- ghc-devs - Haskell.org

RE: [commit: ghc] wip/T14068: Avoid name capture in setJoinResTy (6a50466)
by Simon Peyton Jones 02 Aug '17

02 Aug '17

Joachim Aha! I think my comment on Phab is redundant. You are simply avoiding name capture, which is obviously right; you are not restricting the applicability of the transformation. Needless to say, a Note would help in due course. S | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf | Of git(a)git.haskell.org | Sent: 02 August 2017 04:07 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] wip/T14068: Avoid name capture in setJoinResTy | (6a50466) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : wip/T14068 | Link : | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fghc.ha | skell.org%2Ftrac%2Fghc%2Fchangeset%2F6a5046684f93f1870663119b447fff6ba | eb7a7c4%2Fghc&data=04%7C01%7Csimonpj%40microsoft.com%7Cc45ce3f4deea476 | 9222b08d4d953af74%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6363724 | 00832017356%7CUnknown%7CVW5rbm93bnx7IlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMi | IsIkFOIjoiT3RoZXIifQ%3D%3D%7C- | 1&sdata=29Iajm6oTE3Zm%2Bdl5ocMav7MPw%2FniO9%2BVeuQm38qWXQ%3D&reserved= | 0 | | >--------------------------------------------------------------- | | commit 6a5046684f93f1870663119b447fff6baeb7a7c4 | Author: Joachim Breitner <mail(a)joachim-breitner.de> | Date: Tue Aug 1 22:48:26 2017 -0400 | | Avoid name capture in setJoinResTy | | | >--------------------------------------------------------------- | | 6a5046684f93f1870663119b447fff6baeb7a7c4 | compiler/types/Type.hs | 22 ++++++++++++++++------ | 1 file changed, 16 insertions(+), 6 deletions(-) | | diff --git a/compiler/types/Type.hs b/compiler/types/Type.hs index | 50a35b0..c69d4ff 100644 | --- a/compiler/types/Type.hs | +++ b/compiler/types/Type.hs | @@ -2445,11 +2445,21 @@ setJoinResTy :: Int -- Number of binders to | skip | -> Type -- New type | -- INVARIANT: If any of the first n binders are foralls, those tyvars | cannot | -- appear in the original result type. See isValidJoinPointType. | +-- | +-- When we set the return type under a forall, avoid capture! | setJoinResTy orig_ar new_res_ty orig_ty | - = go orig_ar orig_ty | + = go init_subst orig_ar orig_ty | where | - go 0 _ = new_res_ty | - go n ty | Just (arg_bndr, res_ty) <- splitPiTy_maybe ty | - = mkPiTy arg_bndr (go (n-1) res_ty) | - | otherwise | - = pprPanic "setJoinResTy" (ppr orig_ar <+> ppr orig_ty) | + init_subst :: TCvSubst | + init_subst = mkEmptyTCvSubst (mkInScopeSet (tyCoVarsOfType | + new_res_ty)) | + | + go _ 0 _ = new_res_ty | + go subst n ty | + | Just (t, ty') <- splitForAllTy_maybe ty | + , let (subst', t') = substTyVarBndr subst t | + = mkForAllTy t' Inferred (go subst' (n-1) ty') | + | Just (arg_ty, ty') <- splitFunTy_maybe ty | + , let arg_ty' = substTy subst arg_ty | + = mkFunTy arg_ty' (go subst (n-1) ty') | + | otherwise | + = pprPanic "setJoinResTy" (ppr orig_ar <+> ppr orig_ty) | | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.h | askell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fghc- | commits&data=04%7C01%7Csimonpj%40microsoft.com%7Cc45ce3f4deea4769222b0 | 8d4d953af74%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6363724008320 | 17356%7CUnknown%7CVW5rbm93bnx7IlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFO | IjoiT3RoZXIifQ%3D%3D%7C- | 1&sdata=7%2FAY0EcihvMkjGOHdYdpOog2Roahk0XD%2BunNg54pKFc%3D&reserved=0

1 0

dataToTag# documentation
by David Feuer 02 Aug '17

02 Aug '17

dataToTag# is documented as getting the tag number of an enumeration, which is perfectly reasonable because it's designed to support deriving Enum. But it *appears* to work also for non-enumeration datatypes: dataToTag# Nothing = 0# dataToTag# (Just 3) = 1# Does this actually always work? If so, should that be documented, or is there a realistic possibility that its behavior will change in the future? Additionally: the documentation for dataToTag# urges readers to use GHC.Base.getTag instead. But dataToTag# is exported from the "public" GHC.Exts, whereas getTag is not. Should we add getTag to GHC.Exts, or change the documentation for dataToTag#? David

2 1

code layout of cryptarithm1
by Joachim Breitner 01 Aug '17

01 Aug '17

Hi, if you look at https://perf.haskell.org/ghc/#graph/nofib/time/cryptarithm1 you see that it’s performance varies a lot, sometimes by 10%, between commits should have nothing to do with performance. But most (not all!) of the commits where the performance changes affect the runtime in some way. I conclude that there is something about the code layout that makes this program run faster or slower. If anyone here enjoys tracing and understanding such low-level issues: Here is a nice target to aim at. (I don’t… but I’d appreciate less noise from gipeda :-)) Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata(a)joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

1 0

GHC 8.2 generated code faster than 8.0 but slower than 7.10
by Harendra Kumar 01 Aug '17

01 Aug '17

Unicode normalization library ( https://github.com/harendra-kumar/unicode-transforms) shows around 10% improvement with GHC 8.2.1 when compared to GHC 8.0.1, across most benchmarks. However, it is still somewhat slower when compared to GHC 7.10.3. Here are some results: GHC 7.10.3: benchmarking unicode-transforms-text/NFD/English time 4.823 ms (4.765 ms .. 4.902 ms) benchmarking unicode-transforms-text/NFD/Devanagari time 16.46 ms (16.24 ms .. 16.78 ms) GHC 8.0.1 benchmarking unicode-transforms-text/NFD/English time 6.330 ms (6.232 ms .. 6.439 ms) benchmarking unicode-transforms-text/NFD/Devanagari time 18.02 ms (17.77 ms .. 18.22 ms) GHC 8.2.1 benchmarking unicode-transforms-text/NFD/English time 5.659 ms (5.594 ms .. 5.740 ms) benchmarking unicode-transforms-text/NFD/Devanagari time 16.48 ms (16.30 ms .. 16.69 ms) This library has been optimized with an objective to match or better (in some cases it is better) the performance of the equivalent ICU C++ library (compare only decompose normalization). Some of the last hurdles to match the best case of C++ were in the code generated by GHC. Also, for GHC 7.10 LLVM generated code is significantly faster than GHC native, I have not yet tested LLVM backend with 8.2.1 so not sure if that too has improved correspondingly. I just wanted to share these results with the GHC devs. Thanks for all the good work. Also wondering which specific changes might have caused this improvement. Thanks, Harendra

2 2

Operating on HsSyn
by Simon Peyton Jones 31 Jul '17

31 Jul '17

Devs, Shayan is working away on "Trees that grow"... do keep it on your radar: To: ghc-devs Sent: 25 May 2017 23:49 Do take a look at this: * We propose to re-engineer HsSyn itself. This will touch a lot of code. * But it's very neat, and will bring big long-term advantages * And we can do it a bit at a time The wiki page https://ghc.haskell.org/trac/ghc/wiki/ImplementingTreesThatGrow has the details. It's entirely an internal change, not a change to GHC's specification, so it's independent of the GHC proposals process. But I'd value the opinion of other GHC devs Meanwhile I have a question. When pretty-printing HsSyn we often have a situation like this: data Pass = Parsed | Renamed | Typechecked data HsExpr (p :: Pass) = HsVar (IdP p) | .... type famliy IdP p where IdP Parsed = RdrName IdP Renamed = Name IdP Typechecked = Id instance (Outputable (IdP p)) => Outputable (HsExpr p) where ppr (HsVar v) = ppr v The (ppr v) requires (Outputable (IdP p)), hence the context. Moreover, and more seriously, there are things we just can't pretty-print right now. For example, HsExpr has this data constructor: data HsExpr p = ... | OpApp (LHsExpr p) (LHsExpr p) (PostRn p Fixity) (LHsExpr p) To pretty-print the third argument, we'd need to add instance (Outputable (IdP p), Outputable (PostRn p Fixity)) -- New => Outputable (HsExpr p) where ppr (HsVar v) = ppr v and that gets onerous. So today we never print these annotations, to avoid bloating the instance contexts, which can be painful. It bit me yesterday. We have bitten that bullet for the Data class: look at HsExtension.DataId, which abbreviates the long list of dictionaries: type DataId p = ( Data p , ForallX Data p , Data (NameOrRdrName (IdP p)) , Data (IdP p) , Data (PostRn p (IdP p)) , Data (PostRn p (Located Name)) , Data (PostRn p Bool) , Data (PostRn p Fixity) ,..and nine more... ) Let me note in passing that [wiki:QuantifiedContexts<https://ghc.haskell.org/trac/ghc/wiki/QuantifiedContexts>] would make this somewhat shorter type DataId p = ( Data p , ForallX Data p , Data (NameOrRdrName (IdP p)) , Data (IdP p) , forall t. Data t => Data (PostRn p t)) But we still need one item in this list for each type function, and I am worried about how this scales to the [wiki:ImplementingTreesThatGrow<https://ghc.haskell.org/trac/ghc/wiki/ImplementingTreesThatGrow>] story, when we have a type function for each data constructor -- and there are a lot of data constructors. So I have four questions 1. I think we should probably use a superclass instead of a type synonym class (Data p, ForallX Data p, ....) => DataId p where {} Why? Only one argument to pass, and to pass on to successive calls. I see no downside. 1. Shall we treat Outputable like Data? (I.e. make an abbreviation for a long list of Outputable instances and use it everywhere) 2. I thought of another way to do it (pass a token); see below 1. Are there any other ways? Token passing idea. Perhaps instead of passing lots of functions, we pass a singleton token that encodes the pass, like this: instance (PassC p) => Outputable (HsExpr p) where ppr (HsVar v) = case getPass :: IsPass p of IsParsed -> ppr v IsRenamed -> ppr v IsTypechecked -> ppr v The three ppr's are at different types, of course; that's the point. The infrastructure is something like class PassC p where getPass :: IsPass p data IsPass p where IsParsed :: IsPass Parsed IsRenamed :: IsParsed Renamed IsTypechecked :: IsParsed Typechecked instance PassC Parsed where getPass = IsParsed ...etc... Now we could sweep away all those OutputableX classes, replacing them with dynamic tests on the singletons IsParsed etc. This would have advantages: - Probably faster: there's a dynamic test, but many fewer dictionary arguments and higher-order function dispatch - Only one dictionary to pass; programming is easier. The big downside is that it's not extensible: it works only because we know the three cases. But the "Trees that Grow" story really doesn't scale well to pretty-printing: so maybe we should just give up on that?

4 10

8.2.1: Ord TyCon is gone?
by Levent Erkok 29 Jul '17

29 Jul '17

I'm trying to port some plugin code from GHC 8.0.2 to GHC 8.2.1; alas I'm getting an error suggesting that TyCon is no longer an instance of Ord: Data/SBV/Plugin/Analyze.hs:580:33: error: • No instance for (Ord TyCon) arising from a use of ‘M.lookup’ There are instances for similar types: instance Ord GHC.Types.TyCon -- Defined in ‘GHC.Classes’ • In the expression: k `M.lookup` tcMap In the expression: case k `M.lookup` tcMap of Just knd -> return knd Nothing -> unknown In a case alternative: Just k -> case k `M.lookup` tcMap of Just knd -> return knd Nothing -> unknown | 580 | Just k -> case k `M.lookup` tcMap of | ^^^^^^^^^^^^^^^^^^ I was using TyCon's as a key in a map, but that code is no longer compiling. I've looked through the release notes but couldn't find what the new mechanism is. I'd appreciate if someone can point me in the right direction regarding how to use TyCon as a map key.. Thanks, -Levent.

6 7

[ANNOUNCE] GHC 8.2.1 available
by Ben Gamari 27 Jul '17

27 Jul '17

=============================================== The Glasgow Haskell Compiler -- version 8.2.1 =============================================== The GHC developers are very happy to announce the long-awaited 8.2.1 release of Glasgow Haskell Compiler. Binary and source distributions can be found at https://downloads.haskell.org/~ghc/8.2.1/ This is the second release in the 8.0 series. As such, the focus of this release is performance, stability, and consolidation. Consequently numerous cleanups can be seen throughout the compiler including, * Significant improvements in compiler performance * More robust support for levity polymorphism * Reliable DWARF debugging information * Improved runtime system performance on NUMA systems * Retooling of the cost-center profiler, including support for live streaming of profile data via the GHC event log * Interface file determinism * More robust treatment of join points, enabling significantly better code generation in many cases * Numerous improvements in robustness on Windows * and the resolution of over 500 other tickets In addition, there are a number of new features, * A new, more type-safe type reflection mechanism * The long-awaited Backpack module system * Deriving strategies to disambiguate between GHC's various instance deriving mechanisms * Unboxed sum types, for efficient unpacked representation of sum data types * Compact regions, allowing better control over garbage collection in the presence of large heaps containing many long-lived objects. * Colorful messages and caret diagnostics for more legible errors A more thorough list of the changes in this release can be found in the release notes, https://haskell.org/ghc/docs/8.2.1/docs/html/users_guide/8.2.1-notes.html There are a few changes in release-engineering that should be noted, * Binary distributions for 32-bit CentOS 6.7 have been dropped. Moreover, there are no dedicated CentOS 7.0 distributions as CentOS 7 can use can use Debian 8 binaries. If you would like us to continue to produce 32-bit CentOS 6.7 distributions please let us know. * GHC HQ now builds FreeBSD and OpenBSD distributions for amd64; this comes after many years of these distributions being faithfully provided by Karel Gardas and Pali Gabor Janos, who we should heartily thank for their contributions. GHC HQ building these distributions ourselves will allow us to more quickly ship distributions to users by eliminating the need for a long lag time between source release availability and having all binary distributions available. * There is a technology-preview of an AArch64 Linux binary distribution, as well as an ARM Linux build. AArch64 support is quite preliminary but should be stable in 8.4 thanks to further linker fixes by Moritz Angerman. ARM should be stable. * GHC now tries to use the gold and lld linkers by default. These linkers are significantly faster than the BFD linker implementation that most Linux distributions use by default. If gold or lld are not available GHC will use the system's default linker. GHC can be forced to use the default linker by passing --disable-ld-override to configure. This release has been the result of over a year of hard work by over 150 code contributors. Thanks to everyone who has helped in writing patches, testing, reporting bugs, and offering feedback over the last year. This release cycle was admittedly quite drawn out, significantly longer than expected or desired. While we are confident that the result is worth the wait, we have been steadily working on infrastructure which should help shrink future release cycles and give us better testing between releases. More details on this coming soon. As always, let us know if you encounter trouble. How to get it ~~~~~~~~~~~~~ The easy way is to go to the web page, which should be self-explanatory: http://www.haskell.org/ghc/ We supply binary builds in the native package format for many platforms, and the source distribution is available from the same place. Packages will appear as they are built - if the package for your system isn't available yet, please try again later. Background ~~~~~~~~~~ Haskell is a standard lazy functional programming language. GHC is a state-of-the-art programming suite for Haskell. Included is an optimising compiler generating efficient code for a variety of platforms, together with an interactive system for convenient, quick development. The distribution includes space and time profiling facilities, a large collection of libraries, and support for various language extensions, including concurrency, exceptions, and foreign language interfaces. GHC is distributed under a BSD-style open source license. A wide variety of Haskell related resources (tutorials, libraries, specifications, documentation, compilers, interpreters, references, contact information, links to research groups) are available from the Haskell home page (see below). On-line GHC-related resources ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Relevant URLs on the World-Wide Web: GHC home page http://www.haskell.org/ghc/ GHC developers' home page http://ghc.haskell.org/trac/ghc/ Haskell home page http://www.haskell.org/ Supported Platforms ~~~~~~~~~~~~~~~~~~~ The list of platforms we support, and the people responsible for them, is here: http://ghc.haskell.org/trac/ghc/wiki/Contributors Ports to other platforms are possible with varying degrees of difficulty. The Building Guide describes how to go about porting to a new platform: http://ghc.haskell.org/trac/ghc/wiki/Building Developers ~~~~~~~~~~ We welcome new contributors. Instructions on accessing our source code repository, and getting started with hacking on GHC, are available from the GHC's developer's site run by Trac: http://ghc.haskell.org/trac/ghc/ Mailing lists ~~~~~~~~~~~~~ We run mailing lists for GHC users and bug reports; to subscribe, use the web interfaces at http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-tickets There are several other haskell and ghc-related mailing lists on www.haskell.org; for the full list, see https://mail.haskell.org/cgi-bin/mailman/listinfo Some GHC developers hang out on #haskell on IRC, too: http://www.haskell.org/haskellwiki/IRC_channel Please report bugs using our bug tracking system. Instructions on reporting bugs can be found here: http://www.haskell.org/ghc/reportabug

4 6

Harbormaster update
by Ben Gamari 27 Jul '17

27 Jul '17

Hello everyone, Note that the Harbormaster builders now bootstrap with GHC 8.2.1 and LLVM 3.9. I'll be on the lookout for any failures but ping me if you see Harbormaster fail to build your otherwise perfectly good patch. Cheers, - Ben

1 0

RE: [commit: ghc] master: Improve error messages around kind mismatches. (8e15e3d)
by Simon Peyton Jones 27 Jul '17

27 Jul '17

| typecheck/should_fail/T7368) Solution: track the provenance of | Derived constraints and refuse to float out a constraint that has | an insoluble Derived. For the record, that's not the solution that the patch implements. Instead we just don’t float heterogeneous equalities. I think Richard just forgot to update the commit message. Simon | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf | Of git(a)git.haskell.org | Sent: 27 July 2017 12:50 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] master: Improve error messages around kind | mismatches. (8e15e3d) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : master | Link : | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fghc.ha | skell.org%2Ftrac%2Fghc%2Fchangeset%2F8e15e3d370e9c253ae0dbb330e25b72cb | 00cdb76%2Fghc&data=02%7C01%7Csimonpj%40microsoft.com%7C9993301a429e467 | e29b008d4d4e6350f%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6363675 | 32584262688&sdata=bY%2Bruj0hhXfC23hfMA1ivuAs2Av2%2BY4%2BJK2yyCoPzbA%3D | &reserved=0 | | >--------------------------------------------------------------- | | commit 8e15e3d370e9c253ae0dbb330e25b72cb00cdb76 | Author: Richard Eisenberg <rae(a)cs.brynmawr.edu> | Date: Thu Jun 1 17:27:14 2017 -0400 | | Improve error messages around kind mismatches. | | Previously, when canonicalizing (or unifying, in uType) a | heterogeneous equality, we emitted a kind equality and used the | resulting coercion to cast one side of the heterogeneous equality. | | While sound, this led to terrible error messages. (See the bugs | listed below.) The problem is that using the coercion built from | the emitted kind equality is a bit like a wanted rewriting a | wanted. | The solution is to keep heterogeneous equalities as irreducible. | | See Note [Equalities with incompatible kinds] in TcCanonical. | | This commit also removes a highly suspicious switch to | FM_SubstOnly | when flattening in the kinds of a type variable. I have no idea | why this was there, other than as a holdover from pre-TypeInType. | I've not left a Note because there is simply no reason I can | conceive | of that the FM_SubstOnly should be there. | | One challenge with this patch is that the emitted derived | equalities | might get emitted several times: when a heterogeneous equality is | in an implication and then gets floated out from the implication, | the Derived is present both in and out of the implication. This | causes a duplicate error message. (Test case: | typecheck/should_fail/T7368) Solution: track the provenance of | Derived constraints and refuse to float out a constraint that has | an insoluble Derived. | | Lastly, this labels one test (dependent/should_fail/RAE_T32a) | as expect_broken, because the problem is really #12919. The | different handling of constraints in this patch exposes the error. | | This fixes bugs #11198, #12373, #13530, and #13610. | | test cases: | typecheck/should_fail/{T8262,T8603,tcail122,T12373,T13530,T13610} | | | >--------------------------------------------------------------- | | 8e15e3d370e9c253ae0dbb330e25b72cb00cdb76 | compiler/typecheck/TcCanonical.hs | 296 | ++++++++++++--------- | compiler/typecheck/TcErrors.hs | 75 ++++-- | compiler/typecheck/TcEvidence.hs | 8 +- | compiler/typecheck/TcFlatten.hs | 31 ++- | compiler/typecheck/TcRnTypes.hs | 30 ++- | compiler/typecheck/TcSimplify.hs | 34 ++- | compiler/typecheck/TcType.hs | 10 +- | compiler/typecheck/TcUnify.hs | 28 +- | compiler/types/Type.hs | 4 +- | testsuite/tests/dependent/should_fail/T11471.hs | 2 +- | .../tests/dependent/should_fail/T11471.stderr | 11 +- | testsuite/tests/dependent/should_fail/all.T | 2 +- | testsuite/tests/gadt/gadt7.stderr | 6 +- | .../tests/ghci.debugger/scripts/break012.stdout | 14 +- | .../indexed-types/should_fail/ExtraTcsUntch.stderr | 6 +- | .../tests/indexed-types/should_fail/T5934.stderr | 13 - | testsuite/tests/polykinds/T12593.stderr | 56 ++++ | testsuite/tests/polykinds/T13555.stderr | 21 +- | testsuite/tests/polykinds/T7438.stderr | 6 +- | testsuite/tests/polykinds/T8566.stderr | 2 +- | testsuite/tests/polykinds/T9017.stderr | 10 +- | testsuite/tests/typecheck/should_fail/T12373.hs | 10 + | .../tests/typecheck/should_fail/T12373.stderr | 8 + | testsuite/tests/typecheck/should_fail/T13530.hs | 11 + | .../tests/typecheck/should_fail/T13530.stderr | 7 + | testsuite/tests/typecheck/should_fail/T13610.hs | 11 + | .../tests/typecheck/should_fail/T13610.stderr | 14 + | testsuite/tests/typecheck/should_fail/T5691.stderr | 10 +- | testsuite/tests/typecheck/should_fail/T7368.stderr | 6 +- | .../tests/typecheck/should_fail/T7368a.stderr | 2 +- | testsuite/tests/typecheck/should_fail/T7453.stderr | 48 +--- | testsuite/tests/typecheck/should_fail/T7696.stderr | 4 +- | testsuite/tests/typecheck/should_fail/T8262.stderr | 6 +- | testsuite/tests/typecheck/should_fail/T8603.hs | 4 + | testsuite/tests/typecheck/should_fail/T8603.stderr | 13 +- | testsuite/tests/typecheck/should_fail/all.T | 3 + | .../tests/typecheck/should_fail/tcfail090.stderr | 4 +- | .../tests/typecheck/should_fail/tcfail122.stderr | 8 +- | .../tests/typecheck/should_fail/tcfail123.stderr | 13 +- | .../tests/typecheck/should_fail/tcfail200.stderr | 6 +- | 40 files changed, 537 insertions(+), 316 deletions(-) | | Diff suppressed because of size. To see it, use: | | git diff-tree --root --patch-with-stat --no-color --find-copies- | harder --ignore-space-at-eol --cc | 8e15e3d370e9c253ae0dbb330e25b72cb00cdb76 | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.h | askell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fghc- | commits&data=02%7C01%7Csimonpj%40microsoft.com%7C9993301a429e467e29b00 | 8d4d4e6350f%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6363675325842 | 62688&sdata=JkrPJNfzdjx2Rz%2FfcMq2HlIP%2BlsFIE%2FT68a%2FSUae%2Fzo%3D&r | eserved=0

2 1

Reinstallable lib:ghc (was: [core libraries] Upgradeable base (smallest step forward))
by Herbert Valerio Riedel 27 Jul '17

27 Jul '17

Hi everyone, As part of the recent GHC 8.2.1 release and the associated boot-library publishing to Hackage, I picked up again experimenting with a reinstallable lib:ghc package primarily for the benefit of uploading haddocks for lib:ghc (w/ `--hyperlinked-source`) to Hackage: http://hackage.haskell.org/package/ghc However, this also provides somewhat of a tech-preview of a reinstallable lib:ghc package, which would help mitigate a big issue we had in the past with lib:ghc and its dependencies being frozen to specific versions, and which in combination with cabal's nix-style pkg store that finally provide us with the ability to have multiple instances of the same library versions co-existing in the same package db, is no longer a requirement. So we could have lib:ghc more easily pick dependencies up (but still within reason!) like e.g. lib:text without forcing everyone to the single lib:text version bundled with GHC. And we'd also gain the ability to publish patch-level releases to lib:ghc only inbetween proper minor GHC releases (e.g. to support new versions of library dependencies). Consider this artificial example demo package which requires a different version of binary than the one bundled with GHC 8.2.1: --8<---------------cut here---------------start------------->8--- -- ghc-demo.cabal name: ghc-demo version: 0 build-type: Simple cabal-version: >=2.0 executable ghc-demo main-is: Main.hs build-depends: base ^>= 4.10, ghc ^>= 8.2.1, binary == 0.8.5.0 default-language: Haskell2010 --8<---------------cut here---------------end--------------->8--- Trying to build this will have the cabal solver complain that the installed instance of ghc requires a different (& installed) version of binary, i.e. binary-0.8.5.1. Moreover, the source version of lib:ghc on Hackage has the manual `buildable` cabal flag which defaults to false, and currently prevents lib:ghc to be reinstalled automatically (by enabling an unsatisfiable constraint): --8<---------------cut here---------------start------------->8--- Resolving dependencies... cabal: Could not resolve dependencies: trying: base-4.10.0.0/installed-4.1... (dependency of ghc-demo-0) next goal: ghc (dependency of ghc-demo-0) rejecting: ghc-8.2.1/installed-8.2... (conflict: binary==0.8.5.0, ghc => binary==0.8.5.1/installed-0.8...) trying: ghc-8.2.1 rejecting: ghc-8.2.1:-buildable (conflict: base==4.10.0.0/installed-4.1..., ghc -buildable => base<0) rejecting: ghc-8.2.1:+buildable (manual flag can only be changed explicitly) After searching the rest of the dependency tree exhaustively, these were the goals I've had most trouble fulfilling: ghc-demo, binary, base, ghc, ghc-8.2.1:buildable --8<---------------cut here---------------end--------------->8--- However, if we unlock this by toggling the flag (NB: this only works on linux/x86_64; it's explained later why), e.g. --8<---------------cut here---------------start------------->8--- -- cabal.project file packages: . package ghc flags: +buildable --8<---------------cut here---------------end--------------->8--- The cabal solver would come up with an install-plan: --8<---------------cut here---------------start------------->8--- Resolving dependencies... Build profile: -w ghc-8.2.1 -O1 In order, the following will be built (use -v for more details): - binary-0.8.5.0 {binary-0.8.5.0-7f686cf5c40c3c685540306709c4a2a30d740679f5d31ad196283d242a4d70ac} (lib) (requires download & build) - ghc-boot-8.2.1 {ghc-boot-8.2.1-7930c3d78690c5fa476415d38abbb7717b05cf65e80cc79c12e782c9886507cb} (lib) (requires build) - ghci-8.2.1 {ghci-8.2.1-6292924619a045fd18e9fde0cfa350632672ad7439cb84a5c56ff0e4781346b1} (lib) (requires build) - ghc-8.2.1 {ghc-8.2.1-bc10931b0f169622deb74d84cf10e315730dfd8f2edcf4cad9073d142cf5da0c} (lib) +buildable (requires build) - ghc-demo-0 {ghc-demo-0-inplace-ghc-demo} (exe:ghc-demo) (first run) Configuring binary-0.8.5.0-7f686cf5c40c3c685540306709c4a2a30d740679f5d31ad196283d242a4d70ac (lib) Building binary-0.8.5.0-7f686cf5c40c3c685540306709c4a2a30d740679f5d31ad196283d242a4d70ac (lib) Installing binary-0.8.5.0-7f686cf5c40c3c685540306709c4a2a30d740679f5d31ad196283d242a4d70ac (lib) Finished binary-0.8.5.0-7f686cf5c40c3c685540306709c4a2a30d740679f5d31ad196283d242a4d70ac (lib) Configuring ghc-boot-8.2.1-7930c3d78690c5fa476415d38abbb7717b05cf65e80cc79c12e782c9886507cb (lib) Building ghc-boot-8.2.1-7930c3d78690c5fa476415d38abbb7717b05cf65e80cc79c12e782c9886507cb (lib) Installing ghc-boot-8.2.1-7930c3d78690c5fa476415d38abbb7717b05cf65e80cc79c12e782c9886507cb (lib) Finished ghc-boot-8.2.1-7930c3d78690c5fa476415d38abbb7717b05cf65e80cc79c12e782c9886507cb (lib) Configuring ghci-8.2.1-6292924619a045fd18e9fde0cfa350632672ad7439cb84a5c56ff0e4781346b1 (lib) Building ghci-8.2.1-6292924619a045fd18e9fde0cfa350632672ad7439cb84a5c56ff0e4781346b1 (lib) Installing ghci-8.2.1-6292924619a045fd18e9fde0cfa350632672ad7439cb84a5c56ff0e4781346b1 (lib) Finished ghci-8.2.1-6292924619a045fd18e9fde0cfa350632672ad7439cb84a5c56ff0e4781346b1 (lib) Configuring ghc-8.2.1-bc10931b0f169622deb74d84cf10e315730dfd8f2edcf4cad9073d142cf5da0c (lib) Building ghc-8.2.1-bc10931b0f169622deb74d84cf10e315730dfd8f2edcf4cad9073d142cf5da0c (lib) Installing ghc-8.2.1-bc10931b0f169622deb74d84cf10e315730dfd8f2edcf4cad9073d142cf5da0c (lib) Finished ghc-8.2.1-bc10931b0f169622deb74d84cf10e315730dfd8f2edcf4cad9073d142cf5da0c (lib) Configuring executable 'ghc-demo' for ghc-demo-0.. Preprocessing executable 'ghc-demo' for ghc-demo-0.. Building executable 'ghc-demo' for ghc-demo-0.. [1 of 1] Compiling Main ( Main.hs, /tmp/lll/dist-newstyle/build/x86_64-linux/ghc-8.2.1/ghc-demo-0/c/ghc-demo/build/ghc-demo/ghc-demo-tmp/Main.o ) Linking /tmp/lll/dist-newstyle/build/x86_64-linux/ghc-8.2.1/ghc-demo-0/c/ghc-demo/build/ghc-demo/ghc-demo ... --8<---------------cut here---------------end--------------->8--- So, cabal was able to succesfully produce an executable which uses a reinstalled lib:ghc w/ a different lib:binary version! But now for the limitations: this currently works at best with a GHC 8.2.1 installation whose configuration matches the generated files I manually included in the modified lib:ghc package in http://hackage.haskell.org/package/ghc-8.2.1/src/autogen/ which as you can see from the generated files (see e.g. Config.hs) was a Linux/x86_64/libgmp/internal-libffi/... configuration. ...and this finally brings me to the purpose of why I wrote this email: In order to make lib:ghc properly reinstallable we'd need to either a) have a way to regenerate the files autogen/* via a custom Setup.hs (by the likes of genprimcode or similiar) or the lazy option, b) simply have those files installed in a place we can easily locate (again, from a custom Setup.hs) Does this make any sense; which option do you prefer? Also, I'd like to know if you can think of reasons why or situations when the reinstalled lib:ghc wouldn't work; or other reasons why this is a bad idea. Cheers, hvr On 2016-06-05 at 19:02:29 +0200, Herbert Valerio Riedel wrote: >> As an ultimate goal for a project like this, it would be nice if you could >> do something like upgrade your base package but continue to use the ghc >> package without incompatible types. > > That's one way to approach this. However, I've been experimenting with a > more flexible way: Thanks to cabal-1.24's new nix-based tech-preview > which gets rid of that `--force-reinstall` abomination for good, the > prospect of simply recompiling the `ghc` cabal package becomes more > attractive. > > This would get rid of the primary reason which currently constraints us > to using the package versions bundled in GHC (like e.g. for > `bytestring`) as soon as `ghc` enters an install-plan. > > I'm still investigating this, but early experiments were promising, but > I need to reorganise GHC's source-tree a little bit to know for sure if > this is feasable. But if this as I hope is possible, this could open up > a lot of new possibilities!

3 2