Ross Paterson [Sun, 27 Aug 2006 16:40:44 +0000 (16:40 +0000)]
expand advice on importing these modules
Ross Paterson [Sun, 27 Aug 2006 11:51:40 +0000 (11:51 +0000)]
add Haddock marker
Don Stewart [Sat, 26 Aug 2006 12:43:46 +0000 (12:43 +0000)]
Clarify how one hides Prelude.catch
User feedback indicated that an example was required, of how to hide
Prelude.catch, so add such an example to the docs
Ian Lynagh [Fri, 25 Aug 2006 13:49:36 +0000 (13:49 +0000)]
Workaround for OSes that don't have intmax_t and uintmax_t
OpenBSD (and possibly others) do not have intmax_t and uintmax_t types:
http://www.mail-archive.com/haskell-prime@haskell.org/msg01548.html
so substitute (unsigned) long long if we have them, otherwise
(unsigned) long.
Simon Marlow [Fri, 25 Aug 2006 11:06:10 +0000 (11:06 +0000)]
add docs for par
Ross Paterson [Thu, 24 Aug 2006 14:05:04 +0000 (14:05 +0000)]
document minimal complete definition for Bits
Simon Marlow [Thu, 24 Aug 2006 13:23:11 +0000 (13:23 +0000)]
C regex library bits have moved to the regex-posix package
Esa Ilari Vuokko [Wed, 23 Aug 2006 00:31:26 +0000 (00:31 +0000)]
Add shared Typeable support (ghc only)
Simon Marlow [Thu, 24 Aug 2006 12:12:23 +0000 (12:12 +0000)]
this should have been removed with the previous patch
Simon Marlow [Thu, 24 Aug 2006 09:46:15 +0000 (09:46 +0000)]
remove Text.Regx & Text.Regex.Posix
These are subsumed by the new regex-base, regex-posix and regex-compat
packages.
Don Stewart [Thu, 24 Aug 2006 04:13:26 +0000 (04:13 +0000)]
explicitly tag Data.ByteString rules with the FPS prefix.
Don Stewart [Thu, 24 Aug 2006 01:26:11 +0000 (01:26 +0000)]
Add spec rules for sections in Data.ByteString
Don Stewart [Wed, 23 Aug 2006 14:33:38 +0000 (14:33 +0000)]
Sync Data.ByteString with current stable branch, 0.7
Simon Marlow [Wed, 23 Aug 2006 09:50:59 +0000 (09:50 +0000)]
add notes about why copyFile doesn't remove the target
Simon Marlow [Tue, 22 Aug 2006 12:19:09 +0000 (12:19 +0000)]
copyFile: try removing the target file before opening it for writing
Simon Marlow [Tue, 22 Aug 2006 12:19:09 +0000 (12:19 +0000)]
copyFile: try removing the target file before opening it for writing
Ross Paterson [Mon, 21 Aug 2006 15:21:51 +0000 (15:21 +0000)]
add alternative functors and extra instances
* Alternative class, for functors with a monoid
* instances for Const
* instances for arrows
Simon Marlow [Mon, 21 Aug 2006 13:16:12 +0000 (13:16 +0000)]
generate Haddock docs on all platforms
Ross Paterson [Sat, 19 Aug 2006 17:39:54 +0000 (17:39 +0000)]
remove extra comma from import
Ross Paterson [Fri, 18 Aug 2006 17:03:28 +0000 (17:03 +0000)]
fix docs for withC(A)StringLen
Malcolm.Wallace@cs.york.ac.uk [Fri, 18 Aug 2006 13:08:10 +0000 (13:08 +0000)]
use Haskell'98 compliant indentation in do blocks
Malcolm.Wallace@cs.york.ac.uk [Fri, 18 Aug 2006 13:07:14 +0000 (13:07 +0000)]
use correct names of IOArray operations for nhc98
Ross Paterson [Thu, 17 Aug 2006 23:50:41 +0000 (23:50 +0000)]
add mapMaybe and mapEither, plus WithKey variants
Malcolm.Wallace@cs.york.ac.uk [Thu, 17 Aug 2006 13:55:02 +0000 (13:55 +0000)]
remove Text.Html from nhc98 build
Ross Paterson [Tue, 15 Aug 2006 19:06:09 +0000 (19:06 +0000)]
eliminate more HOST_OS tests
Ross Paterson [Sun, 13 Aug 2006 18:44:35 +0000 (18:44 +0000)]
Hugs only: disable unused process primitives
These were the cause of Hugs bug #30, I think, and weren't used by Hugs anyway.
Ross Paterson [Sat, 12 Aug 2006 10:38:35 +0000 (10:38 +0000)]
markup fix to Data.HashTable
Ross Paterson [Sat, 12 Aug 2006 08:27:02 +0000 (08:27 +0000)]
revert removal of ghcconfig.h from package.conf.in
as it's preprocessed with -undef (pointed out by Esa Ilari Vuokko)
Ross Paterson [Fri, 11 Aug 2006 23:15:21 +0000 (23:15 +0000)]
fix Data.HashTable for non-GHC
Simon Marlow [Fri, 11 Aug 2006 15:23:50 +0000 (15:23 +0000)]
remove deprecated 'withObject'
Simon Marlow [Fri, 11 Aug 2006 15:10:24 +0000 (15:10 +0000)]
Jan-Willem Maessen's improved implementation of Data.HashTable
Rather than incrementally enlarging the hash table, this version
just does it in one go when the table gets too full.
sven.panne@aedion.de [Fri, 11 Aug 2006 14:46:29 +0000 (14:46 +0000)]
Warning police: Make some prototypes from the RTS known
sven.panne@aedion.de [Fri, 11 Aug 2006 14:22:08 +0000 (14:22 +0000)]
Warning police: Removed useless catch-all clause
Ross Paterson [Fri, 11 Aug 2006 12:40:30 +0000 (12:40 +0000)]
reduce dependency on ghcconfig.h
The only remaining use is in cbits/dirUtils.h, which tests solaris2_HOST_OS
(Also System.Info uses ghcplatform.h and several modules import MachDeps.h
to get SIZEOF_* and ALIGNMENT_* from ghcautoconf.h)
Ross Paterson [Thu, 10 Aug 2006 18:29:02 +0000 (18:29 +0000)]
(non-GHC only) track MArray interface change
Simon Marlow [Thu, 10 Aug 2006 11:30:17 +0000 (11:30 +0000)]
move Text.Html to a separate package
Simon Marlow [Thu, 10 Aug 2006 11:28:33 +0000 (11:28 +0000)]
bump version to 2.0
Simon Marlow [Wed, 9 Aug 2006 15:38:10 +0000 (15:38 +0000)]
Remove deprecated Data.FiniteMap and Data.Set interfaces
Ross Paterson [Wed, 9 Aug 2006 12:42:59 +0000 (12:42 +0000)]
move altzone test from ghc to base package
Ross Paterson [Wed, 9 Aug 2006 12:38:12 +0000 (12:38 +0000)]
remove unnecessary #include "ghcconfig.h"
Simon Marlow [Wed, 9 Aug 2006 10:05:48 +0000 (10:05 +0000)]
Change the API of MArray to allow resizable arrays
See #704
The MArray class doesn't currently allow a mutable array to change its
size, because of the pure function
bounds :: (HasBounds a, Ix i) => a i e -> (i,i)
This patch removes the HasBounds class, and adds
getBounds :: (MArray a e m, Ix i) => a i e -> m (i,i)
to the MArray class, and
bounds :: (IArray a e, Ix i) => a i e -> (i,i)
to the IArray class.
The reason that bounds had to be incorporated into the IArray class is
because I couldn't make DiffArray work without doing this. DiffArray
acts as a layer converting an MArray into an IArray, and there was no
way (that I could find) to define an instance of HasBounds for
DiffArray.
Simon Marlow [Tue, 8 Aug 2006 10:07:08 +0000 (10:07 +0000)]
deprecate this module.
Simon Marlow [Mon, 7 Aug 2006 15:55:45 +0000 (15:55 +0000)]
add traceShow (see #474)
Simon Marlow [Mon, 24 Jul 2006 16:02:58 +0000 (16:02 +0000)]
remove spurious 'extern "C" {'
Simon Marlow [Fri, 21 Jul 2006 10:02:25 +0000 (10:02 +0000)]
Fix unsafeIndex for large ranges
simonpj@microsoft [Fri, 14 Jul 2006 12:02:07 +0000 (12:02 +0000)]
Improve documentation of atomically
simonpj@microsoft [Fri, 7 Jul 2006 15:19:01 +0000 (15:19 +0000)]
Add missing method genRange for StdGen (fixes #794)
MERGE TO STABLE
Trac #794 reports (correctly) that the implementation of StdGen
only returns numbers in the range (0..something) rather than
(minBound, maxBound), which is what StdGen's genRange claims.
This commit fixes the problem, by implementing genRange for StdGen
(previously it just used the default method).
Malcolm.Wallace@cs.york.ac.uk [Tue, 11 Jul 2006 16:16:14 +0000 (16:16 +0000)]
disambiguate uses of foldr for nhc98 to compile without errors
Malcolm.Wallace@cs.york.ac.uk [Tue, 11 Jul 2006 16:09:41 +0000 (16:09 +0000)]
make Control.Monad.Instances compilable by nhc98
Lemmih [Sat, 8 Jul 2006 05:55:28 +0000 (05:55 +0000)]
breakpointCond
Simon Marlow [Wed, 5 Jul 2006 14:25:37 +0000 (14:25 +0000)]
UNDO: Merge "unrecognized long opt" fix from 6.4.2
This patch undid the previous patch, "RequireOrder: do not collect
unrecognised options after a non-opt". I asked Sven to revert it, but
didn't get an answer.
See bug #473.
Don Stewart [Mon, 3 Jul 2006 09:18:06 +0000 (09:18 +0000)]
Avoid strictness in accumulator for unpackFoldr
The seq on the accumulator for unpackFoldr will break in the presence of
head/build rewrite rules. The empty list case will be forced, producing
an exception. This is a known issue with seq and rewrite rules that we
just stumbled on to.
Don Stewart [Sun, 2 Jul 2006 08:39:13 +0000 (08:39 +0000)]
Disable unpack/build fusion
unpack/build on bytestrings seems to trigger a bug when interacting with
head/build fusion in GHC.List. The bytestring001 testcase catches it.
I'll investigate further, but best to disable this for now (its not
often used anyway).
Note that with -frules-off or ghc 6.4.2 things are fine. It seems to
have emerged with the recent rules changes.
Don Stewart [Sat, 1 Jul 2006 08:43:45 +0000 (08:43 +0000)]
Import Data.ByteString.Lazy, improve ByteString Fusion, and resync with FPS head
This patch imports the Data.ByteString.Lazy module, and its helpers,
providing a ByteString implemented as a lazy list of strict cache-sized
chunks. This type allows the usual lazy operations to be written on
bytestrings, including lazy IO, with much improved space and time over
the [Char] equivalents.
Don Stewart [Fri, 9 Jun 2006 07:59:24 +0000 (07:59 +0000)]
Wibble in docs for new ForeignPtr functionsn
Ross Paterson [Thu, 22 Jun 2006 17:04:36 +0000 (17:04 +0000)]
comments for Applicative and Traversable
Simon Marlow [Thu, 22 Jun 2006 14:44:46 +0000 (14:44 +0000)]
default to NoBuffering on Windows for a read/write text file
Fixes (works around) #679
Simon Marlow [Thu, 22 Jun 2006 14:44:33 +0000 (14:44 +0000)]
remove dead code
Simon Marlow [Thu, 22 Jun 2006 11:29:11 +0000 (11:29 +0000)]
clarify and expand docs
jeanphilippe.bernardy@gmail.com [Fri, 16 Jun 2006 18:01:21 +0000 (18:01 +0000)]
Add minView and maxView to Map and Set
Ross Paterson [Wed, 14 Jun 2006 11:44:56 +0000 (11:44 +0000)]
add signature for registerDelay
Ross Paterson [Tue, 13 Jun 2006 14:27:04 +0000 (14:27 +0000)]
a few doc comments
Don Stewart [Thu, 8 Jun 2006 01:50:11 +0000 (01:50 +0000)]
Optimised foreign pointer representation, for heap-allocated objects
simonpj@microsoft.com [Mon, 5 Jun 2006 11:58:14 +0000 (11:58 +0000)]
Add the inline function, and many comments
This commit adds the 'inline' function described in the
related patch in the compiler.
I've also added comments about the 'lazy' function.
Ross Paterson [Thu, 25 May 2006 11:16:04 +0000 (11:16 +0000)]
small intro to exceptions
Simon Marlow [Thu, 25 May 2006 09:04:56 +0000 (09:04 +0000)]
export breakpoint
Don Stewart [Thu, 25 May 2006 06:50:12 +0000 (06:50 +0000)]
Merge in changes from fps head. Highlights:
Wed May 24 15:49:38 EST 2006 sjanssen@cse.unl.edu
* instance Monoid ByteString
Wed May 24 15:04:04 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Rearange export lists for the .Char8 modules
Wed May 24 14:59:56 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Implement mapAccumL and reimplement mapIndexed using loopU
Wed May 24 14:47:32 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Change the implementation of the unfoldr(N) functions.
Use a more compact implementation for unfoldrN and change it's behaviour
to only return Just in the case that it actually 'overflowed' the N, so
the boundary case of unfolding exactly N gives Nothing.
Implement unfoldr and Lazy.unfoldr in terms of unfoldrN. Use fibonacci
growth for the chunk size in unfoldr
Wed May 24 08:32:29 EST 2006 sjanssen@cse.unl.edu
* Add unfoldr to ByteString and .Char8
A preliminary implementation of unfoldr.
Wed May 24 01:39:41 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Reorder the export lists to better match the Data.List api
Tue May 23 14:04:32 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* pack{Byte,Char} -> singleton. As per fptools convention
Tue May 23 14:00:51 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* elemIndexLast -> elemIndexEnd
Tue May 23 13:57:34 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* In the search for a more orthogonal api, we kill breakFirst/breakLast,
which were of dubious value
Tue May 23 12:24:09 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Abolish elems. It's name implied it was unpack, but its type didn't. it made no sense
Tue May 23 10:42:09 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Minor doc tidyup. Use haddock markup better.
Tue May 23 11:00:31 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Simplify the join() implementation. Spotted by Duncan.
Simon Marlow [Wed, 24 May 2006 12:18:23 +0000 (12:18 +0000)]
add a way to ask the IO manager thread to exit
Don Stewart [Sat, 20 May 2006 03:04:36 +0000 (03:04 +0000)]
Sync with FPS head, including the following patches:
Thu May 18 15:45:46 EST 2006 sjanssen@cse.unl.edu
* Export unsafeTake and unsafeDrop
Fri May 19 11:53:08 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Add foldl1'
Fri May 19 13:41:24 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Add fuseable scanl, scanl1 + properties
Fri May 19 18:20:40 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Spotted another chance to use unsafeTake,Drop (in groupBy)
Thu May 18 09:24:25 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* More effecient findIndexOrEnd based on the impl of findIndex
Thu May 18 09:22:49 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Eliminate special case in findIndex since it's handled anyway.
Thu May 18 09:19:08 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Add unsafeTake and unsafeDrop
These versions assume the n is in the bounds of the bytestring, saving
two comparison tests. Then use them in varous places where we think this
holds. These cases need double checking (and there are a few remaining
internal uses of take / drop that might be possible to convert).
Not exported for the moment.
Tue May 16 23:15:11 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Handle n < 0 in drop and splitAt. Spotted by QC.
Tue May 16 22:46:22 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Handle n <= 0 cases for unfoldr and replicate. Spotted by QC
Tue May 16 21:34:11 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* mapF -> map', filterF -> filter'
Ross Paterson [Thu, 18 May 2006 15:47:23 +0000 (15:47 +0000)]
haddock fix
Ross Paterson [Thu, 18 May 2006 15:43:16 +0000 (15:43 +0000)]
simplify indexing in Data.Sequence
Simon Marlow [Thu, 18 May 2006 11:33:39 +0000 (11:33 +0000)]
Move Eq, Ord, Show instances for ThreadId to GHC.Conc
Eliminates orphans.
Simon Marlow [Thu, 18 May 2006 11:33:03 +0000 (11:33 +0000)]
Better error handling in the IO manager thread
In particular, handle EBADF just like rts/posix/Select.c, by waking up
all the waiting threads. Other errors are thrown, instead of just
being ignored.
Simon Marlow [Thu, 18 May 2006 10:41:51 +0000 (10:41 +0000)]
#define _REENTRANT 1 (needed to get the right errno on some OSs)
Part 2 of the fix for threaded RTS problems on Solaris and possibly
*BSD (Part 1 was the same change in ghc/includes/Rts.h).
Don Stewart [Thu, 18 May 2006 01:21:54 +0000 (01:21 +0000)]
copyCString* should be in IO. Spotted by Tomasz Zielonka
Duncan Coutts [Wed, 17 May 2006 22:20:44 +0000 (22:20 +0000)]
add import Prelude to get dependencies right for Data/Fixed.hs
Hopefully this fixes parallel builds.
Don Stewart [Wed, 17 May 2006 02:01:50 +0000 (02:01 +0000)]
Fix negative index handling in splitAt, replicate and unfoldrN. Move mapF, filterF -> map', filter' while we're here
Don Stewart [Sat, 13 May 2006 05:17:36 +0000 (05:17 +0000)]
Use our own realloc. Thus reduction functions (like filter) allocate on the Haskell heap. Makes around 10% difference.
Don Stewart [Fri, 12 May 2006 05:07:50 +0000 (05:07 +0000)]
Last two CInt fixes for 64 bit, and bracket writeFile while we're here
Don Stewart [Wed, 10 May 2006 04:33:09 +0000 (04:33 +0000)]
Some small optimisations, generalise the type of unfold
Tue May 9 22:36:29 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Surely the error function should not be inlined.
Tue May 9 22:35:53 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Reorder memory writes for better cache locality.
Tue May 9 23:28:09 EST 2006 Duncan Coutts <duncan.coutts@worc.ox.ac.uk>
* Generalise the type of unfoldrN
The type of unfoldrN was overly constrained:
unfoldrN :: Int -> (Word8 -> Maybe (Word8, Word8)) -> Word8 -> ByteString
if we compare that to unfoldr:
unfoldr :: (b -> Maybe (a, b)) -> b -> [a]
So we can generalise unfoldrN to this type:
unfoldrN :: Int -> (a -> Maybe (Word8, a)) -> a -> ByteString
and something similar for the .Char8 version. If people really do want to
use it a lot with Word8/Char then perhaps we should add a specialise pragma.
Wed May 10 13:26:40 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Add foldl', and thus a fusion rule for length . {map,filter,fold},
that avoids creating an array at all if the end of the pipeline is a 'length' reduction
**END OF DESCRIPTION***
Place the long patch description above the ***END OF DESCRIPTION*** marker.
The first line of this file will be the patch name.
This patch contains the following changes:
M ./Data/ByteString.hs -8 +38
M ./Data/ByteString/Char8.hs -6 +12
Ross Paterson [Wed, 10 May 2006 00:18:26 +0000 (00:18 +0000)]
portable implementation of WordPtr/IntPtr for non-GHC
plus much tweaking of imports to avoid cycles
Simon Marlow [Tue, 9 May 2006 09:26:06 +0000 (09:26 +0000)]
add WordPtr and IntPtr types to Foreign.Ptr, with associated conversions
As suggested by John Meacham.
I had to move the Show instance for Ptr into GHC.ForeignPtr to avoid
recursive dependencies.
Simon Marlow [Tue, 9 May 2006 09:24:27 +0000 (09:24 +0000)]
add CIntPtr, CUIntPtr, CIntMax, CUIntMax types
Simon Marlow [Tue, 9 May 2006 08:27:39 +0000 (08:27 +0000)]
add GHC.Dynamic
Don Stewart [Tue, 9 May 2006 02:34:25 +0000 (02:34 +0000)]
Two things. #if defined(__GLASGOW_HASKELL__) on INLINE [n] pragmas (for jhc). And careful use of INLINE on words/unwords halves runtime for those functions
simonpj@microsoft [Mon, 8 May 2006 14:27:26 +0000 (14:27 +0000)]
Make length a good consumer
Make length into a good consumer. Fixes Trac bug #707.
(Before length simply didn't use foldr.)
simonpj@microsoft [Mon, 8 May 2006 14:25:57 +0000 (14:25 +0000)]
Trim imports
simonpj@microsoft [Mon, 8 May 2006 14:25:07 +0000 (14:25 +0000)]
Make unsafePerformIO lazy
The stricteness analyser used to have a HACK which ensured that NOINLNE things
were not strictness-analysed. The reason was unsafePerformIO. Left to itself,
the strictness analyser would discover this strictness for unsafePerformIO:
unsafePerformIO: C(U(AV))
But then consider this sub-expression
unsafePerformIO (\s -> let r = f x in
case writeIORef v r s of (# s1, _ #) ->
(# s1, r #)
The strictness analyser will now find that r is sure to be eval'd,
and may then hoist it out. This makes tests/lib/should_run/memo002
deadlock.
Solving this by making all NOINLINE things have no strictness info is overkill.
In particular, it's overkill for runST, which is perfectly respectable.
Consider
f x = runST (return x)
This should be strict in x.
So the new plan is to define unsafePerformIO using the 'lazy' combinator:
unsafePerformIO (IO m) = lazy (case m realWorld# of (# _, r #) -> r)
Remember, 'lazy' is a wired-in identity-function Id, of type a->a, which is
magically NON-STRICT, and is inlined after strictness analysis. So
unsafePerformIO will look non-strict, and that's what we want.
Don Stewart [Mon, 8 May 2006 12:23:22 +0000 (12:23 +0000)]
Sync with FPS head.
Mon May 8 10:40:14 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Fix all uses for Int that should be CInt or CSize in ffi imports.
Spotted by Igloo, dcoutts
Mon May 8 16:09:41 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Import nicer loop/loop fusion rule from ghc-ndp
Mon May 8 17:36:07 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Fix stack leak in split on > 60M strings
Mon May 8 17:50:13 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Try same fix for stack overflow in elemIndices
Don Stewart [Mon, 8 May 2006 01:03:11 +0000 (01:03 +0000)]
Fix all uses for Int that should be CInt or CSize in ffi imports. Spotted by Duncan and Ian
Sven Panne [Sun, 7 May 2006 15:50:08 +0000 (15:50 +0000)]
Fixed import list syntax
dons@cse.unsw.edu.au [Sun, 7 May 2006 04:23:01 +0000 (04:23 +0000)]
Faster filterF, filterNotByte
dons@cse.unsw.edu.au [Sun, 7 May 2006 03:30:48 +0000 (03:30 +0000)]
Much faster find, findIndex. Hint from sjanssen
Sven Panne [Sat, 6 May 2006 11:05:19 +0000 (11:05 +0000)]
Merge "unrecognized long opt" fix from 6.4.2
dons@cse.unsw.edu.au [Sat, 6 May 2006 06:10:29 +0000 (06:10 +0000)]
Sat May 6 13:01:34 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Do loopU realloc on the Haskell heap. And add a really tough stress test
Sat May 6 12:28:58 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Use simple, 3x faster concat. Plus QC properties. Suggested by sjanssen and dcoutts
Sat May 6 15:59:31 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* dcoutt's packByte bug squashed
With inlinePerformIO, ghc head was compiling:
packByte 255 `compare` packByte 127
into roughly
case mallocByteString 2 of
ForeignPtr f internals ->
case writeWord8OffAddr# f 0 255 of _ ->
case writeWord8OffAddr# f 0 127 of _ ->
case eqAddr# f f of
False -> case compare (GHC.Prim.plusAddr# f 0)
(GHC.Prim.plusAddr# f 0)
which is rather stunning. unsafePerformIO seems to prevent whatever
magic inlining was leading to this. Only affected the head.
dons@cse.unsw.edu.au [Fri, 5 May 2006 06:08:58 +0000 (06:08 +0000)]
Add array fusion versions of map, filter and foldl
This patch adds fusable map, filter and foldl, using the array fusion
code for unlifted, flat arrays, from the Data Parallel Haskell branch,
after kind help from Roman Leshchinskiy,
Pipelines of maps, filters and folds should now need to walk the
bytestring once only, and intermediate bytestrings won't be constructed.
Ross Paterson [Thu, 4 May 2006 09:30:44 +0000 (09:30 +0000)]
fix for non-GHC
Ross Paterson [Thu, 4 May 2006 09:15:28 +0000 (09:15 +0000)]
use bracket in appendFile (like writeFile)
Simon Marlow [Thu, 4 May 2006 08:45:05 +0000 (08:45 +0000)]
writeFile: close the file on error
Suggested by Ross Paterson, via Neil Mitchell
dons@cse.unsw.edu.au [Wed, 3 May 2006 10:52:59 +0000 (10:52 +0000)]
Sync with FPS head
This patch brings Data.ByteString into sync with the FPS head.
The most significant of which is the new Haskell counting sort.
Changes:
Sun Apr 30 18:16:29 EST 2006 sjanssen@cse.unl.edu
* Fix foldr1 in Data.ByteString and Data.ByteString.Char8
Mon May 1 11:51:16 EST 2006 Don Stewart <dons@cse.unsw.edu.au>
* Add group and groupBy. Suggested by conversation between sjanssen and petekaz on #haskell
Mon May 1 16:42:04 EST 2006 sjanssen@cse.unl.edu
* Fix groupBy to match Data.List.groupBy.
Wed May 3 15:01:07 EST 2006 sjanssen@cse.unl.edu
* Migrate to counting sort.
Data.ByteString.sort used C's qsort(), which is O(n log n). The new algorithm
is O(n), and is faster for strings larger than approximately thirty bytes. We
also reduce our dependency on cbits!