Ticket #4334: lines.dpatch

File lines.dpatch, 52.3 KB (added by daniel.is.fischer, 4 years ago)

patch fixing the space leak without changing strictness

Line 
11 patch for repository /home/dafis/Haskell/Hacking/ghc/libraries/base:
2
3Tue Oct 19 17:23:27 CEST 2010  Daniel Fischer <daniel.is.fischer@web.de>
4  * FIX #4334
5  Make selector thunks visible to GHC to fix a space leak in lines.
6
7New patches:
8
9[FIX #4334
10Daniel Fischer <daniel.is.fischer@web.de>**20101019152327
11 Ignore-this: 92a7c441ff7b9f1dde066479a09435ae
12 Make selector thunks visible to GHC to fix a space leak in lines.
13] {
14hunk ./Data/List.hs 7
15 -- Module      :  Data.List
16 -- Copyright   :  (c) The University of Glasgow 2001
17 -- License     :  BSD-style (see the file libraries/base/LICENSE)
18---
19+--
20 -- Maintainer  :  libraries@haskell.org
21 -- Stability   :  stable
22 -- Portability :  portable
23hunk ./Data/List.hs 40
24    , intersperse       -- :: a -> [a] -> [a]
25    , intercalate       -- :: [a] -> [[a]] -> [a]
26    , transpose         -- :: [[a]] -> [[a]]
27-   
28+
29    , subsequences      -- :: [a] -> [[a]]
30    , permutations      -- :: [a] -> [[a]]
31 
32hunk ./Data/List.hs 341
33 -- Not exported:
34 -- Note that we keep the call to `eq` with arguments in the
35 -- same order as in the reference implementation
36--- 'xs' is the list of things we've seen so far,
37+-- 'xs' is the list of things we've seen so far,
38 -- 'y' is the potential new element
39 elem_by :: (a -> a -> Bool) -> a -> [a] -> Bool
40 elem_by _  _ []         =  False
41hunk ./Data/List.hs 474
42 mapAccumL :: (acc -> x -> (acc, y)) -- Function of elt of input list
43                                     -- and accumulator, returning new
44                                     -- accumulator and elt of result list
45-          -> acc            -- Initial accumulator
46+          -> acc            -- Initial accumulator
47           -> [x]            -- Input list
48           -> (acc, [y])     -- Final accumulator and result list
49 mapAccumL _ s []        =  (s, [])
50hunk ./Data/List.hs 1041
51 -- characters.  The resulting strings do not contain newlines.
52 lines                   :: String -> [String]
53 lines ""                =  []
54+#ifdef __GLASGOW_HASKELL__
55+-- Somehow GHC doesn't detect the selector thunks in the below code,
56+-- so s' keeps a reference to the first line via the pair and we have
57+-- a space leak (cf. #4334).
58+-- So we need to make GHC see the selector thunks with a trick.
59+lines s                 =  cons (case break (== '\n') s of
60+                                    (l, s') -> (l, case s' of
61+                                                    []      -> []
62+                                                    _:s''   -> lines s''))
63+  where
64+    cons ~(h, t)        =  h : t
65+#else
66 lines s                 =  let (l, s') = break (== '\n') s
67                            in  l : case s' of
68                                         []      -> []
69hunk ./Data/List.hs 1057
70                                         (_:s'') -> lines s''
71+#endif
72 
73 -- | 'unlines' is an inverse operation to 'lines'.
74 -- It joins lines, after appending a terminating newline to each.
75}
76
77Context:
78
79[Define SpecConstrAnnotation in GHC.Exts, and import it from there
80simonpj@microsoft.com**20101018135857
81 Ignore-this: 8bf81cbc5787dbb5a3875b5622f67732
82   
83 Reason: avoid having to link the entire ghc package in modules
84 that use compile-time annotations:
85   
86        import GHC.Exts( SpecConstrAnnotation )
87        {-# ANN type T ForceSpecConstr #-}
88   
89 It's a kind of bug that the package exporting SpecConstrAnnotation
90 is linked even though it is only needed at compile time, but putting
91 the data type declaration in GHC.Exts is a simple way to sidestep
92 the problem
93   
94 See See Note [SpecConstrAnnotation] in SpecConstr
95]
96[throwTo: mention interruptible foreign calls
97Simon Marlow <marlowsd@gmail.com>**20101014084220
98 Ignore-this: dbc53d85f870cf649f87186c7185465a
99]
100[remove trailing whitespace
101Simon Marlow <marlowsd@gmail.com>**20101013101906
102 Ignore-this: b8b424540cacbbb3c6d934242e3af795
103]
104[FIX #4381
105Simon Marlow <marlowsd@gmail.com>**20101013101849
106 Ignore-this: f0daa4845eeb444231451b975b71d055
107 Fix scaleFloat by clamping the scaling parameter so that
108 exponent + scale doesn't overflow.
109 
110 Patch by: Daniel Fischer <daniel.is.fischer@web.de>
111]
112[Replaced some throws to throwIOs where the type is IO
113Bas van Dijk <v.dijk.bas@gmail.com>**20100924221340
114 Ignore-this: e74191e4527ae6f7551c95fd41063335
115]
116[Added initial .authorspellings
117Bas van Dijk <v.dijk.bas@gmail.com>**20101005072701
118 Ignore-this: 63628bcabfdd0b7beda4cd37daeccd89
119]
120[Lazier intersperse
121Daniel Fischer <daniel.is.fischer@web.de>**20101002231201
122 Ignore-this: a0fed65930cf19e68b4363381a5ab576
123 A lazier implementation of intersperse, and consequentially intercalate, to
124 avoid space leaks.
125]
126[FIX #4228 (atanh (-1) returns NaN instead of -Infinity)
127ghc@cainnorris.net**20100816213654
128 Ignore-this: dee89c24493e84a02bea711a1c83a73f
129]
130[Make intersectBy lazier
131Daniel Fischer <daniel.is.fischer@web.de>**20100930191731
132 Ignore-this: ef687bc75923434e85c14b57171576aa
133 Add shortcuts to intersectBy for empty list arguments.
134 In addition to being faster in that case, more inputs yield defined results.
135 Treats ticket #4323
136]
137[doc tweak for Directory file type: file names are '\0'-separated
138Simon Marlow <marlowsd@gmail.com>**20100922113811
139 Ignore-this: 96b7b004bd6e5bc3e958ad55bf238ba1
140]
141[documentation for IODeviceType (#4317, edited by me)
142Simon Marlow <marlowsd@gmail.com>**20100915131341
143 Ignore-this: 21c50ca7a189eebcf299523b6e942bae
144]
145[Allow Data.HashTable construction with user-supplied size
146**20100722210726
147 Ignore-this: bd54880bb16a106a992f03b040dc4164
148 
149 This avoids some resizing for users who know they will be inserting a
150 lot of data.
151 
152 http://hackage.haskell.org/trac/ghc/ticket/4193
153]
154[some fixes for hGetBufSome
155Simon Marlow <marlowsd@gmail.com>**20100916113732
156 Ignore-this: 3e596a606c180dc4859ea8f4c9132ca1
157  - fix one case where it was blocking when it shouldn't
158  - a couple of error-message tweaks
159]
160[Windows: map ERROR_NO_DATA to EPIPE, rather than EINVAL
161Simon Marlow <marlowsd@gmail.com>**20100915142618
162 Ignore-this: 9023e5f0542419f225aef26cb6b1d88d
163 WriteFile() returns ERROR_NO_DATA when writing to a pipe that is
164 "closing", however by default the write() wrapper in the CRT maps this
165 to EINVAL so we get confusing things like
166 
167   hPutChar: invalid argument (Invalid Argumnet)
168 
169 when piping the output of a Haskell program into something that closes
170 the pipe early.  This was happening in the testsuite in a few place.
171 
172 The solution is to map ERROR_NO_DATA to EPIPE correctly, as we
173 explicitly check for EPIPE on stdout (in GHC.TopHandler) so we can
174 exit without an error in this case.
175]
176[tighten up parsing of numbers (#1579)
177Simon Marlow <marlowsd@gmail.com>**20100913214733
178 Ignore-this: 3411bf3d2e98cfacb9e0afd11d79e722
179]
180[Add absentError. 
181simonpj@microsoft.com**20100914134639
182 Ignore-this: d0eef5a87e1def4cdbde92a55241c8c4
183 
184 This patch accompanies the HEAD patch:
185 
186   Tue Sep 14 12:38:27 BST 2010  simonpj@microsoft.com
187     * Make absent-arg wrappers work for unlifted types (fix Trac #4306)
188     
189     Previously we were simply passing arguments of unlifted
190     type to a wrapper, even if they were absent, which was
191     stupid.
192     
193     See Note [Absent error Id] in WwLib.
194]
195[Add missing import, fixes build on windows
196simonpj@microsoft.com**20100914122750
197 Ignore-this: 12ece15ef94982ddfbf5f9f7900619da
198]
199[Add a suitable Show instance for TextEncoding (#4273)
200Simon Marlow <marlowsd@gmail.com>**20100913154459
201 Ignore-this: 77f2235460895debd2827f34c42c3435
202]
203[don't fill a finalized handle with an error (see comment)
204Simon Marlow <marlowsd@gmail.com>**20100913153350
205 Ignore-this: c72cdb6898dffa88eca1d781171b2943
206]
207[deriving (Eq, Ord, Read, Show) for Newline and NewlineMode
208Simon Marlow <marlowsd@gmail.com>**20100913153031
209 Ignore-this: 9b9b29bfb7abf5550cfbfa7788f81bf
210]
211[fix warning on Windows
212Simon Marlow <marlowsd@gmail.com>**20100913111536
213 Ignore-this: dacc5448c452daad60ed37a1a5ed096e
214]
215[Put the state-token argument on fill, done, adjust on the RHS
216simonpj@microsoft.com**20100913101832
217 Ignore-this: d228b492de7d4635c026ed24cbc17e34
218 
219 This is so that the functions will inline when
220 applied to their normal (non-state-token) aguments.
221 
222 I forget why I did this, but it seems like the right thing anyway.
223]
224[avoid Foreign.unsafePerformIO
225Ross Paterson <ross@soi.city.ac.uk>**20100909125521
226 Ignore-this: b698101119ffd1bc6311cce0736f745d
227]
228[Remove debugging code accidentally left in
229Simon Marlow <marlowsd@gmail.com>**20100909113331
230 Ignore-this: 906a14176dd37030b8203782a687936b
231]
232[Fix Windows build; patches frmo ezyang
233Ian Lynagh <igloo@earth.li>**20100908123037
234 Ignore-this: 2f02986087edd7da8382221012c27cd0
235]
236[More accurate isatty test for MinGW.
237Edward Z. Yang <ezyang@mit.edu>**20100907154144
238 Ignore-this: 93bdc2b2a8e65a7c4c7d3906bdda01db
239]
240[Fix the build when HAVE_KQUEUE but not HAVE_KEVENT64
241Ian Lynagh <igloo@earth.li>**20100904223703]
242[Fix warnings
243benl@ouroborus.net**20100830044741
244 Ignore-this: 8397aaec7c36046c9ace403e65f32d32
245]
246[fix cache variable name used by FP_SEARCH_LIBS_PROTO
247Ross Paterson <ross@soi.city.ac.uk>**20100819204858
248 Ignore-this: b8113cb3c6f0e03c507297c99d3d82b7
249]
250[Add a missing castPtr (only shows up in -DDEBUG)
251simonpj@microsoft.com**20100815145127
252 Ignore-this: 30b9c42cd3ce7837bdabd254fe66078d
253]
254[Fixed a rounding error in threadDelay
255Johan Tibell <johan.tibell@gmail.com>**20100813124043
256 Ignore-this: 1cb77d0852233ffffb144b134064ee3c
257]
258[export allocaBytesAligned; make allocaArray use the correct alignment (#2917)
259Simon Marlow <marlowsd@gmail.com>**20100812105524
260 Ignore-this: deb6495f7b7b84deaf02b88927a5ba8c
261]
262[deprecate unGetChan and isEmptyChan (see #4154)
263Simon Marlow <marlowsd@gmail.com>**20100705125952
264 Ignore-this: b4e769959f131b2d0001eb7202bc1b92
265]
266[Add type signatures to cope with lack of local generalisation
267simonpj@microsoft.com**20100728124847
268 Ignore-this: d3af9a47c2821c6081bde05a135a92fb
269]
270[Add type signature in local where
271simonpj@microsoft.com**20100727151532
272 Ignore-this: 1c57063ad32d13e0d1ec8daf968bf055
273]
274[Integrated new I/O manager
275Simon Marlow <marlowsd@gmail.com>**20100810082248
276 Ignore-this: ed70a9066ac9b676a446fe99978fef7a
277 (patch originally by Johan Tibell <johan.tibell@gmail.com>, minor merging by me)
278]
279[Add mfilter to Control.Monad
280jon.fairbairn@cl.cam.ac.uk**20090917145616
281 Ignore-this: de4240b60684f3065b29378df3ea98f2
282 
283 Straightforward MonadPlus version of List.filter. I would
284 prefer to call it filter, but the current naming scheme for
285 Control.Monad implies mfilter.
286 
287]
288[move Monad and MonadFix instances for Either from mtl (proposal #4159)
289Ross Paterson <ross@soi.city.ac.uk>**20100729122449
290 Ignore-this: b0f8cd8643679948d1da43bd7c08c5aa
291 
292 The Monad and MonadFix instances for Either (formerly in the mtl
293 package) are moved to Control.Monad.Instances and Control.Monad.Fix
294 respectively.  The Monad instance is still an orphan, to retain Haskell
295 98 compatibility, but the MonadFix instance is together with its class.
296 The Error constraint is removed from both instances, and the default
297 definition of fail is used.
298]
299[Remove egregious ghc-ish from Foreign.Marshal
300Malcolm.Wallace@me.com**20100722075449]
301[add numSparks :: IO Int (#4167)
302Simon Marlow <marlowsd@gmail.com>**20100720153858
303 Ignore-this: 4543f57a7f137f8cae1c3efc5c023a9b
304]
305[add unsafeLocalState from Haskell 2010, and docs
306Simon Marlow <marlowsd@gmail.com>**20100720082819
307 Ignore-this: dcd79fb546ebe29ddff4df279ec2f38
308]
309[docs: mention that Foreign.unsafePerformIO is deprecated
310Simon Marlow <marlowsd@gmail.com>**20100720082804
311 Ignore-this: 4cfebb8f2a1cddc7d15e94e31b2befa4
312 We can't actually deprecate it without introducing a name clash
313 between Foreign.unsafePerformIO and System.IO.Unsafe.unsafePerformIO
314]
315[doc formatting fix
316Simon Marlow <marlowsd@gmail.com>**20100714151347
317 Ignore-this: 255edef607dcd290e198015240b5d125
318]
319[add module intro from Haskell 2010
320Simon Marlow <marlowsd@gmail.com>**20100714115853
321 Ignore-this: 59b5a07507a059ccccdff2dfb6490a27
322]
323[document exception-overriding behaviour in withFile
324Simon Marlow <marlowsd@gmail.com>**20100714104107
325 Ignore-this: f99e641ea2f46d872cb7420a62fa50dc
326]
327[doc: use "finalizer" consistently
328Simon Marlow <marlowsd@gmail.com>**20100714103649
329 Ignore-this: bdfea40f31dc5045fdbc6e12266dda93
330]
331[clarify meaning of bit
332Simon Marlow <marlowsd@gmail.com>**20100714103310
333 Ignore-this: 521b031f1e83ef34ca03d9aa9273df8a
334]
335[note shortcutting behaviour of any/all/elem
336Simon Marlow <marlowsd@gmail.com>**20100714103304
337 Ignore-this: 1605f362ba0712ad1cea1309636f3ea1
338]
339[add cast{C,U}CharToChar and castCharTo{C,U}Char, from Haskell 2010
340Simon Marlow <marlowsd@gmail.com>**20100713132515
341 Ignore-this: 9b1da827016c7b08668078b45964e9de
342]
343[mention that IntPtr and WordPtr can be marshalled to/from intptr_t and uintptr_t
344Simon Marlow <marlowsd@gmail.com>**20100713132403
345 Ignore-this: dcc112a72746ba117a84fa29e71b6800
346]
347[Partial fix for Trac #4136
348simonpj@microsoft.com**20100707135725
349 Ignore-this: 9548eeb3187d9779d4e5c858a0f35354
350 
351 In 'choose' (which is a library function designed specifically
352 to support derived instances of Read), we must match Symbol
353 as well as Ident, for nullary constructors that (wierdly) are
354 symbols.
355]
356[Fix typo in documentation
357Simon Hengel <simon.hengel@wiktory.org>**20100711141648
358 Ignore-this: c052dd8a681832ef598a323ad55eae3a
359]
360[Remove duplicated word in documentation
361Simon Hengel <simon.hengel@wiktory.org>**20100711072703
362 Ignore-this: fb3732dc57be55f14168792f923433
363]
364[Allow nhc98 to cope with recent changes to Control.Exception.
365Malcolm.Wallace@me.com**20100710170940]
366[ New asynchronous exception control API (base parts)
367Simon Marlow <marlowsd@gmail.com>**20100708152735
368 Ignore-this: 71a4811804f04259f1fe739f8863beaf
369   
370 As discussed on the libraries/haskell-cafe mailing lists
371   http://www.haskell.org/pipermail/libraries/2010-April/013420.html
372 
373 This is a replacement for block/unblock in the asychronous exceptions
374 API to fix a problem whereby a function could unblock asynchronous
375 exceptions even if called within a blocked context.
376 
377 The new terminology is "mask" rather than "block" (to avoid confusion
378 due to overloaded meanings of the latter).
379 
380 The following is the new API; the old API is deprecated but still
381 available for the time being.
382 
383 Control.Exception
384 -----------------
385 
386 mask  :: ((forall a. IO a -> IO a) -> IO b) -> IO b
387 mask_ :: IO a -> IO a
388 
389 uninterruptibleMask :: ((forall a. IO a -> IO a) -> IO b) -> IO b
390 uninterruptibleMask_ :: IO a -> IO
391 
392 getMaskingState :: IO MaskingState
393 
394 data MaskingState
395   = Unmasked
396   | MaskedInterruptible
397   | MaskedUninterruptible
398 
399 
400 Control.Concurrent
401 ------------------
402 
403 forkIOUnmasked :: IO () -> IO ThreadId
404]
405[Async-exception safety, and avoid space leaks
406Simon Marlow <marlowsd@gmail.com>**20100708145819
407 Ignore-this: dbfd0e61551e9e7b4fc1c6fe9b9a83de
408 Patch submitted by: Bas van Dijk <v.dijk.bas@gmail.com>
409 Modified slightly by me to remove non-functional changes.
410]
411[Async-exception safety, and avoid space leaks
412Simon Marlow <marlowsd@gmail.com>**20100708103154
413 Ignore-this: 190c3ac8f6633231624da8cf1316588
414 Patch submitted by: Bas van Dijk <v.dijk.bas@gmail.com>
415 Modified slightly by me to remove non-functional changes.
416]
417[Fix a few places where we forgot to close the text codecs (#4029)
418Simon Marlow <marlowsd@gmail.com>**20100702130210
419 Ignore-this: 2e81a4b4cb343181cef34b0f9e2ded47
420 Each time you invoke :load in GHCi it resets the CAFs, including
421 stdin/stdout/stderr, and each of these was allocating a new iconv_t.
422]
423[remove docs from Monad that belonged on the instance for MonadPlus IO
424Simon Marlow <marlowsd@gmail.com>**20100701154203
425 Ignore-this: 59df02542a7ac9421552a2155d848d27
426]
427[docs: unqualify Prelude.IO
428Simon Marlow <marlowsd@gmail.com>**20100701153817
429 Ignore-this: 73b0202876c827e7a5b4a5ce74e724c4
430]
431[unqualify Float and Double
432Simon Marlow <marlowsd@gmail.com>**20100701142727
433 Ignore-this: cbe89d31a00bf49996a33933324fca17
434]
435[extract information about Data.Time from docs for CTime
436Simon Marlow <marlowsd@gmail.com>**20100701142415
437 Ignore-this: c48c9609b8d36e43e033a7bea81d6f17
438]
439[doc typo
440Simon Marlow <marlowsd@gmail.com>**20100701142354
441 Ignore-this: 17a1fd703831c888975ff63fbfa3a9b2
442]
443[peekArray docs: remove mentions of "this version" and "previous version"
444Simon Marlow <marlowsd@gmail.com>**20100701125333
445 Ignore-this: 39a744874258670bd935ba9e38390939
446]
447[doc typo
448Simon Marlow <marlowsd@gmail.com>**20100701124154
449 Ignore-this: 98f5c286e38c2c34c96b05d5e8bc5ad9
450]
451[doc typo
452Simon Marlow <marlowsd@gmail.com>**20100701124128
453 Ignore-this: 10a4314ec7aed336701fc616fb574ebc
454]
455[doc typo
456Simon Marlow <marlowsd@gmail.com>**20100701123715
457 Ignore-this: c4909a7bf7163460ee5d32f58812041e
458]
459[doc wibble: Haskell 98 I/O Error -> 'IOError'
460Simon Marlow <marlowsd@gmail.com>**20100701123612
461 Ignore-this: bf373df781acbc575e4ffe3b7e6059ae
462]
463[doc typo
464Simon Marlow <marlowsd@gmail.com>**20100701123014
465 Ignore-this: 16aaccae48ef3101adf78ea5b0d5a8fd
466]
467[Haddock hacks to fix whitespace consistency
468Simon Marlow <marlowsd@gmail.com>**20100701121631
469 Ignore-this: 61c58dec52a31fd2d3f331a87d2f903f
470]
471[use '==' consistently rather than '->' in examples
472Simon Marlow <marlowsd@gmail.com>**20100701121616
473 Ignore-this: 472b0a05a85d34d9712186040e1636d9
474]
475[doc wibble: remove confusing mention of "Prelude"
476Simon Marlow <marlowsd@gmail.com>**20100701113308
477 Ignore-this: 232283d0096d01cd45e9b3c5c1e63a6d
478]
479[doc wibble: nonstrict -> non-strict
480Simon Marlow <marlowsd@gmail.com>**20100701113253
481 Ignore-this: 4264f0ab23a0835fc13c6e8601d6b743
482]
483[doc whitespace
484Simon Marlow <marlowsd@gmail.com>**20100701112242
485 Ignore-this: 777a95b1d1140c61d3ab95d5eb5809e7
486]
487[move the doc for 'Char' to its new home in ghc-prim:GHC.Types
488Simon Marlow <marlowsd@gmail.com>**20100629134150
489 Ignore-this: 7687db0077a29498349bfb4b44983985
490]
491[doc wibble
492Simon Marlow <marlowsd@gmail.com>**20100629122608
493 Ignore-this: 9a909e5d015332dc445bd9592e6e386d
494]
495[doc updates in System.IO
496Simon Marlow <marlowsd@gmail.com>**20100629122118
497 Ignore-this: 2257ec1cc4cdb8b7804cfa1f3cf32753
498]
499[doc wibble
500Simon Marlow <marlowsd@gmail.com>**20100625134858
501 Ignore-this: 64c50f29df6c389273b818918fe7033a
502]
503[doc wibbles
504Simon Marlow <marlowsd@gmail.com>**20100624154614
505 Ignore-this: b364aad53beea6e741fee2824459b6e8
506]
507[Fix haddock formatting
508Ian Lynagh <igloo@earth.li>**20100625222623]
509[Give nub's complexity in the haddock docs; fixes #4086
510Ian Lynagh <igloo@earth.li>**20100625222059]
511[correct docs for exitWith: only stdout/stderr are flushed, not all Handles
512Simon Marlow <marlowsd@gmail.com>**20100624130506
513 Ignore-this: 33a938dad8f0bc061572e2ec571cacc7
514]
515[fix docs for isSpace
516Simon Marlow <marlowsd@gmail.com>**20100624130444
517 Ignore-this: b35ff080dbb9833176f08e39dbd9ff6d
518]
519[make the hGetBuf/hPutBuf family work with non-FD Handles (#4144)
520Simon Marlow <marlowsd@gmail.com>**20100624130425
521 Ignore-this: 8200f0208a9b1b1cf4824f343d75819a
522]
523[nit in docs for accumArray
524Simon Marlow <marlowsd@gmail.com>**20100622121131
525 Ignore-this: c066a456c40907e767df10c3990f35ff
526]
527[add doc for the ExitCode type
528Simon Marlow <marlowsd@gmail.com>**20100622120930
529 Ignore-this: 99c34332be7f3565da844528b470054a
530]
531[remove extraneous info from docs for Array
532Simon Marlow <marlowsd@gmail.com>**20100622120921
533 Ignore-this: e2a3f5e84fc23eb7bae911f0680e805e
534]
535[add an INLINE to the list version of traverse, to enable fusion
536Simon Marlow <marlowsd@gmail.com>**20100608082531
537 Ignore-this: ea98cdc3308b406bb04c0f7a38c4424b
538]
539[Don't define the C localeEncoding on Windows
540Ian Lynagh <igloo@earth.li>**20100620202342
541 Ignore-this: c4992f6832a391b0cccc5a9b7d643976
542 (it causes warnings, and isn't used)
543]
544[add Applicative instance for Either (proposal #4095)
545Ross Paterson <ross@soi.city.ac.uk>**20100617225110
546 Ignore-this: 50262ec4700dc16efec5755be5b308c5
547 
548 This is not the only possible instance for Either, but this one is
549 compatible with the usual Monad instance.
550]
551[Use libcharset instead of nl_langinfo(CODESET) if possible.
552pho@cielonegro.org**20100519013112
553 Ignore-this: 4c1e278e022a3d276848afc1dcba4425
554 
555 nl_langinfo(CODESET) doesn't always return standardized variations of the encoding names. Use libcharset if possible, which is shipped together with GNU libiconv.
556]
557[Add a note about the interruptibility of throwTo.
558Simon Marlow <marlowsd@gmail.com>**20100615112720
559 Ignore-this: ae9fabe95310d7c364e95f7784793485
560]
561[docs: note that hGetBufNonBlocking isn't non-blocking on Windows
562Simon Marlow <marlowsd@gmail.com>**20100615112547
563 Ignore-this: 4f3e5213e142149affe08c5123d6efea
564]
565[don't depend on Prelude (#4122)
566Simon Marlow <marlowsd@gmail.com>**20100615105631
567 Ignore-this: 1a3fd49b103fe31cbb453f302c18767f
568]
569[Don't depend on Prelude (#4123)
570Simon Marlow <marlowsd@gmail.com>**20100615105401
571 Ignore-this: cc7616d85a1637bc7621b4f2bc181c0e
572]
573[bump version to 4.3.0.0, added instance MonadPlus STM
574Simon Marlow <marlowsd@gmail.com>**20100601144831
575 Ignore-this: 7c3cf7574499c4267372493f2636dc0
576]
577[Moved MonadPlus instance for STM from Control.Monad.STM to GHC.Conc to avoid an orphaned instance
578Bas van Dijk <v.dijk.bas@gmail.com>**20100516160651
579 Ignore-this: 651b852942b2fae2b93f996e39239b8f
580]
581[Added Applicative and Alternative instances for STM
582Bas van Dijk <v.dijk.bas@gmail.com>**20100516171756
583 Ignore-this: 567003bc4040bc97105cda4d31ebf04a
584]
585[expand Foldable instance for Array
586Ross Paterson <ross@soi.city.ac.uk>**20100602212154
587 Ignore-this: 9bd9e9666a9400431eb92352244fe7e7
588]
589[doc comment illustrating Foldable(foldr)
590Ross Paterson <ross@soi.city.ac.uk>**20100527150833
591 Ignore-this: 8f27d889379803f3ba86d6e928428f3c
592]
593[fix syntax in doc comments
594Ross Paterson <ross@soi.city.ac.uk>**20100527150757
595 Ignore-this: cb78da51d60ff6863dc395f1a892c103
596]
597[export hGetBufSome (#4046)
598Simon Marlow <marlowsd@gmail.com>**20100520093538
599 Ignore-this: f467fad9722e27edfad6b3dd75290e7b
600]
601[hWaitForInput: don't try to read from the device (#4078)
602Simon Marlow <marlowsd@gmail.com>**20100517133741
603 Ignore-this: 55ec33b03397380259b91e4ca62207a6
604 readTextDeviceNonBlocking is not non-blocking on Windows
605]
606[hSetEncoding: change the encoding on both read and write sides (#4066)
607Simon Marlow <marlowsd@gmail.com>**20100514124628
608 Ignore-this: 5b9e9caef06356d0296c584159709ebb
609]
610[Correct haddock formatting.
611Adam Vogt <vogt.adam@gmail.com>**20100423022103
612 Ignore-this: d2622339302048fda48080f7d5ce4a2f
613]
614[Fix for hGetBufSome
615Simon Marlow <marlowsd@gmail.com>**20100505135637
616 Ignore-this: 2019680f8fb223956cacfcf0d046f133
617]
618[improve the documentation for throwTo and killThread (#3884)
619Simon Marlow <marlowsd@gmail.com>**20100505135600
620 Ignore-this: ce881d96ddb729acb6ca09c779975e7d
621]
622[elaborate the docs for unsafePerformIO a bit
623Simon Marlow <marlowsd@gmail.com>**20100505101249
624 Ignore-this: 1cec3f67560b672c64c5a0dcf9a79eb7
625]
626[add Typeable instance
627Simon Marlow <marlowsd@gmail.com>**20100504152815
628 Ignore-this: 6d9cf9d62f0ef17fa459bf213a04098
629]
630[Add hGetBufSome, like hGetBuf but can return short reads
631Simon Marlow <marlowsd@gmail.com>**20100504152759
632 Ignore-this: 195c905b43f8d9505029364e2c5b18e
633]
634[Add swap (#3298)
635Simon Marlow <marlowsd@gmail.com>**20100504095339
636 Ignore-this: 13b007dc4594ce252997ec6fa0bbd976
637]
638[inline allocaArray0, to fix withCString benchmark
639Simon Marlow <marlowsd@gmail.com>**20100423124729
640 Ignore-this: 35c96816acc2f3aaf9dd29f7995fa6f0
641]
642[raise asynchronous exceptions asynchronously (#3997)
643Simon Marlow <marlowsd@gmail.com>**20100421094932
644 Ignore-this: 6d987d93d382c0f69c68c326312abd6b
645]
646[add NOINLINE pragmas for stdin/stdout/stderr
647Simon Marlow <marlowsd@gmail.com>**20100421082041
648 Ignore-this: 3fc130268ec786f28d945858d6690986
649]
650[INLINE alloca and malloc
651Simon Marlow <marlowsd@gmail.com>**20100419135333
652 Ignore-this: b218bd611f18721b1505a8c0b9e6a16a
653 See discussion on glasgow-haskell-users:
654   http://www.haskell.org/pipermail/glasgow-haskell-users/2010-April/018740.html
655]
656[Move comment closer to the offending line
657Matthias Kilian <kili@outback.escape.de>**20100419155421
658 Ignore-this: b34a1d7affd66f67d210df2377b585d9
659]
660[Ignore the return code of c_fcntl_write again
661Matthias Kilian <kili@outback.escape.de>**20100415140452
662 Ignore-this: 266d8ba02cc3cb79c85629b3528261c9
663 
664 The return code has been ignored in the past on purpose, because
665 O_NONBLOCK will fail on BSDs for some special files. This fixes the
666 problem mentioned in
667 http://www.haskell.org/pipermail/glasgow-haskell-users/2010-April/018698.html
668 
669]
670[Fix bitrot in IO debugging code
671Ian Lynagh <igloo@earth.li>**20100413134339
672 Also switched to using Haskell Bools (rather than CPP) to en/disable it,
673 so it shouldn't break again in the future.
674]
675[Tiny code tidy-up
676Ian Lynagh <igloo@earth.li>**20100413011147]
677[remove old/wrong comment
678Simon Marlow <marlowsd@gmail.com>**20100325161403
679 Ignore-this: e6e377d44af48c4162d17d55bdf3f821
680]
681[withThread: block asynchronous exceptions before installing exception handler.
682Bas van Dijk <v.dijk.bas@gmail.com>**20100329131624
683 Ignore-this: be5aeb47dbd73807b5f94df11afbb81c
684 Note that I don't unblock the given io computation. Because AFAICS
685 withThread is only called with 'waitFd' which only performs an FFI
686 call which can't receive asynchronous exceptions anyway.
687]
688[runInUnboundThread: block asynchronous exceptions before installing exception handler
689Bas van Dijk <v.dijk.bas@gmail.com>**20100329131549
690 Ignore-this: a00c5e32fe3981ff87bedd367a69051e
691]
692[fix the deprecation message (GHC.IO.Handle.Base -> GHC.IO.Handle)
693Simon Marlow <marlowsd@gmail.com>**20100330121137
694 Ignore-this: 4ca8500a01ac93454507aa8f9dd001f9
695]
696[Make SampleVar an abstract newtype
697Bas van Dijk <v.dijk.bas@gmail.com>**20100318200349
698 Ignore-this: 27939e2a064b75e71cb146117346be30
699]
700[Fix bugs regarding asynchronous exceptions and laziness in Control.Concurrent.SampleVar
701Bas van Dijk <v.dijk.bas@gmail.com>**20100318200104
702 Ignore-this: 7376b2a3afe155daf233a8f1ddc0a7a
703  - Block asynchronous exceptions at the right places
704  - Force thunks before putting them in a MVar
705]
706[Write the thunk 'next' to the MVar
707Bas van Dijk <v.dijk.bas@gmail.com>**20100319125951
708 Ignore-this: dd25636cf220131385ff2fd32493d456
709]
710[change to use STM, fixing 4 things
711Simon Marlow <marlowsd@gmail.com>**20100318104436
712 Ignore-this: 551d30280a7941c08f5c3b14576bdd70
713   1. there was no async exception protection
714   2. there was a space leak (now new value is strict)
715   3. using atomicModifyIORef would be slightly quicker, but can
716      suffer from adverse scheduling issues (see #3838)
717   4. also, the STM version is faster.
718]
719[Tweak docs
720Ian Lynagh <igloo@earth.li>**20100312214129]
721[Fixed dead links in documentation of forkIO
722Bas van Dijk <v.dijk.bas@gmail.com>**20100308222415
723 Ignore-this: 7deb8fd064c867fbede2a6b2e9da4f15
724]
725[Documentation fixes in Control.Exception
726Bas van Dijk <v.dijk.bas@gmail.com>**20100301220442
727 Ignore-this: 761fcba401cbd1f47276ddfc9b5b80f2
728]
729[Plug two race conditions that could lead to deadlocks in the IO manager
730Simon Marlow <marlowsd@gmail.com>**20100225120255
731 Ignore-this: e6983d6b953104d370278ab3e4617e8b
732]
733[FIX #3866: improve documentation of Data.Data.Constr
734jpm@cs.uu.nl**20100224125506
735 Ignore-this: 3818c5d8fee012a3cf322fb455b6e5dc
736]
737[UNDO: Handle NaN, -Infinity and Infinity in the toRational for Float/Double (#3676)
738Simon Marlow <marlowsd@gmail.com>**20100223101603
739 Ignore-this: 78becb2d39b3cd9a1a473a5811ca7d92
740]
741[Put the complexity in the length docs. Fixes trac #3680
742Ian Lynagh <igloo@earth.li>**20100221191425]
743[nhc98 should build Data.Functor.
744Malcolm.Wallace@cs.york.ac.uk**20100221163218]
745[Update the exitWith docs
746Ian Lynagh <igloo@earth.li>**20100213140004
747 Error pointed out by Volker Wysk <vw@volker-wysk.de>
748]
749[Handle NaN, -Infinity and Infinity in the toRational for Float/Double (#3676)
750Simon Marlow <marlowsd@gmail.com>**20100211101955
751 Ignore-this: 261415363303efca265e80290eac5f28
752]
753[For nhc98, import unsafeInterleaveIO rather than defining it here.
754Malcolm.Wallace@cs.york.ac.uk**20100204171021]
755[Stifle warning about unused return value
756benl@cse.unsw.edu.au**20100203025537]
757[fix #3832: use the locale encoding in openTempFile
758Simon Marlow <marlowsd@gmail.com>**20100120211830
759 Ignore-this: df4f778cc5fefb32290c798db722632c
760 Also while I was here fix an XXX: the Handle contained an
761 uninformative string like <fd: 4> for error messages rather than the
762 real file path.
763]
764[Fix the build: export void, so it doesn't give an unused binding warning
765Ian Lynagh <igloo@earth.li>**20100116174451]
766[hIsEOF: don't do any decoding (#3808)
767Simon Marlow <marlowsd@gmail.com>**20100112230317
768 Ignore-this: 6a384dd2d547ffe3ad3762920e5c1671
769]
770[Control.Monad: +void :: f a -> f ()
771gwern0@gmail.com**20100108214455
772 Ignore-this: 4dc07452315f2d1b4941903ff42fc45f
773 See http://hackage.haskell.org/trac/ghc/ticket/3292
774 Turns m a -> m (). Lets one call functions for their side-effects without
775 having to get rid of their return values with '>> return ()'. Very useful
776 in many contexts (parsing, IO etc.); particularly good for 'forkIO' and 'forM_',
777 as they demand return types of 'IO ()' though most interesting IO functions
778 return non-().
779]
780[Replace the implementation of mergesort with a 2x faster one.
781Malcolm.Wallace@cs.york.ac.uk**20091224152014
782 See ticket http://hackage.haskell.org/trac/ghc/ticket/2143.
783]
784[Restore previous Data.Typeable.typeOf*Default implementations for non-ghc.
785Malcolm.Wallace@cs.york.ac.uk**20091223142625
786 Not all compilers have ScopedTypeVariables.
787]
788[Add comments about double bounds-checking, and fast paths for rectangular arrays
789simonpj@microsoft.com**20091218165655
790 Ignore-this: ea0849419dc00927aba4bd410b1cc58d
791 
792 See Note [Double bounds-checking of index values] for the details.
793 
794 The fast paths omit the doubled checks for cases we know about
795]
796[Fix Trac #3245: memoising typeOf
797simonpj@microsoft.com**20091218155117
798 Ignore-this: 5a178a7f2222293c5ee0c3c43bd1b625
799 
800 The performance bug in #3245 was caused by computing the typeRep
801 once for each call of typeOf, rather than once for each dictionary
802 contruction.  (Computing TypeReps is reasonably expensive, because
803 of the hash-consing machinery.)
804 
805 This is readily fixed by putting the TypeRep construction outside
806 the lambda.  (Arguably GHC might have worked that out itself,
807 but it involves floating something between a type lambda and a
808 value lambda, which GHC doesn't currently do. If it happens a lot
809 we could fix that.)
810]
811[Mark 'index' as INLINE in GHC.Arr
812simonpj@microsoft.com**20091216170441
813 Ignore-this: a4df9d8acf496c8e0e9ce5a520509a2a
814 
815 This makes indexing much faster. See Trac #1216
816]
817[Comment the remaining orphan instance modules
818Ian Lynagh <igloo@earth.li>**20091206125021]
819[De-orphan Eq/Ord Float/Double
820Ian Lynagh <igloo@earth.li>**20091205181238]
821[Add comments to "OPTIONS_GHC -fno-warn-orphans" pragmas
822Ian Lynagh <igloo@earth.li>**20091205165854]
823[Data.Either.partitionEithers was insufficiently lazy.
824Malcolm.Wallace@cs.york.ac.uk**20091202032807
825 Ignore-this: 77e1b3288f66608c71458d8a91bcbe12
826 Spotted by Daniel Fischer.
827]
828[fix the docs regarding finalizer guarantees
829Simon Marlow <marlowsd@gmail.com>**20091130144409
830 Ignore-this: d1ab9532c74a002b8075ff60febcbe2d
831]
832[x86_64 requires more stack
833Malcolm.Wallace@cs.york.ac.uk**20091201033745]
834[check for size < 0 in mallocForeignPtrBytes and friends (#3514)
835Simon Marlow <marlowsd@gmail.com>**20091125143822
836 Ignore-this: 91077d01da2bbe1dfed5155e8b40da9
837]
838[hGetContents: close the handle properly on error
839Simon Marlow <marlowsd@gmail.com>**20091125123435
840 Ignore-this: bc37ff678acc6e547dc390285e056eb9
841 
842 When hGetContents caught an error it was closing the handle and then
843 throwing the exception, without updating the handle with the new
844 closed state.  This lead to a double-closed, which was the cause of
845 
846 *** glibc detected *** ./Setup: double free or corruption
847 
848 when iconv_close was called twice on the decoder.
849 
850 See http://hackage.haskell.org/trac/hackage/ticket/609
851]
852[Fix arities of mapFB and zipFB
853Roman Leshchinskiy <rl@cse.unsw.edu.au>**20091126232219
854 Ignore-this: c4e14cd0a92622549c86e67237a40865
855]
856[Remove an unnecessary -fno-warn-orphans flag
857Ian Lynagh <igloo@earth.li>**20091126123404]
858[Tweak layout to work with alternative layout rule
859Ian Lynagh <igloo@earth.li>**20091125232349]
860[Tweak layout to be accepted by the alternative layout rul
861Ian Lynagh <igloo@earth.li>**20091125194147]
862[Make sure zipWithFB has arity 2
863Roman Leshchinskiy <rl@cse.unsw.edu.au>**20091125010003
864 Ignore-this: 4cf60c55666f03d22a9f5a6e07f52d36
865 
866 It gets 2 arguments in the "zipWith" rule but its arity was higher and the new
867 inliner didn't inline it sometimes, for instance here:
868 
869 mpp ::  [Double] -> [Double] -> [Double] -> [Double] -> [Double]
870 mpp as bs cs ds = zipWith (*) (zipWith (+) as bs) (zipWith (+) cs ds)
871 
872 This was a regression vs. 6.10.
873]
874[Remove an old comment
875Ian Lynagh <igloo@earth.li>**20091124134647]
876[De-orphan the Eq/Ord Integer instances
877Ian Lynagh <igloo@earth.li>**20091124133639]
878[Whitespace only
879Ian Lynagh <igloo@earth.li>**20091124133421]
880[Derive some more instances, rather than writing them by hand
881Ian Lynagh <igloo@earth.li>**20091124011747]
882[We can now derive Ord ()
883Ian Lynagh <igloo@earth.li>**20091124011416]
884[De-orphan tuple Eq/Ord instances
885Ian Lynagh <igloo@earth.li>**20091123233343]
886[Control.Exception.Base no longer has any orphans
887Ian Lynagh <igloo@earth.li>**20091123224905]
888[De-orphan the MonadFix ST instance for GHC
889Ian Lynagh <igloo@earth.li>**20091123223544]
890[Rearrange the contents of Control.Monad.ST; no functionality changes
891Ian Lynagh <igloo@earth.li>**20091123222702]
892[De-orphan the Eq/Ord [a] instances
893Ian Lynagh <igloo@earth.li>**20091123215635]
894[De-orphan the Eq/Ord Char instances
895Ian Lynagh <igloo@earth.li>**20091123202253]
896[De-orphan the Eq/Ord Bool instances
897Ian Lynagh <igloo@earth.li>**20091123201817]
898[Move Eq/Ord Ordering instances to de-orphan them
899Ian Lynagh <igloo@earth.li>**20091123194310]
900[Remove ffi warnings for nhc98.
901Malcolm.Wallace@cs.york.ac.uk**20091123063743]
902[Second attempt to fix #1185 (forkProcess and -threaded)
903Simon Marlow <marlowsd@gmail.com>**20091111151915
904 Ignore-this: fa5f5d5e4e080d4b612a37244f937f9c
905 
906 Patch 2/2: first patch is to ghc
907 
908 This time without dynamic linker hacks, instead I've expanded the
909 existing rts/Globals.c to cache more CAFs, specifically those in
910 GHC.Conc.  We were already using this trick for signal handlers, I
911 should have realised before.
912 
913 It's still quite unsavoury, but we can do away with rts/Globals.c in
914 the future when we switch to a dynamically-linked GHCi.
915]
916[Rollback #1185 fix
917Simon Marlow <marlowsd@gmail.com>**20091106140629
918 Ignore-this: cd5667e8474e37e01ba26a1984274811
919 
920 rolling back:
921 
922 Tue Nov  3 16:05:40 GMT 2009  Simon Marlow <marlowsd@gmail.com>
923   * Fix #1185: restart the IO manager after fork()
924   
925   This is the libraries/base part of the patch; there is a corresponding
926   patch to GHC itself.
927   
928   The main change is that we now keep track of the IO manager's ThreadId
929   in a top-level MVar, and ensureIOManagerIsRunning checks whether a
930   previous IO manager thread is alive before starting one.  In the child
931   of fork(), we can hence call ensureIOManagerIsRunning to restart the
932   IO manager.
933 
934     M ./GHC/Conc.lhs -46 +44
935 
936 Wed Nov  4 17:49:45 GMT 2009  Ian Lynagh <igloo@earth.li>
937   * Fix the build on Windows
938 
939     M ./GHC/Conc.lhs -6 +4
940]
941[Fix the build on Windows
942Ian Lynagh <igloo@earth.li>**20091104174945]
943[Fix #1185: restart the IO manager after fork()
944Simon Marlow <marlowsd@gmail.com>**20091103160540
945 Ignore-this: 6dc05464f1500104554637f4759738cc
946 
947 This is the libraries/base part of the patch; there is a corresponding
948 patch to GHC itself.
949 
950 The main change is that we now keep track of the IO manager's ThreadId
951 in a top-level MVar, and ensureIOManagerIsRunning checks whether a
952 previous IO manager thread is alive before starting one.  In the child
953 of fork(), we can hence call ensureIOManagerIsRunning to restart the
954 IO manager.
955]
956[improve the documentation for throwErrnoIfRetry
957Simon Marlow <marlowsd@gmail.com>**20091016112404
958 Ignore-this: b77275cacf730e15757946027168f63e
959]
960[Don't inline unpackFoldrCString ever
961simonpj@microsoft.com**20091029135350
962 Ignore-this: 85d672649b1b776efc7e97500b05d4f9
963]
964[Inline more default methods
965simonpj@microsoft.com**20091029135330
966 Ignore-this: 289c44b0afd6d5631c2a4e0664275ca9
967 
968 Namely Monad: (>>)
969        Eq:    (==), (/=)
970        Num:   (-), negate
971        Real:  quot, rem, div, mod, recip, (/), truncate
972        Float: (**), logBase, sqrt, tan, tanh
973]
974[Move error messages out of INLINEd default methods
975simonpj@microsoft.com**20091029135118
976 Ignore-this: 9e35dc947f94827a3529eb53a41575fd
977 
978 No need to duplicate the error generation!
979]
980[Exploit now-working default-method INLINE pragmas for Data.Bits
981simonpj@microsoft.com**20091029135041
982 Ignore-this: 8adf225f31ca7a3181ee087e9e4fe535
983 
984 * Add INLINE pragmas to default methods for class Bits
985 
986 * Remove redundant instance methods elsewhere, now that
987   the default method will do the job
988]
989[Tidy up and comment imports
990simonpj@microsoft.com**20091029134414
991 Ignore-this: bf2be31035de975d8995e988933cc940
992]
993[Inline foldr and (.) when applied to two arguments not three
994simonpj@microsoft.com**20091029134335
995 Ignore-this: fccb6f3e90e15f44cb465814be85ede2
996 
997 The new INLINE story is (by design) arity-sensitive, so we must
998 put fewer argument on the LHS for foldr and (.)
999]
1000[dirUtils.c no longer available
1001Malcolm.Wallace@cs.york.ac.uk**20091013093833]
1002[Make hGetContents throw an exception if an error is encountered
1003Simon Marlow <marlowsd@gmail.com>**20091012152955
1004 Ignore-this: 9f7a7176193eab25c9daaacd9261f2de
1005 
1006 Strictly speaking this breaks Haskell 98 compatibility, which requires
1007 hGetContents to just end the lazy stream silently if an error is
1008 encountered.  However, for a few reasons we think it will make
1009 everyone's life a bit easier if we make this change
1010 
1011  1. Errors will be a lot more common in GHC 6.12.1, in the form
1012     of Unicode decoding errors.
1013 
1014  2. When Haskell 98 was designed, we didn't know how to throw
1015     exceptions from inside lazy I/O, but now we do.
1016 
1017  3. If anyone is actually relying on the previous behaviour, their
1018     code is arguably broken.
1019]
1020[Re-instate System.Console.Getopt for nhc98 builds.
1021Malcolm.Wallace@cs.york.ac.uk**20091013092843
1022 Although it was split out of base a while back, that change was
1023 reverted for ghc soon afterwards, but nhc98 never noticed.
1024]
1025[Roll back "Another instance of nhc98's strange import semantics."
1026Ian Lynagh <igloo@earth.li>**20091009185618
1027 Fri Oct  9 14:50:51 BST 2009  Malcolm.Wallace@cs.york.ac.uk
1028 GHC (correctly) warns about the unused import, which breaks the validate
1029 build.
1030]
1031[Roll back "Cope with nhc98's (occasionally-strange) import semantics"
1032Ian Lynagh <igloo@earth.li>**20091009184704
1033 Fri Oct  9 14:43:51 BST 2009  Malcolm.Wallace@cs.york.ac.uk
1034 GHC (correctly) warns about the unused import, which breaks the validate
1035 build.
1036]
1037[It seems that nhc98 needs defaulting in Data.Fixed.
1038Malcolm.Wallace@cs.york.ac.uk**20091009135242]
1039[Another instance of nhc98's strange import semantics.
1040Malcolm.Wallace@cs.york.ac.uk**20091009135051]
1041[Make Data.Functor compatible with non-GHC compilers.
1042Malcolm.Wallace@cs.york.ac.uk**20091009134821]
1043[Cope with nhc98's (occasionally-strange) import semantics.
1044Malcolm.Wallace@cs.york.ac.uk**20091009134351]
1045[Fix gratuitous breakage of nhc98 in System.IO.
1046Malcolm.Wallace@cs.york.ac.uk**20091009134001]
1047[Fix gratuitous breakage of nhc98 in Control.Exception.Base.
1048Malcolm.Wallace@cs.york.ac.uk**20091009133615]
1049[Fix gratuitous breakage of non-GHC in Data.Fixed.
1050Malcolm.Wallace@cs.york.ac.uk**20091009133330]
1051[Fix gratuitous breakage for non-GHC in Data.Bits.
1052Malcolm.Wallace@cs.york.ac.uk**20091009133257]
1053[Use UTF-32LE instead of UTF32LE
1054Simon Marlow <marlowsd@gmail.com>**20091006100207
1055 Ignore-this: 7f881e36543d250ef848c9f60d67655a
1056 The latter is not recognised by some iconv implementations.
1057]
1058[Strip any Byte Order Mark (BOM) from the front of decoded streams.
1059Ben.Lippmeier@anu.edu.au*-20090930084229
1060 Ignore-this: d0d0c3ae87b31d71ef1627c8e1786445
1061 When decoding to UTF-32, Solaris iconv inserts a BOM at the front
1062 of the stream, but Linux iconv doesn't.
1063]
1064[use UTF32BE/UTF32LE instead of UCS-4/UCS-4LE
1065Simon Marlow <marlowsd@gmail.com>**20091005101554
1066 Ignore-this: 2aef5e9bec421e714953b7aa1bdfc1b3
1067]
1068[Strip any Byte Order Mark (BOM) from the front of decoded streams.
1069Ben.Lippmeier@anu.edu.au**20090930084229
1070 Ignore-this: d0d0c3ae87b31d71ef1627c8e1786445
1071 When decoding to UTF-32, Solaris iconv inserts a BOM at the front
1072 of the stream, but Linux iconv doesn't.
1073]
1074[Add traceEvent :: String -> IO ()
1075Simon Marlow <marlowsd@gmail.com>**20090925141257
1076 Ignore-this: 8b1888bbf9682ffba13f815b6000e4b1
1077 For emitting an event via the RTS tracing framework
1078]
1079[Fix the error message when flushing the read buffer of a non-seekable Handle
1080Simon Marlow <marlowsd@gmail.com>**20090923090536
1081 Ignore-this: 4342026df93759d99480f4e13f80a492
1082]
1083[Fix #3534: No need to flush the byte buffer when setting binary mode
1084Simon Marlow <marlowsd@gmail.com>**20090923090445
1085 Ignore-this: 625817ed7ae2c12291eb993a99dc640a
1086]
1087[Use let !y = x in .. x .. instead of seq in $! and evaluate (#2273)
1088Simon Marlow <marlowsd@gmail.com>**20090916140454]
1089[make some Applicative functions into methods, and split off Data.Functor (proposal #3335)
1090Ross Paterson <ross@soi.city.ac.uk>**20090915173109
1091 Ignore-this: a0cff4de6dfdbcbd56a66101bc4855a9
1092 
1093 The following functions
1094 
1095     (<$) :: Functor f => a -> f b -> f a
1096     (*>) :: Applicative f => f a -> f b -> f b
1097     (<*) :: Applicative f => f a -> f b -> f a
1098     some :: Alternative f => f a -> f [a]
1099     many :: Alternative f => f a -> f [a]
1100 
1101 are moved into the corresponding classes, with the existing implementations
1102 as default definitions.  This gives people creating instances the option of
1103 defining specialized implementations of these functions, though they should
1104 be equivalent to the default definitions.
1105 
1106 Although (<$) is now a method of the Functor class, it is hidden in the
1107 re-export by the Prelude, Control.Monad and Monad.  The new module
1108 Data.Functor exposes the full class, plus the function (<$>).  These are
1109 also re-exported by Control.Applicative.
1110]
1111[On Windows, use the console code page for text file encoding/decoding.
1112Judah Jacobson <judah.jacobson@gmail.com>**20090913022126
1113 Ignore-this: 86c2f2db8ef92b751599795d3195187b
1114 
1115 We keep all of the code page tables in the module
1116 GHC.IO.Encoding.CodePage.Table.  That file was generated automatically
1117 by running codepages/MakeTable.hs; more details are in the comments at the
1118 start of that script.
1119 
1120 Storing the lookup tables adds about 40KB to each statically linked executable;
1121 this only increases the size of a "hello world" program by about 7%.
1122 
1123 Currently we do not support double-byte encodings (Chinese/Japanese/Korean), since
1124 including those codepages would increase the table size to 400KB.  It will be
1125 straightforward to implement them once the work on library DLLs is finished.
1126]
1127[Fix "init" docs: the input list need not be finite. Fixes trac #3465
1128Ian Lynagh <igloo@earth.li>**20090911210437]
1129[Bump base version to 4.2.0.0
1130Ian Lynagh <igloo@earth.li>**20090911153913]
1131[Address #3310
1132Simon Marlow <marlowsd@gmail.com>**20090830152850
1133 Ignore-this: 40c7f7c171ee299a83092fd360a952b7
1134 
1135  - Rename BlockedOnDeadMVar   -> BlockedIndefinitelyOnMVar
1136  - Rename BlockedIndefinitely -> BlockedIndefinitelyOnSTM
1137  - instance Show BlockedIndefinitelyOnMVar is now
1138      "blocked indefinitely in an MVar operation"
1139  - instance Show BlockedIndefinitelyOnSTM is now
1140      "blocked indefinitely in an STM transaction"
1141 
1142 clients using Control.OldException will be unaffected (the new
1143 exceptions are mapped to the old names).  However, for base4-compat
1144 we'll need to make a version of catch/try that does a similar
1145 mapping.
1146]
1147[Fix unicode conversion for MSB architectures
1148Ben.Lippmeier@anu.edu.au**20090830130028
1149 This fixes the SPARC/Solaris build.
1150]
1151[Fix #3441: detect errors in partial sequences
1152Simon Marlow <marlowsd@gmail.com>**20090830075909
1153 Ignore-this: d12a75d95e0cae5eb1555266810ec281
1154]
1155[Fix hWaitForInput
1156Simon Marlow <marlowsd@gmail.com>**20090827152116
1157 Ignore-this: 2550e911f1a4d4357a5aa8d1764238ce
1158 It was erroneously waiting when there were bytes to decode waiting in
1159 the byte buffer.
1160]
1161[fix debugging code
1162Simon Marlow <marlowsd@gmail.com>**20090827150628
1163 Ignore-this: e1c82fdc19a22e247cd69ff6fa11921d
1164]
1165[Allow for configurable iconv include and library locations.
1166Matthias Kilian <kili@outback.escape.de>**20090826154406
1167 Ignore-this: be95fab611a5534cf184b508964ed498
1168 This should help to fix the build on OpenBSD.
1169]
1170[typo in comment
1171Simon Marlow <marlowsd@gmail.com>**20090826085252
1172 Ignore-this: 1903ee0f354157a6ed3871c100f6b1b9
1173]
1174[un-hide some modules from the Haddock docs
1175Simon Marlow <marlowsd@gmail.com>**20090825152457
1176 Ignore-this: dce6606f93cf977fb24ebe99082dfa62
1177]
1178[Apply fix for #1548, from squadette@gmail.com
1179Simon Marlow <marlowsd@gmail.com>**20090819120700
1180 Ignore-this: 31c237c46a6445f588ed4b8c51bb6231
1181]
1182[improvements to Data.Fixed: instances for Typeable and Data, more predefined types
1183Ashley Yakeley <ashley@semantic.org>**20090812055058
1184 Ignore-this: feeece36d5632f02a05d137d2a39ab78
1185]
1186[Fix "Cabal check" warnings
1187Ian Lynagh <igloo@earth.li>**20090811215856]
1188[Add a GHC.Constants module; fixes trac #3094
1189Ian Lynagh <igloo@earth.li>**20090809183252]
1190[Apply proposal #3393
1191Ian Lynagh <igloo@earth.li>**20090809134717
1192 Add openTempFileWithDefaultPermissions and
1193 openBinaryTempFileWithDefaultPermissions.
1194]
1195[Add some more C wrappers; patch from Krister Walfridsson
1196Ian Lynagh <igloo@earth.li>**20090807200631
1197 Fixes 21 testsuite errors on NetBSD 5.99.
1198]
1199[Fixing configure for autoconf 2.64
1200Alexander Dunlap <alexander.dunlap@gmail.com>**20090805060748
1201 Ignore-this: 992ab91ae3d68c12dbb265776e33e243
1202]
1203[add INLINE toList
1204Ross Paterson <ross@soi.city.ac.uk>**20090806142853
1205 Ignore-this: aba16aabb17d5dca44f15d188945680e
1206 
1207 In anticipation of the fixing of #2353.
1208]
1209[fix a copyright
1210Simon Marlow <marlowsd@gmail.com>**20090805134045
1211 Ignore-this: b0ffbdd38fbba121e8bcba37c4082a60
1212]
1213[Tweak the BufferedIO class to enable a memory-mapped file implementation
1214Simon Marlow <marlowsd@gmail.com>**20090805134036
1215 Ignore-this: ec67d7a0a6d977438deaa342503f77e0
1216 We have to eliminate the assumption that an empty write buffer can be
1217 constructed by setting the buffer pointers to zero: this isn't
1218 necessarily the case when the buffer corresponds to a memory-mapped
1219 file, or other in-memory device implementation.
1220]
1221[Deprecate Control.OldException
1222Ian Lynagh <igloo@earth.li>**20090804143910]
1223[Windows build fix, following RTS tidyup
1224Simon Marlow <marlowsd@gmail.com>**20090803131121
1225 Ignore-this: ce862fb91c2b234211a8757f98690778
1226]
1227[Updates to follow the RTS tidyup
1228Simon Marlow <marlowsd@gmail.com>**20090801220743
1229 Ignore-this: 6e92412df93a66c12d75344053d5634
1230 C functions like isDoubleNaN moved here (primFloat.c)
1231]
1232[Add integer-simple as a build option
1233Ian Lynagh <igloo@earth.li>**20090722013151]
1234[Use shift[LR]Integer in the Bits Integer instance
1235Ian Lynagh <igloo@earth.li>**20090721222440]
1236[depend directly on integer-gmp, rather than indirecting through integer
1237Ian Lynagh <igloo@earth.li>**20090721185228]
1238[Move the instances of Functor and Monad IO to GHC.Base, to avoid orphans
1239Simon Marlow <marlowsd@gmail.com>**20090722102130
1240 Ignore-this: a7d85ac0025d559674249de0108dbcf4
1241]
1242[move "instance Exception Dynamic" so it isn't an orphan
1243Simon Marlow <marlowsd@gmail.com>**20090721093854
1244 Ignore-this: 5ede91ecfec2112c91b699d4de87cd02
1245]
1246[Improve the index checking for array accesses; fixes #2120 #2669
1247Ian Lynagh <igloo@earth.li>**20090719153228
1248 As well as checking that offset we are reading is actually inside the
1249 array, we now also check that it is "in range" as defined by the Ix
1250 instance. This fixes confusing behaviour (#2120) and improves some error
1251 messages (#2669).
1252]
1253[Make chr say what its argument was, if it's a bad argument
1254Ian Lynagh <igloo@earth.li>**20090718151049]
1255[remove unused warning
1256Simon Marlow <marlowsd@gmail.com>**20090715124416
1257 Ignore-this: 31f613654089d0f4a44363946087b41e
1258]
1259[warning fix: -fno-implicit-prelude -> -XNoImplicitPrelude
1260Simon Marlow <marlowsd@gmail.com>**20090715122839
1261 Ignore-this: dc8957249731d5bcb71c01899e5adf2b
1262]
1263[Add hGetEncoding :: Handle -> IO (Maybe TextEncoding)
1264Simon Marlow <marlowsd@gmail.com>**20090715122519
1265 Ignore-this: 14c3eff996db062da1199739781e4708
1266 as suggested during the discussion on the libraries list
1267]
1268[Add more documentation to mkTextEncoding
1269Simon Marlow <marlowsd@gmail.com>**20090715122414
1270 Ignore-this: 97253b2624267df3a246a18121e8ea81
1271 noting that "//IGNORE" and "//TRANSLIT" suffixes can be used with GNU
1272 iconv.
1273]
1274[Add the utf8_bom codec
1275Simon Marlow <marlowsd@gmail.com>**20090715122257
1276 Ignore-this: 1c9396cd805201fe873a39382ced79c7
1277 as suggested during the discussion on the libraries list.
1278]
1279[Export Unicode and newline functionality from System.IO; update Haddock docs
1280Simon Marlow <marlowsd@gmail.com>**20090713113104
1281 Ignore-this: c3f017a555335aa55d106253393f72e2
1282]
1283[add a comment about the non-workingness of CHARBUF_UTF16
1284Simon Marlow <marlowsd@gmail.com>**20090707124406
1285 Ignore-this: 98d00411b68d688b3b4cffc9507b1f35
1286]
1287[Fix build on Windows
1288Ian Lynagh <igloo@earth.li>**20090711004351]
1289[Fix some "warn-unused-do-bind" warnings where we want to ignore the value
1290Ian Lynagh <igloo@earth.li>**20090710204513]
1291[Use throwErrnoIfMinus1_ when calling getrusage
1292Ian Lynagh <igloo@earth.li>**20090710204221]
1293[Remove an unused import
1294Ian Lynagh <igloo@earth.li>**20090710153345]
1295[reportStackOverflow now returns IO ()
1296Ian Lynagh <igloo@earth.li>**20090710153257
1297 It used to do "return undefined" to return IO a.
1298]
1299[GHC.Conc.reportError now returns IO ()
1300Ian Lynagh <igloo@earth.li>**20090710152646
1301 It used to return IO a, by "return undefined".
1302]
1303[Fix some "warn-unused-do-bind" warnings where we want to ignore the value
1304Ian Lynagh <igloo@earth.li>**20090710152526]
1305[Minor SampleVar refactoring
1306Ian Lynagh <igloo@earth.li>**20090710151438]
1307[Fix "warn-unused-do-bind" warnings in GHC/IO/Handle/Text.hs
1308Ian Lynagh <igloo@earth.li>**20090710122905]
1309[Fix some "warn-unused-do-bind" warnings where we just want to ignore the result
1310Ian Lynagh <igloo@earth.li>**20090710005638]
1311[Use the result of writeCharBuf in GHC/IO/Encoding/Latin1.hs too
1312Ian Lynagh <igloo@earth.li>**20090710004032]
1313[Minor code tidyups in GHC.Conc
1314Ian Lynagh <igloo@earth.li>**20090710003801]
1315[Fix "warn-unused-do-bind" warning in GHC.Conc
1316Ian Lynagh <igloo@earth.li>**20090710003530
1317 If we fail to communicate with the IO manager then we print a warning
1318 using debugErrLn from the ghc-prim package.
1319]
1320[Fix "warn-unused-do-bind" warnings in System.Posix.Internals
1321Ian Lynagh <igloo@earth.li>**20090709164546]
1322[Fix "warn-unused-do-bind" warnings where we really do want to ignore the result
1323Ian Lynagh <igloo@earth.li>**20090709163912]
1324[Add back imports needed on Windows
1325Ian Lynagh <igloo@earth.li>**20090707181924]
1326[Remove unused imports
1327Ian Lynagh <igloo@earth.li>**20090707115810]
1328[Remove unused imports from base
1329simonpj@microsoft.com**20090706111842
1330 Ignore-this: f9b5f353e3bb820f787c56d615b28765
1331 
1332 These unused imports are detected by the new unused-import code
1333 
1334]
1335[Use the result of writeCharBuf
1336Simon Marlow <marlowsd@gmail.com>**20090706133303
1337 Ignore-this: 52288dd559bf4c4f313df6197091d935
1338   
1339 This only makes a difference when CHARBUF_UTF16 is in use, which it
1340 normally isn't.  I suspect CHARBUF_UTF16 doesn't currently work for
1341 other reasons (CHARBUF_UTF16 was an experiment before I wrote the
1342 GHC.IO.Encoding.UTF* codecs), but this patch at least makes it
1343 slightly closer to working.
1344]
1345[Remove some cruft from Data.HashTable
1346Ian Lynagh <igloo@earth.li>**20090706181630]
1347[Add 'eof' to Text.ParserCombinators.ReadP
1348simonpj@microsoft.com**20090706111801
1349 Ignore-this: 2aea7b848e00c894761bc4011adaa95d
1350 
1351 Add a ReadP parser that succeeds at the end of input. Very useful!
1352 
1353]
1354[Don't export CLDouble for GHC; fixes trac #2793
1355Ian Lynagh <igloo@earth.li>**20090705155120
1356 We never really supported CLDouble (it was a plain old double underneath),
1357 and pretending that we do does more harm than good.
1358]
1359[a byte between 0x80 and 0xBF is illegal immediately (#3341)
1360Simon Marlow <marlowsd@gmail.com>**20090702081415
1361 Ignore-this: dc19ef59a1a21118d5a7dd38aa2f611c
1362]
1363[avoid a warning
1364Simon Marlow <marlowsd@gmail.com>**20090630084134
1365 Ignore-this: c92a45ee216faf01327feae9fe06d6e2
1366]
1367[Add a wrapper for libiconv.
1368Matthias Kilian <kili@outback.escape.de>**20090629183634
1369 Ignore-this: 23c6047c0d71b745b495cc223574a47f
1370]
1371[#include <sys/times.h> if we have it (should fix build problems)
1372Simon Marlow <marlowsd@gmail.com>**20090629085351
1373 Ignore-this: a35e93b37ca9595c73460243180f4b9d
1374]
1375[set binary mode for existing FDs on Windows (fixes some GHCi test failures)
1376Simon Marlow <marlowsd@gmail.com>**20090626120522
1377 Ignore-this: 580cf636e9c77d8427aff6861d089481
1378]
1379[Move directory-related stuff to the unix package
1380Simon Marlow <marlowsd@gmail.com>**20090625120325
1381 Ignore-this: b997b3cbce0a46ca87ad825bbdc0a411
1382 now that it isn't used on Windows any more.
1383]
1384[TAG 2009-06-25
1385Ian Lynagh <igloo@earth.li>**20090625160056]
1386Patch bundle hash:
1387e2c2f123895a555b7fa56c769daf574727d2dcfe