42:55 by simonmar]
merge before-ghci -> before-ghci-branch-merged into the ghc
(non-compiler) parts of the tree.
simonmar [Tue, 7 Nov 2000 10:38:17 +0000 (10:38 +0000)]
[project @ 2000-11-07 10:38:17 by simonmar]
kill this file once and for all
simonmar [Tue, 7 Nov 2000 10:20:03 +0000 (10:20 +0000)]
[project @ 2000-11-07 10:20:03 by simonmar]
Merge before-ghci -> before-ghci-branch-merged in the non-ghc parts of
the tree.
rrt [Mon, 6 Nov 2000 16:43:28 +0000 (16:43 +0000)]
[project @ 2000-11-06 16:43:28 by rrt]
Updated to reflect reality and be a bit more helpful.
simonmar [Mon, 6 Nov 2000 13:14:52 +0000 (13:14 +0000)]
[project @ 2000-11-06 13:14:52 by simonmar]
build the rts *after* the compiler
simonmar [Mon, 6 Nov 2000 13:13:17 +0000 (13:13 +0000)]
[project @ 2000-11-06 13:13:17 by simonmar]
Change the way source distributions are built.
Now we do a 'make boot', 'make distclean', build a link tree to filter
some files out (eg. .cvsignore), and finally tar up the result. The
process isn't bulletproof, so it makes sense to compare the manifest
against previous versions.
Now you don't have to be in a linked build tree to make a source
distribution.
simonmar [Mon, 6 Nov 2000 13:09:52 +0000 (13:09 +0000)]
[project @ 2000-11-06 13:09:52 by simonmar]
.depend should be in DIST_CLEAN, not MAINTAINER_CLEAN
simonmar [Mon, 6 Nov 2000 12:00:49 +0000 (12:00 +0000)]
[project @ 2000-11-06 12:00:49 by simonmar]
sgmlverb.c is in maintainer-clean
simonmar [Mon, 6 Nov 2000 11:33:22 +0000 (11:33 +0000)]
[project @ 2000-11-06 11:33:22 by simonmar]
config.status should be in distclean
simonpj [Mon, 6 Nov 2000 10:52:00 +0000 (10:52 +0000)]
[project @ 2000-11-06 10:52:00 by simonpj]
Remove FunDeps hi-boot files
simonpj [Mon, 6 Nov 2000 10:50:29 +0000 (10:50 +0000)]
[project @ 2000-11-06 10:50:29 by simonpj]
Remove SpecEnv.hi-boot-5; hasnt been used for ages
simonmar [Mon, 6 Nov 2000 10:09:42 +0000 (10:09 +0000)]
[project @ 2000-11-06 10:09:42 by simonmar]
enhance the clean target for doc subdirs
simonmar [Mon, 6 Nov 2000 10:09:08 +0000 (10:09 +0000)]
[project @ 2000-11-06 10:09:08 by simonmar]
clean more files: %.s for each %.hs/%.lhs, *_stub.{c,h}, *.raw_s.
simonpj [Mon, 6 Nov 2000 08:15:24 +0000 (08:15 +0000)]
[project @ 2000-11-06 08:15:20 by simonpj]
Dealing with instance-decl imports; and removing unnecessary imports
simonpj [Fri, 3 Nov 2000 17:10:58 +0000 (17:10 +0000)]
[project @ 2000-11-03 17:10:57 by simonpj]
More renamer... not in a working state I fear
simonmar [Fri, 3 Nov 2000 17:09:00 +0000 (17:09 +0000)]
[project @ 2000-11-03 17:09:00 by simonmar]
clean up Parser.info in MAINTAINER_CLEAN.
simonmar [Fri, 3 Nov 2000 17:08:36 +0000 (17:08 +0000)]
[project @ 2000-11-03 17:08:36 by simonmar]
"make clean" fixes
simonmar [Fri, 3 Nov 2000 16:54:52 +0000 (16:54 +0000)]
[project @ 2000-11-03 16:54:52 by simonmar]
"make clean" fixes
simonmar [Fri, 3 Nov 2000 16:39:00 +0000 (16:39 +0000)]
[project @ 2000-11-03 16:39:00 by simonmar]
Remove unused files.
simonmar [Fri, 3 Nov 2000 16:30:53 +0000 (16:30 +0000)]
[project @ 2000-11-03 16:30:53 by simonmar]
maintainer-clean should include distclean.
simonmar [Fri, 3 Nov 2000 16:23:39 +0000 (16:23 +0000)]
[project @ 2000-11-03 16:23:37 by simonmar]
"make clean" in here now works as advertised.
sewardj [Fri, 3 Nov 2000 11:36:30 +0000 (11:36 +0000)]
[project @ 2000-11-03 11:36:30 by sewardj]
Finally get CompManager to compile. Also rm some redundant imports.
sewardj [Fri, 3 Nov 2000 11:30:23 +0000 (11:30 +0000)]
[project @ 2000-11-03 11:30:23 by sewardj]
stgSyn/StgInterp_HC_OPTS ---> ghci/StgInterp_HC_OPTS
simonmar [Fri, 3 Nov 2000 11:29:00 +0000 (11:29 +0000)]
[project @ 2000-11-03 11:29:00 by simonmar]
don't generate heimdall.summary unless -heimdall is on.
simonmar [Fri, 3 Nov 2000 11:25:35 +0000 (11:25 +0000)]
[project @ 2000-11-03 11:25:35 by simonmar]
Prepend "./" to the command to run only if the file exists in the
current directory, rather than based on whether the command begins
with a ''/'. This means that 'make HC=ghc-4.08.1' in ghc/tests works
as expected.
simonmar [Fri, 3 Nov 2000 10:42:39 +0000 (10:42 +0000)]
[project @ 2000-11-03 10:42:39 by simonmar]
remind me that -nohi doesn't do anything
simonmar [Fri, 3 Nov 2000 10:38:26 +0000 (10:38 +0000)]
[project @ 2000-11-03 10:38:26 by simonmar]
remind me what this test does
simonmar [Thu, 2 Nov 2000 14:33:15 +0000 (14:33 +0000)]
[project @ 2000-11-02 14:33:15 by simonmar]
Found a better way to do the arch test for $(GhcWithInterpreter), used
it for $(GhcWithNativeCodeGen) too and removed one $(shell...).
simonmar [Thu, 2 Nov 2000 14:27:02 +0000 (14:27 +0000)]
[project @ 2000-11-02 14:27:01 by simonmar]
Cleaning up the configuration/build process.
- New build.mk option: GhcWithInterpreter. It has reasonable
defaults, and shouldn't need to be overriden (you get the interpreter
if you're bootstrapping with ghc 4.09+ on an ELF architecture, at
the moment).
- compilation manager now lives in compiler/compMan. compiler/ghci
contains only interpreter-related files.
- WithGhcHc has gone, it now defaults to $(GHC). This is so that
we can reliably determine the version of $(GHC) using the stuff that
configure tells us. configure gets a new --with-ghc option so you
can specify which ghc to use.
sewardj [Thu, 2 Nov 2000 13:58:45 +0000 (13:58 +0000)]
[project @ 2000-11-02 13:58:44 by sewardj]
Most, but not all changes needed to get CompManager to compile.
simonmar [Thu, 2 Nov 2000 13:36:44 +0000 (13:36 +0000)]
[project @ 2000-11-02 13:36:44 by simonmar]
Remove GhcWithRegisterised since it isn't used. It might be
re-instated later if/when we decide on the right way to implement it.
simonpj [Thu, 2 Nov 2000 10:17:49 +0000 (10:17 +0000)]
[project @ 2000-11-02 10:17:49 by simonpj]
Omitted do-nothing case in SimplCore
simonpj [Wed, 1 Nov 2000 17:15:30 +0000 (17:15 +0000)]
[project @ 2000-11-01 17:15:28 by simonpj]
More renamer commits
Versioning now works properly I think.
The main irritation is that interface files now have fuly-qualified names for
*everything*, even things defined in that module. This is a deficiency in
the pretty printing for interface files. Probable solution: add something
to the SDoc styles. But not today.