Ticket #4336: recip.dpatch

File recip.dpatch, 52.9 KB (added by daniel.is.fischer, 4 years ago)

patch for faster recip for Ratio a

Line 
11 patch for repository /home/dafis/Haskell/Hacking/ghc/libraries/base:
2
3Mon Oct 18 21:36:34 CEST 2010  Daniel Fischer <daniel.is.fischer@web.de>
4  * FIX #4336
5  Avoid superfluous gcd calculation in recip for Ratio a because numerator
6  and denominator are known to be coprime.
7 
8  Also trailing whitespace.
9
10New patches:
11
12[FIX #4336
13Daniel Fischer <daniel.is.fischer@web.de>**20101018193634
14 Ignore-this: fc74219bc3e1d6a8fd8f3e59357acb5c
15 Avoid superfluous gcd calculation in recip for Ratio a because numerator
16 and denominator are known to be coprime.
17 
18 Also trailing whitespace.
19] {
20hunk ./GHC/Real.lhs 9
21 -- Module      :  GHC.Real
22 -- Copyright   :  (c) The University of Glasgow, 1994-2002
23 -- License     :  see libraries/base/LICENSE
24---
25+--
26 -- Maintainer  :  cvs-ghc@haskell.org
27 -- Stability   :  internal
28 -- Portability :  non-portable (GHC Extensions)
29hunk ./GHC/Real.lhs 33
30 infixl 7  /, `quot`, `rem`, `div`, `mod`
31 infixl 7  %
32 
33-default ()              -- Double isn't available yet,
34+default ()              -- Double isn't available yet,
35                         -- and we shouldn't be using defaults anyway
36 \end{code}
37 
38hunk ./GHC/Real.lhs 61
39 infinity   = 1 :% 0
40 notANumber = 0 :% 0
41 
42--- Use :%, not % for Inf/NaN; the latter would
43--- immediately lead to a runtime error, because it normalises.
44+-- Use :%, not % for Inf/NaN; the latter would
45+-- immediately lead to a runtime error, because it normalises.
46 \end{code}
47 
48 
49hunk ./GHC/Real.lhs 194
50 
51     {-# INLINE truncate #-}
52     truncate x          =  m  where (m,_) = properFraction x
53-   
54+
55     round x             =  let (n,r) = properFraction x
56                                m     = if r < 0 then n - 1 else n + 1
57                            in case signum (abs r - 0.5) of
58hunk ./GHC/Real.lhs 202
59                                 0  -> if even n then n else m
60                                 1  -> m
61                                 _  -> error "round default defn: Bad value"
62-   
63+
64     ceiling x           =  if r > 0 then n + 1 else n
65                            where (n,r) = properFraction x
66hunk ./GHC/Real.lhs 205
67-   
68+
69     floor x             =  if r < 0 then n - 1 else n
70                            where (n,r) = properFraction x
71 \end{code}
72hunk ./GHC/Real.lhs 334
73 instance  (Integral a)  => Fractional (Ratio a)  where
74     {-# SPECIALIZE instance Fractional Rational #-}
75     (x:%y) / (x':%y')   =  (x*y') % (y*x')
76-    recip (x:%y)        =  y % x
77+    recip (0:%_)        = error "Ratio.%: zero denominator"
78+    recip (x:%y)
79+        | x < 0         = negate y :% negate x
80+        | otherwise     = y :% x
81     fromRational (x:%y) =  fromInteger x :% fromInteger y
82 
83 instance  (Integral a)  => Real (Ratio a)  where
84hunk ./GHC/Real.lhs 352
85 instance  (Integral a)  => Show (Ratio a)  where
86     {-# SPECIALIZE instance Show Rational #-}
87     showsPrec p (x:%y)  =  showParen (p > ratioPrec) $
88-                           showsPrec ratioPrec1 x .
89+                           showsPrec ratioPrec1 x .
90                            showString " % " .
91                            -- H98 report has spaces round the %
92                            -- but we removed them [May 04]
93hunk ./GHC/Real.lhs 412
94   -> Int                -- ^ the precedence of the enclosing context
95   -> a                  -- ^ the value to show
96   -> ShowS
97-showSigned showPos p x
98+showSigned showPos p x
99    | x < 0     = showParen (p > 6) (showChar '-' . showPos (-x))
100    | otherwise = showPos x
101 
102}
103
104Context:
105
106[Define SpecConstrAnnotation in GHC.Exts, and import it from there
107simonpj@microsoft.com**20101018135857
108 Ignore-this: 8bf81cbc5787dbb5a3875b5622f67732
109   
110 Reason: avoid having to link the entire ghc package in modules
111 that use compile-time annotations:
112   
113        import GHC.Exts( SpecConstrAnnotation )
114        {-# ANN type T ForceSpecConstr #-}
115   
116 It's a kind of bug that the package exporting SpecConstrAnnotation
117 is linked even though it is only needed at compile time, but putting
118 the data type declaration in GHC.Exts is a simple way to sidestep
119 the problem
120   
121 See See Note [SpecConstrAnnotation] in SpecConstr
122]
123[throwTo: mention interruptible foreign calls
124Simon Marlow <marlowsd@gmail.com>**20101014084220
125 Ignore-this: dbc53d85f870cf649f87186c7185465a
126]
127[remove trailing whitespace
128Simon Marlow <marlowsd@gmail.com>**20101013101906
129 Ignore-this: b8b424540cacbbb3c6d934242e3af795
130]
131[FIX #4381
132Simon Marlow <marlowsd@gmail.com>**20101013101849
133 Ignore-this: f0daa4845eeb444231451b975b71d055
134 Fix scaleFloat by clamping the scaling parameter so that
135 exponent + scale doesn't overflow.
136 
137 Patch by: Daniel Fischer <daniel.is.fischer@web.de>
138]
139[Replaced some throws to throwIOs where the type is IO
140Bas van Dijk <v.dijk.bas@gmail.com>**20100924221340
141 Ignore-this: e74191e4527ae6f7551c95fd41063335
142]
143[Added initial .authorspellings
144Bas van Dijk <v.dijk.bas@gmail.com>**20101005072701
145 Ignore-this: 63628bcabfdd0b7beda4cd37daeccd89
146]
147[Lazier intersperse
148Daniel Fischer <daniel.is.fischer@web.de>**20101002231201
149 Ignore-this: a0fed65930cf19e68b4363381a5ab576
150 A lazier implementation of intersperse, and consequentially intercalate, to
151 avoid space leaks.
152]
153[FIX #4228 (atanh (-1) returns NaN instead of -Infinity)
154ghc@cainnorris.net**20100816213654
155 Ignore-this: dee89c24493e84a02bea711a1c83a73f
156]
157[Make intersectBy lazier
158Daniel Fischer <daniel.is.fischer@web.de>**20100930191731
159 Ignore-this: ef687bc75923434e85c14b57171576aa
160 Add shortcuts to intersectBy for empty list arguments.
161 In addition to being faster in that case, more inputs yield defined results.
162 Treats ticket #4323
163]
164[doc tweak for Directory file type: file names are '\0'-separated
165Simon Marlow <marlowsd@gmail.com>**20100922113811
166 Ignore-this: 96b7b004bd6e5bc3e958ad55bf238ba1
167]
168[documentation for IODeviceType (#4317, edited by me)
169Simon Marlow <marlowsd@gmail.com>**20100915131341
170 Ignore-this: 21c50ca7a189eebcf299523b6e942bae
171]
172[Allow Data.HashTable construction with user-supplied size
173**20100722210726
174 Ignore-this: bd54880bb16a106a992f03b040dc4164
175 
176 This avoids some resizing for users who know they will be inserting a
177 lot of data.
178 
179 http://hackage.haskell.org/trac/ghc/ticket/4193
180]
181[some fixes for hGetBufSome
182Simon Marlow <marlowsd@gmail.com>**20100916113732
183 Ignore-this: 3e596a606c180dc4859ea8f4c9132ca1
184  - fix one case where it was blocking when it shouldn't
185  - a couple of error-message tweaks
186]
187[Windows: map ERROR_NO_DATA to EPIPE, rather than EINVAL
188Simon Marlow <marlowsd@gmail.com>**20100915142618
189 Ignore-this: 9023e5f0542419f225aef26cb6b1d88d
190 WriteFile() returns ERROR_NO_DATA when writing to a pipe that is
191 "closing", however by default the write() wrapper in the CRT maps this
192 to EINVAL so we get confusing things like
193 
194   hPutChar: invalid argument (Invalid Argumnet)
195 
196 when piping the output of a Haskell program into something that closes
197 the pipe early.  This was happening in the testsuite in a few place.
198 
199 The solution is to map ERROR_NO_DATA to EPIPE correctly, as we
200 explicitly check for EPIPE on stdout (in GHC.TopHandler) so we can
201 exit without an error in this case.
202]
203[tighten up parsing of numbers (#1579)
204Simon Marlow <marlowsd@gmail.com>**20100913214733
205 Ignore-this: 3411bf3d2e98cfacb9e0afd11d79e722
206]
207[Add absentError. 
208simonpj@microsoft.com**20100914134639
209 Ignore-this: d0eef5a87e1def4cdbde92a55241c8c4
210 
211 This patch accompanies the HEAD patch:
212 
213   Tue Sep 14 12:38:27 BST 2010  simonpj@microsoft.com
214     * Make absent-arg wrappers work for unlifted types (fix Trac #4306)
215     
216     Previously we were simply passing arguments of unlifted
217     type to a wrapper, even if they were absent, which was
218     stupid.
219     
220     See Note [Absent error Id] in WwLib.
221]
222[Add missing import, fixes build on windows
223simonpj@microsoft.com**20100914122750
224 Ignore-this: 12ece15ef94982ddfbf5f9f7900619da
225]
226[Add a suitable Show instance for TextEncoding (#4273)
227Simon Marlow <marlowsd@gmail.com>**20100913154459
228 Ignore-this: 77f2235460895debd2827f34c42c3435
229]
230[don't fill a finalized handle with an error (see comment)
231Simon Marlow <marlowsd@gmail.com>**20100913153350
232 Ignore-this: c72cdb6898dffa88eca1d781171b2943
233]
234[deriving (Eq, Ord, Read, Show) for Newline and NewlineMode
235Simon Marlow <marlowsd@gmail.com>**20100913153031
236 Ignore-this: 9b9b29bfb7abf5550cfbfa7788f81bf
237]
238[fix warning on Windows
239Simon Marlow <marlowsd@gmail.com>**20100913111536
240 Ignore-this: dacc5448c452daad60ed37a1a5ed096e
241]
242[Put the state-token argument on fill, done, adjust on the RHS
243simonpj@microsoft.com**20100913101832
244 Ignore-this: d228b492de7d4635c026ed24cbc17e34
245 
246 This is so that the functions will inline when
247 applied to their normal (non-state-token) aguments.
248 
249 I forget why I did this, but it seems like the right thing anyway.
250]
251[avoid Foreign.unsafePerformIO
252Ross Paterson <ross@soi.city.ac.uk>**20100909125521
253 Ignore-this: b698101119ffd1bc6311cce0736f745d
254]
255[Remove debugging code accidentally left in
256Simon Marlow <marlowsd@gmail.com>**20100909113331
257 Ignore-this: 906a14176dd37030b8203782a687936b
258]
259[Fix Windows build; patches frmo ezyang
260Ian Lynagh <igloo@earth.li>**20100908123037
261 Ignore-this: 2f02986087edd7da8382221012c27cd0
262]
263[More accurate isatty test for MinGW.
264Edward Z. Yang <ezyang@mit.edu>**20100907154144
265 Ignore-this: 93bdc2b2a8e65a7c4c7d3906bdda01db
266]
267[Fix the build when HAVE_KQUEUE but not HAVE_KEVENT64
268Ian Lynagh <igloo@earth.li>**20100904223703]
269[Fix warnings
270benl@ouroborus.net**20100830044741
271 Ignore-this: 8397aaec7c36046c9ace403e65f32d32
272]
273[fix cache variable name used by FP_SEARCH_LIBS_PROTO
274Ross Paterson <ross@soi.city.ac.uk>**20100819204858
275 Ignore-this: b8113cb3c6f0e03c507297c99d3d82b7
276]
277[Add a missing castPtr (only shows up in -DDEBUG)
278simonpj@microsoft.com**20100815145127
279 Ignore-this: 30b9c42cd3ce7837bdabd254fe66078d
280]
281[Fixed a rounding error in threadDelay
282Johan Tibell <johan.tibell@gmail.com>**20100813124043
283 Ignore-this: 1cb77d0852233ffffb144b134064ee3c
284]
285[export allocaBytesAligned; make allocaArray use the correct alignment (#2917)
286Simon Marlow <marlowsd@gmail.com>**20100812105524
287 Ignore-this: deb6495f7b7b84deaf02b88927a5ba8c
288]
289[deprecate unGetChan and isEmptyChan (see #4154)
290Simon Marlow <marlowsd@gmail.com>**20100705125952
291 Ignore-this: b4e769959f131b2d0001eb7202bc1b92
292]
293[Add type signatures to cope with lack of local generalisation
294simonpj@microsoft.com**20100728124847
295 Ignore-this: d3af9a47c2821c6081bde05a135a92fb
296]
297[Add type signature in local where
298simonpj@microsoft.com**20100727151532
299 Ignore-this: 1c57063ad32d13e0d1ec8daf968bf055
300]
301[Integrated new I/O manager
302Simon Marlow <marlowsd@gmail.com>**20100810082248
303 Ignore-this: ed70a9066ac9b676a446fe99978fef7a
304 (patch originally by Johan Tibell <johan.tibell@gmail.com>, minor merging by me)
305]
306[Add mfilter to Control.Monad
307jon.fairbairn@cl.cam.ac.uk**20090917145616
308 Ignore-this: de4240b60684f3065b29378df3ea98f2
309 
310 Straightforward MonadPlus version of List.filter. I would
311 prefer to call it filter, but the current naming scheme for
312 Control.Monad implies mfilter.
313 
314]
315[move Monad and MonadFix instances for Either from mtl (proposal #4159)
316Ross Paterson <ross@soi.city.ac.uk>**20100729122449
317 Ignore-this: b0f8cd8643679948d1da43bd7c08c5aa
318 
319 The Monad and MonadFix instances for Either (formerly in the mtl
320 package) are moved to Control.Monad.Instances and Control.Monad.Fix
321 respectively.  The Monad instance is still an orphan, to retain Haskell
322 98 compatibility, but the MonadFix instance is together with its class.
323 The Error constraint is removed from both instances, and the default
324 definition of fail is used.
325]
326[Remove egregious ghc-ish from Foreign.Marshal
327Malcolm.Wallace@me.com**20100722075449]
328[add numSparks :: IO Int (#4167)
329Simon Marlow <marlowsd@gmail.com>**20100720153858
330 Ignore-this: 4543f57a7f137f8cae1c3efc5c023a9b
331]
332[add unsafeLocalState from Haskell 2010, and docs
333Simon Marlow <marlowsd@gmail.com>**20100720082819
334 Ignore-this: dcd79fb546ebe29ddff4df279ec2f38
335]
336[docs: mention that Foreign.unsafePerformIO is deprecated
337Simon Marlow <marlowsd@gmail.com>**20100720082804
338 Ignore-this: 4cfebb8f2a1cddc7d15e94e31b2befa4
339 We can't actually deprecate it without introducing a name clash
340 between Foreign.unsafePerformIO and System.IO.Unsafe.unsafePerformIO
341]
342[doc formatting fix
343Simon Marlow <marlowsd@gmail.com>**20100714151347
344 Ignore-this: 255edef607dcd290e198015240b5d125
345]
346[add module intro from Haskell 2010
347Simon Marlow <marlowsd@gmail.com>**20100714115853
348 Ignore-this: 59b5a07507a059ccccdff2dfb6490a27
349]
350[document exception-overriding behaviour in withFile
351Simon Marlow <marlowsd@gmail.com>**20100714104107
352 Ignore-this: f99e641ea2f46d872cb7420a62fa50dc
353]
354[doc: use "finalizer" consistently
355Simon Marlow <marlowsd@gmail.com>**20100714103649
356 Ignore-this: bdfea40f31dc5045fdbc6e12266dda93
357]
358[clarify meaning of bit
359Simon Marlow <marlowsd@gmail.com>**20100714103310
360 Ignore-this: 521b031f1e83ef34ca03d9aa9273df8a
361]
362[note shortcutting behaviour of any/all/elem
363Simon Marlow <marlowsd@gmail.com>**20100714103304
364 Ignore-this: 1605f362ba0712ad1cea1309636f3ea1
365]
366[add cast{C,U}CharToChar and castCharTo{C,U}Char, from Haskell 2010
367Simon Marlow <marlowsd@gmail.com>**20100713132515
368 Ignore-this: 9b1da827016c7b08668078b45964e9de
369]
370[mention that IntPtr and WordPtr can be marshalled to/from intptr_t and uintptr_t
371Simon Marlow <marlowsd@gmail.com>**20100713132403
372 Ignore-this: dcc112a72746ba117a84fa29e71b6800
373]
374[Partial fix for Trac #4136
375simonpj@microsoft.com**20100707135725
376 Ignore-this: 9548eeb3187d9779d4e5c858a0f35354
377 
378 In 'choose' (which is a library function designed specifically
379 to support derived instances of Read), we must match Symbol
380 as well as Ident, for nullary constructors that (wierdly) are
381 symbols.
382]
383[Fix typo in documentation
384Simon Hengel <simon.hengel@wiktory.org>**20100711141648
385 Ignore-this: c052dd8a681832ef598a323ad55eae3a
386]
387[Remove duplicated word in documentation
388Simon Hengel <simon.hengel@wiktory.org>**20100711072703
389 Ignore-this: fb3732dc57be55f14168792f923433
390]
391[Allow nhc98 to cope with recent changes to Control.Exception.
392Malcolm.Wallace@me.com**20100710170940]
393[ New asynchronous exception control API (base parts)
394Simon Marlow <marlowsd@gmail.com>**20100708152735
395 Ignore-this: 71a4811804f04259f1fe739f8863beaf
396   
397 As discussed on the libraries/haskell-cafe mailing lists
398   http://www.haskell.org/pipermail/libraries/2010-April/013420.html
399 
400 This is a replacement for block/unblock in the asychronous exceptions
401 API to fix a problem whereby a function could unblock asynchronous
402 exceptions even if called within a blocked context.
403 
404 The new terminology is "mask" rather than "block" (to avoid confusion
405 due to overloaded meanings of the latter).
406 
407 The following is the new API; the old API is deprecated but still
408 available for the time being.
409 
410 Control.Exception
411 -----------------
412 
413 mask  :: ((forall a. IO a -> IO a) -> IO b) -> IO b
414 mask_ :: IO a -> IO a
415 
416 uninterruptibleMask :: ((forall a. IO a -> IO a) -> IO b) -> IO b
417 uninterruptibleMask_ :: IO a -> IO
418 
419 getMaskingState :: IO MaskingState
420 
421 data MaskingState
422   = Unmasked
423   | MaskedInterruptible
424   | MaskedUninterruptible
425 
426 
427 Control.Concurrent
428 ------------------
429 
430 forkIOUnmasked :: IO () -> IO ThreadId
431]
432[Async-exception safety, and avoid space leaks
433Simon Marlow <marlowsd@gmail.com>**20100708145819
434 Ignore-this: dbfd0e61551e9e7b4fc1c6fe9b9a83de
435 Patch submitted by: Bas van Dijk <v.dijk.bas@gmail.com>
436 Modified slightly by me to remove non-functional changes.
437]
438[Async-exception safety, and avoid space leaks
439Simon Marlow <marlowsd@gmail.com>**20100708103154
440 Ignore-this: 190c3ac8f6633231624da8cf1316588
441 Patch submitted by: Bas van Dijk <v.dijk.bas@gmail.com>
442 Modified slightly by me to remove non-functional changes.
443]
444[Fix a few places where we forgot to close the text codecs (#4029)
445Simon Marlow <marlowsd@gmail.com>**20100702130210
446 Ignore-this: 2e81a4b4cb343181cef34b0f9e2ded47
447 Each time you invoke :load in GHCi it resets the CAFs, including
448 stdin/stdout/stderr, and each of these was allocating a new iconv_t.
449]
450[remove docs from Monad that belonged on the instance for MonadPlus IO
451Simon Marlow <marlowsd@gmail.com>**20100701154203
452 Ignore-this: 59df02542a7ac9421552a2155d848d27
453]
454[docs: unqualify Prelude.IO
455Simon Marlow <marlowsd@gmail.com>**20100701153817
456 Ignore-this: 73b0202876c827e7a5b4a5ce74e724c4
457]
458[unqualify Float and Double
459Simon Marlow <marlowsd@gmail.com>**20100701142727
460 Ignore-this: cbe89d31a00bf49996a33933324fca17
461]
462[extract information about Data.Time from docs for CTime
463Simon Marlow <marlowsd@gmail.com>**20100701142415
464 Ignore-this: c48c9609b8d36e43e033a7bea81d6f17
465]
466[doc typo
467Simon Marlow <marlowsd@gmail.com>**20100701142354
468 Ignore-this: 17a1fd703831c888975ff63fbfa3a9b2
469]
470[peekArray docs: remove mentions of "this version" and "previous version"
471Simon Marlow <marlowsd@gmail.com>**20100701125333
472 Ignore-this: 39a744874258670bd935ba9e38390939
473]
474[doc typo
475Simon Marlow <marlowsd@gmail.com>**20100701124154
476 Ignore-this: 98f5c286e38c2c34c96b05d5e8bc5ad9
477]
478[doc typo
479Simon Marlow <marlowsd@gmail.com>**20100701124128
480 Ignore-this: 10a4314ec7aed336701fc616fb574ebc
481]
482[doc typo
483Simon Marlow <marlowsd@gmail.com>**20100701123715
484 Ignore-this: c4909a7bf7163460ee5d32f58812041e
485]
486[doc wibble: Haskell 98 I/O Error -> 'IOError'
487Simon Marlow <marlowsd@gmail.com>**20100701123612
488 Ignore-this: bf373df781acbc575e4ffe3b7e6059ae
489]
490[doc typo
491Simon Marlow <marlowsd@gmail.com>**20100701123014
492 Ignore-this: 16aaccae48ef3101adf78ea5b0d5a8fd
493]
494[Haddock hacks to fix whitespace consistency
495Simon Marlow <marlowsd@gmail.com>**20100701121631
496 Ignore-this: 61c58dec52a31fd2d3f331a87d2f903f
497]
498[use '==' consistently rather than '->' in examples
499Simon Marlow <marlowsd@gmail.com>**20100701121616
500 Ignore-this: 472b0a05a85d34d9712186040e1636d9
501]
502[doc wibble: remove confusing mention of "Prelude"
503Simon Marlow <marlowsd@gmail.com>**20100701113308
504 Ignore-this: 232283d0096d01cd45e9b3c5c1e63a6d
505]
506[doc wibble: nonstrict -> non-strict
507Simon Marlow <marlowsd@gmail.com>**20100701113253
508 Ignore-this: 4264f0ab23a0835fc13c6e8601d6b743
509]
510[doc whitespace
511Simon Marlow <marlowsd@gmail.com>**20100701112242
512 Ignore-this: 777a95b1d1140c61d3ab95d5eb5809e7
513]
514[move the doc for 'Char' to its new home in ghc-prim:GHC.Types
515Simon Marlow <marlowsd@gmail.com>**20100629134150
516 Ignore-this: 7687db0077a29498349bfb4b44983985
517]
518[doc wibble
519Simon Marlow <marlowsd@gmail.com>**20100629122608
520 Ignore-this: 9a909e5d015332dc445bd9592e6e386d
521]
522[doc updates in System.IO
523Simon Marlow <marlowsd@gmail.com>**20100629122118
524 Ignore-this: 2257ec1cc4cdb8b7804cfa1f3cf32753
525]
526[doc wibble
527Simon Marlow <marlowsd@gmail.com>**20100625134858
528 Ignore-this: 64c50f29df6c389273b818918fe7033a
529]
530[doc wibbles
531Simon Marlow <marlowsd@gmail.com>**20100624154614
532 Ignore-this: b364aad53beea6e741fee2824459b6e8
533]
534[Fix haddock formatting
535Ian Lynagh <igloo@earth.li>**20100625222623]
536[Give nub's complexity in the haddock docs; fixes #4086
537Ian Lynagh <igloo@earth.li>**20100625222059]
538[correct docs for exitWith: only stdout/stderr are flushed, not all Handles
539Simon Marlow <marlowsd@gmail.com>**20100624130506
540 Ignore-this: 33a938dad8f0bc061572e2ec571cacc7
541]
542[fix docs for isSpace
543Simon Marlow <marlowsd@gmail.com>**20100624130444
544 Ignore-this: b35ff080dbb9833176f08e39dbd9ff6d
545]
546[make the hGetBuf/hPutBuf family work with non-FD Handles (#4144)
547Simon Marlow <marlowsd@gmail.com>**20100624130425
548 Ignore-this: 8200f0208a9b1b1cf4824f343d75819a
549]
550[nit in docs for accumArray
551Simon Marlow <marlowsd@gmail.com>**20100622121131
552 Ignore-this: c066a456c40907e767df10c3990f35ff
553]
554[add doc for the ExitCode type
555Simon Marlow <marlowsd@gmail.com>**20100622120930
556 Ignore-this: 99c34332be7f3565da844528b470054a
557]
558[remove extraneous info from docs for Array
559Simon Marlow <marlowsd@gmail.com>**20100622120921
560 Ignore-this: e2a3f5e84fc23eb7bae911f0680e805e
561]
562[add an INLINE to the list version of traverse, to enable fusion
563Simon Marlow <marlowsd@gmail.com>**20100608082531
564 Ignore-this: ea98cdc3308b406bb04c0f7a38c4424b
565]
566[Don't define the C localeEncoding on Windows
567Ian Lynagh <igloo@earth.li>**20100620202342
568 Ignore-this: c4992f6832a391b0cccc5a9b7d643976
569 (it causes warnings, and isn't used)
570]
571[add Applicative instance for Either (proposal #4095)
572Ross Paterson <ross@soi.city.ac.uk>**20100617225110
573 Ignore-this: 50262ec4700dc16efec5755be5b308c5
574 
575 This is not the only possible instance for Either, but this one is
576 compatible with the usual Monad instance.
577]
578[Use libcharset instead of nl_langinfo(CODESET) if possible.
579pho@cielonegro.org**20100519013112
580 Ignore-this: 4c1e278e022a3d276848afc1dcba4425
581 
582 nl_langinfo(CODESET) doesn't always return standardized variations of the encoding names. Use libcharset if possible, which is shipped together with GNU libiconv.
583]
584[Add a note about the interruptibility of throwTo.
585Simon Marlow <marlowsd@gmail.com>**20100615112720
586 Ignore-this: ae9fabe95310d7c364e95f7784793485
587]
588[docs: note that hGetBufNonBlocking isn't non-blocking on Windows
589Simon Marlow <marlowsd@gmail.com>**20100615112547
590 Ignore-this: 4f3e5213e142149affe08c5123d6efea
591]
592[don't depend on Prelude (#4122)
593Simon Marlow <marlowsd@gmail.com>**20100615105631
594 Ignore-this: 1a3fd49b103fe31cbb453f302c18767f
595]
596[Don't depend on Prelude (#4123)
597Simon Marlow <marlowsd@gmail.com>**20100615105401
598 Ignore-this: cc7616d85a1637bc7621b4f2bc181c0e
599]
600[bump version to 4.3.0.0, added instance MonadPlus STM
601Simon Marlow <marlowsd@gmail.com>**20100601144831
602 Ignore-this: 7c3cf7574499c4267372493f2636dc0
603]
604[Moved MonadPlus instance for STM from Control.Monad.STM to GHC.Conc to avoid an orphaned instance
605Bas van Dijk <v.dijk.bas@gmail.com>**20100516160651
606 Ignore-this: 651b852942b2fae2b93f996e39239b8f
607]
608[Added Applicative and Alternative instances for STM
609Bas van Dijk <v.dijk.bas@gmail.com>**20100516171756
610 Ignore-this: 567003bc4040bc97105cda4d31ebf04a
611]
612[expand Foldable instance for Array
613Ross Paterson <ross@soi.city.ac.uk>**20100602212154
614 Ignore-this: 9bd9e9666a9400431eb92352244fe7e7
615]
616[doc comment illustrating Foldable(foldr)
617Ross Paterson <ross@soi.city.ac.uk>**20100527150833
618 Ignore-this: 8f27d889379803f3ba86d6e928428f3c
619]
620[fix syntax in doc comments
621Ross Paterson <ross@soi.city.ac.uk>**20100527150757
622 Ignore-this: cb78da51d60ff6863dc395f1a892c103
623]
624[export hGetBufSome (#4046)
625Simon Marlow <marlowsd@gmail.com>**20100520093538
626 Ignore-this: f467fad9722e27edfad6b3dd75290e7b
627]
628[hWaitForInput: don't try to read from the device (#4078)
629Simon Marlow <marlowsd@gmail.com>**20100517133741
630 Ignore-this: 55ec33b03397380259b91e4ca62207a6
631 readTextDeviceNonBlocking is not non-blocking on Windows
632]
633[hSetEncoding: change the encoding on both read and write sides (#4066)
634Simon Marlow <marlowsd@gmail.com>**20100514124628
635 Ignore-this: 5b9e9caef06356d0296c584159709ebb
636]
637[Correct haddock formatting.
638Adam Vogt <vogt.adam@gmail.com>**20100423022103
639 Ignore-this: d2622339302048fda48080f7d5ce4a2f
640]
641[Fix for hGetBufSome
642Simon Marlow <marlowsd@gmail.com>**20100505135637
643 Ignore-this: 2019680f8fb223956cacfcf0d046f133
644]
645[improve the documentation for throwTo and killThread (#3884)
646Simon Marlow <marlowsd@gmail.com>**20100505135600
647 Ignore-this: ce881d96ddb729acb6ca09c779975e7d
648]
649[elaborate the docs for unsafePerformIO a bit
650Simon Marlow <marlowsd@gmail.com>**20100505101249
651 Ignore-this: 1cec3f67560b672c64c5a0dcf9a79eb7
652]
653[add Typeable instance
654Simon Marlow <marlowsd@gmail.com>**20100504152815
655 Ignore-this: 6d9cf9d62f0ef17fa459bf213a04098
656]
657[Add hGetBufSome, like hGetBuf but can return short reads
658Simon Marlow <marlowsd@gmail.com>**20100504152759
659 Ignore-this: 195c905b43f8d9505029364e2c5b18e
660]
661[Add swap (#3298)
662Simon Marlow <marlowsd@gmail.com>**20100504095339
663 Ignore-this: 13b007dc4594ce252997ec6fa0bbd976
664]
665[inline allocaArray0, to fix withCString benchmark
666Simon Marlow <marlowsd@gmail.com>**20100423124729
667 Ignore-this: 35c96816acc2f3aaf9dd29f7995fa6f0
668]
669[raise asynchronous exceptions asynchronously (#3997)
670Simon Marlow <marlowsd@gmail.com>**20100421094932
671 Ignore-this: 6d987d93d382c0f69c68c326312abd6b
672]
673[add NOINLINE pragmas for stdin/stdout/stderr
674Simon Marlow <marlowsd@gmail.com>**20100421082041
675 Ignore-this: 3fc130268ec786f28d945858d6690986
676]
677[INLINE alloca and malloc
678Simon Marlow <marlowsd@gmail.com>**20100419135333
679 Ignore-this: b218bd611f18721b1505a8c0b9e6a16a
680 See discussion on glasgow-haskell-users:
681   http://www.haskell.org/pipermail/glasgow-haskell-users/2010-April/018740.html
682]
683[Move comment closer to the offending line
684Matthias Kilian <kili@outback.escape.de>**20100419155421
685 Ignore-this: b34a1d7affd66f67d210df2377b585d9
686]
687[Ignore the return code of c_fcntl_write again
688Matthias Kilian <kili@outback.escape.de>**20100415140452
689 Ignore-this: 266d8ba02cc3cb79c85629b3528261c9
690 
691 The return code has been ignored in the past on purpose, because
692 O_NONBLOCK will fail on BSDs for some special files. This fixes the
693 problem mentioned in
694 http://www.haskell.org/pipermail/glasgow-haskell-users/2010-April/018698.html
695 
696]
697[Fix bitrot in IO debugging code
698Ian Lynagh <igloo@earth.li>**20100413134339
699 Also switched to using Haskell Bools (rather than CPP) to en/disable it,
700 so it shouldn't break again in the future.
701]
702[Tiny code tidy-up
703Ian Lynagh <igloo@earth.li>**20100413011147]
704[remove old/wrong comment
705Simon Marlow <marlowsd@gmail.com>**20100325161403
706 Ignore-this: e6e377d44af48c4162d17d55bdf3f821
707]
708[withThread: block asynchronous exceptions before installing exception handler.
709Bas van Dijk <v.dijk.bas@gmail.com>**20100329131624
710 Ignore-this: be5aeb47dbd73807b5f94df11afbb81c
711 Note that I don't unblock the given io computation. Because AFAICS
712 withThread is only called with 'waitFd' which only performs an FFI
713 call which can't receive asynchronous exceptions anyway.
714]
715[runInUnboundThread: block asynchronous exceptions before installing exception handler
716Bas van Dijk <v.dijk.bas@gmail.com>**20100329131549
717 Ignore-this: a00c5e32fe3981ff87bedd367a69051e
718]
719[fix the deprecation message (GHC.IO.Handle.Base -> GHC.IO.Handle)
720Simon Marlow <marlowsd@gmail.com>**20100330121137
721 Ignore-this: 4ca8500a01ac93454507aa8f9dd001f9
722]
723[Make SampleVar an abstract newtype
724Bas van Dijk <v.dijk.bas@gmail.com>**20100318200349
725 Ignore-this: 27939e2a064b75e71cb146117346be30
726]
727[Fix bugs regarding asynchronous exceptions and laziness in Control.Concurrent.SampleVar
728Bas van Dijk <v.dijk.bas@gmail.com>**20100318200104
729 Ignore-this: 7376b2a3afe155daf233a8f1ddc0a7a
730  - Block asynchronous exceptions at the right places
731  - Force thunks before putting them in a MVar
732]
733[Write the thunk 'next' to the MVar
734Bas van Dijk <v.dijk.bas@gmail.com>**20100319125951
735 Ignore-this: dd25636cf220131385ff2fd32493d456
736]
737[change to use STM, fixing 4 things
738Simon Marlow <marlowsd@gmail.com>**20100318104436
739 Ignore-this: 551d30280a7941c08f5c3b14576bdd70
740   1. there was no async exception protection
741   2. there was a space leak (now new value is strict)
742   3. using atomicModifyIORef would be slightly quicker, but can
743      suffer from adverse scheduling issues (see #3838)
744   4. also, the STM version is faster.
745]
746[Tweak docs
747Ian Lynagh <igloo@earth.li>**20100312214129]
748[Fixed dead links in documentation of forkIO
749Bas van Dijk <v.dijk.bas@gmail.com>**20100308222415
750 Ignore-this: 7deb8fd064c867fbede2a6b2e9da4f15
751]
752[Documentation fixes in Control.Exception
753Bas van Dijk <v.dijk.bas@gmail.com>**20100301220442
754 Ignore-this: 761fcba401cbd1f47276ddfc9b5b80f2
755]
756[Plug two race conditions that could lead to deadlocks in the IO manager
757Simon Marlow <marlowsd@gmail.com>**20100225120255
758 Ignore-this: e6983d6b953104d370278ab3e4617e8b
759]
760[FIX #3866: improve documentation of Data.Data.Constr
761jpm@cs.uu.nl**20100224125506
762 Ignore-this: 3818c5d8fee012a3cf322fb455b6e5dc
763]
764[UNDO: Handle NaN, -Infinity and Infinity in the toRational for Float/Double (#3676)
765Simon Marlow <marlowsd@gmail.com>**20100223101603
766 Ignore-this: 78becb2d39b3cd9a1a473a5811ca7d92
767]
768[Put the complexity in the length docs. Fixes trac #3680
769Ian Lynagh <igloo@earth.li>**20100221191425]
770[nhc98 should build Data.Functor.
771Malcolm.Wallace@cs.york.ac.uk**20100221163218]
772[Update the exitWith docs
773Ian Lynagh <igloo@earth.li>**20100213140004
774 Error pointed out by Volker Wysk <vw@volker-wysk.de>
775]
776[Handle NaN, -Infinity and Infinity in the toRational for Float/Double (#3676)
777Simon Marlow <marlowsd@gmail.com>**20100211101955
778 Ignore-this: 261415363303efca265e80290eac5f28
779]
780[For nhc98, import unsafeInterleaveIO rather than defining it here.
781Malcolm.Wallace@cs.york.ac.uk**20100204171021]
782[Stifle warning about unused return value
783benl@cse.unsw.edu.au**20100203025537]
784[fix #3832: use the locale encoding in openTempFile
785Simon Marlow <marlowsd@gmail.com>**20100120211830
786 Ignore-this: df4f778cc5fefb32290c798db722632c
787 Also while I was here fix an XXX: the Handle contained an
788 uninformative string like <fd: 4> for error messages rather than the
789 real file path.
790]
791[Fix the build: export void, so it doesn't give an unused binding warning
792Ian Lynagh <igloo@earth.li>**20100116174451]
793[hIsEOF: don't do any decoding (#3808)
794Simon Marlow <marlowsd@gmail.com>**20100112230317
795 Ignore-this: 6a384dd2d547ffe3ad3762920e5c1671
796]
797[Control.Monad: +void :: f a -> f ()
798gwern0@gmail.com**20100108214455
799 Ignore-this: 4dc07452315f2d1b4941903ff42fc45f
800 See http://hackage.haskell.org/trac/ghc/ticket/3292
801 Turns m a -> m (). Lets one call functions for their side-effects without
802 having to get rid of their return values with '>> return ()'. Very useful
803 in many contexts (parsing, IO etc.); particularly good for 'forkIO' and 'forM_',
804 as they demand return types of 'IO ()' though most interesting IO functions
805 return non-().
806]
807[Replace the implementation of mergesort with a 2x faster one.
808Malcolm.Wallace@cs.york.ac.uk**20091224152014
809 See ticket http://hackage.haskell.org/trac/ghc/ticket/2143.
810]
811[Restore previous Data.Typeable.typeOf*Default implementations for non-ghc.
812Malcolm.Wallace@cs.york.ac.uk**20091223142625
813 Not all compilers have ScopedTypeVariables.
814]
815[Add comments about double bounds-checking, and fast paths for rectangular arrays
816simonpj@microsoft.com**20091218165655
817 Ignore-this: ea0849419dc00927aba4bd410b1cc58d
818 
819 See Note [Double bounds-checking of index values] for the details.
820 
821 The fast paths omit the doubled checks for cases we know about
822]
823[Fix Trac #3245: memoising typeOf
824simonpj@microsoft.com**20091218155117
825 Ignore-this: 5a178a7f2222293c5ee0c3c43bd1b625
826 
827 The performance bug in #3245 was caused by computing the typeRep
828 once for each call of typeOf, rather than once for each dictionary
829 contruction.  (Computing TypeReps is reasonably expensive, because
830 of the hash-consing machinery.)
831 
832 This is readily fixed by putting the TypeRep construction outside
833 the lambda.  (Arguably GHC might have worked that out itself,
834 but it involves floating something between a type lambda and a
835 value lambda, which GHC doesn't currently do. If it happens a lot
836 we could fix that.)
837]
838[Mark 'index' as INLINE in GHC.Arr
839simonpj@microsoft.com**20091216170441
840 Ignore-this: a4df9d8acf496c8e0e9ce5a520509a2a
841 
842 This makes indexing much faster. See Trac #1216
843]
844[Comment the remaining orphan instance modules
845Ian Lynagh <igloo@earth.li>**20091206125021]
846[De-orphan Eq/Ord Float/Double
847Ian Lynagh <igloo@earth.li>**20091205181238]
848[Add comments to "OPTIONS_GHC -fno-warn-orphans" pragmas
849Ian Lynagh <igloo@earth.li>**20091205165854]
850[Data.Either.partitionEithers was insufficiently lazy.
851Malcolm.Wallace@cs.york.ac.uk**20091202032807
852 Ignore-this: 77e1b3288f66608c71458d8a91bcbe12
853 Spotted by Daniel Fischer.
854]
855[fix the docs regarding finalizer guarantees
856Simon Marlow <marlowsd@gmail.com>**20091130144409
857 Ignore-this: d1ab9532c74a002b8075ff60febcbe2d
858]
859[x86_64 requires more stack
860Malcolm.Wallace@cs.york.ac.uk**20091201033745]
861[check for size < 0 in mallocForeignPtrBytes and friends (#3514)
862Simon Marlow <marlowsd@gmail.com>**20091125143822
863 Ignore-this: 91077d01da2bbe1dfed5155e8b40da9
864]
865[hGetContents: close the handle properly on error
866Simon Marlow <marlowsd@gmail.com>**20091125123435
867 Ignore-this: bc37ff678acc6e547dc390285e056eb9
868 
869 When hGetContents caught an error it was closing the handle and then
870 throwing the exception, without updating the handle with the new
871 closed state.  This lead to a double-closed, which was the cause of
872 
873 *** glibc detected *** ./Setup: double free or corruption
874 
875 when iconv_close was called twice on the decoder.
876 
877 See http://hackage.haskell.org/trac/hackage/ticket/609
878]
879[Fix arities of mapFB and zipFB
880Roman Leshchinskiy <rl@cse.unsw.edu.au>**20091126232219
881 Ignore-this: c4e14cd0a92622549c86e67237a40865
882]
883[Remove an unnecessary -fno-warn-orphans flag
884Ian Lynagh <igloo@earth.li>**20091126123404]
885[Tweak layout to work with alternative layout rule
886Ian Lynagh <igloo@earth.li>**20091125232349]
887[Tweak layout to be accepted by the alternative layout rul
888Ian Lynagh <igloo@earth.li>**20091125194147]
889[Make sure zipWithFB has arity 2
890Roman Leshchinskiy <rl@cse.unsw.edu.au>**20091125010003
891 Ignore-this: 4cf60c55666f03d22a9f5a6e07f52d36
892 
893 It gets 2 arguments in the "zipWith" rule but its arity was higher and the new
894 inliner didn't inline it sometimes, for instance here:
895 
896 mpp ::  [Double] -> [Double] -> [Double] -> [Double] -> [Double]
897 mpp as bs cs ds = zipWith (*) (zipWith (+) as bs) (zipWith (+) cs ds)
898 
899 This was a regression vs. 6.10.
900]
901[Remove an old comment
902Ian Lynagh <igloo@earth.li>**20091124134647]
903[De-orphan the Eq/Ord Integer instances
904Ian Lynagh <igloo@earth.li>**20091124133639]
905[Whitespace only
906Ian Lynagh <igloo@earth.li>**20091124133421]
907[Derive some more instances, rather than writing them by hand
908Ian Lynagh <igloo@earth.li>**20091124011747]
909[We can now derive Ord ()
910Ian Lynagh <igloo@earth.li>**20091124011416]
911[De-orphan tuple Eq/Ord instances
912Ian Lynagh <igloo@earth.li>**20091123233343]
913[Control.Exception.Base no longer has any orphans
914Ian Lynagh <igloo@earth.li>**20091123224905]
915[De-orphan the MonadFix ST instance for GHC
916Ian Lynagh <igloo@earth.li>**20091123223544]
917[Rearrange the contents of Control.Monad.ST; no functionality changes
918Ian Lynagh <igloo@earth.li>**20091123222702]
919[De-orphan the Eq/Ord [a] instances
920Ian Lynagh <igloo@earth.li>**20091123215635]
921[De-orphan the Eq/Ord Char instances
922Ian Lynagh <igloo@earth.li>**20091123202253]
923[De-orphan the Eq/Ord Bool instances
924Ian Lynagh <igloo@earth.li>**20091123201817]
925[Move Eq/Ord Ordering instances to de-orphan them
926Ian Lynagh <igloo@earth.li>**20091123194310]
927[Remove ffi warnings for nhc98.
928Malcolm.Wallace@cs.york.ac.uk**20091123063743]
929[Second attempt to fix #1185 (forkProcess and -threaded)
930Simon Marlow <marlowsd@gmail.com>**20091111151915
931 Ignore-this: fa5f5d5e4e080d4b612a37244f937f9c
932 
933 Patch 2/2: first patch is to ghc
934 
935 This time without dynamic linker hacks, instead I've expanded the
936 existing rts/Globals.c to cache more CAFs, specifically those in
937 GHC.Conc.  We were already using this trick for signal handlers, I
938 should have realised before.
939 
940 It's still quite unsavoury, but we can do away with rts/Globals.c in
941 the future when we switch to a dynamically-linked GHCi.
942]
943[Rollback #1185 fix
944Simon Marlow <marlowsd@gmail.com>**20091106140629
945 Ignore-this: cd5667e8474e37e01ba26a1984274811
946 
947 rolling back:
948 
949 Tue Nov  3 16:05:40 GMT 2009  Simon Marlow <marlowsd@gmail.com>
950   * Fix #1185: restart the IO manager after fork()
951   
952   This is the libraries/base part of the patch; there is a corresponding
953   patch to GHC itself.
954   
955   The main change is that we now keep track of the IO manager's ThreadId
956   in a top-level MVar, and ensureIOManagerIsRunning checks whether a
957   previous IO manager thread is alive before starting one.  In the child
958   of fork(), we can hence call ensureIOManagerIsRunning to restart the
959   IO manager.
960 
961     M ./GHC/Conc.lhs -46 +44
962 
963 Wed Nov  4 17:49:45 GMT 2009  Ian Lynagh <igloo@earth.li>
964   * Fix the build on Windows
965 
966     M ./GHC/Conc.lhs -6 +4
967]
968[Fix the build on Windows
969Ian Lynagh <igloo@earth.li>**20091104174945]
970[Fix #1185: restart the IO manager after fork()
971Simon Marlow <marlowsd@gmail.com>**20091103160540
972 Ignore-this: 6dc05464f1500104554637f4759738cc
973 
974 This is the libraries/base part of the patch; there is a corresponding
975 patch to GHC itself.
976 
977 The main change is that we now keep track of the IO manager's ThreadId
978 in a top-level MVar, and ensureIOManagerIsRunning checks whether a
979 previous IO manager thread is alive before starting one.  In the child
980 of fork(), we can hence call ensureIOManagerIsRunning to restart the
981 IO manager.
982]
983[improve the documentation for throwErrnoIfRetry
984Simon Marlow <marlowsd@gmail.com>**20091016112404
985 Ignore-this: b77275cacf730e15757946027168f63e
986]
987[Don't inline unpackFoldrCString ever
988simonpj@microsoft.com**20091029135350
989 Ignore-this: 85d672649b1b776efc7e97500b05d4f9
990]
991[Inline more default methods
992simonpj@microsoft.com**20091029135330
993 Ignore-this: 289c44b0afd6d5631c2a4e0664275ca9
994 
995 Namely Monad: (>>)
996        Eq:    (==), (/=)
997        Num:   (-), negate
998        Real:  quot, rem, div, mod, recip, (/), truncate
999        Float: (**), logBase, sqrt, tan, tanh
1000]
1001[Move error messages out of INLINEd default methods
1002simonpj@microsoft.com**20091029135118
1003 Ignore-this: 9e35dc947f94827a3529eb53a41575fd
1004 
1005 No need to duplicate the error generation!
1006]
1007[Exploit now-working default-method INLINE pragmas for Data.Bits
1008simonpj@microsoft.com**20091029135041
1009 Ignore-this: 8adf225f31ca7a3181ee087e9e4fe535
1010 
1011 * Add INLINE pragmas to default methods for class Bits
1012 
1013 * Remove redundant instance methods elsewhere, now that
1014   the default method will do the job
1015]
1016[Tidy up and comment imports
1017simonpj@microsoft.com**20091029134414
1018 Ignore-this: bf2be31035de975d8995e988933cc940
1019]
1020[Inline foldr and (.) when applied to two arguments not three
1021simonpj@microsoft.com**20091029134335
1022 Ignore-this: fccb6f3e90e15f44cb465814be85ede2
1023 
1024 The new INLINE story is (by design) arity-sensitive, so we must
1025 put fewer argument on the LHS for foldr and (.)
1026]
1027[dirUtils.c no longer available
1028Malcolm.Wallace@cs.york.ac.uk**20091013093833]
1029[Make hGetContents throw an exception if an error is encountered
1030Simon Marlow <marlowsd@gmail.com>**20091012152955
1031 Ignore-this: 9f7a7176193eab25c9daaacd9261f2de
1032 
1033 Strictly speaking this breaks Haskell 98 compatibility, which requires
1034 hGetContents to just end the lazy stream silently if an error is
1035 encountered.  However, for a few reasons we think it will make
1036 everyone's life a bit easier if we make this change
1037 
1038  1. Errors will be a lot more common in GHC 6.12.1, in the form
1039     of Unicode decoding errors.
1040 
1041  2. When Haskell 98 was designed, we didn't know how to throw
1042     exceptions from inside lazy I/O, but now we do.
1043 
1044  3. If anyone is actually relying on the previous behaviour, their
1045     code is arguably broken.
1046]
1047[Re-instate System.Console.Getopt for nhc98 builds.
1048Malcolm.Wallace@cs.york.ac.uk**20091013092843
1049 Although it was split out of base a while back, that change was
1050 reverted for ghc soon afterwards, but nhc98 never noticed.
1051]
1052[Roll back "Another instance of nhc98's strange import semantics."
1053Ian Lynagh <igloo@earth.li>**20091009185618
1054 Fri Oct  9 14:50:51 BST 2009  Malcolm.Wallace@cs.york.ac.uk
1055 GHC (correctly) warns about the unused import, which breaks the validate
1056 build.
1057]
1058[Roll back "Cope with nhc98's (occasionally-strange) import semantics"
1059Ian Lynagh <igloo@earth.li>**20091009184704
1060 Fri Oct  9 14:43:51 BST 2009  Malcolm.Wallace@cs.york.ac.uk
1061 GHC (correctly) warns about the unused import, which breaks the validate
1062 build.
1063]
1064[It seems that nhc98 needs defaulting in Data.Fixed.
1065Malcolm.Wallace@cs.york.ac.uk**20091009135242]
1066[Another instance of nhc98's strange import semantics.
1067Malcolm.Wallace@cs.york.ac.uk**20091009135051]
1068[Make Data.Functor compatible with non-GHC compilers.
1069Malcolm.Wallace@cs.york.ac.uk**20091009134821]
1070[Cope with nhc98's (occasionally-strange) import semantics.
1071Malcolm.Wallace@cs.york.ac.uk**20091009134351]
1072[Fix gratuitous breakage of nhc98 in System.IO.
1073Malcolm.Wallace@cs.york.ac.uk**20091009134001]
1074[Fix gratuitous breakage of nhc98 in Control.Exception.Base.
1075Malcolm.Wallace@cs.york.ac.uk**20091009133615]
1076[Fix gratuitous breakage of non-GHC in Data.Fixed.
1077Malcolm.Wallace@cs.york.ac.uk**20091009133330]
1078[Fix gratuitous breakage for non-GHC in Data.Bits.
1079Malcolm.Wallace@cs.york.ac.uk**20091009133257]
1080[Use UTF-32LE instead of UTF32LE
1081Simon Marlow <marlowsd@gmail.com>**20091006100207
1082 Ignore-this: 7f881e36543d250ef848c9f60d67655a
1083 The latter is not recognised by some iconv implementations.
1084]
1085[Strip any Byte Order Mark (BOM) from the front of decoded streams.
1086Ben.Lippmeier@anu.edu.au*-20090930084229
1087 Ignore-this: d0d0c3ae87b31d71ef1627c8e1786445
1088 When decoding to UTF-32, Solaris iconv inserts a BOM at the front
1089 of the stream, but Linux iconv doesn't.
1090]
1091[use UTF32BE/UTF32LE instead of UCS-4/UCS-4LE
1092Simon Marlow <marlowsd@gmail.com>**20091005101554
1093 Ignore-this: 2aef5e9bec421e714953b7aa1bdfc1b3
1094]
1095[Strip any Byte Order Mark (BOM) from the front of decoded streams.
1096Ben.Lippmeier@anu.edu.au**20090930084229
1097 Ignore-this: d0d0c3ae87b31d71ef1627c8e1786445
1098 When decoding to UTF-32, Solaris iconv inserts a BOM at the front
1099 of the stream, but Linux iconv doesn't.
1100]
1101[Add traceEvent :: String -> IO ()
1102Simon Marlow <marlowsd@gmail.com>**20090925141257
1103 Ignore-this: 8b1888bbf9682ffba13f815b6000e4b1
1104 For emitting an event via the RTS tracing framework
1105]
1106[Fix the error message when flushing the read buffer of a non-seekable Handle
1107Simon Marlow <marlowsd@gmail.com>**20090923090536
1108 Ignore-this: 4342026df93759d99480f4e13f80a492
1109]
1110[Fix #3534: No need to flush the byte buffer when setting binary mode
1111Simon Marlow <marlowsd@gmail.com>**20090923090445
1112 Ignore-this: 625817ed7ae2c12291eb993a99dc640a
1113]
1114[Use let !y = x in .. x .. instead of seq in $! and evaluate (#2273)
1115Simon Marlow <marlowsd@gmail.com>**20090916140454]
1116[make some Applicative functions into methods, and split off Data.Functor (proposal #3335)
1117Ross Paterson <ross@soi.city.ac.uk>**20090915173109
1118 Ignore-this: a0cff4de6dfdbcbd56a66101bc4855a9
1119 
1120 The following functions
1121 
1122     (<$) :: Functor f => a -> f b -> f a
1123     (*>) :: Applicative f => f a -> f b -> f b
1124     (<*) :: Applicative f => f a -> f b -> f a
1125     some :: Alternative f => f a -> f [a]
1126     many :: Alternative f => f a -> f [a]
1127 
1128 are moved into the corresponding classes, with the existing implementations
1129 as default definitions.  This gives people creating instances the option of
1130 defining specialized implementations of these functions, though they should
1131 be equivalent to the default definitions.
1132 
1133 Although (<$) is now a method of the Functor class, it is hidden in the
1134 re-export by the Prelude, Control.Monad and Monad.  The new module
1135 Data.Functor exposes the full class, plus the function (<$>).  These are
1136 also re-exported by Control.Applicative.
1137]
1138[On Windows, use the console code page for text file encoding/decoding.
1139Judah Jacobson <judah.jacobson@gmail.com>**20090913022126
1140 Ignore-this: 86c2f2db8ef92b751599795d3195187b
1141 
1142 We keep all of the code page tables in the module
1143 GHC.IO.Encoding.CodePage.Table.  That file was generated automatically
1144 by running codepages/MakeTable.hs; more details are in the comments at the
1145 start of that script.
1146 
1147 Storing the lookup tables adds about 40KB to each statically linked executable;
1148 this only increases the size of a "hello world" program by about 7%.
1149 
1150 Currently we do not support double-byte encodings (Chinese/Japanese/Korean), since
1151 including those codepages would increase the table size to 400KB.  It will be
1152 straightforward to implement them once the work on library DLLs is finished.
1153]
1154[Fix "init" docs: the input list need not be finite. Fixes trac #3465
1155Ian Lynagh <igloo@earth.li>**20090911210437]
1156[Bump base version to 4.2.0.0
1157Ian Lynagh <igloo@earth.li>**20090911153913]
1158[Address #3310
1159Simon Marlow <marlowsd@gmail.com>**20090830152850
1160 Ignore-this: 40c7f7c171ee299a83092fd360a952b7
1161 
1162  - Rename BlockedOnDeadMVar   -> BlockedIndefinitelyOnMVar
1163  - Rename BlockedIndefinitely -> BlockedIndefinitelyOnSTM
1164  - instance Show BlockedIndefinitelyOnMVar is now
1165      "blocked indefinitely in an MVar operation"
1166  - instance Show BlockedIndefinitelyOnSTM is now
1167      "blocked indefinitely in an STM transaction"
1168 
1169 clients using Control.OldException will be unaffected (the new
1170 exceptions are mapped to the old names).  However, for base4-compat
1171 we'll need to make a version of catch/try that does a similar
1172 mapping.
1173]
1174[Fix unicode conversion for MSB architectures
1175Ben.Lippmeier@anu.edu.au**20090830130028
1176 This fixes the SPARC/Solaris build.
1177]
1178[Fix #3441: detect errors in partial sequences
1179Simon Marlow <marlowsd@gmail.com>**20090830075909
1180 Ignore-this: d12a75d95e0cae5eb1555266810ec281
1181]
1182[Fix hWaitForInput
1183Simon Marlow <marlowsd@gmail.com>**20090827152116
1184 Ignore-this: 2550e911f1a4d4357a5aa8d1764238ce
1185 It was erroneously waiting when there were bytes to decode waiting in
1186 the byte buffer.
1187]
1188[fix debugging code
1189Simon Marlow <marlowsd@gmail.com>**20090827150628
1190 Ignore-this: e1c82fdc19a22e247cd69ff6fa11921d
1191]
1192[Allow for configurable iconv include and library locations.
1193Matthias Kilian <kili@outback.escape.de>**20090826154406
1194 Ignore-this: be95fab611a5534cf184b508964ed498
1195 This should help to fix the build on OpenBSD.
1196]
1197[typo in comment
1198Simon Marlow <marlowsd@gmail.com>**20090826085252
1199 Ignore-this: 1903ee0f354157a6ed3871c100f6b1b9
1200]
1201[un-hide some modules from the Haddock docs
1202Simon Marlow <marlowsd@gmail.com>**20090825152457
1203 Ignore-this: dce6606f93cf977fb24ebe99082dfa62
1204]
1205[Apply fix for #1548, from squadette@gmail.com
1206Simon Marlow <marlowsd@gmail.com>**20090819120700
1207 Ignore-this: 31c237c46a6445f588ed4b8c51bb6231
1208]
1209[improvements to Data.Fixed: instances for Typeable and Data, more predefined types
1210Ashley Yakeley <ashley@semantic.org>**20090812055058
1211 Ignore-this: feeece36d5632f02a05d137d2a39ab78
1212]
1213[Fix "Cabal check" warnings
1214Ian Lynagh <igloo@earth.li>**20090811215856]
1215[Add a GHC.Constants module; fixes trac #3094
1216Ian Lynagh <igloo@earth.li>**20090809183252]
1217[Apply proposal #3393
1218Ian Lynagh <igloo@earth.li>**20090809134717
1219 Add openTempFileWithDefaultPermissions and
1220 openBinaryTempFileWithDefaultPermissions.
1221]
1222[Add some more C wrappers; patch from Krister Walfridsson
1223Ian Lynagh <igloo@earth.li>**20090807200631
1224 Fixes 21 testsuite errors on NetBSD 5.99.
1225]
1226[Fixing configure for autoconf 2.64
1227Alexander Dunlap <alexander.dunlap@gmail.com>**20090805060748
1228 Ignore-this: 992ab91ae3d68c12dbb265776e33e243
1229]
1230[add INLINE toList
1231Ross Paterson <ross@soi.city.ac.uk>**20090806142853
1232 Ignore-this: aba16aabb17d5dca44f15d188945680e
1233 
1234 In anticipation of the fixing of #2353.
1235]
1236[fix a copyright
1237Simon Marlow <marlowsd@gmail.com>**20090805134045
1238 Ignore-this: b0ffbdd38fbba121e8bcba37c4082a60
1239]
1240[Tweak the BufferedIO class to enable a memory-mapped file implementation
1241Simon Marlow <marlowsd@gmail.com>**20090805134036
1242 Ignore-this: ec67d7a0a6d977438deaa342503f77e0
1243 We have to eliminate the assumption that an empty write buffer can be
1244 constructed by setting the buffer pointers to zero: this isn't
1245 necessarily the case when the buffer corresponds to a memory-mapped
1246 file, or other in-memory device implementation.
1247]
1248[Deprecate Control.OldException
1249Ian Lynagh <igloo@earth.li>**20090804143910]
1250[Windows build fix, following RTS tidyup
1251Simon Marlow <marlowsd@gmail.com>**20090803131121
1252 Ignore-this: ce862fb91c2b234211a8757f98690778
1253]
1254[Updates to follow the RTS tidyup
1255Simon Marlow <marlowsd@gmail.com>**20090801220743
1256 Ignore-this: 6e92412df93a66c12d75344053d5634
1257 C functions like isDoubleNaN moved here (primFloat.c)
1258]
1259[Add integer-simple as a build option
1260Ian Lynagh <igloo@earth.li>**20090722013151]
1261[Use shift[LR]Integer in the Bits Integer instance
1262Ian Lynagh <igloo@earth.li>**20090721222440]
1263[depend directly on integer-gmp, rather than indirecting through integer
1264Ian Lynagh <igloo@earth.li>**20090721185228]
1265[Move the instances of Functor and Monad IO to GHC.Base, to avoid orphans
1266Simon Marlow <marlowsd@gmail.com>**20090722102130
1267 Ignore-this: a7d85ac0025d559674249de0108dbcf4
1268]
1269[move "instance Exception Dynamic" so it isn't an orphan
1270Simon Marlow <marlowsd@gmail.com>**20090721093854
1271 Ignore-this: 5ede91ecfec2112c91b699d4de87cd02
1272]
1273[Improve the index checking for array accesses; fixes #2120 #2669
1274Ian Lynagh <igloo@earth.li>**20090719153228
1275 As well as checking that offset we are reading is actually inside the
1276 array, we now also check that it is "in range" as defined by the Ix
1277 instance. This fixes confusing behaviour (#2120) and improves some error
1278 messages (#2669).
1279]
1280[Make chr say what its argument was, if it's a bad argument
1281Ian Lynagh <igloo@earth.li>**20090718151049]
1282[remove unused warning
1283Simon Marlow <marlowsd@gmail.com>**20090715124416
1284 Ignore-this: 31f613654089d0f4a44363946087b41e
1285]
1286[warning fix: -fno-implicit-prelude -> -XNoImplicitPrelude
1287Simon Marlow <marlowsd@gmail.com>**20090715122839
1288 Ignore-this: dc8957249731d5bcb71c01899e5adf2b
1289]
1290[Add hGetEncoding :: Handle -> IO (Maybe TextEncoding)
1291Simon Marlow <marlowsd@gmail.com>**20090715122519
1292 Ignore-this: 14c3eff996db062da1199739781e4708
1293 as suggested during the discussion on the libraries list
1294]
1295[Add more documentation to mkTextEncoding
1296Simon Marlow <marlowsd@gmail.com>**20090715122414
1297 Ignore-this: 97253b2624267df3a246a18121e8ea81
1298 noting that "//IGNORE" and "//TRANSLIT" suffixes can be used with GNU
1299 iconv.
1300]
1301[Add the utf8_bom codec
1302Simon Marlow <marlowsd@gmail.com>**20090715122257
1303 Ignore-this: 1c9396cd805201fe873a39382ced79c7
1304 as suggested during the discussion on the libraries list.
1305]
1306[Export Unicode and newline functionality from System.IO; update Haddock docs
1307Simon Marlow <marlowsd@gmail.com>**20090713113104
1308 Ignore-this: c3f017a555335aa55d106253393f72e2
1309]
1310[add a comment about the non-workingness of CHARBUF_UTF16
1311Simon Marlow <marlowsd@gmail.com>**20090707124406
1312 Ignore-this: 98d00411b68d688b3b4cffc9507b1f35
1313]
1314[Fix build on Windows
1315Ian Lynagh <igloo@earth.li>**20090711004351]
1316[Fix some "warn-unused-do-bind" warnings where we want to ignore the value
1317Ian Lynagh <igloo@earth.li>**20090710204513]
1318[Use throwErrnoIfMinus1_ when calling getrusage
1319Ian Lynagh <igloo@earth.li>**20090710204221]
1320[Remove an unused import
1321Ian Lynagh <igloo@earth.li>**20090710153345]
1322[reportStackOverflow now returns IO ()
1323Ian Lynagh <igloo@earth.li>**20090710153257
1324 It used to do "return undefined" to return IO a.
1325]
1326[GHC.Conc.reportError now returns IO ()
1327Ian Lynagh <igloo@earth.li>**20090710152646
1328 It used to return IO a, by "return undefined".
1329]
1330[Fix some "warn-unused-do-bind" warnings where we want to ignore the value
1331Ian Lynagh <igloo@earth.li>**20090710152526]
1332[Minor SampleVar refactoring
1333Ian Lynagh <igloo@earth.li>**20090710151438]
1334[Fix "warn-unused-do-bind" warnings in GHC/IO/Handle/Text.hs
1335Ian Lynagh <igloo@earth.li>**20090710122905]
1336[Fix some "warn-unused-do-bind" warnings where we just want to ignore the result
1337Ian Lynagh <igloo@earth.li>**20090710005638]
1338[Use the result of writeCharBuf in GHC/IO/Encoding/Latin1.hs too
1339Ian Lynagh <igloo@earth.li>**20090710004032]
1340[Minor code tidyups in GHC.Conc
1341Ian Lynagh <igloo@earth.li>**20090710003801]
1342[Fix "warn-unused-do-bind" warning in GHC.Conc
1343Ian Lynagh <igloo@earth.li>**20090710003530
1344 If we fail to communicate with the IO manager then we print a warning
1345 using debugErrLn from the ghc-prim package.
1346]
1347[Fix "warn-unused-do-bind" warnings in System.Posix.Internals
1348Ian Lynagh <igloo@earth.li>**20090709164546]
1349[Fix "warn-unused-do-bind" warnings where we really do want to ignore the result
1350Ian Lynagh <igloo@earth.li>**20090709163912]
1351[Add back imports needed on Windows
1352Ian Lynagh <igloo@earth.li>**20090707181924]
1353[Remove unused imports
1354Ian Lynagh <igloo@earth.li>**20090707115810]
1355[Remove unused imports from base
1356simonpj@microsoft.com**20090706111842
1357 Ignore-this: f9b5f353e3bb820f787c56d615b28765
1358 
1359 These unused imports are detected by the new unused-import code
1360 
1361]
1362[Use the result of writeCharBuf
1363Simon Marlow <marlowsd@gmail.com>**20090706133303
1364 Ignore-this: 52288dd559bf4c4f313df6197091d935
1365   
1366 This only makes a difference when CHARBUF_UTF16 is in use, which it
1367 normally isn't.  I suspect CHARBUF_UTF16 doesn't currently work for
1368 other reasons (CHARBUF_UTF16 was an experiment before I wrote the
1369 GHC.IO.Encoding.UTF* codecs), but this patch at least makes it
1370 slightly closer to working.
1371]
1372[Remove some cruft from Data.HashTable
1373Ian Lynagh <igloo@earth.li>**20090706181630]
1374[Add 'eof' to Text.ParserCombinators.ReadP
1375simonpj@microsoft.com**20090706111801
1376 Ignore-this: 2aea7b848e00c894761bc4011adaa95d
1377 
1378 Add a ReadP parser that succeeds at the end of input. Very useful!
1379 
1380]
1381[Don't export CLDouble for GHC; fixes trac #2793
1382Ian Lynagh <igloo@earth.li>**20090705155120
1383 We never really supported CLDouble (it was a plain old double underneath),
1384 and pretending that we do does more harm than good.
1385]
1386[a byte between 0x80 and 0xBF is illegal immediately (#3341)
1387Simon Marlow <marlowsd@gmail.com>**20090702081415
1388 Ignore-this: dc19ef59a1a21118d5a7dd38aa2f611c
1389]
1390[avoid a warning
1391Simon Marlow <marlowsd@gmail.com>**20090630084134
1392 Ignore-this: c92a45ee216faf01327feae9fe06d6e2
1393]
1394[Add a wrapper for libiconv.
1395Matthias Kilian <kili@outback.escape.de>**20090629183634
1396 Ignore-this: 23c6047c0d71b745b495cc223574a47f
1397]
1398[#include <sys/times.h> if we have it (should fix build problems)
1399Simon Marlow <marlowsd@gmail.com>**20090629085351
1400 Ignore-this: a35e93b37ca9595c73460243180f4b9d
1401]
1402[set binary mode for existing FDs on Windows (fixes some GHCi test failures)
1403Simon Marlow <marlowsd@gmail.com>**20090626120522
1404 Ignore-this: 580cf636e9c77d8427aff6861d089481
1405]
1406[Move directory-related stuff to the unix package
1407Simon Marlow <marlowsd@gmail.com>**20090625120325
1408 Ignore-this: b997b3cbce0a46ca87ad825bbdc0a411
1409 now that it isn't used on Windows any more.
1410]
1411[TAG 2009-06-25
1412Ian Lynagh <igloo@earth.li>**20090625160056]
1413Patch bundle hash:
1414f77240a2b18f1bbc9982b3a4fb6c969d80837bb2