ghc-hetmet.git
24 years ago[project @ 1999-10-05 10:30:26 by simonmar]
simonmar [Tue, 5 Oct 1999 10:30:35 +0000 (10:30 +0000)]
[project @ 1999-10-05 10:30:26 by simonmar]
Build System Cleanup
--------------------

This commit is intended to clear the way for automatic RPM building.
In particular, the idea is that 'make install' doesn't build anything,
so you can do

make install prefix=/some/tmp/prefix

to install everything in /some/tmp/prefix, but leave the hardwired-in
install directories the same.

For scripts that depend on the install dir (currently just the GHC
driver and mkdependHS), we now build two versions: <script>-inplace,
which is used for running the script from the build tree, and
<script>, which is the to-be-installed version.

NOTE: binary distributions are now a little bit trickier to build.
You *must* include the line "BIN_DIST=1" in your build.mk if you
intend to make a binary distribution from the current build tree.
This is because certain scripts have to be built differently, and we
don't rebuild them when doing 'make binary-dist' anymore (since 'make
binary-dist' just does a 'make install' with a re-targetted prefix,
just like the RPM builder).

Other changes

- the binary-dist machinery is now all in fptools/Makefile
- removed a gratuitous $(package)-$(version) level of
  directories from the binary distribution.
- binary distributions are now placed under the package
  name, rather than fptools/fptools.
- various other minor cleanups.

24 years ago[project @ 1999-10-05 10:22:40 by simonmar]
simonmar [Tue, 5 Oct 1999 10:22:42 +0000 (10:22 +0000)]
[project @ 1999-10-05 10:22:40 by simonmar]
--with-ghc-hc is now --with-hc (to be used as the "locally installed
Haskell compiler" for all projects).

24 years ago[project @ 1999-10-05 09:58:13 by simonmar]
simonmar [Tue, 5 Oct 1999 09:58:13 +0000 (09:58 +0000)]
[project @ 1999-10-05 09:58:13 by simonmar]
Remove hstags for now.

24 years ago[project @ 1999-10-05 09:04:30 by kglynn]
kglynn [Tue, 5 Oct 1999 09:04:30 +0000 (09:04 +0000)]
[project @ 1999-10-05 09:04:30 by kglynn]
(keving)

Much simplified and beautified CPR analysis code.  (And also much
shorter, we'd better write this up before it disappears).

Added (constant) functions to the abstract domain. Note that
Fun^n Bot (n >= 1) == Bot and likewise for Top

Treats divergent computations as Bot (rather than Top as previous) so
non-divergent paths dominate which allows us to generate more accurate
CPR info (see e.g. chr).

We use the result of strictness analysis to tell us if an Id is
divergent (when applied to sufficient args), therefore we should run
after the strictness analysis pass.

24 years ago[project @ 1999-10-05 09:02:30 by simonmar]
simonmar [Tue, 5 Oct 1999 09:02:39 +0000 (09:02 +0000)]
[project @ 1999-10-05 09:02:30 by simonmar]
Flatten out the tuple of bounds in the Array, MutableArray and
ByteArray datatypes.  This improves performance of heavy array
manipulations quite significantly.

24 years ago[project @ 1999-10-04 16:14:34 by simonmar]
simonmar [Mon, 4 Oct 1999 16:14:34 +0000 (16:14 +0000)]
[project @ 1999-10-04 16:14:34 by simonmar]
Service signal handlers if we get an EINTR from select(2).

24 years ago[project @ 1999-10-04 16:13:18 by simonmar]
simonmar [Mon, 4 Oct 1999 16:13:18 +0000 (16:13 +0000)]
[project @ 1999-10-04 16:13:18 by simonmar]
allow raising exceptions in threads blocked on I/O or time delays

24 years ago[project @ 1999-10-04 15:00:49 by simonmar]
simonmar [Mon, 4 Oct 1999 15:00:49 +0000 (15:00 +0000)]
[project @ 1999-10-04 15:00:49 by simonmar]
typo

24 years ago[project @ 1999-10-04 14:03:47 by simonmar]
simonmar [Mon, 4 Oct 1999 14:03:50 +0000 (14:03 +0000)]
[project @ 1999-10-04 14:03:47 by simonmar]
cleanup

24 years ago[project @ 1999-10-04 13:57:41 by simonmar]
simonmar [Mon, 4 Oct 1999 13:57:41 +0000 (13:57 +0000)]
[project @ 1999-10-04 13:57:41 by simonmar]
hstags doesn't work any more - disconnect it

24 years ago[project @ 1999-10-01 08:41:06 by sof]
sof [Fri, 1 Oct 1999 08:41:06 +0000 (08:41 +0000)]
[project @ 1999-10-01 08:41:06 by sof]
avoid giving 'sub's empty parameter lists

24 years ago[project @ 1999-09-30 16:20:14 by sof]
sof [Thu, 30 Sep 1999 16:20:14 +0000 (16:20 +0000)]
[project @ 1999-09-30 16:20:14 by sof]
For strftime()'s format string, stay within the bounds of ISO/ANSI C

24 years ago[project @ 1999-09-30 15:50:02 by sof]
sof [Thu, 30 Sep 1999 15:50:04 +0000 (15:50 +0000)]
[project @ 1999-09-30 15:50:02 by sof]
Simplified the interface between CPUTime.getCPUTime and getCPUTime()

24 years ago[project @ 1999-09-30 14:25:46 by sof]
sof [Thu, 30 Sep 1999 14:25:47 +0000 (14:25 +0000)]
[project @ 1999-09-30 14:25:46 by sof]
Common up Hugs&ghc implementation of addToClockTime, clearing up some potential signed vs. unsigned problems in the process

24 years ago[project @ 1999-09-30 12:42:25 by sof]
sof [Thu, 30 Sep 1999 12:42:26 +0000 (12:42 +0000)]
[project @ 1999-09-30 12:42:25 by sof]
primGetClockTime was incorrectly calling prim_getClockTime; fixed dormant bug in Show instance for ClockTime

24 years ago[project @ 1999-09-30 12:38:44 by sof]
sof [Thu, 30 Sep 1999 12:38:44 +0000 (12:38 +0000)]
[project @ 1999-09-30 12:38:44 by sof]
removed prim_getClockTime

24 years ago[project @ 1999-09-30 12:35:50 by sof]
sof [Thu, 30 Sep 1999 12:35:50 +0000 (12:35 +0000)]
[project @ 1999-09-30 12:35:50 by sof]
more of same

24 years ago[project @ 1999-09-30 12:35:04 by sof]
sof [Thu, 30 Sep 1999 12:35:04 +0000 (12:35 +0000)]
[project @ 1999-09-30 12:35:04 by sof]
don't re-define O_NOCTTY

24 years ago[project @ 1999-09-26 16:07:38 by sof]
sof [Sun, 26 Sep 1999 16:07:38 +0000 (16:07 +0000)]
[project @ 1999-09-26 16:07:38 by sof]
Doc'ed -fwarn-missing-fields

24 years ago[project @ 1999-09-26 16:05:34 by sof]
sof [Sun, 26 Sep 1999 16:05:35 +0000 (16:05 +0000)]
[project @ 1999-09-26 16:05:34 by sof]
Record construction tests

24 years ago[project @ 1999-09-26 16:01:08 by sof]
sof [Sun, 26 Sep 1999 16:01:12 +0000 (16:01 +0000)]
[project @ 1999-09-26 16:01:08 by sof]
Increased friendliness re: record construction a little:

  * constructions that fail to mention one or more strict
    fields are now flagged as an error, which the Report demands.
  * Optionally warn about other missing fields. -fwarn-missing-fields
    takes you there, and it is in currently in the '-W' set of
    warnings.

24 years ago[project @ 1999-09-22 11:53:33 by sof]
sof [Wed, 22 Sep 1999 11:53:33 +0000 (11:53 +0000)]
[project @ 1999-09-22 11:53:33 by sof]
At startup time, install a SIGINT termination handler which calls
shutdownHaskellAndExit(), if invoked.

24 years ago[project @ 1999-09-21 14:28:14 by sof]
sof [Tue, 21 Sep 1999 14:28:14 +0000 (14:28 +0000)]
[project @ 1999-09-21 14:28:14 by sof]
Addr is now a _casm_ free zone + added indexStablePtrOffAddr

24 years ago[project @ 1999-09-21 09:46:04 by sof]
sof [Tue, 21 Sep 1999 09:46:04 +0000 (09:46 +0000)]
[project @ 1999-09-21 09:46:04 by sof]
v1.2 of Show Addr instance, pad out to fixed length

24 years ago[project @ 1999-09-21 09:19:57 by sof]
sof [Tue, 21 Sep 1999 09:19:57 +0000 (09:19 +0000)]
[project @ 1999-09-21 09:19:57 by sof]
when mangling gcc output, make sure we distinguish between a \ at EOL and a \ in DOS paths

24 years ago[project @ 1999-09-21 09:02:05 by sof]
sof [Tue, 21 Sep 1999 09:02:05 +0000 (09:02 +0000)]
[project @ 1999-09-21 09:02:05 by sof]
updated

24 years ago[project @ 1999-09-21 09:01:38 by sof]
sof [Tue, 21 Sep 1999 09:01:38 +0000 (09:01 +0000)]
[project @ 1999-09-21 09:01:38 by sof]
Reg. tests for h{Set,Get}Posn + hSeek

24 years ago[project @ 1999-09-21 09:00:46 by sof]
sof [Tue, 21 Sep 1999 09:00:46 +0000 (09:00 +0000)]
[project @ 1999-09-21 09:00:46 by sof]
Testing the sanity of Show on Addr

24 years ago[project @ 1999-09-21 08:59:46 by sof]
sof [Tue, 21 Sep 1999 08:59:47 +0000 (08:59 +0000)]
[project @ 1999-09-21 08:59:46 by sof]
Moved Show instance for Addr from PrelAddr to Addr & made it display
the Addr in hex (+ have it cope with ptrs with MSB set.)

24 years ago[project @ 1999-09-20 11:18:05 by simonmar]
simonmar [Mon, 20 Sep 1999 11:18:05 +0000 (11:18 +0000)]
[project @ 1999-09-20 11:18:05 by simonmar]
Fix for compiling w/ ghc-2.10

24 years ago[project @ 1999-09-20 10:22:40 by sof]
sof [Mon, 20 Sep 1999 10:22:40 +0000 (10:22 +0000)]
[project @ 1999-09-20 10:22:40 by sof]
Added 'updateIORef :: IORef a -> (a->a) -> IO ()'

24 years ago[project @ 1999-09-20 10:18:29 by simonmar]
simonmar [Mon, 20 Sep 1999 10:18:30 +0000 (10:18 +0000)]
[project @ 1999-09-20 10:18:29 by simonmar]
- fix bug in setSocketOption__ which meant that trying to set
  SO_REUSEADDR on Linux (and possibly other OS's) didn't work.

- add rudimentary non-blocking connect support.

24 years ago[project @ 1999-09-20 08:36:35 by panne]
panne [Mon, 20 Sep 1999 08:36:35 +0000 (08:36 +0000)]
[project @ 1999-09-20 08:36:35 by panne]
Need sys/types.h for off_t

24 years ago[project @ 1999-09-19 19:30:26 by sof]
sof [Sun, 19 Sep 1999 19:30:26 +0000 (19:30 +0000)]
[project @ 1999-09-19 19:30:26 by sof]
exts/ is now also a _ccall_ free zone.

24 years ago[project @ 1999-09-19 19:30:04 by sof]
sof [Sun, 19 Sep 1999 19:30:04 +0000 (19:30 +0000)]
[project @ 1999-09-19 19:30:04 by sof]
* Re-exported IO.HandlePosn, i.e.,

         type HandlePosition = Integer
 data HandlePosn = HandlePosn Handle HandlePosition

* Added hTell :: Handle -> HandlePosition
  (merely a wrapper for IO.hGetPosn )

* Added hSetBinaryMode :: Handle -> Bool -> IO Bool
  for dynamically changing the 'translation mode' of a Handle.
  This stuff is only useful on platforms that make a distinction
  between text and binary files (e.g., Win32)

24 years ago[project @ 1999-09-19 19:27:10 by sof]
sof [Sun, 19 Sep 1999 19:27:10 +0000 (19:27 +0000)]
[project @ 1999-09-19 19:27:10 by sof]
IOExts.hSetBinaryMode wrapper

24 years ago[project @ 1999-09-19 19:26:57 by sof]
sof [Sun, 19 Sep 1999 19:26:57 +0000 (19:26 +0000)]
[project @ 1999-09-19 19:26:57 by sof]
updates

24 years ago[project @ 1999-09-19 19:26:14 by sof]
sof [Sun, 19 Sep 1999 19:26:14 +0000 (19:26 +0000)]
[project @ 1999-09-19 19:26:14 by sof]
wibble

24 years ago[project @ 1999-09-19 19:25:24 by sof]
sof [Sun, 19 Sep 1999 19:25:24 +0000 (19:25 +0000)]
[project @ 1999-09-19 19:25:24 by sof]
Allow seeking on devices and beyond EOF

24 years ago[project @ 1999-09-19 19:22:14 by sof]
sof [Sun, 19 Sep 1999 19:22:14 +0000 (19:22 +0000)]
[project @ 1999-09-19 19:22:14 by sof]
Have sizeof_time_t report the size in *words*

24 years ago[project @ 1999-09-19 19:21:22 by sof]
sof [Sun, 19 Sep 1999 19:21:22 +0000 (19:21 +0000)]
[project @ 1999-09-19 19:21:22 by sof]
unify two near-identical impls of toClockSec()

24 years ago[project @ 1999-09-19 19:20:50 by sof]
sof [Sun, 19 Sep 1999 19:20:50 +0000 (19:20 +0000)]
[project @ 1999-09-19 19:20:50 by sof]
* drop the restriction that seeks cannot be made on devices & beyond
  EOFs. If the underlying lseek() doesn't like us doing either, it'll
  let us know.
* When asking for the current position under Win32, take into account
  that lseek() reports the _untranslated_ position, so adjust the resulting
  position by scanning input buffer looking for \n's (and treat them
  as if \r\n.)

24 years ago[project @ 1999-09-19 19:15:26 by sof]
sof [Sun, 19 Sep 1999 19:15:26 +0000 (19:15 +0000)]
[project @ 1999-09-19 19:15:26 by sof]
re-organised the initialisation of the 'flags' a little; record whether a file handle is in 'binary' translation mode or not.

24 years ago[project @ 1999-09-19 19:12:39 by sof]
sof [Sun, 19 Sep 1999 19:12:42 +0000 (19:12 +0000)]
[project @ 1999-09-19 19:12:39 by sof]
Drop the use of _ccall_, _casm_ and lit-lits in std/, "foreign import" is
the future.

24 years ago[project @ 1999-09-19 19:10:54 by sof]
sof [Sun, 19 Sep 1999 19:10:54 +0000 (19:10 +0000)]
[project @ 1999-09-19 19:10:54 by sof]
nullAddr: avoid the use of a lit-lit

24 years ago[project @ 1999-09-17 11:38:18 by sof]
sof [Fri, 17 Sep 1999 11:38:18 +0000 (11:38 +0000)]
[project @ 1999-09-17 11:38:18 by sof]
Consistently use .dll_o for toplevel object files that are to be included with all DLL-based apps

24 years ago[project @ 1999-09-17 11:29:51 by sof]
sof [Fri, 17 Sep 1999 11:29:51 +0000 (11:29 +0000)]
[project @ 1999-09-17 11:29:51 by sof]
expected (std)output of show001

24 years ago[project @ 1999-09-17 11:23:33 by sof]
sof [Fri, 17 Sep 1999 11:23:33 +0000 (11:23 +0000)]
[project @ 1999-09-17 11:23:33 by sof]
Testing Show for Maybes and Eithers

24 years ago[project @ 1999-09-17 11:22:53 by sof]
sof [Fri, 17 Sep 1999 11:22:53 +0000 (11:22 +0000)]
[project @ 1999-09-17 11:22:53 by sof]
Fixed bogus Show instances for Maybe and Either

24 years ago[project @ 1999-09-17 10:43:51 by sof]
sof [Fri, 17 Sep 1999 10:43:54 +0000 (10:43 +0000)]
[project @ 1999-09-17 10:43:51 by sof]
DLL building tweaks, including terrible hack to compile PrelMain

24 years ago[project @ 1999-09-17 10:43:07 by sof]
sof [Fri, 17 Sep 1999 10:43:07 +0000 (10:43 +0000)]
[project @ 1999-09-17 10:43:07 by sof]
posix is not for Win32

24 years ago[project @ 1999-09-17 09:38:33 by sof]
sof [Fri, 17 Sep 1999 09:38:35 +0000 (09:38 +0000)]
[project @ 1999-09-17 09:38:33 by sof]
Non-blocking I/O isn't supported on 'pure' Win32 platforms.

24 years ago[project @ 1999-09-17 09:16:57 by simonpj]
simonpj [Fri, 17 Sep 1999 09:16:57 +0000 (09:16 +0000)]
[project @ 1999-09-17 09:16:57 by simonpj]
[This is the commit message that should have gone with the
 typechecker commit a few minutes ago, whose message said
~/tmp/msg1
]

Improve location and context information on Insts (notably Dicts),
and thereby improve error messages that relate to dictionaries.

This arose from Marc van Doguen's suggestion.

24 years ago[project @ 1999-09-17 09:15:22 by simonpj]
simonpj [Fri, 17 Sep 1999 09:15:44 +0000 (09:15 +0000)]
[project @ 1999-09-17 09:15:22 by simonpj]
This bunch of commits represents work in progress on inlining and
worker/wrapper stuff.

Currently, I think it makes the compiler slightly worse than 4.04, for
reasons I don't yet understand.  But it means that Simon and I can
both peer at what is going on.

* Substantially improve handling of coerces in worker/wrapper

* exprIsDupable for an application (f e1 .. en) wasn't calling exprIsDupable
  on the arguments!!  So applications with few, but large, args were being dupliated.

* sizeExpr on an application wasn't doing a nukeScrutDiscount on the arg of
  an application!!  So bogus discounts could accumulate from arguments!

* Improve handling of INLINE pragmas in calcUnfoldingGuidance.  It was really
  wrong before

24 years ago[project @ 1999-09-17 09:13:14 by simonpj]
simonpj [Fri, 17 Sep 1999 09:13:17 +0000 (09:13 +0000)]
[project @ 1999-09-17 09:13:14 by simonpj]
/home/simonpj/tmp/msg1

24 years ago[project @ 1999-09-17 09:11:20 by simonpj]
simonpj [Fri, 17 Sep 1999 09:11:39 +0000 (09:11 +0000)]
[project @ 1999-09-17 09:11:20 by simonpj]
Remove SST.lhs

24 years ago[project @ 1999-09-17 09:09:55 by simonpj]
simonpj [Fri, 17 Sep 1999 09:09:55 +0000 (09:09 +0000)]
[project @ 1999-09-17 09:09:55 by simonpj]
Add drvrun007, a test that crashed Hugs

24 years ago[project @ 1999-09-16 19:54:37 by sof]
sof [Thu, 16 Sep 1999 19:54:37 +0000 (19:54 +0000)]
[project @ 1999-09-16 19:54:37 by sof]
toInt wibble

24 years ago[project @ 1999-09-16 19:37:58 by sof]
sof [Thu, 16 Sep 1999 19:37:58 +0000 (19:37 +0000)]
[project @ 1999-09-16 19:37:58 by sof]
Added expected stdout to reg. test for Float&Double exceptional IEEE values.
(the numbers were produced on a PII & with the MS C run-time -- don't be
surprised if you see slightly different numbers on your box.)

24 years ago[project @ 1999-09-16 19:26:01 by sof]
sof [Thu, 16 Sep 1999 19:26:01 +0000 (19:26 +0000)]
[project @ 1999-09-16 19:26:01 by sof]
checking Main.main's type + newtype def that caused a ghc-X (X<4.03, I think) TC crash

24 years ago[project @ 1999-09-16 18:40:05 by sof]
sof [Thu, 16 Sep 1999 18:40:05 +0000 (18:40 +0000)]
[project @ 1999-09-16 18:40:05 by sof]
Added Tree example from (older) versions of the Report

24 years ago[project @ 1999-09-16 18:37:50 by sof]
sof [Thu, 16 Sep 1999 18:37:50 +0000 (18:37 +0000)]
[project @ 1999-09-16 18:37:50 by sof]
Extended to show why changing the defn of ReadS (and read) is a good idea

24 years ago[project @ 1999-09-16 17:55:57 by sof]
sof [Thu, 16 Sep 1999 17:55:57 +0000 (17:55 +0000)]
[project @ 1999-09-16 17:55:57 by sof]
newtypes with a labelled field caused DS trouble sometime back

24 years ago[project @ 1999-09-16 17:38:12 by sof]
sof [Thu, 16 Sep 1999 17:38:12 +0000 (17:38 +0000)]
[project @ 1999-09-16 17:38:12 by sof]
Bunch of H98 tests I had lying about

24 years ago[project @ 1999-09-16 17:28:06 by sof]
sof [Thu, 16 Sep 1999 17:28:06 +0000 (17:28 +0000)]
[project @ 1999-09-16 17:28:06 by sof]
Don't insist that {-# OPTIONS ... #-} pragma is followed by \n

24 years ago[project @ 1999-09-16 17:23:38 by sof]
sof [Thu, 16 Sep 1999 17:23:38 +0000 (17:23 +0000)]
[project @ 1999-09-16 17:23:38 by sof]
Doc'ed the (retirement?) home of fromInt and toInt

24 years ago[project @ 1999-09-16 17:14:33 by sof]
sof [Thu, 16 Sep 1999 17:14:33 +0000 (17:14 +0000)]
[project @ 1999-09-16 17:14:33 by sof]
Handle -syslib com for non-installed builds ; dLL_ifs.hi is not a normal .hi file

24 years ago[project @ 1999-09-16 17:11:38 by sof]
sof [Thu, 16 Sep 1999 17:11:38 +0000 (17:11 +0000)]
[project @ 1999-09-16 17:11:38 by sof]
mkVanillaModule: sigh, special treatment for module Main.

24 years ago[project @ 1999-09-16 17:09:54 by sof]
sof [Thu, 16 Sep 1999 17:09:54 +0000 (17:09 +0000)]
[project @ 1999-09-16 17:09:54 by sof]
minor readIface tidyup

24 years ago[project @ 1999-09-16 14:32:02 by sof]
sof [Thu, 16 Sep 1999 14:32:02 +0000 (14:32 +0000)]
[project @ 1999-09-16 14:32:02 by sof]
Since 'gcc -E' is the CPP we're now using, make good use of its
support for directly generating dependencies.

24 years ago[project @ 1999-09-16 13:14:38 by simonmar]
simonmar [Thu, 16 Sep 1999 13:14:43 +0000 (13:14 +0000)]
[project @ 1999-09-16 13:14:38 by simonmar]
Cleanup of non-blocking I/O

- file descriptors are now always set to non-blocking mode.

- we don't do an inputReady operation on descriptors before
  attempting to read from them any more.

- the non-blocking flag on Handles has gone.

- the {set,clear}[Conn]NonBlockingFlag() functions have gone.

- the socket operations have been made to work properly with threads:
  accept is now non-blocking (it does a threadWaitRead instead of
  blocking), and the file descriptors returned by accept are set to
  non-blocking mode.

Win32 will need some adjustments, no doubt.

24 years ago[project @ 1999-09-16 12:29:55 by simonmar]
simonmar [Thu, 16 Sep 1999 12:29:55 +0000 (12:29 +0000)]
[project @ 1999-09-16 12:29:55 by simonmar]
Fix for DEBUG && !PROFILING

24 years ago[project @ 1999-09-16 10:25:31 by sof]
sof [Thu, 16 Sep 1999 10:25:31 +0000 (10:25 +0000)]
[project @ 1999-09-16 10:25:31 by sof]
mingw32 support

24 years ago[project @ 1999-09-16 09:17:25 by sof]
sof [Thu, 16 Sep 1999 09:17:25 +0000 (09:17 +0000)]
[project @ 1999-09-16 09:17:25 by sof]
NumExts.showListWith doc'ed

24 years ago[project @ 1999-09-16 08:53:34 by simonmar]
simonmar [Thu, 16 Sep 1999 08:53:34 +0000 (08:53 +0000)]
[project @ 1999-09-16 08:53:34 by simonmar]
Main trunk is now 4.05.

24 years ago[project @ 1999-09-16 08:40:16 by sof]
sof [Thu, 16 Sep 1999 08:40:16 +0000 (08:40 +0000)]
[project @ 1999-09-16 08:40:16 by sof]
Minor Win32 changes

24 years ago[project @ 1999-09-16 08:37:08 by sof]
sof [Thu, 16 Sep 1999 08:37:08 +0000 (08:37 +0000)]
[project @ 1999-09-16 08:37:08 by sof]
Magic Constant Removal

24 years ago[project @ 1999-09-16 08:35:38 by sof]
sof [Thu, 16 Sep 1999 08:35:38 +0000 (08:35 +0000)]
[project @ 1999-09-16 08:35:38 by sof]
Minor tweaks to do with DLL building

24 years ago[project @ 1999-09-16 08:33:54 by sof]
sof [Thu, 16 Sep 1999 08:33:54 +0000 (08:33 +0000)]
[project @ 1999-09-16 08:33:54 by sof]
Disable timer ticks on 'pure' Win32 platforms unless really needed
(for profiling) -- no reason to enable it elsewhere as the Win32 bits
doesn't currently support thread{Delay,Wait{Read,Write}}#

24 years ago[project @ 1999-09-16 08:29:01 by sof]
sof [Thu, 16 Sep 1999 08:29:01 +0000 (08:29 +0000)]
[project @ 1999-09-16 08:29:01 by sof]
Move DllMain() into separate file + it doesn't call startupHaskell()
any longer upon loading of the DLL. That is the task of clients of
the RTS.

24 years ago[project @ 1999-09-15 13:54:16 by sof]
sof [Wed, 15 Sep 1999 13:54:16 +0000 (13:54 +0000)]
[project @ 1999-09-15 13:54:16 by sof]
Add cmp_thread

24 years ago[project @ 1999-09-15 13:50:14 by sof]
sof [Wed, 15 Sep 1999 13:50:14 +0000 (13:50 +0000)]
[project @ 1999-09-15 13:50:14 by sof]
Code that trundles over stable_ptr_table need to start at index 1,
as index 0 is unused (and may contain garbage.)

24 years ago[project @ 1999-09-15 13:48:25 by sof]
sof [Wed, 15 Sep 1999 13:48:26 +0000 (13:48 +0000)]
[project @ 1999-09-15 13:48:25 by sof]
When constructing vanilla modules from ModuleNames, consult the HiMaps
to make sure we're using the 'right' kind of Module (i.e., it has got
the DLL flag correctly set.)

24 years ago[project @ 1999-09-15 13:46:38 by sof]
sof [Wed, 15 Sep 1999 13:46:38 +0000 (13:46 +0000)]
[project @ 1999-09-15 13:46:38 by sof]
mkVanillaModule: made DLL savvy

24 years ago[project @ 1999-09-15 13:46:28 by simonmar]
simonmar [Wed, 15 Sep 1999 13:46:29 +0000 (13:46 +0000)]
[project @ 1999-09-15 13:46:28 by simonmar]
Cost-centre heap profiling and symbol table heap profiling are now
merged into one file: ProfHeap.c.

24 years ago[project @ 1999-09-15 13:45:14 by simonmar]
simonmar [Wed, 15 Sep 1999 13:45:20 +0000 (13:45 +0000)]
[project @ 1999-09-15 13:45:14 by simonmar]
Heap Profiling
--------------

This commit adds simple hp2ps-style heap profiling.  Just the '-h'
option is supported so far.

Heap profiles will show the top two cost centres on each cost centre
stack.

Time/Alloc profiles have been extended to aggregate the costs on a
per-cost-centre basis and show the table of highest consuming cost
centres before the full table of cost centre stacks.

Profiles work best if '-caf-all' is used, especially when using
one of the '-auto' flags.

24 years ago[project @ 1999-09-15 08:00:11 by simonpj]
simonpj [Wed, 15 Sep 1999 08:00:11 +0000 (08:00 +0000)]
[project @ 1999-09-15 08:00:11 by simonpj]
Extra words about unsafePerformIO

24 years ago[project @ 1999-09-14 12:16:36 by simonmar]
simonmar [Tue, 14 Sep 1999 12:16:39 +0000 (12:16 +0000)]
[project @ 1999-09-14 12:16:36 by simonmar]
A couple of fixes and cleanups to ticky-ticky profiling:

- remove UPD_EXISTING (doesn't make sense)
- add UPD_CON_IN_PLACE, now that we have in-place updates
- clean up the output a little.

24 years ago[project @ 1999-09-14 08:38:08 by simonmar]
simonmar [Tue, 14 Sep 1999 08:38:08 +0000 (08:38 +0000)]
[project @ 1999-09-14 08:38:08 by simonmar]
gcc 2.95 on Sparc changed the assembly output slightly.  This should
fix it.

24 years ago[project @ 1999-09-13 13:48:01 by simonmar]
simonmar [Mon, 13 Sep 1999 13:48:02 +0000 (13:48 +0000)]
[project @ 1999-09-13 13:48:01 by simonmar]
GHC's web pages have moved to http://www.haskell.org/ghc/.

24 years ago[project @ 1999-09-13 11:02:22 by sof]
sof [Mon, 13 Sep 1999 11:02:22 +0000 (11:02 +0000)]
[project @ 1999-09-13 11:02:22 by sof]
updated

24 years ago[project @ 1999-09-13 11:02:08 by sof]
sof [Mon, 13 Sep 1999 11:02:08 +0000 (11:02 +0000)]
[project @ 1999-09-13 11:02:08 by sof]
Only use initUserSignals() if you've got it defined

24 years ago[project @ 1999-09-13 11:01:05 by sof]
sof [Mon, 13 Sep 1999 11:01:07 +0000 (11:01 +0000)]
[project @ 1999-09-13 11:01:05 by sof]
GMP only depends on libiberty for the implementation of random().
Avoid this dep. on a (pure) Win32 platform.

24 years ago[project @ 1999-09-13 08:28:45 by sof]
sof [Mon, 13 Sep 1999 08:28:45 +0000 (08:28 +0000)]
[project @ 1999-09-13 08:28:45 by sof]
Win32 version of awaitEvent(). Easy :)

24 years ago[project @ 1999-09-13 08:14:51 by simonmar]
simonmar [Mon, 13 Sep 1999 08:14:51 +0000 (08:14 +0000)]
[project @ 1999-09-13 08:14:51 by simonmar]
oops, back out last change.  It was already there.

24 years ago[project @ 1999-09-13 08:14:00 by simonmar]
simonmar [Mon, 13 Sep 1999 08:14:00 +0000 (08:14 +0000)]
[project @ 1999-09-13 08:14:00 by simonmar]
Fix building with profiling.

24 years ago[project @ 1999-09-12 19:18:22 by sof]
sof [Sun, 12 Sep 1999 19:18:22 +0000 (19:18 +0000)]
[project @ 1999-09-12 19:18:22 by sof]
In case of a partial writes, buffer pointer wasn't being adjusted/used.

24 years ago[project @ 1999-09-12 16:30:34 by sof]
sof [Sun, 12 Sep 1999 16:30:34 +0000 (16:30 +0000)]
[project @ 1999-09-12 16:30:34 by sof]
DLL install support

24 years ago[project @ 1999-09-12 16:27:00 by sof]
sof [Sun, 12 Sep 1999 16:27:00 +0000 (16:27 +0000)]
[project @ 1999-09-12 16:27:00 by sof]
If DLL support enabled, add import libs to INSTALL_LIBS + DLL to INSTALL_PROGS

24 years ago[project @ 1999-09-12 16:24:46 by sof]
sof [Sun, 12 Sep 1999 16:24:46 +0000 (16:24 +0000)]
[project @ 1999-09-12 16:24:46 by sof]
inputReady() and mingw - always return 1, not 0