Ticket #1789: trac1789

File trac1789, 63.2 KB (added by SamAnklesaria, 5 years ago)

patch bundle

Line 
11 patch for repository http://darcs.haskell.org/ghc:
2
3Fri Jun 18 10:06:49 CDT 2010  [email protected]
4  * trac #1789 (warnings for missing import lists)
5
6New patches:
7
8[trac #1789 (warnings for missing import lists)
9[email protected]**20100618150649
10 Ignore-this: b0b0b1e048fbca0817c1e6fade1153fa
11] {
12hunk ./compiler/main/DynFlags.hs 168
13    | Opt_WarnIncompletePatterns
14    | Opt_WarnIncompletePatternsRecUpd
15    | Opt_WarnMissingFields
16+   | Opt_WarnMissingImportList
17    | Opt_WarnMissingMethods
18    | Opt_WarnMissingSigs
19    | Opt_WarnNameShadowing
20hunk ./compiler/main/DynFlags.hs 1500
21   ( "warn-incomplete-patterns",         Opt_WarnIncompletePatterns, const Supported ),
22   ( "warn-incomplete-record-updates",   Opt_WarnIncompletePatternsRecUpd, const Supported ),
23   ( "warn-missing-fields",              Opt_WarnMissingFields, const Supported ),
24+  ( "warn-missing-import-lists",        Opt_WarnMissingImportList, const Supported ),
25   ( "warn-missing-methods",             Opt_WarnMissingMethods, const Supported ),
26   ( "warn-missing-signatures",          Opt_WarnMissingSigs, const Supported ),
27   ( "warn-name-shadowing",              Opt_WarnNameShadowing, const Supported ),
28hunk ./compiler/rename/RnNames.lhs 73
29             when (notNull prel_imports) $ addWarn (implicitPreludeWarn)
30           )
31 
32-         stuff1 <- mapM (rnImportDecl this_mod) (prel_imports ++ ordinary)
33-         stuff2 <- mapM (rnImportDecl this_mod) source
34+         stuff1 <- mapM (rnImportDecl this_mod implicit_prelude) (prel_imports ++ ordinary)
35+         stuff2 <- mapM (rnImportDecl this_mod implicit_prelude) source
36          let (decls, rdr_env, imp_avails,hpc_usage) = combine (stuff1 ++ stuff2)
37          return (decls, rdr_env, imp_avails,hpc_usage)
38 
39hunk ./compiler/rename/RnNames.lhs 89
40                    imp_avails1 `plusImportAvails` imp_avails2,
41                   hpc_usage1 || hpc_usage2)
42 
43-rnImportDecl  :: Module
44+rnImportDecl  :: Module -> Bool
45              -> LImportDecl RdrName
46              -> RnM (LImportDecl Name, GlobalRdrEnv, ImportAvails,AnyHpcUsage)
47 
48hunk ./compiler/rename/RnNames.lhs 93
49-rnImportDecl this_mod (L loc (ImportDecl loc_imp_mod_name mb_pkg want_boot
50+rnImportDecl this_mod implicit_prelude (L loc (ImportDecl loc_imp_mod_name mb_pkg want_boot
51                                          qual_only as_mod imp_details))
52   = setSrcSpan loc $ do
53 
54hunk ./compiler/rename/RnNames.lhs 107
55        imp_mod_name = unLoc loc_imp_mod_name
56        doc = ppr imp_mod_name <+> ptext (sLit "is directly imported")
57 
58+    case imp_details of
59+        (Just _) -> return ()
60+        Nothing -> when (not (implicit_prelude && imp_mod_name == pRELUDE_NAME)) $
61+            ifOptM Opt_WarnMissingImportList (addWarn (missingImportListWarn imp_mod_name))
62+
63     iface <- loadSrcInterface doc imp_mod_name want_boot mb_pkg
64 
65        -- Compiler sanity check: if the import didn't say
66hunk ./compiler/rename/RnNames.lhs 1455
67 nullModuleExport mod
68   = ptext (sLit "The export item `module") <+> ppr mod <> ptext (sLit "' exports nothing")
69 
70+missingImportListWarn :: ModuleName -> SDoc
71+missingImportListWarn mod
72+  = ptext (sLit "The module") <+> quotes (ppr mod) <+> ptext (sLit "is missing an import list")
73+
74 moduleWarn :: ModuleName -> WarningTxt -> SDoc
75 moduleWarn mod (WarningTxt txt)
76   = sep [ ptext (sLit "Module") <+> quotes (ppr mod) <> ptext (sLit ":"),
77hunk ./docs/users_guide/flags.xml 1132
78          </row>
79 
80          <row>
81+           <entry><option>-fwarn-missing-import-lists</option></entry>
82+           <entry>warn when explicit imports lack an import list</entry>
83+           <entry>dynamic</entry>
84+           <entry><option>-fnowarn-missing-import-lists</option></entry>
85+         </row>
86+
87+         <row>
88            <entry><option>-fwarn-missing-methods</option></entry>
89            <entry>warn when class methods are undefined</entry>
90            <entry>dynamic</entry>
91}
92
93Context:
94
95[Fix bindisttest Makefile
96Ian Lynagh <[email protected]>**20100616205611
97 Ignore-this: 39cd352152422f378572fc3859c5a377
98]
99[Remove some more unused make variables
100Ian Lynagh <[email protected]>**20100616180519]
101[Convert some more variable names to FOO_CMD, for consistency
102Ian Lynagh <[email protected]>**20100616175916]
103[Rename some variables from FOO to FOO_CMD
104Ian Lynagh <[email protected]>**20100616161108
105 This fixes a problem with commands like gzip, where if $GZIP is exported
106 in the environment, then when make runs a command it'll put the Makefile
107 variable's value in the environment. But gzip treats $GZIP as arguments
108 for itself, so when we run gzip it thinks we're giving it "gzip" as an
109 argument.
110]
111[Make the "show" target work anywhere in the build tree
112Ian Lynagh <[email protected]>**20100616122910
113 Ignore-this: 299d40cbe16112accd9f14e56fa12158
114]
115[Change ghc-pwd's license to a string Cabal recognises
116Ian Lynagh <[email protected]>**20100615204015
117 Ignore-this: c935b6ad7f605aab0168997a90b40fc6
118]
119[fix warning
120Simon Marlow <[email protected]>**20100604205933
121 Ignore-this: 2aaa4ed6a8b9ae1e39adc4696aaf14a3
122]
123[--install-signal-handles=no does not affect the timer signal (#1908)
124Simon Marlow <[email protected]>**20100527214627
125 Ignore-this: b0c51f1abdb159dc360662485095a11a
126]
127[Small optimisation: allocate nursery blocks contiguously
128Simon Marlow <[email protected]>**20100509194928
129 Ignore-this: e650e99e9ea9493d2efb245d565beef4
130 This lets automatic prefetching work better, for a tiny performance boost
131]
132[fix -fforce-recomp setting: module is PrimOp, not PrimOps
133Simon Marlow <[email protected]>**20100507084507
134 Ignore-this: f76e0d9b643682ec0e8fb7d91afdea68
135]
136[it should be an error to use relative directories (#4134)
137Simon Marlow <[email protected]>**20100615151740
138 Ignore-this: 2068021701832e018ca41b22877921d5
139]
140[missing include-dirs or library-dirs is only a warning now (#4104)
141Simon Marlow <[email protected]>**20100615151702
142 Ignore-this: e3114123cef147bbd28ccb64581a1afb
143]
144[fix #3822: desugaring case command in arrow notation
145Ross Paterson <[email protected]>**20100615225110
146 Ignore-this: 477d6c460b4174b94b4cd113fa5b9d19
147 
148 Get the set of free variables from the generated case expression:
149 includes variables in the guards and decls that were missed before,
150 and is also a bit simpler.
151]
152[Deprecate the -fvia-C flag; trac #3232
153Ian Lynagh <[email protected]>**20100615151836
154 Ignore-this: c2452b2648bf7e44546465c1b964fce
155]
156[Avoid using the new ~~ perl operator in the mangler
157Ian Lynagh <[email protected]>**20100615151236
158 Ignore-this: 709a7ba4e514b1596841b3ba7e5c6cc
159]
160[stmAddInvariantToCheck: add missing init of invariant->lock (#4057)
161Simon Marlow <[email protected]>**20100615123643
162 Ignore-this: 3b132547fa934cecf71a846db2a5f70e
163]
164[Add new LLVM code generator to GHC. (Version 2)
165David Terei <[email protected]>**20100615094714
166 Ignore-this: 4dd2fe5854b64a3f0339d484fd5c238
167 
168 This was done as part of an honours thesis at UNSW, the paper describing the
169 work and results can be found at:
170 
171 http://www.cse.unsw.edu.au/~pls/thesis/davidt-thesis.pdf
172 
173 A Homepage for the backend can be found at:
174 
175 http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/Backends/LLVM
176 
177 Quick summary of performance is that for the 'nofib' benchmark suite, runtimes
178 are within 5% slower than the NCG and generally better than the C code
179 generator.  For some code though, such as the DPH projects benchmark, the LLVM
180 code generator outperforms the NCG and C code generator by about a 25%
181 reduction in run times.
182 
183]
184[Fix Trac #4127: build GlobalRdrEnv in GHCi correctly
185[email protected]**20100615070626
186 Ignore-this: d907e3bfa7882878cea0af172aaf6e84
187 
188 GHCi was building its GlobalRdrEnv wrongly, so that the
189 gre_par field was bogus.  That in turn fooled the renamer.
190 The fix is easy: use the right function!  Namely, call
191 RnNames.gresFromAvail rather than availsToNameSet.
192]
193[Comments, and improvement to pretty-printing of HsGroup
194[email protected]**20100615070409
195 Ignore-this: ec8358f2485370b20226a97ec84e9024
196]
197[Don't reverse bindings in rnMethodBinds (fix Trac #4126)
198[email protected]**20100614163935
199 Ignore-this: a6ffbb5af6f51b142ed0aeae8ee5e3a9
200]
201[Fix Trac #4120: generate a proper coercion when unifying forall types
202[email protected]**20100614134311
203 Ignore-this: 601592bb505305f1954cbe730f168da4
204 
205 This was just a blatant omission, which hasn't come up before.
206 Easily fixed, happily.
207]
208[Use mkFunTy to ensure that invariants are respected
209[email protected]**20100614134159
210 Ignore-this: 67dcada7a4e8d9927581cd77af71b6f
211]
212[Remove redundant debug code
213[email protected]**20100601154151
214 Ignore-this: e6ff11c04c631cf6aac73788cbcf02b5
215]
216[Fix Trac #4099: better error message for type functions
217[email protected]**20100531140413
218 Ignore-this: 3f53ca98cf770577818b9c0937482577
219 
220 Now we only want about "T is a type function and might not be
221 injective" when matchin (T x) against (T y), which is the case
222 that is really confusing.
223]
224[Gruesome fix in CorePrep to fix embarassing Trac #4121
225[email protected]**20100614132726
226 Ignore-this: fe82d15474afaac3e6133adfd7a7e055
227 
228 This is a long-lurking bug that has been flushed into
229 the open by other arity-related changes.  There's a
230 long comment
231 
232      Note [CafInfo and floating]
233 
234 to explain. 
235 
236 I really hate the contortions we have to do through to keep correct
237 CafRef information on top-level binders.  The Right Thing, I believe,
238 is to compute CAF and arity information later, and merge it into the
239 interface-file information when the latter is generated.
240 
241 But for now, this hackily fixes the problem.
242]
243[Fix a bug in CorePrep that meant output invariants not satisfied
244[email protected]**20100531150013
245 Ignore-this: d34eb36d8877d3caf1cf2b20de426abd
246 
247 In cpePair I did things in the wrong order so that something that
248 should have been a CprRhs wasn't.  Result: a crash in CoreToStg.
249 Fix is easy, and I added more informative type signatures too.
250]
251[Robustify the treatement of DFunUnfolding
252[email protected]**20100531145332
253 Ignore-this: 8f5506ada4d89f6ab8ad1e8c3ffb09ba
254 
255 See Note [DFun unfoldings] in CoreSyn.  The issue here is that
256 you can't tell how many dictionary arguments a DFun needs just
257 from looking at the Arity of the DFun Id: if the dictionary is
258 represented by a newtype the arity might include the dictionary
259 and value arguments of the (single) method.
260 
261 So we need to record the number of arguments need by the DFun
262 in the DFunUnfolding itself.  Details in
263    Note [DFun unfoldings] in CoreSyn
264]
265[Fix spelling in comment
266[email protected]**20100614132259
267 Ignore-this: bbf0d55f2e5f10ef9c74592c12f9201c
268]
269[Update docs on view patterns
270[email protected]**20100614074801
271 Ignore-this: 8617b9078800d4942d71f142a5b6c831
272]
273[Fix printing of splices; part of #4124
274Ian Lynagh <[email protected]>**20100613154838
275 Just putting parens around non-atomic expressions isn't sufficient
276 for splices, as only the $x and $(e) forms are valid input.
277]
278[In ghci, catch IO exceptions when calling canonicalizePath
279Ian Lynagh <[email protected]>**20100613134627
280 We now get an exception if the path doesn't exist
281]
282[Whitespace only
283Ian Lynagh <[email protected]>**20100612213119]
284[Whitespace only
285Ian Lynagh <[email protected]>**20100612165450]
286[Update ghci example output in user guide; patch from YitzGale in #4111
287Ian Lynagh <[email protected]>**20100612162250]
288[Fix #4131 missing UNTAG_CLOSURE in messageBlackHole()
289[email protected]**20100611044614]
290[messageBlackHole: fix deadlock bug caused by a missing 'volatile'
291Simon Marlow <[email protected]>**20100610080636
292 Ignore-this: 3cda3054bb45408aa9bd2d794b69c938
293]
294[Pass --no-tmp-comp-dir to Haddock (see comment)
295Simon Marlow <[email protected]>**20100604083214
296 Ignore-this: bfa4d74038637bd149f4d878b4eb8a87
297]
298[Track changes to DPH libs
299Roman Leshchinskiy <[email protected]>**20100607052903
300 Ignore-this: 4dbc3f8418af3e74b3fc4f9a9dfe7764
301]
302[Track changes to DPH libs
303Roman Leshchinskiy <[email protected]>**20100607012642
304 Ignore-this: 5d4e498171a3c57ab02621bfaea82cff
305]
306[In ghc-pkg, send warnings to stderr
307Ian Lynagh <[email protected]>**20100606161726
308 Ignore-this: 56927d13b5e1c1ce2752734f0f9b665b
309]
310[Re-add newlines to enable layout for multi-line input.
311Ian Lynagh <[email protected]>**20100602180737
312 Patch from Adam Vogt <[email protected]>
313 Partial fix for #3984
314]
315[Don't use unnecessary parens when printing types (Fix Trac 4107)
316[email protected]**20100604110143
317 Ignore-this: a833714ab13013c4345b222f4e87db1d
318 
319    f :: Eq a => a -> a
320 rather than
321    f :: (Eq a) => a -> a
322]
323[Track DPH library changes
324Roman Leshchinskiy <[email protected]>**20100604005728
325 Ignore-this: 32bc2fbea6ad975e89545d4c42fd7c30
326]
327[fix --source-entity option passed to Haddock: we needed to escape a #
328Simon Marlow <[email protected]>**20100603125459
329 Ignore-this: d52ae6188b510c482bcebb23f0e553ae
330]
331[__stg_EAGER_BLACKHOLE_INFO -> __stg_EAGER_BLACKHOLE_info (#4106)
332Simon Marlow <[email protected]>**20100602091419
333 Ignore-this: 293315ac8f86fd366b8d61992ecc7961
334]
335[Add xhtml package (a new dependency of Haddock; not installed/shipped)
336Simon Marlow <[email protected]>**20100602090101
337 Ignore-this: af0ac8b91abe98f7fdb624ea0a4dee20
338]
339[Use UserInterrupt rather than our own Interrupted exception (#4100)
340Simon Marlow <[email protected]>**20100602082345
341 Ignore-this: 1909acf2f452593138b9f85024711714
342]
343[Add the global package DB to ghc --info (#4103)
344Simon Marlow <[email protected]>**20100602082233
345 Ignore-this: fd5c0e207e70eb0f62606c45dc5b8124
346]
347[rts/sm/GC.c: resize_generations(): Remove unneeded check of number of generations.
348Marco Túlio Gontijo e Silva <[email protected]>**20100528115612
349 Ignore-this: 6f1bea62917c01c7adac636146132c97
350 
351 This "if" is inside another "if" which checks for RtsFlags.GcFlags.generations
352 > 1, so testing this again is redundant, assuming the number of generations
353 won't change during program execution.
354]
355[rts/sm/BlockAlloc.c: Small comment correction.
356Marco Túlio Gontijo e Silva <[email protected]>**20100526205839
357 Ignore-this: bd2fcd4597cc872d80b0e2eeb1c3998a
358]
359[rts/sm/GC.c: Annotate constants.
360Marco Túlio Gontijo e Silva <[email protected]>**20100526205707
361 Ignore-this: f232edb89383564d759ed890a18f602f
362]
363[includes/rts/storage/GC.h: generation_: n_words: Improve comment.
364Marco Túlio Gontijo e Silva <[email protected]>**20100526204615
365 Ignore-this: f5d5feefa8f7b552303978f1804fea23
366]
367[Add PPC_RELOC_LOCAL_SECTDIFF support; patch from PHO in #3654
368Ian Lynagh <[email protected]>**20100601204211
369 Ignore-this: 51293b7041cdce3ce7619ef11cf7ceb
370]
371[powerpc-apple-darwin now supports shared libs
372Ian Lynagh <[email protected]>**20100601173325]
373[PIC support for PowerPC
374[email protected]**20100508143900
375 Ignore-this: 3673859a305398c4acae3f4d7c997615
376 
377 PPC.CodeGen.getRegister was not properly handling PicBaseReg.
378 It seems working with this patch, but I'm not sure this change is correct.
379]
380[Vectoriser: only treat a function as scalar if it actually computes something
381Roman Leshchinskiy <[email protected]>**20100601045630
382 Ignore-this: e5d99a6ddb62052e3520094a5af47552
383]
384[Add a release notes file for 6.14.1
385Ian Lynagh <[email protected]>**20100530171117
386 Ignore-this: 1941e6d3d1f4051b69ca2f17a1cf84d6
387]
388[Check dblatex actually creates the files we tell it to
389Ian Lynagh <[email protected]>**20100530171043
390 Ignore-this: ccc72caea2313be05cbac59bb54c0603
391 If it fails, it still exits successfully.
392]
393[Add darwin to the list of OSes for which we use mmap
394Ian Lynagh <[email protected]>**20100529145016
395 Ignore-this: a86d12a3334aaaafc86f7af9dbb0a7ae
396 Patch from Barney Stratford
397]
398[Simplify the CPP logic in rts/Linker.c
399Ian Lynagh <[email protected]>**20100529144929
400 Ignore-this: 1288f5b752cc1ab8b1c90cfd0ecfdf68
401]
402[Fix validate on OS X
403Ian Lynagh <[email protected]>**20100529154726]
404[OS X x86_64 fix from Barney Stratford
405Ian Lynagh <[email protected]>**20100529122440]
406[OS X 64 installer fixes from Barney Stratford
407Ian Lynagh <[email protected]>**20100528234935]
408[fix warning
409Simon Marlow <[email protected]>**20100525155812
410 Ignore-this: f34eee3fe3d89579fd8d381c91ced750
411]
412[Fix doc bugs (#4071)
413Simon Marlow <[email protected]>**20100525155728
414 Ignore-this: aa25be196de567de360075022a1942f7
415]
416[Make sparks into weak pointers (#2185)
417Simon Marlow <[email protected]>**20100525150435
418 Ignore-this: feea0bb5006007b82c932bc3006124d7
419 The new strategies library (parallel-2.0+, preferably 2.2+) is now
420 required for parallel programming, otherwise parallelism will be lost.
421]
422[If you say 'make' or 'make stage=2' here, pretend we're in the ghc dir
423Simon Marlow <[email protected]>**20100525085301
424 Ignore-this: 78b740337aa460915c812cbbcdae5321
425]
426[Another attempt to get these #defines right
427Simon Marlow <[email protected]>**20100525154313
428 Ignore-this: 460ca0c47d81cd25eae6542114f67899
429 Apparently on Solaris it is an error to omit _ISOC99_SOURCE when using
430 _POSIX_C_SOURCE==200112L.
431]
432[Add configure flags for the location of GMP includes/library; fixes #4022
433Ian Lynagh <[email protected]>**20100525221616
434 Ignore-this: fc3060caf995d07274ec975eeefbdf3e
435]
436[Refactor pretty printing of TyThings to fix Trac #4015
437[email protected]**20100525153126
438 Ignore-this: 8f15053b7554f62caa84201d2e4976d2
439]
440[When haddocking, we need the dependencies to have been built
441Ian Lynagh <[email protected]>**20100525145830
442 as haddock loads the .hi files with the GHC API.
443]
444[Fix profiling output; spotted by jlouis
445Ian Lynagh <[email protected]>**20100525111217
446 We were outputing the number of words allocated in a column titled "bytes".
447]
448[Improve printing of TyThings; fixes Trac #4087
449[email protected]**20100525114045
450 Ignore-this: da2a757a533454bba80b9b77cc5a771
451]
452[Spelling in comments
453[email protected]**20100525114001
454 Ignore-this: 270f3da655e526cf04e27db7a01e29c0
455]
456[Refactor (again) the handling of default methods
457[email protected]**20100525113910
458 Ignore-this: 6686f6cdb878d57abf6b49fec64fcbb1
459 
460 This patch fixes Trac #4056, by
461 
462  a) tidying up the treatment of default method names
463  b) removing the 'module' argument to newTopSrcBinder
464 
465 The details aren't that interesting, but the result
466 is much tidier. The original bug was a 'nameModule' panic,
467 caused by trying to find the module of a top-level name.
468 But TH quotes generate Internal top-level names that don't
469 have a module, and that is generally a good thing. 
470 
471 Fixing that in turn led to the default-method refactoring,
472 which also makes the Name for a default method be handled
473 in the same way as other derived names, generated in BuildTyCl
474 via a call newImplicitBinder.  Hurrah.
475]
476[Don't do SpecConstr on NOINLINE things (Trac #4064)
477[email protected]**20100525112807
478 Ignore-this: 452be0a2cef0042fb67275c2827b5f72
479 
480 Since the RULE from specialising gets the same Activation as
481 the inlining for the Id itself there's no point in specialising
482 a NOINLINE thing, because the rule will be permanently switched
483 off.
484 
485 See Note [Transfer activation] in SpecConstr
486 and Note [Auto-specialisation and RULES] in Specialise.
487]
488[Change our #defines to work on FreeBSD too
489Simon Marlow <[email protected]>**20100524105828
490 Ignore-this: b23ede46211e67859206c0ec57d6a86f
491 With glibc, things like _POSIX_C_SOURCE and _ISOC99_SOURCE are
492 additive, but on FreeBSD they are mutually exclusive.  However, it
493 turns out we only need to define _POSIX_C_SOURCE and _XOPEN_SOURCE to
494 get all the C99 stuff we need too, so there's no need for any #ifdefs.
495 
496 Submitted by: Gabor PALI <[email protected]>
497]
498[Add a missing UNTAG_CLOSURE, causing bus errors on Sparc
499Simon Marlow <[email protected]>**20100524105547
500 Ignore-this: a590b5391d6f05d50c8c088456c3c166
501 We just about got away with this on x86 which isn't
502 alignment-sensitive.  The result of the memory load is compared
503 against a few different values, but there is a fallback case that
504 happened to be the right thing when the pointer was tagged.  A good
505 bug to find, nonetheless.
506]
507[Add wiki links
508Simon Marlow <[email protected]>**20100520095953
509 Ignore-this: c22f126cde166e6207922b2eb51d29e3
510]
511[the 'stage=0' trick to disable all compiler builds stopped working; fix it
512Simon Marlow <[email protected]>**20100520104455
513 Ignore-this: bb6fae9056471612c8dbf06916188c33
514]
515[Comments and formatting only
516[email protected]**20100524014021
517 Ignore-this: 64579c38154728b632e358bec751cc0b
518]
519[Core prettyprinter fixes. Patch from Tim Chevalier. Fixes #4085
520Ian Lynagh <[email protected]>**20100522225048]
521[Correct install-name for dynamic Darwin rts
522[email protected]**20100508151155
523 Ignore-this: 6d31716c8c113dcb46e9cb925c4201df
524]
525[Fix the RTS debug_p build
526Ian Lynagh <[email protected]>**20100522163127]
527[Unset $CFLAGS for "GNU non-executable stack" configure test; fixes #3889
528Ian Lynagh <[email protected]>**20100521165005
529 With gcc 4.4 we get
530     Error: can't resolve `.note.GNU-stack' {.note.GNU-stack section} - `.Ltext0' {.text section}
531 when running gcc with the -g flag. To work around this we unset
532 CFLAGS when running the test.
533]
534[Don't run "set -o igncr" before configuring libffi
535Ian Lynagh <[email protected]>**20100520162918
536 Ignore-this: 489fa94df23f2adf4ff63c8ede2c0794
537 It used to make the build work on cygwin, but now it breaks it instead:
538     config.status: creating include/Makefile
539     gawk: ./confLqjohp/subs.awk:1: BEGIN {\r
540     gawk: ./confLqjohp/subs.awk:1: ^ backslash not last character on line
541     config.status: error: could not create include/Makefile
542     make[2]: *** [libffi/stamp.ffi.configure-shared] Error 1
543     make[1]: *** [all] Error 2
544]
545[Stop passing -Wl,-macosx_version_min to gcc
546Ian Lynagh <[email protected]>**20100520154003
547 Fixes a build failure on OS X 10.6. When linking
548     rts/dist/build/libHSrts-ghc6.13.20100519.dylib
549 we got
550     ld: symbol dyld_stub_binding_helper not defined (usually in crt1.o/dylib1.o/bundle1.o)
551     collect2: ld returned 1 exit status
552]
553[Fix build on FreeBSD; patch from Gabor PALI
554Ian Lynagh <[email protected]>**20100519140552]
555[Fix package shadowing order (#4072)
556Simon Marlow <[email protected]>**20100519104617
557 Ignore-this: 26ea5e4bb5dff18618b807a54c7d6ebb
558 
559 Later packages are supposed to shadow earlier ones in the stack,
560 unless the ordering is overriden with -package-id flags.
561 Unfortunately an earlier fix for something else had sorted the list of
562 packages so that it was in lexicographic order by installedPackageId,
563 and sadly our test (cabal/shadow) didn't pick this up because the
564 lexicographic ordering happened to work for the test.  I've now fixed
565 the test so it tries both orderings.
566]
567[Set more env variables when configuring libffi
568Ian Lynagh <[email protected]>**20100518185014
569 We now tell it where to find ld, nm and ar
570]
571[Set the location of ar to be the in-tree ar on Windows
572Ian Lynagh <[email protected]>**20100518181556]
573[Change another / to </> to avoid building paths containing \/
574Ian Lynagh <[email protected]>**20100518172015
575 This will hopefully fix #2889.
576]
577[Fix #4074 (I hope).
578Simon Marlow <[email protected]>**20100518113214
579 Ignore-this: 73cd70f5bc6f5add5247b61985c03fc1
580 
581 1. allow multiple threads to call startTimer()/stopTimer() pairs
582 2. disable the timer around fork() in forkProcess()
583 
584 A corresponding change to the process package is required.
585]
586[we don't have a gcc-lib in LIB_DIR any more
587Simon Marlow <[email protected]>**20100401102351
588 Ignore-this: f41acd2d8f8e6763aa8bd57a0b44a7e4
589]
590[In validate, use gmake if available; based on a patch from Gabor PALI
591Ian Lynagh <[email protected]>**20100517200654]
592[Remove duplicate "./configure --help" output; fixes #4075
593Ian Lynagh <[email protected]>**20100516141206]
594[Update various 'sh boot's to 'perl boot'
595Ian Lynagh <igloo@earth.li>**20100516122609
596 Spotted by Marco Túlio Gontijo e Silva
597]
598[add missing initialisation for eventBufMutex
599Simon Marlow <marlowsd@gmail.com>**20100514094943
600 Ignore-this: 7f75594a8cb54fbec5aebd46bb959f45
601]
602[Undo part of #4003 patch
603Simon Marlow <marlowsd@gmail.com>**20100513142017
604 Ignore-this: cb65db86a38a7e5ccee9f779e489d104
605 We still need the workaround for when compiling HEAD with 6.12.2
606 
607]
608[Fix makefile loop (#4050)
609pho@cielonegro.org**20100507140707
610 Ignore-this: 3a1cb13d0600977e74d17ac26cbef83d
611 
612 The libtool creates "libffi.dylib" and "libffi.5.dylib" but not "libffi.5.0.9.dylib". Having it in libffi_DYNAMIC_LIBS causes an infinite makefile loop.
613]
614[fix !TABLES_NEXT_TO_CODE
615Simon Marlow <marlowsd@gmail.com>**20100510151934
616 Ignore-this: fccb859b114bef1c3122c98e60af51
617]
618[looksLikeModuleName: allow apostrophe in module names (#4051)
619Simon Marlow <marlowsd@gmail.com>**20100510094741
620 Ignore-this: df9348f3ba90608bec57257b47672985
621]
622[add the proper library dependencies for GhcProfiled=YES
623Simon Marlow <marlowsd@gmail.com>**20100506122118
624 Ignore-this: 6236993aa308ab5b5e1e5ea5f65982a
625]
626[Fix Trac #4003: fix the knot-tying in checkHiBootIface
627simonpj@microsoft.com**20100511075026
628 Ignore-this: a9ce2a318386fdc8782848df84592002
629 
630 I had incorrectly "optimised" checkHiBootIface so that it forgot
631 to update the "knot-tied" type environment.
632 
633 This patch fixes the HEAD
634]
635[Re-engineer the derived Ord instance generation code (fix Trac #4019)
636simonpj@microsoft.com**20100510133333
637 Ignore-this: 8fe46e4dad27fbee211a7928acf372c2
638   
639 As well as fixing #4019, I rejigged the way that Ord instances are
640 generated, which should make them faster in general.  See the
641 Note [Generating Ord instances].
642 
643 I tried to measure the performance difference from this change, but
644 the #4019 fix only removes one conditional branch per iteration, and
645 I couldn't measure a consistent improvement.  But still, tihs is
646 better than before.
647]
648[Make arity of INLINE things consistent
649simonpj@microsoft.com**20100510133005
650 Ignore-this: 15e7abf803d1dcb3f4ca760d2d939d0d
651 
652 We eta-expand things with INLINE pragmas;
653 see Note [Eta-expanding INLINE things].
654 
655 But I eta-expanded it the wrong amount when the function
656 was overloaded.  Ooops.
657]
658[Compacting GC fix, we forgot to thread the new bq field of StgTSO.
659Simon Marlow <marlowsd@gmail.com>**20100510082325
660 Ignore-this: a079c8446e2ad53efff6fd95d0f3ac80
661]
662[Add version constraints for the boot packages; fixes trac #3852
663Ian Lynagh <igloo@earth.li>**20100509175051
664 When using the bootstrapping compiler, we now explicitly constrain
665 the version of boot packages (Cabal, extensible-exceptions, etc) to the
666 in-tree version, so that the build system is less fragile should the
667 user have a newer version installed for the bootstrapping compiler.
668]
669[Don't include inter-package dependencies when compiling with stage 0; #4031
670Ian Lynagh <igloo@earth.li>**20100509130511
671 This fixes a problem when building with GHC 6.12 on Windows, where
672 dependencies on stage 0 (bootstrapping compiler) packages have absolute
673 paths c:/ghc/..., and make gets confused by the colon.
674]
675[Add a ghc.mk for bindisttest/
676Ian Lynagh <igloo@earth.li>**20100508223911]
677[Move some make variables around so they are available when cleaning
678Ian Lynagh <igloo@earth.li>**20100508212405]
679[Optimise checkremove a bit
680Ian Lynagh <igloo@earth.li>**20100508202006]
681[Improve the bindisttest Makefile
682Ian Lynagh <igloo@earth.li>**20100508195450]
683[Add tools to test that cleaning works properly
684Ian Lynagh <igloo@earth.li>**20100508194105]
685[Tweak the ghc-pkg finding code
686Ian Lynagh <igloo@earth.li>**20100508125815
687 It now understand the ghc-stage[123] names we use in-tree, and it won't
688 go looking for any old ghc-pkg if it can't find the one that matches
689 ghc.
690]
691[Add a way to show what cleaning would be done, without actually doing it
692Ian Lynagh <igloo@earth.li>**20100508122438]
693[Tidy up the "rm" flags in the build system
694Ian Lynagh <igloo@earth.li>**20100508115745]
695[Fix crash in nested callbacks (#4038)
696Simon Marlow <marlowsd@gmail.com>**20100507093222
697 Ignore-this: cade85e361534ce711865a4820276388
698 Broken by "Split part of the Task struct into a separate struct
699 InCall".
700]
701[Add $(GhcDynamic) knob, set to YES to get stage2 linked with -dynamic
702Simon Marlow <marlowsd@gmail.com>**20100428205241
703 Ignore-this: 1db8bccf92099785ecac39aebd27c92d
704 Default currently NO.
705 
706 Validate passed with GhcDynamic=YES on x86/Linux here.
707 
708 The compiler is currently slower on x86 when linked -dynamic,
709 because the GC inner loop has been adversely affected by -fPIC, I'm
710 looking into how to fix it.
711]
712[omit "dyn" from the way appended to the __stginit label
713Simon Marlow <marlowsd@gmail.com>**20100428204914
714 Ignore-this: 14183f3defa9f2bde68fda6729b740bc
715 When GHCi is linked dynamically, we still want to be able to load
716 non-dynamic object files.
717]
718[improvements to findPtr(), a neat hack for browsing the heap in gdb
719Simon Marlow <marlowsd@gmail.com>**20100506115427
720 Ignore-this: ac57785bb3e13b97a5945f753f068738
721]
722[Fix +RTS -G1
723Simon Marlow <marlowsd@gmail.com>**20100506110739
724 Ignore-this: 86a5de39a94d3331a4ee1213f82be497
725]
726[Enable the "redundant specialise pragmas" warning; fixes trac #3855
727Ian Lynagh <igloo@earth.li>**20100506175351]
728[Find the correct external ids when there's a wrapper
729simonpj@microsoft.com**20100506164135
730 Ignore-this: 636266407b174b05b2b8646cc73062c0
731 
732 We were failing to externalise the wrapper id for a function
733 that had one.
734]
735[Add a comment about pattern coercions
736simonpj@microsoft.com**20100506164027
737 Ignore-this: 17428089f3df439f65d892e23e8ed61a
738]
739[Comments only
740simonpj@microsoft.com**20100506163829
741 Ignore-this: 169167b6463873ab173cc5750c5be469
742]
743[Make a missing name in mkUsageInfo into a panic
744simonpj@microsoft.com**20100506163813
745 Ignore-this: b82ff1b8bf89f74f146db7cb5cc4c4d7
746 
747 We really want to know about this!
748]
749[Refactoring of hsXxxBinders
750simonpj@microsoft.com**20100506163737
751 Ignore-this: 97c6667625262b160f9746f7bea1c980
752 
753 This patch moves various functions that extract the binders
754 from a HsTyClDecl, HsForeignDecl etc into HsUtils, and gives
755 them consistent names.
756]
757[Fix Trac #3966: warn about useless UNPACK pragmas
758simonpj@microsoft.com**20100506163337
759 Ignore-this: 5beb24b686eda6113b614dfac8490df1
760 
761 Warning about useless UNPACK pragmas wasn't as easy as I thought.
762 I did quite a bit of refactoring, which improved the code by refining
763 the types somewhat.  In particular notice that in DataCon, we have
764 
765     dcStrictMarks   :: [HsBang]
766     dcRepStrictness :: [StrictnessMarks]
767 
768 The former relates to the *source-code* annotation, the latter to
769 GHC's representation choice.
770]
771[Make tcg_dus behave more sanely; fixes a mkUsageInfo panic
772simonpj@microsoft.com**20100506162719
773 Ignore-this: d000bca15b0e127e297378ded1bfb81b
774 
775 The tcg_dus field used to contain *uses* of type and class decls,
776 but not *defs*.  That was inconsistent, and it really went wrong
777 for Template Haskell bracket.  What happened was that
778  foo = [d| data A = A
779                   f :: A -> A
780                   f x = x |]
781 would find a "use" of A when processing the top level of the module,
782 which in turn led to a mkUsageInfo panic in MkIface.  The cause was
783 the fact that the tcg_dus for the nested quote didn't have defs for
784 A.
785]
786[Add a HsExplicitFlag to SpliceDecl, to improve Trac #4042
787simonpj@microsoft.com**20100506161523
788 Ignore-this: e4e563bac2fd831cc9e94612f5b4fa9d
789 
790 The issue here is that
791 
792     g :: A -> A
793     f
794     data A = A
795 
796 is treated as if you'd written $(f); that is the call of
797 f is a top-level Template Haskell splice.  This patch
798 makes sure that we *first* check the -XTemplateHaskellFlag
799 and bleat about a parse error if it's off.  Othewise we
800 get strange seeing "A is out of scope" errors.
801]
802[Change an assert to a warn
803simonpj@microsoft.com**20100506161111
804 Ignore-this: 739a4fb4c7940376b0f2c8ad52a1966c
805 
806 This is in the constraint simplifier which I'm about
807 to rewrite, so I'm hoping the assert isn't fatal!
808]
809[Tidy up debug print a little
810simonpj@microsoft.com**20100506161027
811 Ignore-this: bd5492878e06bee1cddcbb3fc4df66d8
812]
813[Remove useless UNPACK pragmas
814simonpj@microsoft.com**20100506161012
815 Ignore-this: 3e5ab1a7cf58107034412a798bc214e5
816]
817[Add WARNM2 macro, plus some refactoring
818simonpj@microsoft.com**20100506160808
819 Ignore-this: 2ab4f1f0b5d94be683036e77aec09255
820]
821[Use -Wwarn for the binary package, becuase it has redundant UNPACK pragmas
822simonpj@microsoft.com**20100506160750
823 Ignore-this: cf0d3a11473e28bfce9602e716e69a5f
824]
825[Fix Trac #3966: warn about unused UNPACK pragmas
826simonpj@microsoft.com**20100409201812
827 Ignore-this: c96412596b39c918b5fb9b3c39ce2119
828]
829[Fix Trac #3953: fail earlier when using a bogus quasiquoter
830simonpj@microsoft.com**20100409201748
831 Ignore-this: ef48e39aa932caed538643985234f043
832]
833[Fix Trac #3965: tighten conditions when deriving Data
834simonpj@microsoft.com**20100409184420
835 Ignore-this: 96f7d7d2da11565d26b465d7d0497ac9
836 
837 It's tricky to set up the context for a Data instance.  I got it wrong
838 once, and fixed it -- hence the "extra_constraints" in
839 TcDeriv.inferConstraints. 
840 
841 But it still wasn't right!  The tricky bit is that dataCast1 is only
842 generated when T :: *->*, and dataCast2 when T :: *->*->*. (See
843 the code in TcGenDeriv for dataCastX.
844]
845[Fix Trac #3964: view patterns in DsArrows
846simonpj@microsoft.com**20100409165557
847 Ignore-this: d823c182831d5e2e592e995b16180e2f
848 
849 Just a missing case; I've eliminated the catch-all so
850 that we get a warning next time we extend HsPat
851]
852[Fix Trac #3955: renamer and type variables
853simonpj@microsoft.com**20100409163710
854 Ignore-this: bd5ec64d76c0f583bf5f224792bf294c
855 
856 The renamer wasn't computing the free variables of a type declaration
857 properly.  This patch refactors a bit, and makes it more robust,
858 fixing #3955 and several other closely-related bugs.  (We were
859 omitting some free variables and that could just possibly lead to a
860 usage-version tracking error.
861]
862[Layout only
863simonpj@microsoft.com**20100409163506
864 Ignore-this: 1f14990b5aa0b9821b84452fb34e9f41
865]
866[Give a better deprecated message for INCLUDE pragmas; fixes #3933
867Ian Lynagh <igloo@earth.li>**20100506130910
868 We now have a DeprecatedFullText constructor, so we can override the
869 "-#include is deprecated: " part of the warning.
870]
871[De-haddock a comment that confuses haddock
872Ian Lynagh <igloo@earth.li>**20100506123607]
873[Fix comment to not confuse haddock
874Ian Lynagh <igloo@earth.li>**20100506113642]
875[Detect EOF when trying to parse a string in hp2ps
876Ian Lynagh <igloo@earth.li>**20100506000830]
877[Make the demand analyser sdd demands for strict constructors
878simonpj@microsoft.com**20100505200936
879 Ignore-this: eb32632adbc354eb7a5cf884c263e0d3
880 
881 This opportunity was spotted by Roman, and is documented in
882 Note [Add demands for strict constructors] in DmdAnal.
883]
884[Fix interaction of exprIsCheap and the lone-variable inlining check
885simonpj@microsoft.com**20100505200723
886 Ignore-this: f3cb65085c5673a99153d5d7b6559ab1
887 
888 See Note [Interaction of exprIsCheap and lone variables] in CoreUnfold
889 
890 This buglet meant that a nullary definition with an INLINE pragma
891 counter-intuitively didn't get inlined at all.  Roman identified
892 the bug.
893]
894[Matching cases in SpecConstr and Rules
895simonpj@microsoft.com**20100505200543
896 Ignore-this: f5c28c780fbf8badce84c6fdc9aa1779
897 
898 This patch has zero effect.  It includes comments,
899 a bit of refactoring, and a tiny bit of commment-out
900 code go implement the "matching cases" idea below.
901 
902 In the end I've left it disabled because while I think
903 it does no harm I don't think it'll do any good either.
904 But I didn't want to lose the idea totally. There's
905 a thread called "Storable and constant memory" on
906 the libraries@haskell.org list (Apr 2010) about it.
907 
908 Note [Matching cases]
909 ~~~~~~~~~~~~~~~~~~~~~
910 {- NOTE: This idea is currently disabled.  It really only works if
911          the primops involved are OkForSpeculation, and, since
912         they have side effects readIntOfAddr and touch are not.
913         Maybe we'll get back to this later .  -}
914   
915 Consider
916    f (case readIntOffAddr# p# i# realWorld# of { (# s#, n# #) ->
917       case touch# fp s# of { _ ->
918       I# n# } } )
919 This happened in a tight loop generated by stream fusion that
920 Roman encountered.  We'd like to treat this just like the let
921 case, because the primops concerned are ok-for-speculation.
922 That is, we'd like to behave as if it had been
923    case readIntOffAddr# p# i# realWorld# of { (# s#, n# #) ->
924    case touch# fp s# of { _ ->
925    f (I# n# } } )
926]
927[Comments only
928simonpj@microsoft.com**20100504163629
929 Ignore-this: 3be12df04714aa820bce706b5dc8a9cb
930]
931[Comments only
932simonpj@microsoft.com**20100504163529
933 Ignore-this: 791e2fd39c7d880ce1dc80ebdf3a5398
934]
935[Comments only
936simonpj@microsoft.com**20100504163457
937 Ignore-this: f19e9ffeb3d65770b1595bca5f97a59d
938]
939[Comments only (about type families)
940simonpj@microsoft.com**20100417145032
941 Ignore-this: dd39425ef2155d52dbf55a4d5fd97cb8
942]
943[Fix hp2ps when the .hp file has large string literals
944Ian Lynagh <igloo@earth.li>**20100505191921]
945[In build system, call package-config after including package data
946Ian Lynagh <igloo@earth.li>**20100504225035
947 Otherwise the $1_$2_HC_OPTS variable gets clobbered.
948]
949[runghc: flush stdout/stderr on an exception (#3890)
950Simon Marlow <marlowsd@gmail.com>**20100505133848
951 Ignore-this: 224c1898cec64cb1c94e0d7033e7590e
952]
953[Remove the Unicode alternative for ".." (#3894)
954Simon Marlow <marlowsd@gmail.com>**20100505121202
955 Ignore-this: 2452cd67281667106f9169747b6d784f
956]
957[tidyup; no functional changes
958Simon Marlow <marlowsd@gmail.com>**20100505115015
959 Ignore-this: d0787e5cdeef1dee628682fa0a46019
960]
961[Make the running_finalizers flag task-local
962Simon Marlow <marlowsd@gmail.com>**20100505114947
963 Ignore-this: 345925d00f1dca203941b3c5d84c90e1
964 Fixes a bug reported by Lennart Augustsson, whereby we could get an
965 incorrect error from the RTS about re-entry from a finalizer,
966]
967[add a MAYBE_GC() in killThread#, fixes throwto003(threaded2) looping
968Simon Marlow <marlowsd@gmail.com>**20100505114746
969 Ignore-this: efea04991d6feed04683a42232fc85da
970]
971[Allow filepath-1.2.*
972Simon Marlow <marlowsd@gmail.com>**20100505101139
973 Ignore-this: 1b5580cd9cd041ec48f40cd37603326a
974]
975[BlockedOnMsgThrowTo is possible in resurrectThreads (#4030)
976Simon Marlow <marlowsd@gmail.com>**20100505094534
977 Ignore-this: ac24a22f95ffeaf480187a1620fdddb2
978]
979[Don't raise a throwTo when the target is masking and BlockedOnBlackHole
980Simon Marlow <marlowsd@gmail.com>**20100505094506
981 Ignore-this: 302616931f61667030d77ddfbb02374e
982]
983[Fix build with GHC 6.10
984Ian Lynagh <igloo@earth.li>**20100504180302
985 In GHC 6.10, intersectionWith is (a -> b -> a) instead of (a -> b -> c),
986 so we need to jump through some hoops to get the more general type.
987]
988[The libffi patches are no longer needed
989Ian Lynagh <igloo@earth.li>**20100504171603]
990[Use the in-tree windres; fixes trac #4032
991Ian Lynagh <igloo@earth.li>**20100504170941]
992[Print unfoldings on lambda-bound variables
993Simon PJ <simonpj@microsoft.com>**20100503181822
994 Ignore-this: 2fd5a7502cc6273d96258e0914f0f8cd
995 
996 ...in the unusual case where they have one;
997 see Note [Case binders and join points] in Simplify.lhs
998]
999[Replace FiniteMap and UniqFM with counterparts from containers.
1000Milan Straka <fox@ucw.cz>**20100503171315
1001 Ignore-this: a021972239163dbf728284b19928cebb
1002 
1003 The original interfaces are kept. There is small performance improvement:
1004 - when compiling for five nofib, we get following speedups:
1005     Average                -----           -2.5%
1006     Average                -----           -0.6%
1007     Average                -----           -0.5%
1008     Average                -----           -5.5%
1009     Average                -----          -10.3%
1010 - when compiling HPC ten times, we get:
1011     switches                          oldmaps   newmaps
1012     -O -fasm                          117.402s  116.081s (98.87%)
1013     -O -fasm -fregs-graph             119.993s  118.735s (98.95%)
1014     -O -fasm -fregs-iterative         120.191s  118.607s (98.68%)
1015]
1016[Make the demand analyser take account of lambda-bound unfoldings
1017Simon PJ <simonpj@microsoft.com>**20100503151630
1018 Ignore-this: 2ee8e27d4df2debfc79e6b8a17c32bc1
1019 
1020 This is a long-standing lurking bug. See Note [Lamba-bound unfoldings]
1021 in DmdAnal.
1022 
1023 I'm still not really happy with this lambda-bound-unfolding stuff.
1024]
1025[Fix dynamic libs on OS X, and enable them by default
1026Ian Lynagh <igloo@earth.li>**20100503150302]
1027[Switch back to using bytestring from the darcs repo; partially fixes #3855
1028Ian Lynagh <igloo@earth.li>**20100502113458]
1029[Fix some cpp warnings when building on FreeBSD; patch from Gabor PALI
1030Ian Lynagh <igloo@earth.li>**20100428150700]
1031[Fix "make 2"
1032Ian Lynagh <igloo@earth.li>**20100427162212
1033 The new Makefile logic was enabling the stage 1 rules when stage=2,
1034 so "make 2" was rebuilding stage 1.
1035]
1036[Inplace programs depend on their shell wrappers
1037Ian Lynagh <igloo@earth.li>**20100427160038]
1038[--make is now the default (#3515), and -fno-code works with --make (#3783)
1039Simon Marlow <marlowsd@gmail.com>**20100427122851
1040 Ignore-this: 33330474fa4703f32bf9997462b4bf3c
1041 If the command line contains any Haskell source files, then we behave
1042 as if --make had been given.
1043 
1044 The meaning of the -c flag has changed (back): -c now selects one-shot
1045 compilation, but stops before linking.  However, to retain backwards
1046 compatibility, -c is still allowed with --make, and means the same as
1047 --make -no-link.  The -no-link flag has been un-deprecated.
1048 
1049 -fno-code is now allowed with --make (#3783); the fact that it was
1050 disabled before was largely accidental, it seems.  We also had some
1051 regressions in this area: it seems that -fno-code was causing a .hc
1052 file to be emitted in certain cases.  I've tidied up the code, there
1053 was no need for -fno-code to be a "mode" flag, as far as I can tell.
1054 
1055 -fno-code does not emit interface files, nor does it do recompilation
1056 checking, as suggested in #3783.  This would make Haddock emit
1057 interface files, for example, and I'm fairly sure we don't want to do
1058 that.  Compiling with -fno-code is pretty quick anyway, perhaps we can
1059 get away without recompilation checking.
1060]
1061[remove duplicate docs for -e in --help output (#4010)
1062Simon Marlow <marlowsd@gmail.com>**20100426140642
1063 Ignore-this: 187ff893ba8ffa0ec127867a7590e38d
1064]
1065[workaround for #4003, fixes HEAD build with 6.12.2
1066Simon Marlow <marlowsd@gmail.com>**20100426103428
1067 Ignore-this: c4bc445dc8052d4e6efef3f1daf63562
1068]
1069[Make sure all the clean rules are always included
1070Ian Lynagh <igloo@earth.li>**20100424181823
1071 In particular, this fixes a problem where stage3 bits weren't being cleaned
1072]
1073[Correct the name of the amd64/FreeBSD platform in PlatformSupportsSharedLibs
1074Ian Lynagh <igloo@earth.li>**20100424132830
1075 We weren't getting sharedlibs on amd64/FreeBSD because of this
1076]
1077[Include DPH docs in bindists
1078Ian Lynagh <igloo@earth.li>**20100424123101]
1079[reinstate eta-expansion during SimplGently, to fix inlining of sequence_
1080Simon Marlow <marlowsd@gmail.com>**20100423124853
1081 Ignore-this: 4fa0fd5bafe0d6b58fc81076f50d5f8d
1082]
1083[fix 64-bit value for W_SHIFT, which thankfully appears to be not used
1084Simon Marlow <marlowsd@gmail.com>**20100422213605
1085 Ignore-this: 525c062d2456c224ec8d0e083edd3b55
1086]
1087[Add missing constant folding and optimisation for unsigned division
1088Simon Marlow <marlowsd@gmail.com>**20100422213443
1089 Ignore-this: fb10d1cda0852fab0cbcb47247498fb3
1090 Noticed by Denys Rtveliashvili <rtvd@mac.com>, see #4004
1091]
1092[Fix the GHC API link in the main doc index.html
1093Ian Lynagh <igloo@earth.li>**20100422213226]
1094[Give the right exit code in darcs-all
1095Ian Lynagh <igloo@earth.li>**20100421171339
1096 Our END block was calling system, which alters $?. So now we save and
1097 restore it.
1098]
1099[Use StgWord64 instead of ullong
1100Ian Lynagh <igloo@earth.li>**20100421162336
1101 This patch also fixes ullong_format_string (renamed to showStgWord64)
1102 so that it works with values outside the 32bit range (trac #3979), and
1103 simplifies the without-commas case.
1104]
1105[Implement try10Times in Makefile
1106Ian Lynagh <igloo@earth.li>**20100420165909
1107 Avoid using seq, as FreeBSD has jot instead.
1108]
1109[Fix crash in non-threaded RTS on Windows
1110Simon Marlow <marlowsd@gmail.com>**20100420122125
1111 Ignore-this: 28b0255a914a8955dce02d89a7dfaca
1112 The tso->block_info field is now overwritten by pushOnRunQueue(), but
1113 stg_block_async_info was assuming that it still held a pointer to the
1114 StgAsyncIOResult.  We must therefore save this value somewhere safe
1115 before putting the TSO on the run queue.
1116]
1117[Expand the scope of the event_buf_mutex to cover io_manager_event
1118Simon Marlow <marlowsd@gmail.com>**20100420122026
1119 Ignore-this: 185a6d84f7d4a35997f10803f6dacef1
1120 I once saw a failure that I think was due to a race on
1121 io_manager_event, this should fix it.
1122]
1123[Flags -auto and -auto-all operate only on functions not marked INLINE.
1124Milan Straka <fox@ucw.cz>**20100331191050
1125 Ignore-this: 3b63580cfcb3c33d62ad697c36d94d05
1126]
1127[Spelling correction for LANGUAGE pragmas
1128Max Bolingbroke <batterseapower@hotmail.com>**20100413192825
1129 Ignore-this: 311b51ba8d43f6c7fd32f48db9a88dee
1130]
1131[Update the user guide so it talks about the newer "do rec" notation everywhere
1132Ian Lynagh <igloo@earth.li>**20100416205416
1133 Some of the problems highlighted in trac #3968.
1134]
1135[Fix typo
1136Ian Lynagh <igloo@earth.li>**20100416205412]
1137[Fix Trac #3950: unifying types of different kinds
1138simonpj@microsoft.com**20100412151845
1139 Ignore-this: d145b9de5ced136ef2c39f3ea4a04f4a
1140 
1141 I was assuming that the unifer only unified types of the
1142 same kind, but now we can "defer" unsolved constraints that
1143 invariant no longer holds.  Or at least is's more complicated
1144 to ensure. 
1145 
1146 This patch takes the path of not assuming the invariant, which
1147 is simpler and more robust.  See
1148 Note [Mismatched type lists and application decomposition]
1149]
1150[Fix Trac #3943: incorrect unused-variable warning
1151simonpj@microsoft.com**20100412151630
1152 Ignore-this: 52459f2b8b02c3cb120abe674dc9a060
1153 
1154 In fixing this I did the usual little bit of refactoring
1155]
1156[Convert boot and boot-pkgs to perl
1157Ian Lynagh <igloo@earth.li>**20100415143919
1158 This stops us having to worry about sh/sed/... portability.
1159]
1160[Use $(MAKE), not make, when recursively calling make
1161Ian Lynagh <igloo@earth.li>**20100415121453]
1162[Remove the ghc_ge_609 makefile variables
1163Ian Lynagh <igloo@earth.li>**20100412235658
1164 They are now guaranteed to be YES
1165]
1166[Increase the minimum version number required to 6.10 in configure.ac
1167Ian Lynagh <igloo@earth.li>**20100412235313]
1168[The bootstrapping compiler is now required to be > 609
1169Ian Lynagh <igloo@earth.li>**20100409161046]
1170[Handle IND_STATIC in isRetainer
1171Ian Lynagh <igloo@earth.li>**20100409104207
1172 IND_STATIC used to be an error, but at the moment it can happen
1173 as isAlive doesn't look through IND_STATIC as it ignores static
1174 closures. See trac #3956 for a program that hit this error.
1175]
1176[Add Data and Typeable instances to HsSyn
1177David Waern <david.waern@gmail.com>**20100330011020
1178 Ignore-this: c3f2717207b15539fea267c36b686e6a
1179 
1180 The instances (and deriving declarations) have been taken from the ghc-syb
1181 package.
1182]
1183[Fix for derefing ThreadRelocated TSOs in MVar operations
1184Simon Marlow <marlowsd@gmail.com>**20100407092824
1185 Ignore-this: 94dd7c68a6094eda667e2375921a8b78
1186]
1187[sanity check fix
1188Simon Marlow <marlowsd@gmail.com>**20100407092746
1189 Ignore-this: 9c18cd5f5393e5049015ca52e62a1269
1190]
1191[get the reg liveness right in the putMVar# heap check
1192Simon Marlow <marlowsd@gmail.com>**20100407092724
1193 Ignore-this: b1ba07a59ecfae00e9a1f8391741abc
1194]
1195[initialise the headers of MSG_BLACKHOLE objects properly
1196Simon Marlow <marlowsd@gmail.com>**20100407081712
1197 Ignore-this: 183dcd0ca6a395d08db2be12b02bdd79
1198]
1199[initialise the headers of MVAR_TSO_QUEUE objects properly
1200Simon Marlow <marlowsd@gmail.com>**20100407081514
1201 Ignore-this: 4b4a2f30cf2fb69ca4128c41744687bb
1202]
1203[undo debugging code
1204Simon Marlow <marlowsd@gmail.com>**20100406142740
1205 Ignore-this: 323c2248f817b6717c19180482fc4b00
1206]
1207[putMVar#: fix reg liveness in the heap check
1208Simon Marlow <marlowsd@gmail.com>**20100406135832
1209 Ignore-this: cddd2c7807ac7612c9b2c4c0d384d284
1210]
1211[account for the new BLACKHOLEs in the GHCi debugger
1212Simon Marlow <marlowsd@gmail.com>**20100406133406
1213 Ignore-this: 4d4aeb4bbada3f50dc1fb0123f565e8f
1214]
1215[don't forget to deRefTSO() in tryWakeupThread()
1216Simon Marlow <marlowsd@gmail.com>**20100406130411
1217 Ignore-this: 171d57c4f8653835dec0b69f9be9881c
1218]
1219[Fix bug in popRunQueue
1220Simon Marlow <marlowsd@gmail.com>**20100406091453
1221 Ignore-this: 9d3cec8f18f5c5cbd51751797386eb6f
1222]
1223[fix bug in migrateThread()
1224Simon Marlow <marlowsd@gmail.com>**20100401105840
1225 Ignore-this: 299bcf0d1ea0f8865f3e845eb93d2ad3
1226]
1227[Remove the IND_OLDGEN and IND_OLDGEN_PERM closure types
1228Simon Marlow <marlowsd@gmail.com>**20100401093519
1229 Ignore-this: 95f2480c8a45139835eaf5610217780b
1230 These are no longer used: once upon a time they used to have different
1231 layout from IND and IND_PERM respectively, but that is no longer the
1232 case since we changed the remembered set to be an array of addresses
1233 instead of a linked list of closures.
1234]
1235[Change the representation of the MVar blocked queue
1236Simon Marlow <marlowsd@gmail.com>**20100401091605
1237 Ignore-this: 20a35bfabacef2674df362905d7834fa
1238 
1239 The list of threads blocked on an MVar is now represented as a list of
1240 separately allocated objects rather than being linked through the TSOs
1241 themselves.  This lets us remove a TSO from the list in O(1) time
1242 rather than O(n) time, by marking the list object.  Removing this
1243 linear component fixes some pathalogical performance cases where many
1244 threads were blocked on an MVar and became unreachable simultaneously
1245 (nofib/smp/threads007), or when sending an asynchronous exception to a
1246 TSO in a long list of thread blocked on an MVar.
1247 
1248 MVar performance has actually improved by a few percent as a result of
1249 this change, slightly to my surprise.
1250 
1251 This is the final cleanup in the sequence, which let me remove the old
1252 way of waking up threads (unblockOne(), MSG_WAKEUP) in favour of the
1253 new way (tryWakeupThread and MSG_TRY_WAKEUP, which is idempotent).  It
1254 is now the case that only the Capability that owns a TSO may modify
1255 its state (well, almost), and this simplifies various things.  More of
1256 the RTS is based on message-passing between Capabilities now.
1257]
1258[eliminate some duplication with a bit of CPP
1259Simon Marlow <marlowsd@gmail.com>**20100330154355
1260 Ignore-this: 838f7d341f096ca14c86ab9c81193e36
1261]
1262[Make ioManagerDie() idempotent
1263Simon Marlow <marlowsd@gmail.com>**20100401100705
1264 Ignore-this: a5996b43cdb2e2d72e6e971d7ea925fb
1265 Avoids screeds of "event buffer overflowed; event dropped" in
1266 conc059(threaded1).
1267]
1268[Move a thread to the front of the run queue when another thread blocks on it
1269Simon Marlow <marlowsd@gmail.com>**20100329144521
1270 Ignore-this: c518ff0d41154680edc811d891826a29
1271 This fixes #3838, and was made possible by the new BLACKHOLE
1272 infrastructure.  To allow reording of the run queue I had to make it
1273 doubly-linked, which entails some extra trickiness with regard to
1274 GC write barriers and suchlike.
1275]
1276[remove non-existent MUT_CONS symbols
1277Simon Marlow <marlowsd@gmail.com>**20100330152600
1278 Ignore-this: 885628257a9d03f2ece2a754d993014a
1279]
1280[change throwTo to use tryWakeupThread rather than unblockOne
1281Simon Marlow <marlowsd@gmail.com>**20100329144613
1282 Ignore-this: 10ad4965e6c940db71253f1c72218bbb
1283]
1284[tiny GC optimisation
1285Simon Marlow <marlowsd@gmail.com>**20100329144551
1286 Ignore-this: 9e095b9b73fff0aae726f9937846ba92
1287]
1288[New implementation of BLACKHOLEs
1289Simon Marlow <marlowsd@gmail.com>**20100329144456
1290 Ignore-this: 96cd26793b4e6ab9ddd0d59aae5c2f1d
1291 
1292 This replaces the global blackhole_queue with a clever scheme that
1293 enables us to queue up blocked threads on the closure that they are
1294 blocked on, while still avoiding atomic instructions in the common
1295 case.
1296 
1297 Advantages:
1298 
1299  - gets rid of a locked global data structure and some tricky GC code
1300    (replacing it with some per-thread data structures and different
1301    tricky GC code :)
1302 
1303  - wakeups are more prompt: parallel/concurrent performance should
1304    benefit.  I haven't seen anything dramatic in the parallel
1305    benchmarks so far, but a couple of threading benchmarks do improve
1306    a bit.
1307 
1308  - waking up a thread blocked on a blackhole is now O(1) (e.g. if
1309    it is the target of throwTo).
1310 
1311  - less sharing and better separation of Capabilities: communication
1312    is done with messages, the data structures are strictly owned by a
1313    Capability and cannot be modified except by sending messages.
1314 
1315  - this change will utlimately enable us to do more intelligent
1316    scheduling when threads block on each other.  This is what started
1317    off the whole thing, but it isn't done yet (#3838).
1318 
1319 I'll be documenting all this on the wiki in due course.
1320 
1321]
1322[Fix warnings (allow pushOnRunQueue() to not be inlined)
1323Simon Marlow <marlowsd@gmail.com>**20100401114559
1324 Ignore-this: f40bfbfad70a5165a946d11371605b7d
1325]
1326[remove out of date comment
1327Simon Marlow <marlowsd@gmail.com>**20100401105853
1328 Ignore-this: 26af88dd418ee0bcda7223b3b7e4e8d2
1329]
1330[tidy up spacing in stderr traces
1331Simon Marlow <marlowsd@gmail.com>**20100326163122
1332 Ignore-this: 16558b0433a274be217d4bf39aa4946
1333]
1334[Fix an assertion that was not safe when running in parallel
1335Simon Marlow <marlowsd@gmail.com>**20100325143656
1336 Ignore-this: cad08fb8900eb3a475547af0189fcc47
1337]
1338[Never jump directly to a thunk's entry code, even if it is single-entry
1339Simon Marlow <marlowsd@gmail.com>**20100325114847
1340 Ignore-this: 938da172c06a97762ef605c8fccfedf1
1341 I don't think this fixes any bugs as we don't have single-entry thunks
1342 at the moment, but it could cause problems for parallel execution if
1343 we ever did re-introduce update avoidance.
1344]
1345[Rename forgotten -dverbose-simpl to -dverbose-core2core in the docs.
1346Milan Straka <fox@ucw.cz>**20100331153626
1347 Ignore-this: 2da58477fb96e1cfb80f37dddd7c422c
1348]
1349[Add -pa and -V to the documentation of time profiling options.
1350Milan Straka <fox@ucw.cz>**20100329191121
1351 Ignore-this: be74d216481ec5a19e5f40f85e6e3d65
1352]
1353[Keep gcc 4.5 happy
1354Simon Marlow <marlowsd@gmail.com>**20100330120425
1355 Ignore-this: 7811878cc2bd1ce9cfbb5bf102fe3454
1356]
1357[Fix warning compiling Linker.c for PPC Mac
1358naur@post11.tele.dk**20100403182355
1359 Ignore-this: e2d2448770c9714ce17dd6cf3e297063
1360 The warning message eliminated is:
1361 > rts/Linker.c:4756:0:
1362 >      warning: nested extern declaration of 'symbolsWithoutUnderscore'
1363]
1364[Fix error compiling AsmCodeGen.lhs for PPC Mac (mkRtsCodeLabel)
1365naur@post11.tele.dk**20100403181656
1366 Ignore-this: deb7524ea7852a15a2ac0849c8c82f74
1367 The error messages eliminated are:
1368 > compiler/nativeGen/AsmCodeGen.lhs:875:31:
1369 >     Not in scope: `mkRtsCodeLabel'
1370 > compiler/nativeGen/AsmCodeGen.lhs:879:31:
1371 >     Not in scope: `mkRtsCodeLabel'
1372 > compiler/nativeGen/AsmCodeGen.lhs:883:31:
1373 >     Not in scope: `mkRtsCodeLabel'
1374]
1375[Fix error compiling AsmCodeGen.lhs for PPC Mac (DestBlockId)
1376naur@post11.tele.dk**20100403180643
1377 Ignore-this: 71e833e94ed8371b2ffabc2cf80bf585
1378 The error message eliminated is:
1379 > compiler/nativeGen/AsmCodeGen.lhs:637:16:
1380 >     Not in scope: data constructor `DestBlockId'
1381]
1382[Fix boot-pkgs's sed usage to work with Solaris's sed
1383Ian Lynagh <igloo@earth.li>**20100401153441]
1384[Pass "-i org.haskell.GHC" to packagemaker when building the OS X installer
1385Ian Lynagh <igloo@earth.li>**20100331144707
1386 This seems to fix this failure:
1387 [...]
1388 ** BUILD SUCCEEDED **
1389 rm -f -f GHC-system.pmdoc/*-contents.xml
1390 /Developer/usr/bin/packagemaker -v --doc GHC-system.pmdoc\
1391              -o /Users/ian/to_release/ghc-6.12.1.20100330/GHC-6.12.1.20100330-i386.pkg
1392 2010-03-31 15:08:15.695 packagemaker[13909:807] Setting to : 0 (null)
1393 2010-03-31 15:08:15.709 packagemaker[13909:807] Setting to : 0 org.haskell.glasgowHaskellCompiler.ghc.pkg
1394 2010-03-31 15:08:15.739 packagemaker[13909:807] relocate: (null) 0
1395 2010-03-31 15:08:15.740 packagemaker[13909:807] *** Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '*** -[NSXMLDocument initWithXMLString:options:error:]: nil argument'
1396 2010-03-31 15:08:15.741 packagemaker[13909:807] Stack: (
1397     2511962091,
1398     2447007291,
1399     2511961547,
1400     2511961610,
1401     2432803204,
1402     453371,
1403     447720,
1404     436209,
1405     435510,
1406     9986,
1407     9918
1408 )
1409 make[1]: *** [framework-pkg] Trace/BPT trap
1410 make: *** [framework-pkg] Error 2
1411]
1412[Use machdepCCOpts when compiling the file to toggle -(no-)rtsopts
1413Ian Lynagh <igloo@earth.li>**20100331161302
1414 Should fix toggling on OS X "Snow Leopard". Diagnosed by Roman Leshchinskiy.
1415]
1416[Avoid a non-portable use of tar reported by Roman Leshchinskiy
1417Ian Lynagh <igloo@earth.li>**20100330145802]
1418[Don't install EXTRA_PACKAGES by default
1419Simon Marlow <marlowsd@gmail.com>**20100330142714
1420 Ignore-this: d4cc8f87a6de8d9d1d6dc9b77130b3
1421]
1422[fix a non-portable printf format
1423Simon Marlow <marlowsd@gmail.com>**20100330134437
1424 Ignore-this: d41c23c54ec29654cb2049de1e588570
1425]
1426[avoid single quote in #error
1427Simon Marlow <marlowsd@gmail.com>**20100330120346
1428 Ignore-this: 663f39e7a27fead2f648fbf22d345bb4
1429]
1430[use FMT_Word64 instead of locally-defined version
1431Simon Marlow <marlowsd@gmail.com>**20100330114650
1432 Ignore-this: 82697b8095dffb3a8e196c687006ece0
1433]
1434[remove old/unused DotnetSupport and GhcLibsWithUnix
1435Simon Marlow <marlowsd@gmail.com>**20100330123732
1436 Ignore-this: c68814868b3671abdc369105bbeafe6c
1437]
1438[fix return type cast in f.i.wrapper when using libffi (#3516)
1439Simon Marlow <marlowsd@gmail.com>**20100329154220
1440 Ignore-this: f898eb8c9ae2ca2009e539735b92c438
1441 
1442 Original fix submitted by
1443   Sergei Trofimovich <slyfox@community.haskell.org>
1444 modified by me:
1445  - exclude 64-bit types
1446  - compare uniques, not strings
1447  - #include "ffi.h" is conditional
1448]
1449[libffi: install 'ffitarget.h' header as sole 'ffi.h' is unusable
1450Simon Marlow <marlowsd@gmail.com>**20100329135734
1451 Ignore-this: f9b555ea289d8df1aa22cb6faa219a39
1452 Submitted by: Sergei Trofimovich <slyfox@community.haskell.org>
1453 Re-recorded against HEAD.
1454]
1455[avoid a fork deadlock (see comments)
1456Simon Marlow <marlowsd@gmail.com>**20100329132329
1457 Ignore-this: 3377f88b83bb3b21e42d7fc5f0d866f
1458]
1459[tidy up the end of the all_tasks list after forking
1460Simon Marlow <marlowsd@gmail.com>**20100329132253
1461 Ignore-this: 819d679875be5f344e816210274d1c29
1462]
1463[Add a 'setKeepCAFs' external function (#3900)
1464Simon Marlow <marlowsd@gmail.com>**20100329110036
1465 Ignore-this: ec532a18cad4259a09847b0b9ae2e1d2
1466]
1467[Explicitly check whether ar supports the @file syntax
1468Ian Lynagh <igloo@earth.li>**20100329123325
1469 rather than assuming that all GNU ar's do.
1470 Apparently OpenBSD's older version doesn't.
1471]
1472[Fix the format specifier for Int64/Word64 on Windows
1473Ian Lynagh <igloo@earth.li>**20100327182126
1474 mingw doesn't understand %llu/%lld - it treats them as 32-bit rather
1475 than 64-bit. We use %I64u/%I64d instead.
1476]
1477[Fix the ghci startmenu item
1478Ian Lynagh <igloo@earth.li>**20100326235934
1479 I'm not sure what changed, but it now doesn't work for me without
1480 the "Start in" field being set.
1481]
1482[Fix paths to docs in "Start Menu" entries in Windows installer; fixes #3847
1483Ian Lynagh <igloo@earth.li>**20100326155917]
1484[Add a licence file for the Windows installer to use
1485Ian Lynagh <igloo@earth.li>**20100326155130]
1486[Add gcc-g++ to the inplace mingw installation; fixes #3893
1487Ian Lynagh <igloo@earth.li>**20100326154714]
1488[Add the licence file to the Windows installer. Fixes #3934
1489Ian Lynagh <igloo@earth.li>**20100326152449]
1490[Quote the paths to alex and happy in configure
1491Ian Lynagh <igloo@earth.li>**20100325143449
1492 Ignore-this: d6d6e1a250f88985bbeea760e63a79db
1493]
1494[Use </> rather than ++ "/"
1495Ian Lynagh <igloo@earth.li>**20100325133237
1496 This stops us generating paths like
1497     c:\foo\/ghc460_0/ghc460_0.o
1498 which windres doesn't understand.
1499]
1500[Append $(exeext) to utils/ghc-pkg_dist_PROG
1501Ian Lynagh <igloo@earth.li>**20100324233447
1502 Fixes bindist creation
1503]
1504[A sanity check
1505Simon Marlow <marlowsd@gmail.com>**20100325110500
1506 Ignore-this: 3b3b76d898c822456857e506b7531e65
1507]
1508[do_checks: do not set HpAlloc if the stack check fails
1509Simon Marlow <marlowsd@gmail.com>**20100325110328
1510 Ignore-this: 899ac8c29ca975d03952dbf4608d758
1511 
1512 This fixes a very rare heap corruption bug, whereby
1513 
1514  - a context switch is requested, which sets HpLim to zero
1515    (contextSwitchCapability(), called by the timer signal or
1516    another Capability).
1517 
1518  - simultaneously a stack check fails, in a code fragment that has
1519    both a stack and a heap check.
1520 
1521 The RTS then assumes that a heap-check failure has occurred and
1522 subtracts HpAlloc from Hp, although in fact it was a stack-check
1523 failure and retreating Hp will overwrite valid heap objects.  The bug
1524 is that HpAlloc should only be set when Hp has been incremented by the
1525 heap check.  See comments in rts/HeapStackCheck.cmm for more details.
1526 
1527 This bug is probably incredibly rare in practice, but I happened to be
1528 working on a test that triggers it reliably:
1529 concurrent/should_run/throwto001, compiled with -O -threaded, args 30
1530 300 +RTS -N2, run repeatedly in a loop.
1531]
1532[comments and formatting only
1533Simon Marlow <marlowsd@gmail.com>**20100325104617
1534 Ignore-this: c0a211e15b5953bb4a84771bcddd1d06
1535]
1536[Change how perl scripts get installed; partially fixes #3863
1537Ian Lynagh <igloo@earth.li>**20100324171422
1538 We now regenerate them when installing, which means the path for perl
1539 doesn't get baked in
1540]
1541[Pass the location of gcc in the ghc wrapper script; partially fixes #3863
1542Ian Lynagh <igloo@earth.li>**20100324171408
1543 This means we don't rely on baking a path to gcc into the executable
1544]
1545[Quote the ar path in configure
1546Ian Lynagh <igloo@earth.li>**20100324162043]
1547[Remove unused cUSER_WAY_NAMES cUSER_WAY_OPTS
1548Ian Lynagh <igloo@earth.li>**20100324145048]
1549[Remove unused cCONTEXT_DIFF
1550Ian Lynagh <igloo@earth.li>**20100324145013]
1551[Remove unused cEnableWin32DLLs
1552Ian Lynagh <igloo@earth.li>**20100324144841]
1553[Remove unused cGHC_CP
1554Ian Lynagh <igloo@earth.li>**20100324144656]
1555[Fix the build for non-GNU-ar
1556Ian Lynagh <igloo@earth.li>**20100324132907]
1557[Tweak the Makefile code for making .a libs; fixes trac #3642
1558Ian Lynagh <igloo@earth.li>**20100323221325
1559 The main change is that, rather than using "xargs ar" we now put
1560 all the filenames into a file, and do "ar @file". This means that
1561 ar adds all the files at once, which works around a problem where
1562 files with the same basename in a later invocation were overwriting
1563 the existing file in the .a archive.
1564]
1565[Enable shared libraries on Windows; fixes trac #3879
1566Ian Lynagh <igloo@earth.li>**20100320231414
1567 Ignore-this: c93b35ec5b7a7fa6ddb286d17a616216
1568]
1569[Add the external core PDF to the new build system
1570Ian Lynagh <igloo@earth.li>**20100321161909]
1571[Allow specifying $threads directly when validating
1572Ian Lynagh <igloo@earth.li>**20100321112835]
1573[Remove LazyUniqFM; fixes trac #3880
1574Ian Lynagh <igloo@earth.li>**20100320213837]
1575[UNDO: slight improvement to scavenging ...
1576Simon Marlow <marlowsd@gmail.com>**20100319153413
1577 Ignore-this: f0ab581c07361f7b57eae02dd6ec893c
1578 
1579 Accidnetally pushed this patch which, while it validates, isn't
1580 correct.
1581 
1582 rolling back:
1583 
1584 Fri Mar 19 11:21:27 GMT 2010  Simon Marlow <marlowsd@gmail.com>
1585   * slight improvement to scavenging of update frames when a collision has occurred
1586 
1587     M ./rts/sm/Scav.c -19 +15
1588]
1589[slight improvement to scavenging of update frames when a collision has occurred
1590Simon Marlow <marlowsd@gmail.com>**20100319112127
1591 Ignore-this: 6de2bb9614978975f17764a0f259d9bf
1592]
1593[Don't install the utf8-string package
1594Ian Lynagh <igloo@earth.li>**20100317212709]
1595[Don't use -Bsymbolic when linking the RTS
1596Ian Lynagh <igloo@earth.li>**20100316233357
1597 This makes the RTS hooks work when doing dynamic linking
1598]
1599[Fix Trac #3920: Template Haskell kinds
1600simonpj@microsoft.com**20100317123519
1601 Ignore-this: 426cac7920446e04f3cc30bd1d9f76e2
1602 
1603 Fix two places where we were doing foldl instead of foldr
1604 after decomposing a Kind.  Strange that the same bug appears
1605 in two quite different places!
1606]
1607[copy_tag_nolock(): fix write ordering and add a write_barrier()
1608Simon Marlow <marlowsd@gmail.com>**20100316143103
1609 Ignore-this: ab7ca42904f59a0381ca24f3eb38d314
1610 
1611 Fixes a rare crash in the parallel GC.
1612 
1613 If we copy a closure non-atomically during GC, as we do for all
1614 immutable values, then before writing the forwarding pointer we better
1615 make sure that the closure itself is visible to other threads that
1616 might follow the forwarding pointer.  I imagine this doesn't happen
1617 very often, but I just found one case of it: in scavenge_stack, the
1618 RET_FUN case, after evacuating ret_fun->fun we then follow it and look
1619 up the info pointer.
1620]
1621[Add sliceP mapping to vectoriser builtins
1622benl@ouroborus.net**20100316060517
1623 Ignore-this: 54c3cafff584006b6fbfd98124330aa3
1624]
1625[Comments only
1626benl@ouroborus.net**20100311064518
1627 Ignore-this: d7dc718cc437d62aa5b1b673059a9b22
1628]
1629[TAG 2010-03-16
1630Ian Lynagh <igloo@earth.li>**20100316005137
1631 Ignore-this: 234e3bc29e2f26cc59d7b03d780cc352
1632]
1633Patch bundle hash:
16347c78f0cd481036754b50ec92af0da175c2414a33