New patches: [Fix freeHaskellFunctionPtr for Darwin/i386 Aaron Tomb **20071029202636] { hunk ./rts/Adjustor.c 1032 -#elif defined(x86_HOST_ARCH) && defined(darwin_HOST_OS) +#elif defined(i386_HOST_ARCH) && defined(darwin_HOST_OS) } Context: [Fix build with GHC < 6.4 Ian Lynagh **20071029183155] [Set interfacedir (using $topdir, not $httptopdir) Ian Lynagh **20071029174825] [Teach ghc-pkg about $httptopdir Ian Lynagh **20071029161130] [Fill in holes in the release notes Ian Lynagh **20071028225253] [installPackage needs to treat $httptopdir the same as $topdir Ian Lynagh **20071028134534] [Define and use $httptopdir for the haddock docs locations Ian Lynagh **20071028123552] [We need to copy .buildinfo files into the bindists Ian Lynagh **20071028131752] [Fix imports Ian Lynagh **20071027231844] [Fix the build with GHC < 6.4 (foldl1' didn't exist) Ian Lynagh **20071027210526] [MERGED: Comments only Ian Lynagh **20071027170004 Sat Oct 27 08:46:42 PDT 2007 simonpj@microsoft.com ] [Make 'improvement' work properly in TcSimplify simonpj@microsoft.com**20071027155459 (Please merge this, and the preceding handful from me to the 6.8 branch.) This patch fixes a serious problem in the type checker, whereby TcSimplify was going into a loop because it thought improvement had taken place, but actually the unificataion was actually deferred. We thereby fix Trac #1781, #1783, #1795, and #1797! In fixing this I found what a mess TcSimplify.reduceContext is! We need to fix this. The main idea is to replace the "improvement flag" in Avails with a simpler and more direct test: have any of the mutable type variables in the (zonked) 'given' or 'irred' constraints been filled in? This test uses the new function TcMType.isFilledMetaTyVar; the test itself is towards the end of reduceContext. I fixed a variety of other infelicities too, and left some ToDos. ] [An implication constraint can abstract over EqInsts simonpj@microsoft.com**20071027155433] [In an AbsBinds, the 'dicts' can include EqInsts simonpj@microsoft.com**20071027154903 An AbsBinds abstrats over evidence, and the evidence can be both Dicts (class constraints, implicit parameters) and EqInsts (equality constraints). So we need to - use varType rather than idType - use instToVar rather than instToId - use zonkDictBndr rather than zonkIdBndr in zonking It actually all worked before, but gave warnings. ] [More notes simonpj@microsoft.com**20071027154702] [Add anyM to IOEnv simonpj@microsoft.com**20071027154551] [Add a note to NOTES simonpj@microsoft.com**20071027100220] [Adding hpc documentation about sum and map, push to STABLE. andy@unsafeperformio.com**20071025050341] [Fixing typo in runtime documentation for hpc, push to stable andy@unsafeperformio.com**20071025045456] [default to installing runhaskell and hsc2hs again, but provide knobs to turn them off Simon Marlow **20071025084222] [Fix a whole heap of speling errrs in the docs Josef Svenningsson **20071007213858] [We need to install-docs when making the Windows bindist Ian Lynagh **20071027144333] [We need to set _way=* in rts/ both when making and installing bindists Ian Lynagh **20071027142914 This is a hack, but it means we get libHSrts*.a etc rather than just libHSrts.a. ] [arrows is no longer an extralib Ian Lynagh **20071027123656] [Only build/install the man page if XSLTPROC is defined Ian Lynagh **20071027122155] [install the Cabal docs, and make them show up in a binary distribution Simon Marlow **20071026122456] [cp => $(CP) Simon Marlow **20071026111054] [get rid of the html subdirectory under share/doc/ghc/users_guide Simon Marlow **20071026110919] [fix the links in the library documentation index Simon Marlow **20071025152245] [binary-dist: Makefile-vars needs HADDOCK_DOCS=YES Simon Marlow **20071025135816] [FIX #1791: fail with out-of-heap when allocating more than the max heap size in one go Simon Marlow **20071024095420 Normally the out-of-heap check is performed post-GC, but there are cases where we can detect earlier that we definitely have exhausted the heap size limit. ] [patch from #1782; fixes check-packages target on Solaris Simon Marlow **20071022133337] [Refer to "boot" libs, not "core" libs Ian Lynagh **20070914132658] [Fix more warnings Simon Marlow **20071023131351] [FIX BUILD (on 32-bit platforms): hs_hpc_module() type mismatch Simon Marlow **20071023082233] [second attempt to fix C compiler warnings with -fhpc Simon Marlow **20071019133243 The hs_hpc_module() prototype in RtsExternal.h didn't match its usage: we were passing StgWord-sized parameters but the prototype used C ints. I think it accidentally worked because we only ever passed constants that got promoted. The constants unfortunately were sometimes negative, which caused the C compiler to emit warnings. I suspect PprC.pprHexVal may be wrong to emit negative constants in the generated C, but I'm not completely sure. Anyway, it's easy to fix this in CgHpc, which is what I've done. ] [Change some ints to unsigned ints Simon Marlow **20071018095503 Fixes some gratuitous warnings when compiling via C with -fhpc ] [Fix typo that prevented zonking of rhs of EqInsts Manuel M T Chakravarty **20071018131040 MERGE TO STABLE ] [Fix ghc package in bindists; it wasn't adding the depenedency on readline Ian Lynagh **20071024120633] [Correct a comment Ian Lynagh **20071024114549] [Fix installing the ghc package .hi files in a bindist Ian Lynagh **20071024114219] [Build the manpage when building, not when installing Ian Lynagh **20071024112914] [Hack to make sure we get all the RTS ways in bindists Ian Lynagh **20071024004155] [Fix installing the documentation in the bindists Ian Lynagh **20071023234624] [FIX #1784: EM_AMD64 and EM_X86_64 might both be defined to the same value Simon Marlow **20071019110223] [fix -fbreak-on-exception for unregsterised Simon Marlow **20071018110621] [fix :print when !tablesNextToCode Simon Marlow **20071018105340] [fix breakpoints in unregisterised mode Simon Marlow **20071018101929] [implement #1468, :browse on its own uses the currently-loaded module Simon Marlow **20071019115751] [warning removal Simon Marlow **20071003172715] [FIX #1450: asynchronous exceptions are now printed by +RTS -xc Simon Marlow **20071018134951] [recordMutable: test for gen>0 before calling recordMutableCap Simon Marlow **20071017125657 For some reason the C-- version of recordMutable wasn't verifying that the object was in an old generation before attempting to add it to the mutable list, and this broke maessen_hashtab. This version of recordMutable is only used in unsafeThaw#. ] [fix warning on Windows Simon Marlow **20071017121645] [Improving the combine mode for hpc andy@galois.com**20071014171009 we now have Processing Coverage files: sum Sum multiple .tix files in a single .tix file combine Combine two .tix files in a single .tix file map Map a function over a single .tix file Where sum joins many .tix files, combine joins two files (with extra functionality possible), and map just applied a function to single .tix file. These changes were improvements driven by hpc use cases. END OF DESCRIPTION*** Place the long patch description above the ***END OF DESCRIPTION*** marker. The first line of this file will be the patch name. This patch contains the following changes: M ./utils/hpc/Hpc.hs -1 +3 M ./utils/hpc/HpcCombine.hs -33 +84 M ./utils/hpc/HpcFlags.hs -11 +59 ] [FIX #1759 while respecting the ticks andy@galois.com**20071015033319] [Don't barf on error message with non-tc tyvars Manuel M T Chakravarty **20071018060336 MERGE TO STABLE ] [Fix #1662: do not simplify constraints for vanilla pattern matches simonpj@microsoft.com**20071016124710 See Note [Arrows and patterns] in TcPat. This fixes Trac 1662. Test is arrows/should_compile/arrowpat.hs Please merge ] [Some more traceTcs Manuel M T Chakravarty **20070929121941] [Tell Cabal it's version number when we build it. Duncan Coutts **20071018094742] [First draft of 6.8.1 release notes Ian Lynagh **20071016150731] [-ftype-families -> -XTypeFamilies Ian Lynagh **20071016123010] [Fix #1709: do not expose the worker for a loop-breaker simonpj@microsoft.com**20071016131840 The massive 'Uni' program produced a situation in which a function that had a worker/wrapper split was chosen as a loop breaker. If the worker is exposed in the interface file, then an importing module may go into an inlining loop: see comments on TidyPgm.tidyWorker. This patch fixes the inlining bug. The code that gives rise to this bizarre case is still not good (it's a bunch of implication constraints and we are choosing a bad loop breaker) but the first thing is to fix the bug. It's rather hard to produce a test case! Please merge to the 6.8 branch. ] [Fix DoCon: Another try at getting extractResults right simonpj@microsoft.com**20071012162325 For some reason TcSimplify.extractResults is quite difficult to get right. This is another attempt; finally I think I have it. Strangely enough, it's only Sergey's DoCon program that shows up the bug, which manifested as a failure in the Simplifier lookupRecBndr $dGCDRing{v a1Lz} [lid] But it was due to extractResults producing multiple bindings for the same dictionary. Please merge this to the stable branch (after previous patches to TcSimplify though). ] [Fix Trac #1759: do not let ticks get in the way of spotting trivially-true guards simonpj@microsoft.com**20071010164731 GHC spots that an 'otherwise' guard is true, and uses that knowledge to avoid reporting spurious missing-pattern or overlaps with -Wall. The HPC ticks were disguising the 'otherwise', which led to this failure. Now we check. The key change is defining DsGRHSs.isTrueLHsExpr. Test is ds062 ] [comment-out "use vars" in 3 places (see #1739) Simon Marlow **20071008115740] [Improve pretty-printing of splices in HsSyn simonpj@microsoft.com**20071010123726] [we need to #include "Stg.h" first, we can't rely on GHC to inject it Simon Marlow **20071010153244 This fixes the unreg build, and in general building the RTS code via-C. I'm not sure at what stage this was broken, but I think it was working accidentally before. ] [Fix Trac #1680; check for unboxed tuples in TcType.marshalableTyCon simonpj@microsoft.com**20071011123426] [FIX BUILD (when compiling base via C): declare n_capabilities Simon Marlow **20071010103704] [MERGED: FIX #1681: withBreakAction had too large a scope in runStmt Ian Lynagh **20071012151253 Simon Marlow **20071010085820 ] [Update documentation for win32 DLL linking Clemens Fruhwirth **20071010074415] [Fix Trac #1746: make rule-matching work properly with Cast expressions simonpj@microsoft.com**20070929104406 The Cast case of the rule-matcher was simply wrong. This patch fixes it; see Trac #1746. I also fixed the rule generation in SpecConstr to generate a wild-card for the cast expression, which we don't want to match on. This makes the rule more widely applicable; it wasn't the cause of the bug. ] [export n_capabilities, see #1733 Simon Marlow **20071009142701] [export stopTimer(), we need this in the unix package Simon Marlow **20070912200057] [FIX #1743, create a fresh unique for each Id we bind at a breakpoint Simon Marlow **20071009142554] [also acquire/release task->lock across fork() Simon Marlow **20070927091331 further attempt to fix #1391 on MacOS ] [attempt to fix #1391, hold locks across fork() and initialize them in the child Simon Marlow **20070914145519] [also call initMutex on every task->lock, see #1391 Simon Marlow **20071009122409] [error message fix (#1758) Simon Marlow **20071008134958] [FIX validate for PPC Mac OS X - RegAllocStats.hs Thorkil Naur **20071005144105] [FIX validate for PPC Mac OS X - RegAllocLinear.hs Thorkil Naur **20071005143607] [FIX validate for PPC Mac OS X - Linker.c Thorkil Naur **20071005144908] [FIX validate for PPC Mac OS X - Evac.h Thorkil Naur **20071005144454] [MERGED: FIX #1748: -main-is wasn't handling the case of a single hierarchical module Ian Lynagh **20071012143757 Simon Marlow **20071008131305 test case is driver062.5 ] [FIX: add missing case to OccName.isSymOcc David Waern **20071002143459] [mention what SCC stands for Simon Marlow **20071011135736] [Fix Trac #1755; check for stage errors in TH quoted Names simonpj@microsoft.com**20071010150250 There are a number of situations in which you aren't allowed to use a quoted Name in a TH program, such as \x -> 'x But we weren't checking for that! Now we are. Merge to stable branch. Test is TH_qname. ] [Fix Trac #1678; be more careful about catching and reporting exceptions in spliced TH monadic computations simonpj@microsoft.com**20071010145705 Many of the new lines are comments to explain the slightly-convoluted in which exceptions get propagated out of the Q monad. This fixes Trac 1679; test is TH_runIO (as well as the exising TH_fail). Please merge ] [FIX: tidy up TcSimplify following equality constraints additions simonpj@microsoft.com**20071010093334 The combination of "type refinement" for GADTs and the new equality constraints has made TcSimplify rather complicated. And wrong: it generated bogus code for cholewo-eval. This patch is still far from entirely satisfactory. There are too many places where invariants are unclear, and the code is still a bit of a mess. But I believe it's better, and it passes the regression tests! So I think it's good enough for the 6.8 release. Please merge. The main changes are: - get rid of extractLocalResults (which was always suspicious) - instead, treat the 'refinement' along with 'givens', by adding a field to RedEnv, red_reft which travels with red_givens - I also reworked extractResults a bit, which looked wrong to me This entailed changing the Given constructor in Avail to take an Inst rather than a TcId ] [Comments and debug output only simonpj@microsoft.com**20070927110842] [mkIfaceExports: sort the children of AvailTC Simon Marlow **20071002114917 This fixes a problem with spurious recompilations: each time a module was recompiled, the order of the children would change, causing extra recompilation. MERGE TO STABLE ] [FIX BUILD FD_SETSIZE signed jochemberndsen@dse.nl**20070927132649 On FreeBSD FD_SETSIZE is unsigned. Cast it to a signed int for portability. ] [FIX BUILD addDLL returns const char* jochemberndsen@dse.nl**20070927132619 addDLL returns const char*, not just a char*. Fix compiler warning ] [FIX BUILD `set -o igncr'-issue on FreeBSD jochemberndsen@dse.nl**20070926203750 `set -o igncr' does not work on non-cygwin-systems. Fail silently if this command does not work, instead of aborting the build. ] [Change DOCOPTIONS pragma to DOC_OPTIONS David Waern **20071002143849 MERGE TO STABLE ] [FIX: parsing of doc options David Waern **20071002143713 Lexing of the doc options pragma was changed, but but no change was made to the parser to reflect that. This patch fixes this problem. MERGE TO STABLE ] [FIX: mkWWcpr takes open alg types into account Manuel M T Chakravarty **20071002130407 - This fixed the failures of GMapAssoc and GMapTop for optmising ways MERGE TO STABLE ] [FIX #1738: KPush rule of FC must take dataConEqTheta into account Manuel M T Chakravarty **20071001154343 MERGE TO STABLE ] [FIX #1729: Don't try to expand syn families with -XLiberalTypeSynonyms Manuel M T Chakravarty **20070929122624 MERGE TO STABLE ] [FIX: Make boxy splitters aware of type families Manuel M T Chakravarty **20070928225541 MERGE TO STABLE ] [Finally, I managed to squash an infamous bug in :print Pepe Iborra **20070927151300 It turns out the newtype handling code in :print was slipping non mutable Tyvars in the types reconstructed. The error message eventually produced was rather obscure: [src/Tp.hs:75:28-64] *MainTp> :p x *** Exception: No match in record selector Var.tcTyVarDetails [src/Tp.hs:75:28-64] *MainTp> Due to non mutable tyvars, unifyType was failing. A well placed assertion in the unifyType code would have made my life much easier. Which reminds me I should install a -ddump-* system in the RTTI subsystem, or future hackers will run away in swearing. MERGE TO STABLE ] [Be a bit more flexible in terminal identification for do_bold Pepe Iborra **20070927141549 In Os X for instance, by default we have TERM=xterm-color MERGE TO STABLE ] [further stub filename fix: I managed to break non-stubdir -fvia-C compilation Simon Marlow **20070927102539] [MERGED: html_installed_root shouldn't contain $$pkgid Ian Lynagh **20070927131012 This actually didn't break anything, as the shell expanded $pkgid to the empty string, but it was still wrong. ] [esacpe backslashes in the filename in the .rc file Simon Marlow **20070912111658] [FIX -stubdir bug: the .hc file was #including the wrong _stub.h filename Simon Marlow **20070926134539 Using -stubdir together with hierarchical modules, -fvia-C, and --make is essentially broken in 6.6.x. Recently discovered by Cabal's use of -stubdir. Test cases: driver027/driver028 (I've updated them to use -fvia-C, in order to test for this bug). ] [fix #1734, panic in :show modules after load failure Simon Marlow **20070926100732] [Fixing #1340, adding HPC Documentation andy@galois.com**20070926055331] [Be more specific about file-header pragmas simonpj@microsoft.com**20070924162847 Document the rules for pragmas that must occur at the top of a file. Please merge this patch ] [When haddocking, other packages are in ../$$pkgid not ../$$pkg Ian Lynagh **20070923193344] [Unbreak "dist" target for fresh trees sven.panne@aedion.de**20070923094358 The previous hack to include Parser.hs in source distros broke the possibility of doing a "make dist" in a fresh tree, i.e. one which has just been checked out and configured, but *not* built. Of course you will need Happy for such a source distro later, but at least the freedom to do this is important. The ultimate goal should be that something like "make dist" will work in a freshly checked out tree, with no prerequisite steps (this is very common in most projects). We should move towards that goal, not away from it... MERGE TO STABLE ] [Notice when C modules have changed when deciding whether or not to link Ian Lynagh **20070923181620 Based on a patch from Lemmih ] [Whitespace changes only Ian Lynagh **20070923174242] [Remove remaining bits of bindist "make in-place" Ian Lynagh **20070923163640] [Add an entry for strings treated as ISO-8859-1 to the users guide bug list Ian Lynagh **20070923131845] [Define and use PprTyThing.pprTypeForUser simonpj@microsoft.com**20070911085123 When printing types for the user, the interactive UI often wants to leave foralls implicit. But then (as Claus points out) we need to be careful about name capture. For example with this source program class C a b where op :: forall a. a -> b we were erroneously displaying the class in GHCi (with suppressed foralls) thus: class C a b where op :: a -> b which is utterly wrong. This patch fixes the problem, removes GHC.dropForAlls (which is dangerous), and instead supplies PprTyThing.pprTypeForUser, which does the right thing. ] [MERGED: Move OPTIONS pragmas above comments Ian Lynagh **20070921205741] [Fix building with compilers which don't have an integer for a patch level Ian Lynagh **20070921165316] [Removed duplicate entry for derbugging flag -ddump-tc from the user guide v.dijk.bas@gmail.com**20070920121306] [make stamp.inplace-gcc-lib copy $(LD) instead of $(GccDir)ld.exe to avoid building problem shelarcy **20070920130159] [Overhaul of the rewrite rules Manuel M T Chakravarty **20070915074119 - Cleaned up and simplified rules - Added many missing cases - The rules OccursCheck and swap have been eliminated and integrate with the other rules; ie, Subst and Unify perform the occurs check themselves and they can deal with left-to-right and right-to-left oriented rewrites. This makes the code simpler and more efficient. - Also added comments. ] [MERGED: GHCi debugger: Added a -fprint-evld-with-show flag Ian Lynagh **20070920200212 Pepe Iborra **20070912153505 The flag enables the use of Show instances in :print. By default they are not used anymore ] [Remove a redundant part of distrib/Makefile's "make install" Ian Lynagh **20070920183903] [FIX #1677; poor error message for misspelled module declaration Simon Marlow **20070911085452] [Fix Trac #1718: interaction of error, unlifted tuples, and casts simonpj@microsoft.com**20070920120049 See Note [Float coercions (unlifted)] in Simplify Test is simpl018 ] [Fix --print-docdir for relocatable builds; fixes #1226 Ian Lynagh **20070919140100] [Use $(RelocatableBuild) rather than $(Windows) Ian Lynagh **20070919140037] [Fix exponential-time behaviour with type synonyms; rename -XPartiallyAppliedTypeSynonyms to -XLiberalTypeSynonyms simonpj@microsoft.com**20070919171207 Fixes exponential behaviour present in GHC 6.6! I renamed the flag because the old (not very old) name wasn't describing what it does. ] [FIX #1688: Givens in checkLoop are not that rigid after all Manuel M T Chakravarty **20070919150738 - This patch re-instates the policy of 6.6.1 to zonk the given constraints in the simplifier loop. MERGE TO STABLE ] [FIX #1713: watch out for type families in splitAppTy functions Manuel M T Chakravarty **20070919122011 MERGE TO STABLE ] [Make the error message mentioning -XPatternSignatures spell the flag correctly simonpj@microsoft.com**20070919142821] [Hack to get haskell-src's Parser.hs into source distributions Ian Lynagh **20070919194222] [Catch any exceptions thrown by getEnv; fixes #1704 Ian Lynagh **20070918215752] [Give Windows its own definitions of docdir (and htmldir etc) Ian Lynagh **20070918201249] [Include build.mk or validate{,-settings}.mk through custom-settings.mk Ian Lynagh **20070918200637] [Install extra-gcc-opts in bindists Ian Lynagh **20070918194839] [Writing out .tix file only if you are the original process, not a child. andy@galois.com**20070917230641 This lets us gain coverage on programs that use fork, like xmonad. * To be merged to STABLE * ] [removing the functions hs_hpc_read and hs_hpc_write inside Hpc.c, they are dead code. andy@galois.com**20070917230201 * To be merged to STABLE. * ] [Add validate.mk to .darcs-boring Ian Lynagh **20070918184509] [avoid platform dependencies: my_uintptr_t ==> StgWord Simon Marlow **20070917120156] [add support for EM_AMD64 elf machine type, openbsd/amd64 ghci works Don Stewart **20070916034845] [patch Linker.c for amd64/openbsd Don Stewart **20070916012728] [Merge of register allocator patches. Ben.Lippmeier@anu.edu.au**20070918102256 This is a conflict avoiding merge of several patches which finish off the new register allocator. After this patch all tests should go through with both -fregs-graph and -fregs-iterative. Mon Sep 17 14:26:14 BST 2007 Ben.Lippmeier@anu.edu.au * Tune coalescing in non-iterative register allocator If iterative coalescing isn't turned on, then do a single aggressive coalescing pass for the first build/color cycle and then back off to conservative coalescing for subseqent passes. Aggressive coalescing is a cheap way to eliminate lots of the reg-reg moves, but it can make the graph less colorable - if we turn it on for every pass then allocation for code with a large amount of register pressure (ie SHA1) doesn't converge in a sensible number of cycles. Mon Sep 17 13:42:41 BST 2007 Ben.Lippmeier@anu.edu.au * Bugfix to iterative coalescer Coalescences must be applied to the unsimplified graph in the same order they were found by the iterative coalescing algorithm - otherwise the vreg rewrite mapping will be wrong and badness will ensue. Mon Sep 17 12:30:24 BST 2007 Ben.Lippmeier@anu.edu.au * Add -dasm-lint When -dasm-lint is turned on the register conflict graph is checked for internal consistency after each build/color pass. Make sure that all edges point to valid nodes, that nodes are colored differently to their neighbours, and if the graph is supposed to be colored, that all nodes actually have a color. Fri Sep 14 17:42:34 BST 2007 Ben.Lippmeier@anu.edu.au * Count CmmTops processed so far in the native code generator To help with debugging / nicer -ddump-asm-regalloc-stages Fri Sep 14 17:14:08 BST 2007 Ben.Lippmeier@anu.edu.au * Change spill cost function back to inverse length of live range. On further testing it turns out that using Chaitin's spill cost formula of counting the number of uses of a var and then dividing by the degree of the node isn't actually a win. Perhaps because this isn't counting the number of actual spills inserted in the output code. This would be worth revisiting if other work manages to increase the register pressure in the native code. Fri Sep 14 10:39:07 BST 2007 Ben.Lippmeier@anu.edu.au * Better cleaning of spills in spill cleaner Track what slots each basic block reloads from. When cleaning spill instructions we can use this information to decide whether the slot spilled to will ever be read from on this path. Thu Sep 13 16:54:07 BST 2007 Ben.Lippmeier@anu.edu.au * Better calculation of spill costs / selection of spill candidates. Use Chaitin's formula for calculation of spill costs. Cost to spill some vreg = (num writes + num reads) / degree of node With 2 extra provisos: 1) Don't spill vregs that live for only 1 instruction. 2) Always prefer to spill vregs that live for a number of instructions more than 10 times the number of vregs in that class. Proviso 2 is there to help deal with basic blocks containing very long live ranges - SHA1 has live ranges > 1700 instructions. We don't ever try to keep these long lived ranges in regs at the expense of others. Because stack slots are allocated from a global pool, and there is no slot coalescing yet, without this condition the allocation of SHA1 dosn't converge fast enough and eventually runs out of stack slots. Prior to this patch we were just choosing to spill the range with the longest lifetime, so we didn't bump into this particular problem. ] [Keep valgrind happy when calling timer_create sven.panne@aedion.de**20070916111927 Fill all of the sigevent structure with zeroes before individual fields are set. Although not strictly necessary, this keeps tools like valgrind from complaining about passing uninitialized data, which is a good thing. MERGE TO STABLE ] [Make DESTDIR work again sven.panne@aedion.de**20070916084339 installPackage is a horror, using --force is probably not the perfect way to make DESTDIR functionality work again, but given the current state of affairs, I can't find a quick and clean solution. MERGE TO STABLE ] [Clean stage/ghc-inplace.c Ian Lynagh **20070917094300] [MERGED: Resurrect the "lib" subdirectory in the installation tree, it was somehow lost Ian Lynagh **20070917094212 Sun Sep 16 09:41:22 BST 2007 sven.panne@aedion.de * Resurrect the "lib" subdirectory in the installation tree, it was somehow lost Having all package directory directly in the toplevel installation directory would be a bad thing because of potential name clashes, and it wasn't done this way earlier, so reinstantiate the old layout. MERGE TO STABLE ] [Unbreak "clean" and "distclean" targets when there is no testsuite sven.panne@aedion.de**20070915085845] [distclean: extra-gcc-opts, testsuite Simon Marlow **20070914125444] [Another attempt at getting bindists working everywhere Ian Lynagh **20070916005328] [MERGED: Better modelling of newtypes in the Term datatype Ian Lynagh **20070914203411 Pepe Iborra **20070912165855 This helps to get pretty printing right, nested newtypes were not being shown correctly by :print ] [Use sed to make stage-specific ghc-inplace.c's Ian Lynagh **20070914164821 When trying to do it with -D we were having trouble getting the double quotes through in all Windows configurations. ] [Remove --export-all-symbols for DLL linking, it is default and prevents us from using .def files Clemens Fruhwirth **20070831104624] [Fix copy+paste-o, spotted by Simon Marlow Ian Lynagh **20070914142650] [Add a --names-only flag for list --simple-output Ian Lynagh **20070907181944 We use this in the testsuite to find out which libraries we should run the tests from. ] [fix install-docs for non-html docs Simon Marlow **20070914093738] [Remove call to validateGraph that snuck in in last patch Ben.Lippmeier@anu.edu.au**20070914092504 Also some comment / var name wibbles ] [Correct the branch versino number (6.8.0 rather than 6.8) Ian Lynagh **20070913181649] [More library installation path fiddling Ian Lynagh **20070913174242] [libsubdir for libraries is now always $$pkgid Ian Lynagh **20070913173926] [Define RelocatableBuild variable Ian Lynagh **20070913155331 default YES on Windows, NO otherwise. ] [We need to thread lots more paths through installPackage to make bindists work Ian Lynagh **20070913154830] [Update .darcs-boring Ian Lynagh **20070913015704] [Add some more bits to the boring file Ian Lynagh **20070907212208] [STABLE ONLY: Fix build Ian Lynagh **20070913004230] [MERGED: GHCi debugger: new flag -fbreak-on-error Ian Lynagh **20070912141427 Pepe Iborra **20070911101443 This flag works like -fbreak-on-exception, but only stops on uncaught exceptions. ] [STABLE only: Use -fno-warn-orphans in compiler/ if the compiler supports it Ian Lynagh **20070912140917 This was done as part of a larger patch in the HEAD. ] [Give push-all the ability to pull with a --pull flag Ian Lynagh **20070912140743 OK, so the name is a bit wrong now... ] [Handle doc-index*.html, not just doc-index.html Ian Lynagh **20070912135407 haddock sometimes makes doc-index-A.html etc files. Not sure why it doesn't for me. Patch from Judah Jacobson. ] [Fix repeated section name in documentation. judah.jacobson@gmail.com**20070907211616] [Fix installation code Ian Lynagh **20070912113954] [updating hpc toolkit andy@galois.com**20070908051600 The hpc overlay has been ported from hpc-0.4 The new API for readMix is now used. ] [Fixing hpc to allow use of hash function to seperate source files on source path andy@galois.com**20070907223357] [Fixing Hpc's Finite Map compat lib for ghc 6.2.1 andy@galois.com**20070908055320] [Further tweaking of haddock doc installation Ian Lynagh **20070911224734 On Windows we now always use a path beginning $topdir/ so bindists are relocatable. We also now tell "Setup configure" where we are putting the documentation, and tell installPackage to override as little as possible. ] [make in-place has bitrotted, so don't advertise or support it Ian Lynagh **20070911202202] [Synched documentation links with current directory layout sven.panne@aedion.de**20070911071801 Somehow the "html" subdirs are gone, this change was not completely intentional, but it is nice, anyway. Those subdirs never served any real purpose... MERGE TO STABLE ] [Refactoring & documenting the Term pprinter used by :print Pepe Iborra **20070911180411] [Custom printer for the Term datatype that won't output TypeRep values Pepe Iborra *-20070911151454 The term pretty printer used by :print shouldn't output the contents of TypeRep values, e.g. inside Dynamic values ] [Try and rewrite reloads to reg-reg moves in the spill cleaner Ben.Lippmeier@anu.edu.au**20070911173833] [Don't try and coalesce nodes with themselves Ben.Lippmeier@anu.edu.au**20070911151211] [Try and allocate vregs spilled/reloaded from some slot to the same hreg Ben.Lippmeier@anu.edu.au**20070911145054] [Better handling of live range joins via spill slots in spill cleaner Ben.Lippmeier@anu.edu.au**20070911130247] [Make sure to coalesce all the nodes found during iterative scanning Ben.Lippmeier@anu.edu.au**20070910162909] [Add iterative coalescing to graph coloring allocator Ben.Lippmeier@anu.edu.au**20070907172315 Iterative coalescing interleaves conservative coalesing with the regular simplify/scan passes. This increases the chance that nodes will be coalesced as they will have a lower degree than at the beginning of simplify. The end result is that more register to register moves will be eliminated in the output code, though the iterative nature of the algorithm makes it slower compared to non-iterative coloring. Use -fregs-iterative for graph coloring allocation with iterative coalescing -fregs-graph for non-iterative coalescing. The plan is for iterative coalescing to be enabled with -O2 and have a quicker, non-iterative algorithm otherwise. The time/benefit tradeoff between iterative and not is still being tuned - optimal graph coloring is NP-hard, afterall.. ] [Cure space leak in coloring register allocator Ben.Lippmeier@anu.edu.au**20070906131522 We now do a deep seq on the graph after it is 'built', but before coloring. Without this, the colorer will just force bits of it and the heap will fill up with half evaluated pieces of graph from previous build/spill stages and zillions of apply thunks. ] [Small improvement to GraphColor.selectColor Ben.Lippmeier@anu.edu.au**20070906103347 When selecting a color for a node, try and avoid using colors that conflicting nodes prefer. Not sure if this'll make much difference, but it was easy enough to add.. ] [Custom printer for the Term datatype that won't output TypeRep values Pepe Iborra **20070911151454 The term pretty printer used by :print shouldn't output the contents of TypeRep values, e.g. inside Dynamic values ] [Documentation for -fbreak-on-error Pepe Iborra **20070911101944] [Remove obsolete -fdebugging flag Pepe Iborra **20070907135857 A left over from the 1st GHCi debugger prototype ] [Tiny optimisation/simplification to FunDeps.grow simonpj@microsoft.com**20070907101133] [Warning police simonpj@microsoft.com**20070907101103] [Add clarifying comments simonpj@microsoft.com**20070907101046] [Fix zonking in mkExports simonpj@microsoft.com**20070906090346 I'd missed zonk, so that an error message for missing type signature read (unhelpfully) main/GHC.hs:1070:0: Warning: Definition but no type signature for `upsweep'' Inferred type: upsweep' :: forall t1. t The trouble was that 't' hadn't been zonked. Push to the stable branch ] [Cleaned up version of Tom's unflattened skolemOccurs Manuel M T Chakravarty **20070910083457] [Made TcTyFuns warning clean Manuel M T Chakravarty **20070907121113] [fix for Simple9 Tom Schrijvers **20070906171703 No longer include non-indexed arguments in lookup of matching type function clause. By including non-indexed (additional) arguments, the lookup always fails. ] [Improved error messages for higher-rank equality contexts Manuel M T Chakravarty **20070907101901] [FIX: Type families test Simple14 Manuel M T Chakravarty **20070907092217] [FIX #903: mkWWcpr: not a product Simon Marlow **20070910103830 This fixes the long-standing bug that prevents some code with mutally-recursive modules from being compiled with --make and -O, including GHC itself. See the comments for details. There are some additional cleanups that were forced/enabled by this patch: I removed importedSrcLoc/importedSrcSpan: it wasn't adding any useful information, since a Name already contains its defining Module. In fact when re-typechecking an interface file we were wrongly replacing the interesting SrcSpans in the Names with boring importedSrcSpans, which meant that location information could degrade after reloading modules. Also, recreating all these Names was a waste of space/time. ] [The RTS is Haddock-less, tell make about it sven.panne@aedion.de**20070910065759 MERGE TO STABLE ] [Include package documentation, n-th attempt... sven.panne@aedion.de**20070909154053 MERGE TO STABLE ] [Yet another attempt to get the paths for the installed documentation correct sven.panne@aedion.de**20070909124922 MERGE TO STABLE ] [Add a "show" target here, too, quite useful for debugging the build process sven.panne@aedion.de**20070909123813 MERGE TO STABLE ] [Never try to build Haddock docs in ghc/compiler, even with HADDOCK_DOCS=YES sven.panne@aedion.de**20070909123401 MERGE TO STABLE ] [Removed install-dirs target, it is unnecessary and leads to stray empty directories sven.panne@aedion.de**20070909121157 MERGE TO STABLE ] [Removed install-dirs from phony targets, it is unused sven.panne@aedion.de**20070909102815 MERGE TO STABLE ] [Nicer GHCi debugger underlining Pepe Iborra **20070910145319 Improved the underlining of blocks. With this patch it does: Stopped at break020.hs:(6,20)-(7,29) _result :: t1 () = _ 5 vv 6 in_another_decl _ = do line1 0 7 line2 0 ^^ 8 Instead of Stopped at break020.hs:(6,20)-(7,29) _result :: t1 () = _ 5 6 in_another_decl _ = do line1 0 ^^ 7 line2 0 ^^ 8 ] [FIX #1669 (GHCi debugger underlining is in the wrong place) Pepe Iborra **20070910142129 We weren't taking into account the offset added by the line numbers: Stopped at break020.hs:10:2-8 _result :: IO () = _ 9 main = do 10 line1 0 ^^^^^^^ 11 line2 0 This patch adjusts that ] [Add a BeConservative setting to the make system Ian Lynagh **20070910133528 If it is set, we don't try to use clock_gettime ] [Add a crucial missing ; Ian Lynagh **20070908231024] [Use := for PACKAGE_TARNAME, no reason for not doing so sven.panne@aedion.de**20070908124645 MERGE TO STABLE ] [Removed unused oldincludedir, things are already complicated enough sven.panne@aedion.de**20070908124448 MERGE TO STABLE ] [Added comment about GNU coding standards/autoconf history sven.panne@aedion.de**20070908123317 MERGE TO STABLE ] [We seem to use Outputable unconditionally these days sven.panne@aedion.de**20070908142712 MERGE TO STABLE ] [Make various assertions work when !DEBUG Ian Lynagh **20070908003112] [Fix assertions in RtClosureInspect Ian Lynagh **20070907233330] [In ASSERT and friends, use all the expressions we are passed even if !DEBUG Ian Lynagh **20070907233324 Otherwise we may get unused variable warnings. GHC should optimise them all out for us. ] [Don't put directories for unbuildable libraries in bindists Ian Lynagh **20070907210037 We think there is a library there, and then installPackage fails to install it. ] [Rejig boot Ian Lynagh **20070907122847 find on Windows doesn't understand -L, so stop trying to be clever and just autoreconf everything. Also, print out the names of directories as we autoreconf them, so that if autoreconfing one breaks then we know which one it was. ] [Fix publishing Ian Lynagh **20070907121414 Paths like c:/foo/bar get misinterpreted by rsync (really SSH?), as it thinks we want /foo/bar on the machine c. ] [Set do_bold based on $TERM, not platform Ian Lynagh **20070906175535] [Make installPackage install settings from the [package].buildinfo file. judah.jacobson@gmail.com**20070906010044 M ./libraries/installPackage.hs -1 +14 ] [Wibble some variable definitions to fix installation of bindists Ian Lynagh **20070906140430] [Remove hardtop_plat/FPTOOLS_TOP_ABS_PLATFORM Ian Lynagh **20070906122036 They are now the same as hardtop/FPTOOLS_TOP_ABS, so use those instead. Also removed some substitutions of / for \, as we now use a Haskell program to find the top path, and it only makes paths with /s in. ] [warning police Pepe Iborra **20070906102417] [Cleanup of equality rewriting and no swapInsts for wanteds Manuel M T Chakravarty **20070906115818 - Removed code duplication - Added comments - Took out swapInsts for wanteds. With the recent extension to swapInsts it does mess up error messages if applied to wanteds and i should not be necessary. NB: The code actually shrunk. Line increase is due to comments. ] [Remove EqInsts from addSCs to avoid -DDEBUG warnings Manuel M T Chakravarty **20070906095102] [EqInst related clean up Manuel M T Chakravarty **20070906095018 - Remove some unused and some superflous functions - Add comments regarding ancestor equalities - Tidied ancestor equality computation - Replace some incorrect instToId by instToVar (but there are still some bad ones around as we still get warnings with -DDEBUG) - Some cleaned up layout NB: Code growth is just due to more comments. ] [Remove dead code in TcSimplify Manuel M T Chakravarty **20070906031719] [Fix -DDEBUG warning Manuel M T Chakravarty **20070906023914] [also swap for variables in completion algorithm Tom Schrijvers **20070905134426] [Set GhcBootLibs=YES in mk/validate-settings.mk Ian Lynagh **20070906113629] [Quote all the arguments to installPackage Ian Lynagh **20070905232959 Makes it obvious what's going on if any are empty. ] [FIX #1465, error messages could sometimes say things like "A.T doesn't match A.T" Simon Marlow **20070906093744 This turned out to be a black hole, however we believe we now have a plan that does the right thing and shouldn't need to change again. Error messages will only ever refer to a name in an unambiguous way, falling back to :. if no unambiguous shorter variant can be found. See HscTypes.mkPrintUnqualified for the details. Earlier hacks to work around this problem have been removed (TcSimplify). ] [fix error in .hi-boot-6 Simon Marlow **20070905112503] [Improve GraphColor.colorScan Ben.Lippmeier@anu.edu.au**20070905164401 Testing whether a node in the conflict graph is trivially colorable (triv) is still a somewhat expensive operation. When we find a triv node during scanning, even though we remove it and its edges from the graph, this is unlikely to to make the nodes we've just scanned become triv - so there's not much point re-scanning them right away. Scanning now takes place in passes. We scan the whole graph for triv nodes and remove all the ones found in a batch before rescanning old nodes. Register allocation for SHA1.lhs now takes (just) 40% of total compile time with -O2 -fregs-graph on x86 ] [Fix OS X warnings Roman Leshchinskiy **20070906004831] [Declare ctime_r on Mac OS Roman Leshchinskiy **20070906001613 On Mac OS, ctime_r is not declared in time.h if _POSIX_C_SOURCE is defined. We work around this by providing a declaration ourselves. ] [FIX #1651: use family instances during interactive typechecking Manuel M T Chakravarty **20070905130244] [Updates to work with latest cabal. Duncan Coutts **20070906131726] [Add an OPTIONS -w pragma to utils/genprimopcode/Lexer.xx Ian Lynagh **20070905184808 SPJ reports that it has warnings (=> errors with -Werror) on Windows. ] [Build settings for validation are now in mk/validate-settings.mk Ian Lynagh **20070905184614] [Don't give warnings in compat Ian Lynagh **20070905182923 There are lots of warnings in here due to things like modules being imported that, in some versions of GHC, aren't used. Thus we don't give any warnings in here, and therefore validating with -Werror won't make the build fail. An alternative would be to do SRC_HC_OPTS := $(filter-out -Werror,$(SRC_HC_OPTS)) but if warnings are expected then there is little point in spewing them out anyway. On the other hand, there aren't any warnings for me (GHC 6.6 on Linux/amd64), so perhaps it would be worth fixing them instead. ] [Refactor, improve, and document the deriving mechanism simonpj@microsoft.com**20070905170730 This patch does a fairly major clean-up of the code that implements 'deriving. * The big changes are in TcDeriv, which is dramatically cleaned up. In particular, there is a clear split into a) inference of instance contexts for deriving clauses b) generation of the derived code, given a context Step (a) is skipped for standalone instance decls, which have an explicitly provided context. * The handling of "taggery", which is cooperative between TcDeriv and TcGenDeriv, is cleaned up a lot * I have added documentation for standalone deriving (which was previously wrong). * The Haskell report is vague on exactly when a deriving clause should succeed. Prodded by Conal I have loosened the rules slightly, thereyb making drv015 work again, and documented the rules in the user manual. I believe this patch validates ok (once I've update the test suite) and can go into the 6.8 branch. ] [Further documentation about mdo, suggested by Benjamin Franksen simonpj@microsoft.com**20070829083349] [Refactor MachRegs.trivColorable to do unboxed accumulation Ben.Lippmeier@anu.edu.au**20070905125219 trivColorable was soaking up total 31% time, 41% alloc when compiling SHA1.lhs with -O2 -fregs-graph on x86. Refactoring to use unboxed accumulators and walk directly over the UniqFM holding the set of conflicts reduces this to 17% time, 6% alloc. ] [change of representation for GenCmm, GenCmmTop, CmmProc Norman Ramsey **20070905164802 The type parameter to a C-- procedure now represents a control-flow graph, not a single instruction. The newtype ListGraph preserves the current representation while enabling other representations and a sensible way of prettyprinting. Except for a few changes in the prettyprinter the new compiler binary should be bit-for-bit identical to the old. ] [enable and slay warnings in cmm/Cmm.hs Norman Ramsey **20070905164646] [fix warnings Simon Marlow **20070905114205] [FIX #1650: ".boot modules interact badly with the ghci debugger" Simon Marlow **20070905104716 In fact hs-boot files had nothing to do with it: the problem was that GHCi would forget the breakpoint information for a module that had been reloaded but not recompiled. It's amazing that we never noticed this before. The ModBreaks were in the ModDetails, which was the wrong place. When we avoid recompiling a module, ModDetails is regenerated from ModIface by typecheckIface, and at that point it has no idea what the ModBreaks should be, so typecheckIface made it empty. The right place for the ModBreaks to go is with the Linkable, which is retained when compilation is avoided. So now I've placed the ModBreaks in with the CompiledByteCode, which also makes it clear that only byte-code modules have breakpoints. This fixes break022/break023 ] [Fix boot: it was avoiding autoreconfing Simon Marlow **20070905101419 Two problems here: find needs to dereference symbolic links (-L option, I really hope that's portable), and we need to notice when aclocal.m4 is updated. Somehow I think this was easier when it just always ran autoreconf... what was wrong with that? ] [don't generate .hi-boot/.o-boot files in GHCi Simon Marlow **20070904141231] [refactoring only Simon Marlow **20070904141209] [completion for modules in 'import M' Simon Marlow **20070904104458] [make the GhcThreaded setting lazy, because GhcUnregisterised might not be set yet Simon Marlow **20070904101729] [Typo Ian Lynagh **20070905161402] [Fix bindist creation on Windows Ian Lynagh **20070905161354] [Fix up bindist creation and publishing Ian Lynagh **20070905160641] [{Enter,Leave}CriticalSection imports should be outside #ifdef __PIC__ Simon Marlow **20070905084941] [warning police Ben.Lippmeier@anu.edu.au**20070905094509] [Do conservative coalescing in register allocator Ben.Lippmeier@anu.edu.au**20070903163404 Avoid coalescing nodes in the register conflict graph if the new node will not be trivially colorable. Also remove the front end aggressive coalescing pass. For typical Haskell code the graph coloring allocator now does about as well as the linear allocator. For code with a large amount of register pressure it does much better, but takes longer. For SHA1.lhs from darcs on x86 spills reloads reg-reg-moves inserted inserted left in code compile-time linear 1068 1311 229 7.69(s) graph 387 902 340 16.12(s) ] [Use dlsym on OS X if available Roman Leshchinskiy **20070905052213 On OS X 10.4 and newer, we have to use dlsym because the old NS* interface has been deprecated. The patch checks for HAVE_DLFCN_H instead of switching on the OS version. There is one additional quirk: although OS X prefixes global symbols with an underscore, dlsym expects its argument NOT to have a leading underscore. As a hack, we simply strip it off in lookupSymbol. Something a bit more elaborate might be cleaner. ] [bug fix in Decomp step of completion algorithm for given equations Tom Schrijvers **20070904123945] [fix of wanted equational class context Tom Schrijvers **20070904080014 Previously failed to account for equational class context for wanted dictionary contraints, e.g. wanted C a in class a ~ Int => C a instance C Int should give rise to wanted a ~ Int and consequently discharge a ~ Int by unifying a with Int and then discharge C Int with the instance. All ancestor equalities are taken into account. ] [FIX: Correct Leave/EnterCriticalSection imports Manuel M T Chakravarty **20070905010217] [Don't hardwire the build path into the Haddock docs sven.panne@aedion.de**20070904172855 Formerly, the ghc-pkg was called to get the HTML dirs for other packages, but of course doing this at *build* time is totally wrong. Now we use a relative path, just like before. This is probably not perfect, but much better than before. As a sidenote: Cabal calls the relevant flag "html-location", ghc-pkg calls the field "haddock-html", and Haddock itself uses it as part of "read-interface". Too much creativity is sometimes a bad thing... ] [put the @N suffix on stdcall foreign calls in .cmm code Simon Marlow **20070904142853 This applies to EnterCriticalSection and LeaveCriticalSection in the RTS ] [Set datarootdir to the value configure gives us (if any) so datadir works Ian Lynagh **20070905013239 We then set datarootdir to something else later on so that things still work when configure doesn't set it. ] [Add a -Warn flag Ian Lynagh **20070904141028] [Always turn on -Wall -Werror when compiling the compiler, even for stage 1 Ian Lynagh **20070904140324] [Fix CodingStyle#Warnings URLs Ian Lynagh **20070904140115] [OPTIONS_GHC overrides the command-line, not the other way around Simon Marlow **20070904100623] [fix cut-and-pasto Simon Marlow **20070904100526] [FIX #1651: unBox types when deferring unification Manuel M T Chakravarty **20070904072542 - This fixes the first part of #1651; ie, the panic in ghci. ] [Better error message for unsolvable equalities Manuel M T Chakravarty **20070903074528] [Use := rather than = when assigning make variables to avoid cycles Ian Lynagh **20070903235117] [Don't use autoconf's datarootdir as <2.60 doesn't have it Ian Lynagh **20070903234504] [Use OPTIONS rather than OPTIONS_GHC for pragmas Ian Lynagh **20070903233903 Older GHCs can't parse OPTIONS_GHC. This also changes the URL referenced for the -w options from WorkingConventions#Warnings to CodingStyle#Warnings for the compiler modules. ] [fix build (sorry, forgot to push with previous patch) Simon Marlow **20070903200615] [remove debugging code Simon Marlow **20070903200003] [Fix building RTS with gcc 2.*; declare all variables at the top of a block Ian Lynagh **20070903165847 Patch from Audrey Tang. ] [Change the version number to 6.8 on the branch Ian Lynagh **20070903172054] [TAG ghc-6.8 branched 2007-09-03 Ian Lynagh **20070903155416] Patch bundle hash: 5772327383521db295c5ebfd76e21714dfc9d670