Ticket #1962: 1962.patch

File 1962.patch, 25.2 KB (added by igloo, 7 years ago)
2New patches:
4[Tweak installation so it works with Solaris's sh
5Ian Lynagh <[email protected]>**20071209140724] {
6hunk ./compiler/Makefile 961
7-       set -e; for d in stage$(stage)/*/; do $(INSTALL_DIR) $(BIN_DIST_DIR)/compiler/$$d; done
8+       set -e; for d in `ls -1d stage$(stage)/*/`; do $(INSTALL_DIR) $(BIN_DIST_DIR)/compiler/$$d; done
13[BIN_DIST_INST_SUBDIR Needs to be defined in config.mk so ./Makefile can see it
14Ian Lynagh <[email protected]>**20071207121317]
15[Improve eta reduction, to reduce Simplifier iterations
16[email protected]**20071203150039
18 I finally got around to investigating why the Simplifier was sometimes
19 iterating so often.  There's a nice example in Text.ParserCombinators.ReadPrec,
20 which produced:
22 NOTE: Simplifier still going after 3 iterations; bailing out.  Size = 339
23 NOTE: Simplifier still going after 3 iterations; bailing out.  Size = 339
24 NOTE: Simplifier still going after 3 iterations; bailing out.  Size = 339
26 No progress is being made.  It turned out that an interaction between
27 eta-expansion, casts, and eta reduction was responsible. The change is
28 small and simple, in SimplUtils.mkLam: do not require the body to be
29 a Lam when floating the cast outwards. 
31 I also discovered a missing side condition in the same equation, so fixing
32 that is good too.  Now there is no loop when compiling ReadPrec.
34 Should do a full nofib run though.
37[MERGED: fix race conditions in sandboxIO (#1583, #1922, #1946)
38Ian Lynagh <[email protected]>**20071206150509
39 Simon Marlow <[email protected]>**20071204114444
40  using the new block-inheriting forkIO (#1048)
42[Make eta reduction check more carefully for bottoms (fix Trac #1947)
43[email protected]**20071204145803
45 Eta reduction was wrongly transforming
46        f = \x. f x
47 to
48        f = f
50 Solution: don't trust f's arity information; instead look at its
51 unfolding.  See Note [Eta reduction conditions]
53 Almost all the new lines are comments!
57[protect console handler against concurrent access (#1922)
58Simon Marlow <[email protected]>**20071204153918]
59[forkIO starts the new thread blocked if the parent is blocked (#1048)
60Simon Marlow <[email protected]>**20071204110947]
61[Workaround for #1959: assume untracked names have changed
62Simon Marlow <[email protected]>**20071206092349
63 This fixes the 1959 test, but will do more recompilation than is
64 strictly necessary (but only when -O is on).  Still, more
65 recompilation is better than segfaults, link errors or other random
66 breakage.
68[FIX part of #1959: declaration versions were not being incremented correctly
69Simon Marlow <[email protected]>**20071206084556
70 We were building a mapping from ModuleName to [Occ] from the usage
71 list, using the usg_mod field as the key.  Unfortunately, due to a
72 very poor naming decision, usg_mod is actually the module version, not
73 the ModuleName.  usg_name is the ModuleName.  Since Version is also an
74 instance of Uniquable, there was no type error: all that happened was
75 lookups in the map never succeeded.  I shall rename the fields of
76 Usage in a separate patch.
78 This doesn't completely fix #1959, but it gets part of the way there.
80 I have to take partial blame as the person who wrote this fragment of
81 code in late 2006 (patch "Interface file optimisation and removal of
82 nameParent").
84[FIX #1110: hackery also needed when running gcc for CPP
85Simon Marlow <[email protected]>**20071205150230]
86[move FP_FIND_ROOT after the "GHC is required" check
87Simon Marlow <[email protected]>**20071205101814]
88[Change --shared to -shared in Win32 DLL docs
89[email protected]**20071204154023]
90[#include ../includes/MachRegs.h rather than just MachRegs.h
91Ian Lynagh <[email protected]>**20071205170335
92 This fixes building on NixOS. I'm not sure why it worked everywhere else,
93 but not on NixOS, before.
95[Fix the libraries Makefile
96Ian Lynagh <[email protected]>**20071205125015
97     x && y
98 is not the same as
99     if x; then y; fi
100 as the latter doesn't fail when x fails
102[Copy hscolour.css into dist/... so it gets installed with the library docs
103Ian Lynagh <[email protected]>**20071205013703]
104[Add the hscolour.css from hscolour 1.8
105Ian Lynagh <[email protected]>**20071205011733]
106[MERGED: Reorganise TcSimplify (again); FIX Trac #1919
107Ian Lynagh <[email protected]>**20071203140443
108 [email protected]**20071128173146
110  This was a bit tricky.  We had a "given" dict like (d7:Eq a); then it got
111  supplied to reduceImplication, which did some zonking, and emerged with
112  a "needed given" (d7:Eq Int). That got everything confused.
114  I found a way to simplify matters significantly.  Now reduceContext
115     - first deals with methods/literals/dictionaries
116     - then deals with implications
117  Separating things in this way not only made the bug go away, but
118  eliminated the need for the recently-added "needed-givens" results returned
119  by checkLoop.  Hurrah.
121  It's still a swamp.  But it's a bit better.
123[FIX #1843: Generate different instructions on PPC
124Ian Lynagh <[email protected]>**20071203123237
125 The old ones caused lots of
126     unknown scattered relocation type 4
127 errors. Patch from Chris Kuklewicz.
129[Fix bindist creation: readline/config.mk is gone
130Ian Lynagh <[email protected]>**20071203123031]
131[Don't default to stripping binaries when installing
132Ian Lynagh <[email protected]>**20071202195817]
133[Add package version bumps to the release notes
134Ian Lynagh <[email protected]>**20071201180437]
135[Move file locking into the RTS, fixing #629, #1109
136Simon Marlow <[email protected]>**20071120140859
137 File locking (of the Haskell 98 variety) was previously done using a
138 static table with linear search, which had two problems: the array had
139 a fixed size and was sometimes too small (#1109), and performance of
140 lockFile/unlockFile was suboptimal due to the linear search.
141 Also the algorithm failed to count readers as required by Haskell 98
142 (#629).
144 Now it's done using a hash table (provided by the RTS).  Furthermore I
145 avoided the extra fstat() for every open file by passing the dev_t and
146 ino_t into lockFile.  This and the improvements to the locking
147 algorithm result in a healthy 20% or so performance increase for
148 opening/closing files (see openFile008 test).
150[FIX #1744: ignore the byte-order mark at the beginning of a file
151Simon Marlow <[email protected]>**20071130101100]
152[FIX #1914: GHCi forgot all the modules that were loaded before an error
153Simon Marlow <[email protected]>**20071130130734]
154[Update cabal version number used in bootstrapping
155Duncan Coutts <[email protected]>**20071129144259]
156[Change name of DOC_OPTIONS pragma to OPTIONS_HADDOCK
157David Waern <[email protected]>**20071130092506]
158[Update ANNOUNCE for 6.8.2
159Ian Lynagh <[email protected]>**20071128224343]
160[Add 6.8.2 release notes
161Ian Lynagh <[email protected]>**20071128222302]
162[FIX Trac #1935: generate superclass constraints for derived classes
163[email protected]**20071128150541
165 This bug only reports a problem with phantom types, but actually
166 there was quite a long-standing and significant omission in the
167 constraint generation for derived classes.  See
168 Note [Superclasses of derived instance] in TcDeriv.
170 The test deriving-1935 tests both cases.
174[add comment
175Simon Marlow <[email protected]>**20071128111417]
176[FIX #1916: don't try to convert float constants to int in CMM optimizer
177Bertram Felgenhauer <[email protected]>**20071122095513]
178[give a more useful message when the static flags have not been initialised (#1938)
179Simon Marlow <[email protected]>**20071127135435]
180[Avoid making Either String an instance of Monad in the Haddock parser
181David Waern <[email protected]>**20071114204050]
182[FIX BUILD (with GHC 6.2.x): update .hi-boot file
183Simon Marlow <[email protected]>**20071116101227]
184[Fix build
185David Waern <[email protected]>**20071114125842
186 I had forgot to update HaddockLex.hi-boot-6, so the build with 6.2.2
187 failed. This fixes that.
189[Merge from Haddock: Add <<url>> for images
190David Waern <[email protected]>**20071112220537
191 A merge of this patch:
193   Mon Aug  7 16:22:14 CEST 2006  Simon Marlow <[email protected]>
194     * Add <<url>> for images
195     Submitted by: Lennart Augustsson
197 Please merge to the 6.8.2 branch.
199[Merge from Haddock: Modify lexing of /../
200David Waern <[email protected]>**20071112023856
202   Tue Aug 28 11:19:54 CEST 2007  Simon Marlow <[email protected]>
203     * Modify lexing of /../
204     This makes /../ more like '..', so that a single / on a line doesn't
205     trigger a parse error.  This should reduce the causes of accidental
206     parse errors in Haddock comments; apparently stray / characters are
207     a common source of failures.
209 Please merge this to the 6.8.2 branch.
211[Merge from Haddock: allow blank lines inside code blocks
212David Waern <[email protected]>**20071112013439
214   Tue Jan  9 14:14:34 CET 2007  Simon Marlow <[email protected]>
215     * allow blank lines inside a @...@ code block
217 Please merge this to the 6.8.2 branch
219[Merge of a patch from the old Haddock branch:
220David Waern <[email protected]>**20071112013143
222   Fri Jan  5 12:13:41 CET 2007  Simon Marlow <[email protected]>
223     * Fix up a case of extra vertical space after a code block
225 Please merge this to the 6.8.2 branch
227[MERGED: fix stage 1 compilation
228Ian Lynagh <[email protected]>**20071128171145
229 Simon Marlow <[email protected]>**20071106142057
231[MERGED: GHC API: add checkAndLoadModule
232Ian Lynagh <[email protected]>**20071128151112
233 Simon Marlow <[email protected]>**20071106140121
234  Does what the name suggests: it performs the function of both
235  checkModule and load on that module, avoiding the need to process each
236  module twice when checking a batch of modules.  This will make Haddock
237  and ghctags much faster.
239  Along with this is the beginnings of a refactoring of the HscMain
240  interface.  HscMain now exports functions for separately running the
241  parser, typechecher, and generating ModIface and ModDetails.
242  Eventually the plan is to complete this interface and use it to
243  replace the existing one.
245[Fix Trac #1913: check data const for derived types are in scope
246[email protected]**20071121151428
248 When deriving an instance, the data constructors should all be in scope.
249 This patch checks the condition.
253[MERGED: canonicalise the path to HsColour
254Ian Lynagh <[email protected]>**20071127205104
255 Simon Marlow <[email protected]>**20071126141614
257[FIX #1925: the interpreter was not maintaining tag bits correctly
258Simon Marlow <[email protected]>**20071127122614
259 See comment for details
261[Rebuild utils with the stage1 compiler when making a bindist; fixes trac #1860
262Ian Lynagh <[email protected]>**20071127203959
263 This is a bit unpleasant, as "make binary-dist" really shouldn't actually
264 build anything, but it works.
266[Remove the --print-docdir flag
267Ian Lynagh <[email protected]>**20071127195605
268 It wasn't doing the right thing for bindists. Let's rethink...
270[MERGE: Tidy and trim the type environment in mkBootModDetails
271Simon Marlow <[email protected]>**20071123153556
273 Should fix Trac #1833
275 We were failing to trim the type envt in mkBootModDetails, so several
276 functions all called (*), for example, were getting into the interface.
277 Result chaos.  It only actually bites when we do the retyping-loop thing,
278 which is why it's gone so long without a fix.
280[Consistently put www. on the front of haskell.org in URLs
281Ian Lynagh <[email protected]>**20071126215256]
282[Fix some more URLs
283Ian Lynagh <[email protected]>**20071126214147]
284[Tweak some URLs
285Ian Lynagh <[email protected]>**20071126194148]
286[Fix some links
287Ian Lynagh <[email protected]>**20071126184406]
288[Copy gmp stamps into bindists, so we don't try and rebuild gmp
289Ian Lynagh <[email protected]>**20071125211919]
290[On Windows, Delete the CriticalSection's we Initialize
291Ian Lynagh <[email protected]>**20071125125845]
292[On Windows, add a start menu link to the flag reference
293Ian Lynagh <[email protected]>**20071125124429]
294[Remove html/ from the paths we put in the start menu on Windows
295Ian Lynagh <[email protected]>**20071125124150]
296[Make ":" in GHCi repeat the last command
297Ian Lynagh <[email protected]>**20071124231857
298 It used to be a synonym for ":r" in 6.6.1, but this wasn't documented or
299 known about by the developers. In 6.8.1 it was accidentally broken.
300 This patch brings it back, but as "repeat the last command", similar to
301 pressing enter in gdb. This is almost as good for people who want it to
302 reload, and means that it can also be used to repeat commands like :step.
304[Don't make a library documentation prologue
305Ian Lynagh <[email protected]>**20071124211943
306 It's far too large now, and no-one complained when 6.8.1 didn't have one.
308[Don't put package version numbers in links in index.html
309Ian Lynagh <[email protected]>**20071124211629]
310[Define install-strip in Makefile
311Ian Lynagh <[email protected]>**20071124205037]
312[Define install-strip in distrib/Makefile
313Ian Lynagh <[email protected]>**20071124204803]
314[Install gmp from bindists; fixes trac #1848
315Ian Lynagh <[email protected]>**20071124185240]
316[(native gen) fix code generated for GDTOI on x86_32
317Bertram Felgenhauer <[email protected]>**20071121063942
318 See trac #1910.
320[Put library docs in a $pkg, rather than $pkgid, directory; fixes trac #1864
321Ian Lynagh <[email protected]>**20071124171220]
322[Copy the INSTALL hack from mk/config.mk.in into distrib/Makefile-bin-vars.in
323Ian Lynagh <[email protected]>**20071124163028
324 configure will set INSTALL to ./install-sh if it can't find it in the path,
325 so we need to replace the . with the path to our root.
327[Make install-sh executable /before/ we try to find it
328Ian Lynagh <[email protected]>**20071124162450]
329[Set mandir consistently to $(prefix)/share/man, as per trac #1879
330Ian Lynagh <[email protected]>**20071123235933]
331[If we have hscolour then make source code links in teh haddock docs
332Ian Lynagh <[email protected]>**20071123211559]
333[Document --info in the +RTS -? help
334Ian Lynagh <[email protected]>**20071123204352]
335[FIX #1910: fix code generated for GDTOI on x86_32
336Bertram Felgenhauer <[email protected]>*-20071121102627]
337[Add -dcore-lint when validating libraries
338[email protected]**20071105164733]
339[Fix Trac #1909: type of map in docs
340[email protected]**20071120160152]
341[Two small typos in the flags summary (merge to 6.8 branch)
342[email protected]**20071119134639]
343[FIX #1910: fix code generated for GDTOI on x86_32
344Bertram Felgenhauer <[email protected]>**20071121102627]
345[FIX #1847 (improve :browse! docs, fix unqual)
346[email protected]**20071108013147
348 - add example to docs, explain how to interpret
349   output of `:browse! Data.Maybe`
350 - print unqualified names according to current
351   context, not the context of the target module
354[FIX Trac #1825: standalone deriving Typeable
355[email protected]**20071120125732
357 Standalone deriving of typeable now requires you to say
358        instance Typeable1 Maybe
359 which is exactly the shape of instance decl that is generated
360 by a 'deriving( Typeable )' clause on the data type decl.
362 This is a bit horrid, but it's the only consistent way, at least
363 for now.  If you say something else, the error messages are helpful.
365 MERGE to 6.8 branch
368[FIX Trac #1806: test for correct arity for datacon in infix pattern patch
369[email protected]**20071119114301
371 Happily the fix is easy; pls merge
374[MERGED: FIX #1715: egregious bug in ifaceDeclSubBndrs
375Ian Lynagh <[email protected]>**20071122123551
376 [email protected]**20071120111723
378  ifaceDeclSubBndrs didn't have an IfaceSyn case; but with type
379  families an IfaceSyn can introduce subordinate binders.  Result:
380  chaos.
382  The fix is easy though.  Merge to 6.8 branch.
384[Improve the situation for Trac #959: civilised warning instead of a trace msg
385[email protected]**20071119122938
387 This doesn't fix the root cause of the bug, but it makes the report
388 more civilised, and points to further info.
391[MERGED: Try to manage the size of the text rendered for ':show bindings'
392Pepe Iborra <[email protected]>**20071119111243
393 Pepe Iborra <[email protected]>**20071114231601] {
395[MERGED: wibble
396Pepe Iborra <[email protected]>**20071119111109
397   Pepe Iborra <[email protected]>**20071114233356] {
399[MERGED: Make the Term ppr depth aware
400Ian Lynagh <[email protected]>**20071119111053
401 Pepe Iborra <[email protected]>**20071114183417] {
403[MERGED: GHCi debugger: added a new flag, -fno-print-binding-contents
404Pepe Iborra <[email protected]>**20071119105634]
405[MERGED: Print binding contents in :show bindings
406Ian Lynagh <[email protected]>**20071118173321
407 Pepe Iborra <[email protected]>**20071006123952] {
409[MERGE: Zonk quantified tyvars with skolems + Rejig the error messages a bit; fixes a minor bug
410[email protected]**20071121143346
412 This merges to the 6.8 branch two HEAD patches (names above).  The
413 former "Zonk..." was for some reason partly in the branch already,
414 which is what led to the complications.
416 Here are the original patches:
418 Fri Oct 19 12:56:53 BST 2007  Manuel M T Chakravarty <[email protected]>
419   * Zonk quantified tyvars with skolems
421   We used to zonk quantified type variables to regular TyVars.  However, this
422   leads to problems.  Consider this program from the regression test suite:
424     eval :: Int -> String -> String -> String
425     eval 0 root actual = evalRHS 0 root actual
427     evalRHS :: Int -> a
428     evalRHS 0 root actual = eval 0 root actual
430   It leads to the deferral of an equality
432     (String -> String -> String) ~ a
434   which is propagated up to the toplevel (see TcSimplify.tcSimplifyInferCheck).
435   In the meantime `a' is zonked and quantified to form `evalRHS's signature.
436   This has the *side effect* of also zonking the `a' in the deferred equality
437   (which at this point is being handed around wrapped in an implication
438   constraint).
440   Finally, the equality (with the zonked `a') will be handed back to the
441   simplifier by TcRnDriver.tcRnSrcDecls calling TcSimplify.tcSimplifyTop.
442   If we zonk `a' with a regular type variable, we will have this regular type
443   variable now floating around in the simplifier, which in many places assumes to
444   only see proper TcTyVars.
446   We can avoid this problem by zonking with a skolem.  The skolem is rigid
447   (which we requirefor a quantified variable), but is still a TcTyVar that the
448   simplifier knows how to deal with.
450 Thu Nov  1 17:50:22 GMT 2007  [email protected]
451   * Rejig the error messages a bit; fixes a minor bug
453   The type checker was only reporting the first message if an equality
454   failed to match.  This patch does a bit of refactoring and fixes the
455   bug, which was in the bogus use of eqInstMisMatch
456   in tcSimplify.report_no_instances.b
458   This is really a bug in 6.8 too, so this would be good to merge across
459   to the 6.8 branch.
463[MERGE to 6.8: Refactor error recovery slightly
466 Slightly rejigged from HEAD:
468   Mostly this patch is refacoring, but it also avoids post-tc zonking if
469   the typechecker found errors.  This is good because otherwise with
470   DEBUG you can get the "Inventing strangely-kinded TyCon" warning.
473[Improve pretty-printing for HsSyn
475[warning removal
476Simon Marlow <simonmar@microsoft.com>**20071009105138]
477[TcTyFuns: remove some duplicate code
478Manuel M T Chakravarty <chak@cse.unsw.edu.au>**20071004142315]
479[Fix deferring on tyvars in TcUnify.subFunTys
480Manuel M T Chakravarty <chak@cse.unsw.edu.au>**20071018044352]
481[TcUnify.subFunTys must take type families into account
482Manuel M T Chakravarty <chak@cse.unsw.edu.au>**20071017114326
483 * A bug reported by Andrew Appleyard revealed that subFunTys did take
484   neither type families nor equalities into account.  In a fairly obscure
485   case there was also a coercion ignored.
488Manuel M T Chakravarty <chak@cse.unsw.edu.au>**20071003145715]
489[FIX #1791 on stable branch
490Simon Marlow <simonmar@microsoft.com>**20071114163945
491 The 6.8 branch needed this fix in another place
493[remove an incorrect assertion
494Simon Marlow <simonmar@microsoft.com>**20071016151829]
495[remove --define-name from the --help usage message (#1596)
496Simon Marlow <simonmar@microsoft.com>**20071114153417
499[Attempt at fixing #1873, #1360
500Simon Marlow <simonmar@microsoft.com>**20071116152148
502 I think I figured out a reasonable way to manage the GHCi context,
503 comments welcome.
505 Rule 1: external package modules in the context are persistent.  That
506 is, when you say 'import Data.Maybe' it survives over :load, :add,
507 :reload and :cd.
509 Rule 2: :load and :add remove all home-package modules from the
510 context and add the rightmost target, as a *-module if possible.  This
511 is as before, and makes sense for :load because we're starting a new
512 program; the old home-package modules don't make sense any more.  For
513 :add, it usually does what you want, because the new target will
514 become the context.
516 Rule 3: any modules from the context that fail to load during a
517 :reload are remembered, and re-added to the context at the next
518 successful :reload.
520 Claus' suggestion about adding the "remembered" modules to the prompt
521 prefixed with a ! is implemented but commented out.  I couldn't
522 decide whether it was useful or confusing.
524 One difference that people might notice is that after a :reload where
525 there were errors, GHCi would previously dump you in the most recent
526 module that it loaded.  Now it dumps you in whatever subset of the
527 current context still makes sense, and in the common case that will
528 probably be {Prelude}.
530[on Windows, install to a directory with spaces (test for #1828)
531Simon Marlow <simonmar@microsoft.com>**20071115155327]
532[FIX #1828: installing to a patch with spaces in
533Simon Marlow <simonmar@microsoft.com>**20071115155747
534 We have to pass the path to gcc when calling windres, which itself
535 might have spaces in.  Furthermore, we have to pass the path to gcc's
536 tools to gcc.  This means getting the quoting right, and after much
537 experimentation and reading of the windres sources I found something
538 that works: passing --use-temp-files to windres makes it use its own
539 implementation of quoting instead of popen(), and this does what we
540 want.  Sigh.
542[FIX #1679: crash on returning from a foreign call
543Simon Marlow <simonmar@microsoft.com>**20071115131635
544 We forgot to save a pointer to the BCO over the foreign call.  Doing
545 enough allocation and GC during the call could provoke a crash.
547[Make pprNameLoc more robust in absence of loc information
548Pepe Iborra <mnislaih@gmail.com>**20071114233343]
549[Wibble to fix Trac #1901 (shorten messsage slightly)
551[FIX Trac #1901: check no existential context in H98 mode
553[Improve documentation of data type declarations (Trac #1901)
555[Improve links from flag reference to the relevant section; and improve doc of RankN flags
557[Documentation only - fix typo in flags reference
558Tim Chevalier <chevalier@alum.wellesley.edu>**20071115055748]
559[Make SpecConstr work again
562 In a typo I'd written env instead of env', and as a result RULES are
563 practically guaranteed not to work in a recursive group.  This pretty
564 much kills SpecConstr in its tracks!
566 Well done Kenny Lu for spotting this.  The fix is easy.
568 Merge into 6.8 please.
573[Accept x86_64-*-freebsd* as well as amd64-*-freebsd* in configure.ac
574Ian Lynagh <igloo@earth.li>**20071117154502
575 Patch from Brian P. O'Hanlon
577[FIX #1837: emit deprecated message for unversioned dependencies
578Simon Marlow <simonmar@microsoft.com>**20071114153010]
579[FIX Trac 1888; duplicate INLINE pragmas
582 There are actually three things here
583 - INLINE pragmas weren't being pretty-printed properly
584 - They were being classified into too-narrow boxes by eqHsSig
585 - They were being printed in to much detail by hsSigDoc
587 All easy.  Test is rnfail048.
590[Fix Trac 1865: GHCi debugger crashes with :print
591Pepe Iborra <mnislaih@gmail.com>**20071113170113]
592[FIX Trac 1662: actually check for existentials in proc patterns
595 I'd fixed the bug for code that should be OK, but had forgotten to
596 make the test for code that should be rejected!
598 Test is arrowfail004
601[FIX #1653 (partially): add -X flags to completion for :set
602Simon Marlow <simonmar@microsoft.com>**20071113153257]
603[Fix #782, #1483, #1649: Unicode GHCi input
604Simon Marlow <simonmar@microsoft.com>**20071114151411
605 GHCi input is now treated universally as UTF-8, except for the Windows
606 console where we do the correct conversion from the current code
607 page (see System.Win32.stringToUnicode).
609 That leaves non-UTF-8 locales on Unix as unsupported, but (a) we only
610 accept source files in UTF-8 anyway, and (b) UTF-8 is quite ubiquitous
611 as the default locale.
614[comments only: point to relevant bug reports
615Simon Marlow <simonmar@microsoft.com>**20070924103323]
616[Remove ex-extralibs from libraries/Makefile
617Ian Lynagh <igloo@earth.li>**20071111213618]
618[Remove the X11 and HGL libraries from extralibs
619Ian Lynagh <igloo@earth.li>**20071111213447
620 Don Stewart, X11 maintainer, requested we remove X11, and HGL depends on it
621 on Linux (and we don't try to build HGL on Windows).
623[catch up with removal of config.mk in the readline package
624Simon Marlow <simonmar@microsoft.com>**20071107095952]
625[TAG 2007-11-11
626Ian Lynagh <igloo@earth.li>**20071111164452]
627Patch bundle hash: